(PDF) VISUAL FIELD DIGEST

6th edition VISUAL FIELD DIGEST A guide to perimetry and the Octopus perimeter 4.1 2.1 -0.8 6.7 2.6 0.4 1.0 0.4 0.7 0.6 Lyne Racette, Monika Fischer, Hans Bebie, Gábor Holló, Chris A. Johnson, Chota Matsumoto Illustrated by Philip Earnhart 6th edition VISUAL FIELD DIGEST A guide to perimetry and the Octopus perimeter Lyne Racette, Monika Fischer, Hans Bebie, Gábor Holló, Chris A. Johnson, Chota Matsumoto Illustrated by Philip Earnhart Editor: Haag-Streit AG, Köniz, Switzerland 6 Edition, published 2016 Layout, cover design & illustrations by Philip Earnhart ISBN 978-3-033-05854-5 Copyright © 2016 HAAG-STREIT AG Haag-Streit AG allows the use of this publication for personal or academic use under the conditions that (i) it is used without commercial purpose and (ii) the content is reproduced exactly as the original by mentioning Haag-Streit AG, Switzerland as the owner of the copyright. Non-academic, non-personal or commercial users might only use this publication in whole or in part after a written authorization by the copyright holder. Trademark statement “Haag-Streit”, “900” and “Octopus” are either registered trademarks or trademarks of Haag-Streit Holding AG. The following are either registered trademarks or trademarks of Carl Zeiss Meditec: “Guided Progression Analysis”, “GPA”, “Humphrey”, “HFA”, “SITA”, “SITA Fast”, “SITA Standard”, “Visual Field Index”, and “VFI”. III AUTHORS & CONTRIBUTORS AUTHORS CONTRIBUTORS LYNE RACETTE, PhD JONATHAN MYERS, MD Assistant Professor Co-Director of Glaucoma Service Eugene and Marilyn Glick Eye Institute, Wills Eye Hospital Department of Ophthalmology Philadelphia, USA Indiana University School of Medicine Indianapolis, USA RAMANARASIAH RANGARAJ, MD Consultant Surgeon & Head of MONIKA FISCHER, MSc the Department of Ophthalmology Market Manager Perimetry Premier Eye Care And Surgical Center Haag-Streit AG Chennai, INDIA Köniz, SWITZERLAND FIONA ROWE, PhD HANS BEBIE, PhD Professor of Orthoptics Professor Emeritus Orthoptics and Health Services Research Institute for Theoretical Physics Institute of Psychology, Health and Society University of Bern University of Liverpool Bern, SWITZERLAND Liverpool, UNITED KINGDOM GÁBOR HOLLÓ, MD, PhD, DSc SONOKO TAKADA MD, PhD Professor of Ophthalmology Adjunct Assistant Professor Director of Glaucoma and Perimetry Unit Department of Ophthalmology Department of Ophthalmology Kindai University Faculty of Medicine Semmelweis University Osaka-Sayama, JAPAN Budapest, HUNGARY CHRIS A. JOHNSON, PhD, DSc ILLUSTRATOR Professor of Ophthalmology Director of Visual Field Reading Center PHILIP EARNHART University of Iowa Art, Communication & Design Iowa City, USA earnhart.ch Biel, SWITZERLAND CHOTA MATSUMOTO, MD, PhD Professor of Ophthalmology Department of Ophthalmology, Kindai University Faculty of Medicine Osaka-Sayama, JAPAN Preface V PREFACE Since the publication of the 5th edition of the Visual Field about these recent advancements and retains the com- Digest in 2004, clinicians’ expectations regarding visual prehensiveness of past editions. ield testing and analysis have signi icantly increased. In today’s busy and fast-paced clinics, maximizing the Furthermore, this 6th edition puts a stronger emphasis trade-off between accuracy of results, test duration and on the challenges and possible pitfalls associated with the effort required from both the patient and the examiner visual ield testing in clinical practice and provides guid- is more important than ever before. ance on how to overcome them. While this edition builds on the previous versions, its format has been updated While the basic testing principles used in perimetry with the intention of making visual ield testing accessible today have remained largely unchanged since the intro- to everyone, including clinicians, residents, researchers, duction of the manual Goldmann perimeter in 1946, examiners, students and those without previous knowl- Octopus perimeters have pioneered numerous import- edge of perimetry. Much effort has been invested in ant changes in perimetry. The development of the irst creating instructive igures to support the key points of automated perimeter, the Octopus 201, by Fankhauser, the text. Spahr and Jenni in 1974, opened the door for automated perimetric testing as we know it today. Further, semi- We wish to thank Philip Earnhart for creating the igures automation in kinetic perimetry, irst introduced for the and graphics that beautifully illustrate this book and Koo- Octopus perimeter nearly 20 years ago, has facilitated sha Ramezani for proofreading the inal version. Further- kinetic testing. more, this project would not have been possible without the unfailing support of Haag-Streit AG, for which we are Since then, knowledge on how to best select, perform grateful. Finally, we should like to thank our contributors and interpret perimetric tests in clinical practice has for providing us with the clinical cases used throughout increased considerably. Normative databases, global the book to illustrate various aspects of perimetry and indices such as Mean Defect, the Defect Curve and many for sharing their knowledge with us. other useful tools for analyzing the measured sensitivity thresholds have been irst introduced on Octopus perime- We hope that this book on perimetry in general, and on ters, before becoming worldwide standards in visual ield the Octopus perimeter in particular, is not only compre- interpretation. hensive, but also enjoyable to read for anybody inter- ested in visual ield testing. We are convinced that the Since the last edition of this book 12 years ago, several information shared in the pages ahead will be useful to advances in perimetric testing with Octopus perimeters clinicians and ultimately to their patients, whose sight have been achieved. EyeSuite Progression Analysis has we care deeply about. We wish you an enriching and been developed and is a powerful tool for assessing pro- pleasant reading experience. gression. In addition, both Cluster Analysis and Polar Analysis are helpful features for establishing a relationship Lyne Racette, Monika Fischer, Hans Bebie, Gábor Holló, between functional and structural results. This new edition Chris A. Johnson, Chota Matsumoto of the Visual Field Digest provides in-depth information September 2016 VI Table of contents TABLE OF CONTENTS 1. INTRODUCTION 1 WHY READ THIS BOOK 1 WHO SHOULD READ THIS BOOK 1 HOW TO READ THIS BOOK 2 CONTENT AT A GLANCE 3 2. WHAT IS PERIMETRY 7 INTRODUCTION 7 Perimetry – A standard test in ophthalmology 7 THE NORMAL VISUAL FIELD 8 Spatial extent of the visual ield 8 Sensitivity to light in the visual ield 9 The hill of vision – A visualization of visual function 10 MEASURING SENSITIVITY TO LIGHT ACROSS THE VISUAL FIELD 11 Perimetry allows quanti ication of abnormal sensitivity to light 11 The perimetric test 12 DISPLAY OF SENSITIVITY THRESHOLDS 14 The decibel scale used in Perimetry 14 Graphic display of sensitivity thresholds 16 CHALLENGES IN VISUAL FIELD TESTING AND INTERPRETATION 17 Perimetric testing has low resolution 17 Normal sensitivities depend on age and test location 18 Perimetry has objective and subjective components 20 Normal luctuation depends on test locations and disease severity 20 Clinical standard for visual function testing 22 3. HOW TO PERFORM PERIMETRY YOU CAN TRUST 25 INTRODUCTION 25 Perimetry – A subjective test 25 Perimetry – Need for a team approach 26 The important role of the doctor 26 The important role of the visual ield examiner 27 HOW TO PERFORM VISUAL FIELD TESTING 28 Setting up the perimeter 28 Placing an adequate trial lens 29 Instructing the patient 29 Setting up and positioning the patient 31 Monitoring the patient during the examination 35 Table of contents VII COMMON PITFALLS TO AVOID 37 Inconsistent patient behavior 37 Mistakes in the set-up procedure 40 External obstructions blocking stimuli from reaching the retina 43 Clinical relevance of untrustworthy visual ields 45 4. KEY EXAMINATION PARAMETERS 47 FIXED EXAMINATION PARAMETERS 47 PATIENT-SPECIFIC EXAMINATION PARAMETERS 48 Type of perimetry: Static or kinetic perimetry 48 Stimulus type: Standard or non-conventional 51 Test pattern 54 Test strategy 56 5. SELECTING A TEST PATTERN 59 INTRODUCTION 59 TEST PATTERNS FOR GLAUCOMA 60 Typical visual ield defects in glaucoma 60 Standard test pattern in glaucoma care 62 Alternative test patterns for the central 30° 64 Further test patterns for glaucoma 65 TEST PATTERNS FOR NEUROLOGICAL VISUAL FIELD LOSS 67 Typical visual ield defects in neuro-ophthalmic conditions 67 Thorough assessment of neurological visual ield defects 69 TEST PATTERNS FOR RETINOPATHIES 70 Typical visual ield defects in retinopathies 70 Test patterns for the macula 71 Test patterns for the full ield 72 TEST PATTERNS FOR VISUAL ABILITY TESTING 74 Test patterns for visual ability to drive 74 Test patterns for blepharoptosis 76 Test patterns for visual impairment 78 6. SELECTING A TEST STRATEGY 81 INTRODUCTION 81 QUANTITATIVE STRATEGIES 82 Normal strategy 83 Dynamic strategy 85 Low-vision strategy 86 Tendency-Oriented-Perimetry (TOP) strategy 87 QUALITATIVE STRATEGIES 90 1-Level Test strategy (two-zone strategy) 91 Screening strategy 92 2-Level Test strategy (three-zone strategy) 94 RECOMMENDATIONS ON KEY EXAMINATION PARAMETERS 96 VIII Table of contents 7. OVERVIEW OF VISUAL FIELD REPRESENTATIONS 99 INTRODUCTION 99 RELATIONSHIP AMONG OCTOPUS VISUAL FIELD REPRESENTATIONS 99 REPRESENTATIONS DISPLAYING SENSITIVITY THRESHOLDS 101 Values 101 Grayscale of Values 102 REPRESENTATIONS BASED ON COMPARISON WITH NORMAL 103 Comparison 103 Grayscale of Comparison 105 Probabilities 107 Defect Curve 109 Cluster Analysis 110 Polar Analysis 113 REPRESENTATIONS BASED ON COMPARISON WITH NORMAL, CORRECTED FOR DIFFUSE DEFECT 115 Corrected Comparison 115 Corrected Probabilities 117 Corrected Cluster analysis 118 GLOBAL INDICES 118 Mean Sensitivity (MS) 119 Mean Defect (MD) 119 Square root of Loss Variance (sLV) 120 Corrected square root of Loss Variance (CsLV) 121 Diffuse Defect (DD) 121 Local Defect (LD) 122 RELIABILITY INDICES 123 False Positive (FP) answers 123 False Negative (FN) answers 124 Reliability Factor (RF) 124 Short-term Fluctuation (SF) 124 8. CLINICAL INTERPRETATION OF A VISUAL FIELD 127 INTRODUCTION 127 STEP-BY-STEP INTERPRETATION OF A VISUAL FIELD 134 Overview of step-by-step work low 134 Step 1 – Con irm patient and examination parameters 135 Step 2 – Determine whether the visual ield can be trusted 136 Step 3 – Identify diffuse visual ield defects 140 Step 4 – Distinguish between normal and abnormal visual ields 145 Step 5 – Assess shape and depth of defect 149 Step 6 – Assess cluster defects in glaucoma 152 Step 7 – Where to look for a structural defects 155 Step 8 – Assess severity 159 Table of contents IX 9. INTERPRETATION OF VISUAL FIELD PROGRESSION 165 INTRODUCTION 165 ASSESSMENT OF GLOBAL VISUAL FIELD CHANGE 168 Change of Mean Defect (MD) as a measure of global change 168 Trend analysis for the visualization of change 168 Using probabilities to distinguish between stable and changing visual ield series 171 MD Trend Analysis 174 Interpretation of MD Trend Analysis 176 Selection of adequate visual ields for analysis 176 DISTINCTION BETWEEN LOCAL AND DIFFUSE CHANGE 178 Importance of distinction between local and diffuse change 178 Use of Diffuse Defect index (DD) to identify diffuse change 179 Use of Local Defect index (LD) to identify local change 180 Use of square root of Loss Variance (sLV) to identify local change 180 Clinical interpretation of 4 Global Trend Analyses 181 CLUSTER TREND AND CORRECTED CLUSTER TREND ANALYSIS 183 Importance of assessing cluster progression in glaucoma 183 Cluster Trend and Corrected Cluster Trend Analysis 183 POLAR TREND ANALYSIS 187 Importance of establishing a relationship between structural and functional progression 187 Use of polar trend analysis to assist in the detection of glaucomatous structural progression 187 10. NON-CONVENTIONAL PERIMETRY 193 INTRODUCTION 193 FUNCTION-SPECIFIC PERIMETRY 194 Rational for using function-speci ic perimetry 194 Use of function-speci ic perimetry in clinical practice 195 Pulsar Perimetry 196 Flicker Perimetry 198 Short Wavelength Automated Perimetry (SWAP) 200 STIMULUS V FOR PATIENTS WITH LOW VISION 201 11. KINETIC PERIMETRY 205 WHAT IS KINETIC PERIMETRY 205 Limitations of static perimetry 205 Description of kinetic perimetry 207 WHY PERFORM KINETIC PERIMETRY 210 Bene its of kinetic perimetry 210 Limitations of kinetic perimetry 212 HOW TO PERFORM KINETIC PERIMETRY 214 The Goldmann perimeter: Kinetic visual ield testing 214 Key decisions in kinetic perimetry 215 Stimulus types 216 General testing methodologies 218 Step-by-Step example of kinetic perimetry 227 Automation of kinetic perimetry 230 X Table of contents 12. TRANSITIONING TO A DIFFERENT PERIMETER MODEL 235 INTRODUCTION 235 GENERAL ASPECTS OF TRANSITIONING 236 Measured sensitivity thresholds cannot be compared across different perimeter models 236 Device-speci ic normative databases allow comparison of sensitivity losses between devices 237 Import of existing data to ensure continuity 239 Managing patient-related luctuation 243 SPECIFIC ASPECTS RELATED TO TRANSITIONING FROM THE HUMPHREY FIELD ANALYZER 243 Selection of test parameters 243 Interpretation of a single visual ield 244 Interpretation of visual ield progression 250 13. CLINICAL CASES 255 INTRODUCTION 255 GLAUCOMA – SINGLE FIELD 257 GLAUCOMA – TREND 266 NEUROLOGICAL DISEASES 272 RETINAL DISEASES 280 INDEX 285 1 CHAPTER 1 INTRODUCTION WHY READ THIS BOOK The irst Visual Field Digest was published in 1983 and published (e.g., Global Trend Analysis, Cluster Trend has been used as a guide to perimetry and the Octopus Analysis and Polar Trend Analysis), this edition places a perimeter by thousands of Octopus users ever since. stronger emphasis on the clinical application of perim- This 6th edition is a completely revised version of the etry compared to previous editions. All key concepts are 5th edition published in 2004. Not only does it contain illustrated to facilitate understanding. This allows any updates on features developed since the last edition was reader to easily and quickly grasp the key information. WHO SHOULD READ THIS BOOK This book has been written for any current or future care professionals and provides many practical tips and eye care professionals who perform or interpret visual tricks to get even more out of their perimetric testing. ield examinations as part of their diagnostic routine. And last, it has been written for researchers and expert This group not only includes clinicians in optometry users of perimetry who are interested in the scienti ic and ophthalmology, but also visual ield examiners who background of perimetry and the Octopus perimeter. administer perimetric tests to patients. While this book provides in-depth information about A wide range of users will ind useful information in the design and use of the Octopus perimeters, it is also this book. It has been created for students with limited very useful reading for users of other perimeter brands, knowledge in perimetry and therefore explains funda- as the fundamental concepts of perimetry are compara- mentals in perimetry in an easy to understand manner. ble among perimeter brands and are illustrated in this In addition, it has been composed for experienced eye book in an easy to understand way. 2 Chapter 1 | Introduction HOW TO READ THIS BOOK To cater to the needs of readers with different experi- For more experienced users, individual chapters or sec- ence levels as well as different learning styles, this book tions in this book can also be read individually, as each can be read in several ways. chapter is structured in a way that it is self-explanatory, or if not, a clear reference to another chapter is given. For students and inexperienced users in perimetry, this book is structured in a way that, when read from begin- To find and understand key information quickly, all ning to end, it allows the content to be followed with essential concepts are graphically illustrated to support minimal prior knowledge. For this reason, the book a quick understanding of the concept. With more than starts with fundamentals of perimetry such as, what the 200 graphics available in this book, it is thus possible test does, how to administer the test and how to choose to grasp key information just by looking at the graphics test parameters, before moving on to visual ield inter- and reading the captions. pretation and special topics like kinetic perimetry or function-speci ic perimetry. To tie the learning to real If several choices or methods are compared, overview clinical situations, this book concludes with a case pre- tables are provided for quick comparison between sentation section. them. Sometimes, in-depth background information KEY ELEMENTS USED IN THIS BOOK Key examination parameters | Patient-specific examination parameters 51 52 Key examination parameters | Patient-specific examination parameters TABLE Provides a quick overview and contrasts different concepts/methods COMPARISON BETWEEN STATIC AND KINETIC PERIMETRY TABLE 4-1 BOX 4B GOLDMANN SIZES I TO V STATIC KINETIC Š‡•‹œ‡…‘˜‡–‹‘•—•‡†–‘†ƒ›–‘†‡•…”‹„‡ƒ’‡”‹- ‡–”‹…•–‹—Ž—•ƒŽŽ™‡”‡†‡”‹˜‡†ˆ”‘–Š‡™‘”‘ˆ BLIND SPOT V 1.7° ”‘ˆ‡••‘”ƒ• ‘Ž†ƒǡ™Š‘†‡˜‡Ž‘’‡†–Š‡ ‘Ž†- ADVANTAGES Ž‹‹…ƒŽ‰‘Ž†•–ƒ†ƒ”† ‹‰Š•’ƒ–‹ƒŽ”‡•‘Ž—–‹‘ ƒ’‡”‹‡–‡”‹ͳͻͶ͸Ǥ‡†‡ϐ‹‡†•–ƒ†ƒ”†•‹œ‡• ˆ‘”’‡”‹‡–”‹…•–‹—Ž‹ǡƒ†–Š‡ ‘Ž†ƒ•‹œ‡•–‘ 20 IV 0.8° ‹‰Š’”‡…‹•‹‘•‡•‹–‹˜‹–›–Š”‡•Š‘Ž†• ƒ•–’‡”‹’Š‡”ƒŽ–‡•–‹‰ ƒ”‡•–‹ŽŽ™‹†‡Ž›—•‡†Ǥƒ…Š•–‡’…‘””‡•’‘†•–‘ƒ …Šƒ‰‡‹†‹ƒ‡–‡”„›ƒˆƒ…–‘”‘ˆʹƒ†‹ƒ”‡ƒ„›ƒ Fully automated ”‘˜‹†‡•‹ˆ‘”ƒ–‹‘ƒ„‘—–‘–Š‡”˜‹•—ƒŽˆ—…–‹‘• ˆƒ…–‘”‘ˆͶǤ‹œ‡‹••‡˜‡”ƒŽ–‹‡••ƒŽŽ‡”–Šƒ–Š‡ III 0.43° ’Š›•‹‘Ž‘‰‹…ƒŽ„Ž‹†•’‘–ǡ•‘–Šƒ–‹–™ƒ•…‘•‹†‡”‡† WHAT IT IS BEST ƒŽŽ…Šƒ‰‡•‹•‡•‹–‹˜‹–›–Š”‡•Š‘Ž†• ƒŽŽ…Šƒ‰‡•‹•’ƒ–‹ƒŽ‡š–‡–‘ˆƒ†‡ˆ‡…– –‘„‡ƒƒ……—”ƒ–‡‡ƒ•—”‡‡–•‹œ‡Ǥ II 0.2° AT DETECTING Šƒ‰‡•‹–Š‡…‡–”ƒŽƒ”‡ƒ ‡”‹’Š‡”ƒŽ…Šƒ‰‡• The Goldmann stimuli I to V are presented in relation Remaining vision in advanced diseases I 0.1° to the size of the physiological blind spot. BOX COMMON USES Žƒ—…‘ƒ ‡—”‘Ǧ‘’Š–ŠƒŽ‘Ž‘‰‹…ƒŽ…‘†‹–‹‘• Provides expert Macular diseases ‡”‹’Š‡”ƒŽ”‡–‹ƒ†‹•‡ƒ•‡• knowledge Visual ability testing ‘™˜‹•‹‘ FUNCTION-SPECIFIC PERIMETRY Š‹Ž†”‡ —…–‹‘Ǧ•’‡…‹ϐ‹…’‡”‹‡–”›—•‡•†‹ˆˆ‡”‡–•–‹—Ž—•–›’‡• „ƒ…‰”‘—† ȋŠ‘”–Ǧƒ˜‡Ž‡‰–Š —–‘ƒ–‡† ‡”‹‡–”›ǡ –‘ •–‹—Žƒ–‡ †‹ˆˆ‡”‡– ˜‹•—ƒŽ ˆ—…–‹‘• ȋ‡Ǥ‰Ǥǡ ‘–‹‘ǡ ‘” ‘” ȌǢ ƒ ™Š‹–‡ ϐŽ‹…‡”‹‰ •–‹—Ž—• ‘ ƒ ™Š‹–‡ „ƒ…- …‘Ž‘” ˜‹•‹‘Ȍǡ „—– –Š‡› ƒŽŽ Šƒ˜‡ –Š‡ •ƒ‡ ’—”’‘•‡ǣ ‰”‘—†ȋ Ž‹…‡”‡”‹‡–”›ȌǢ‘”ƒ’—Ž•ƒ–‹‰•–‹—Ž—•™‹–Š ‡ƒ•—”‹‰ƒ•—„•‡–‘ˆ–Š‡˜‹•—ƒŽ•›•–‡‹†‹˜‹†—ƒŽŽ›ǡ–‘ …‘…‡–”‹… ”‹‰• …Šƒ‰‹‰ „‘–Š •’ƒ–‹ƒŽ ”‡•‘Ž—–‹‘ ƒ† STIMULUS TYPE: STANDARD OR NON-CONVENTIONAL ‰‡–‘”‡•‡•‹–‹˜‡”‡•’‘•‡•ˆ‘”‡ƒ”Ž›†‹•‡ƒ•‡†‡–‡…–‹‘Ǥ …‘–”ƒ•–ȋ—Ž•ƒ”‡”‹‡–”›ȌǤŠ‡›ƒ”‡†‡•…”‹„‡†‹‘”‡ ‹ˆˆ‡”‡–…–‘’—•’‡”‹‡–‡”‘†‡Ž•‘ˆˆ‡”†‹ˆˆ‡”‡–ˆ—…- †‡–ƒ‹Ž‹Šƒ’–‡”ͳͲǤ STANDARD WHITE-ON-WHITE PERIMETRY –‹‘Ǧ•’‡…‹ϐ‹…•–‹—Ž‹ȋFIG 4-3Ȍǣƒ„Ž—‡•–‹—Ž—•‘ƒ›‡ŽŽ‘™ Š‡ •–ƒ†ƒ”† ’‡”‹‡–”‹… •–‹—Ž—• ‹• ™Š‹–‡ ‘ ƒ ™Š‹–‡ Ž‹‰Š– ƒŽŽ‘™• ˜‹•—ƒŽ ϐ‹‡Ž† –‡•–‹‰ ˆ”‘ ‡ƒ”Ž› –‘ ƒ†˜ƒ…‡† „ƒ…‰”‘—†ǡƒ†–Š‹•–›’‡‘ˆ’‡”‹‡–”›‹•…‘‘Ž›”‡- †‹•‡ƒ•‡ǡ ‹Ǥ‡Ǥǡ ‹– ‘ˆˆ‡”• ƒ Žƒ”‰‡ †›ƒ‹… –‡•–‹‰ ”ƒ‰‡Ǥ › FUNCTION-SPECIFIC PERIMETRY ˆ‡””‡†–‘ƒ•™Š‹–‡Ǧ‘Ǧ™Š‹–‡’‡”‹‡–”›ǡ‘”–ƒ†ƒ”†—–‘- …‘˜‡–‹‘ǡ–Š‡•–ƒ†ƒ”†•–‹—Ž—•—•‡†‹•”‘—†ǡ™‹–Šƒ ƒ–‡†‡”‹‡–”›ȋȌǤ †‹ƒ‡–‡”‘ˆͲǤͶ͵ιǡ™Š‹…Š‹•ƒŽ•‘–Š‡ ‘Ž†ƒ•–‹—Ž—•ǡ •‹œ‡ǡ„ƒ•‡†‘–Š‡†‡ϐ‹‹–‹‘„›”‘ˆ‡••‘”ƒ• ‘Ž†- Š‡™Š‹–‡…‘Ž‘—”•–‹—Ž—•‘ˆˆ‡”•–Š‡ƒ†˜ƒ–ƒ‰‡‘ˆ•–‹- ƒǤ ‘”‘”‡‹ˆ‘”ƒ–‹‘‘ ‘Ž†ƒ•–‹—Ž—••‹œ- —Žƒ–‹‰ƒŽŽ†‹ˆˆ‡”‡–”‡–‹ƒŽ…‡ŽŽ–›’‡•Ǥ•ƒ”‡•—Ž–ǡ™Š‹–‡ es, refer to BOX 4BǤ Time 1 Time 2 ON OFF SWAP Flicker Pulsar TEXT FIGURE Full presentation Illustrates key FIGURE 4-3 Stimuli of function-specific perimetry from left to right: Short Wavelength Automated Perimetry (SWAP), Flicker of a topic Perimetry and Pulsar Perimetry. information in an easy to understand way FIGURE 1-1 To accommodate the preferences of different readers, different structural elements are used in the Visual Field Digest. To highlight key information, there are Figures and Tables; to provide a full overview of a topic, there is full text; and to provide in-depth expert knowledge, there are Boxes. Content at a glance 3 is of interest to some readers, but not crucial for good fere with the low of the book. The elements described clinical practice. Such information is provided in a light above are shown in FIG 1-1. blue box and can be read for interest but does not inter- CONTENT AT A GLANCE In this section, a brief overview of the content of each chapter is presented. CHAPTER 2 – WHAT IS PERIMETRY? Chapter 2 provides essential information on perimetry provides a general introduction on how the data is dis- as a technology which is valid for any perimeter brand. played, and highlights common challenges associated It shows how and why visual ield testing is performed, with visual ield testing. CHAPTER 3 – HOW TO PERFORM PERIMETRY YOU CAN TRUST Chapter 3 focuses on information relevant to visual ield falls in perimetry such as learning effects, fatigue effects, technicians and those people instructing them. It stresses set-up errors and artifacts are presented, along with the the importance of the visual ield technician in obtaining procedures for avoiding these problems. How to detect trustworthy visual ield results and explains the essential whether a visual ield is trustworthy is later presented steps of visual ield testing. In a second part, common pit- in Chapter 8. CHAPTER 4 – KEY EXAMINATION PARAMETERS Chapter 4 focuses on ixed examination parameters and The idea is to provide an introduction to what these pa- the key patient-speci ic parameters a clinician needs rameters are and how to make appropriate testing deci- to decide about. Key questions to be answered regard- sions. The key parameters will be described in depth in ing patient-speci ic test parameters are the following: subsequent chapters. 1) Static or kinetic perimetry? 2) Which stimulus type? 3) Which test pattern? 4) Which strategy? 4 Chapter 1 | Introduction CHAPTER 5 – SELECTING A TEST PATTERN Chapter 5 presents all available test patterns on Octo- Performance evaluations such as driving and visual dis- pus perimeters. The chapter is organized according to ability tests as well as ptosis test patterns are described pathology or test (i.e., it starts with glaucoma, and con- towards the end of the chapter. tinues with neuro-ophthalmic and retinal diseases). CHAPTER 6 – SELECTING A TEST STRATEGY Chapter 6 presents all available test strategies on Octo- clinician in selecting one of the various quantitative or pus perimeters and shows that there is always a trade-off qualitative test strategies. between test duration and accuracy in order to guide the CHAPTER 7 – OVERVIEW OF VISUAL FIELD REPRESENTATIONS Chapter 7 introduces all visual ield representations avail- in each representation and further information about the able on Octopus perimeters and shows their respective design of the representation. For clinicians, this chapter relationships. Further, each representation is explained in can serve as a glossary. detail, including a clear de inition of all the symbols used CHAPTER 8 – CLINICAL INTERPRETATION OF A VISUAL FIELD Chapter 8 is a key chapter in this book, guiding clinicians Further, this chapter highlights those representations through visual ield interpretation in an easy to follow most useful in answering speci ic clinical questions, and work low. It starts by showing 6 visual ield examples shows how to interpret these representations in clinical and their respective representations across all stages of practice. Clinical examples are frequently provided to disease to provide a graphical reference on what visual illustrate the bene its of each respective representation ield results look like in a given situation. The same cases in a certain clinical situation. are also provided as a poster that can be removed from the book as a reference in daily clinical practice. CHAPTER 9 – INTERPRETATION OF VISUAL FIELD PROGRESSION Chapter 9 focuses on the use of EyeSuite Progression determine whether a visual ield series is stable or not. Analysis to assess visual ield progression. It explains Further, it shows the bene its and interpretation of the the fundamentals of the trend analysis approach used to various trend representations, including Global Trend Content at a glance 5 Analysis, Cluster Trend Analysis and Polar Trend Analy- the visual ield in which progression is occurring and, in sis, which not only allow it to be determined whether a case of glaucoma, where to look for a spatial relationship visual ield series is progressing and at which rate, but with structural results. also whether progression is diffuse or local, the area of CHAPTER 10 – NON-CONVENTIONAL PERIMETRY Chapter 10 focuses on other stimulus types besides the formation about Pulsar, SWAP and Flicker perimetry. The standard Goldmann size III used in perimetry. The chap- chapter then concludes with the bene its of using a larger ter starts with function-speci ic perimetry designed for stimulus V for low-vision patients. early glaucoma detection and provides background in- CHAPTER 11 – KINETIC PERIMETRY Chapter 11 focuses on kinetic perimetry. Similar to the presented and illustrated in a real clinical case. Towards static perimetry chapter, the basic examination parame- the end, the bene its of different levels of automation are ters and when to choose each one are discussed. Gener- also discussed. al approaches on how to perform kinetic perimetry are CHAPTER 12 – TRANSITIONING TO A DIFFERENT PERIMETER MODEL Chapter 12 focuses on speci ic challenges associated rimeter models and shows the impact of patient-related with transitioning from one perimeter model to another. luctuation. To support a smooth transition from an HFA It focuses both on the transition to a different Octopus perimeter to an Octopus perimeter, guidance in relation model, as well as the transition from a Humphrey to an to known HFA perimeter terminologies is provided on Octopus model. It highlights the importance of normative the selection of test parameters as well as the interpreta- databases for minimizing the differences between pe- tion of the perimetric result. CHAPTER 13 – CLINICAL CASES To support the interpretation of visual ield results in contain key patient information, as well as visual ield clinical practice, 23 clinical cases are presented, show- results and other relevant diagnostic results such as IOP, ing typical visual ields of patients with glaucoma, neu- fundus images, OCT scans and MRIs. ro-ophthalmic disease and retinal disease. All these cases 6 7 CHAPTER 2 WHAT IS PERIMETRY? INTRODUCTION PERIMETRY – A STANDARD TEST IN OPHTHALMOLOGY Perimetry is a standard method used in ophthalmol- nose glaucoma, but it is also often used to assess visu- ogy and optometry to assess a patient’s visual ield. al loss resulting from retinal diseases, as well as optic It provides a measure of the patient’s visual function nerve, chiasmal or post-chiasmal damage due to trauma, throughout their ield of vision. The devices used to per- stroke, compression and tumors. form this evaluation are called perimeters. Perimetry is performed for several reasons: 1) detection of pathol- Additionally, perimetry is used regularly for visual ability ogies; 2) evaluation of disease status; 3) follow-up of testing. Its most common use is to test a person’s visual pathologies over time to determine progression or dis- ability to drive. Furthermore, it is used to provide a ease stability; 4) determination of ef icacy of treatment quantitative measure of visual function in order to de- and 5) visual ability testing. termine eligibility for a pension for visual impairment, and also to assess the bene its of ptosis surgery. Any pathology along the visual pathway usually results in a loss of visual function. Perimetry can identify de- In sum, perimetry is a universally available diagnostic viations from normal, and consequently the associated method to assess a patient’s visual ield or visual function. pathologies. Perimetry is most commonly used to diag- 8 Chapter 2 | What is perimetry? THE NORMAL VISUAL FIELD SPATIAL EXTENT OF THE VISUAL FIELD The visual ield of a person is de ined as the area in nothing can be seen). The extent of the visual ield is an which a person can see at a given moment relative to essential part of one’s visual function, because a con- the direction of ixation, without head or eye movement stricted visual ield has a signi icant negative impact on (i.e., it de ines the boundaries of the area beyond which activities of daily living, and as a result on quality of life. SPATIAL EXTENT OF A NORMAL VISUAL FIELD A) MONOCULAR VISUAL FIELD SUPERIOR SUPERIOR NASAL (right eye) (right eye) NASAL TEMPORAL (right eye) Fixation TEMPORAL (right eye) INFERIOR INFERIOR B) BINOCULAR VISUAL FIELD SUPERIOR SUPERIOR TEMPORAL (left eye) TEMPORAL (left eye) TEMPORAL Fixation (right eye) TEMPORAL (right eye) INFERIOR INFERIOR FIGURE 2-1 The monocular visual field of one eye is limited by the eye socket, nose, brow and cheekbones (A). The binocu- lar visual field of two eyes overlaps in the central area (B). The normal visual field 9 The visual ield of one eye is called the monocular visual In people with normal vision, the visual ield is binoc- ield (FIG 2-1A). Its spatial extent in people with normal ular (FIG 2-1B). This means that it contains input from vision is limited by the facial anatomy of the person, both eyes, with integration and mapping of information with the eye socket, nose, brow and cheekbones, which from the two eyes, allowing for stereo acuity and depth outlines the limits of the visual ield. On average, the perception. Visual information in the central 60 degrees monocular visual ield extends from 60° nasally to ap- of the visual ield is processed by both eyes. proximately 90° or more temporally, and from approxi- mately 60° superiorly to 70° inferiorly. SENSITIVITY TO LIGHT IN THE VISUAL FIELD The area in which a person can see (extent of the visual level hanging from the ceiling. In that room, only a few ield) does not suf ice to describe a person’s vision. It is people can see. As the light intensity of the bulb is in- also important to have a measure of sensitivity to light. creased, an increasing number of people will be able to But what is a person’s sensitivity to light? One can see in the room. The people who could see even the very imagine a room in which 100 people are present. The dim light bulb have a very high sensitivity to light, while room is dim, with an adjustable light bulb at its lowest the others have a lower sensitivity to light (FIG 2-2). SENSITIVITY TO LIGHT High Dim light Sensitivity Light to light intensity Low Bright light FIGURE 2-2 This figure illustrates the inverse relationship between light intensity and sensitivity to light. A person who can perceive a very dim light has a very high sensitivity to light, while a person who can only perceive very bright lights has low sensitivity to light. 10 Chapter 2 | What is perimetry? THE HILL OF VISION – A VISUALIZATION OF VISUAL FUNCTION Sensitivity to light is not uniform across the spatial ex- the x- and y-axes representing the visual ield locations tent of the visual ield and depends on location within and the z-axis representing the sensitivity to light. Since the visual ield. For normal eyes and in typical daytime this representation resembles a hill, it is commonly re- illumination, sensitivity is highest in the central area of ferred to as the hill of vision, which is a visualization of the visual ield and decreases gradually towards the pe- a person’s visual function. Areas within the hill of vision riphery. To visualize this, sensitivities across the visual represent areas of seeing, and areas outside the hill of vi- ield can be drawn as a three-dimensional graph, with sion represent areas of non-seeing (FIG 2-3). HILL OF VISION Sensitivity to light SUPERIOR Fixation Blind Spot TEMPORAL 70˚ NASAL 80˚ 90˚ INFERIOR FIGURE 2-3 The hill of vision is a three-dimensional representation of the visual field, with the x- and y-axes showing the spatial extent of the visual field using radial coordinates, and the z-axis showing sensitivity to light. Its name stems from the fact that normal sensitivity to light is higher at the center than in the periphery, so that normal vision in this representation resembles a hill. Measuring sensitivity to light across the visual field 11 MEASURING SENSITIVITY TO LIGHT ACROSS THE VISUAL FIELD PERIMETRY ALLOWS QUANTIFICATION OF ABNORMAL SENSITIVITY TO LIGHT Deviations from the normal hill of vision provide valu- can be either constrictions of the boundaries of the visual able clues regarding visual ield loss and the underlying ield, or depressions of sensitivity. Such depressions can pathologies. The pattern and shape of visual loss can be be present throughout the visual ield (widespread low- identi ied by investigating deviations from the normal hill ering of sensitivity), or localized in speci ic areas of the of vision. Differences in the visual ield between the two visual ield (scotomas). It is thus desirable to quantify a eyes can also be identi ied by inspecting deviations from patient’s hill of vision with high accuracy and to identify the normal hill of vision. These deviations from normal its deviation from a normal hill of vision (FIG 2-4). PERIMETRY ALLOWS DETECTION OF ABNORMAL SENSITIVITY TO LIGHT Sensitivity to light Normal Hill of Vision Pathological Hill of Vision FIGURE 2-4 Pathologies affecting sensitivity to light result in an altered hill of vision for the patient. The deviation from the normal hill of vision provides valuable information regarding the nature and severity of the pathology. 12 Chapter 2 | What is perimetry? THE PERIMETRIC TEST Perimetry accurately quanti ies a patient’s sensitivity to stimulus anywhere in their visual ield by pressing a re- light throughout the visual ield in a systematic, highly sponse button. Conceptually and to simplify things, one standardized manner. To assess the visual ield, a hemi- can imagine that at the irst location the luminance of spheric cupola is typically used to project small light the stimulus is increased from the “off” position to the stimuli across the entire area of the visual ield. These dimmest level of an adjustable light bulb. If the patient stimuli, and the uniform background onto which the cannot see the stimulus when it is off or very dim, anoth- stimuli are projected, are highly standardized in terms of er stimulus is shown later, at a higher level of light inten- shape, size, color, light intensity and duration, to ensure sity. Once the stimulus reaches a certain light intensity, high reproducibility. The most commonly used test con- the patient can see it and presses the button. It should ditions project a round, white stimulus on a background, be noted that the stimulus is always turned off before the which is also white, but dimmer than the stimulus. The next stimulus is presented. luminance (i.e., the re lected light intensity) of the stim- ulus can be altered from very low to very high. More This minimum light intensity that can be seen de ines the detailed information on key examination parameters is patient’s sensitivity to light (i.e., the threshold between provided in Chapter 4. non-seeing and seeing) (FIG 2-5). Due to this evaluation method, in perimetry the word threshold is often used, To perform a perimetric test, patients are asked to sit instead of sensitivity to light. For ease of understanding, in front of the cupola with their head stabilized, to ix- “sensitivity threshold” is the term used throughout ate onto a target in the center, and to indicate seeing a this book. SENSITIVITY THRESHOLDS Dim = Seen Stimulus = Not seen No Do you see the stimulus? No THRESHOLD SENSITIVITY No Fixation Yes Yes Yes Stimulus Yes Bright Stimulus FIGURE 2-5 The sensitivity threshold between seeing and non-seeing for stimuli of different intensity presented against a fixed background illumination at a given location in the visual field provides one data point on the hill of vision. Measuring sensitivity to light across the visual field 13 The sensitivity threshold at the irst test location provides across the visual field ( FIG 2-6B). By connecting the the irst data point to characterize the hill of vision (FIG sensitivity thresholds at all tested locations, a patient’s 2-6A). To determine the patient’s hill of vision, the afore- hill of vision can be drawn (FIG 2-6C). mentioned procedure is then repeated at many locations DRAWING THE HILL OF VISION FROM THE SENSITIVITY THRESHOLDS A) SENSITIVITY THRESHOLD OF FIRST LOCATION Do you see the stimulus? Sensitivity threshold of first location Sensitivity threshold Fixation Stimulus B) SENSITIVITY THRESHOLDS AT DIFFERENT LOCATIONS Sensitivity thresholds at all tested locations Stimulus Sensitivity threshold Do you see here? Fixation Do you see there? C) SENSITIVITY THRESHOLDS AT ALL TESTED LOCATIONS Sensitivity thresholds Stimulus at all tested locations Sensitivity threshold Do you see here? Fixation Do you see there? FIGURE 2-6 The hill of vision can be drawn from the individually determined sensitivity thresholds at each location. 14 Chapter 2 | What is perimetry? While the process used to determine sensitivity thresh- perimetry and they will be discussed in depth in Chapters 4, olds is easy to understand, it would be much too time-con- 5 and 6. Additionally, the order of stimulus presentation is suming to test each location of the hill of vision in this randomized throughout the visual ield, to avoid patients manner. Therefore, more ef icient strategies are used in becoming accustomed to a certain presentation pattern. DISPLAY OF SENSITIVITY THRESHOLDS THE DECIBEL SCALE USED IN PERIMETRY In clinical practice, visual ield information needs to be perimetric stimulus that the device can display, whereas easy to interpret and should directly correspond to the values close to 32 dB represent normal foveal vision for clinical situation. For that purpose, perimetry employs a 20-year-old person. While the decibel scale is intuitive the decibel scale, with its unit of measurement being to understand and use in clinical practice, the underlying the decibel (dB). The decibel range depends on perim- considerations and formulas are less intuitive and of lim- eter type and typically ranges from 0 dB to approxi- ited relevance for clinical practice. For those interested, mately 32 dB in the fovea. A sensitivity threshold of 0 dB they are explained in BOX 2A. means that a patient is not able to see the most intense BOX 2A THE RATIONALE FOR THE USE OF THE DECIBEL SCALE The intensity of the light that is re lected on the perimetric surface is called luminance and can be measured objectively with a light meter. It is expressed in candelas per meter squared (cd/m ) or in the older unit, the apostilb (asb), with 1 cd/m corresponding to 3.14 asb. The measurement indicates light lux per unit area. In theory, sensitivity thresholds could be expressed in luminance units. While this would be correct, it would be impractical in clinical practice for the following reasons: 1. Large number of discrete luminance levels The human eye can adjust to a large range of luminance levels over at least 3-4 orders of magnitude (e.g., from almost 0 asb to 10,000 asb in normal daytime lighting conditions). This would make certain threshold values very large and impractical to display. 2. The relationship between visual function and luminance is not linear Visual function is not linear with regard to the light intensity levels. For example, while an increase of 90 asb is likely to be noticed when luminance is increased from 10 to 100 asb, this same absolute increase in luminance (90 asb) would hardly be noticeable when luminance is increased from 1,000 to 1,090 asb. 3. Inverse relationship between luminance and sensitivity to light There is an inverse relationship between stimulus luminance and a patient’s sensitivity to light. A patient with high sensitivity to light only needs a stimulus with low luminance to be able to see it, while a patient with low sensitivity to light needs a stimulus with high luminance. For clinical Display of sensitivity thresholds 15 use, a scale de ining visual ield loss as low and good vision as high would be more intuitive than the inverse luminance scale. 4. Lack of definition of complete visual field loss Since luminance and sensitivity to light are inversely related, complete visual ield loss would be a very high luminance number. This number would be limited by the maximum stimulus the peri- meter is able to display, potentially resulting in large differences between different perimeter models. THE DEFINITION OF SENSITIVITY TO LIGHT USING THE DECIBEL SCALE The decibel scale addresses all of these issues and uses luminance levels solely as input variables. The relationship between the decibel scale and the luminance scale in apostilbs is shown below. RELATIONSHIP BETWEEN SENSITIVITY TO LIGHT AND LUMINANCE SENSITIVITY TO LIGHT STIMULUS LUMINANCE (SENSITIVITY THRESHOLD) Decibels (dB) Apostilb (asb) 40 0.4 Foveal normal ~32.8 dB sensitivity for 30 4.0 20-year-old person 20 40 10 400 0 4,000 The decibel scale is used to express sensitivity to light. This ϔigure shows the relationship between sensitivity to light and luminance. The maximum stimulus brightness, which is used as a default in recent Octopus perimeter models, is 4,000 asb. It is a logarithmic scale and is inversely related to the linear luminance scale in apostilbs (asb). Note that the maximum stimulus brightness might be different in different perimeter models. The sensitivity to light in decibels is de ined using the formula below dB = 10 * log (Lmax/L) where dB is the sensitivity threshold, Lmax is the maximum luminance the perimeter can display, and L is the luminance of the stimulus at the threshold (both expressed in apostilbs). The logarithmic scale is used to address the large range of luminance values and to relate this range more linearly to visual function. To address the inverse relationship between luminance and sensitivity to light, the inverse of luminance (1/L) is used in the formula; and to make sure that near complete visual ield loss equals 0 dB, which is intuitive, the maximum stimulus luminance Lmax is added to the equation. Since 0 dB refers to the maximum intensity that the perimeter can produce, its interpretation in terms of stimulus luminance may be different for various visual ield devices. This should be kept in mind when switching between different perimeter models. Chapter 12 will focus on how to deal with differences between perimeters in clinical practice. 16 Chapter 2 | What is perimetry? GRAPHIC DISPLAY OF SENSITIVITY THRESHOLDS The three-dimensional hill of vision contains large function from the three-dimensional representation. amounts of information. It may therefore be challenging Cartographers face similar challenges when displaying to appropriately display all aspects of a patient’s visual three-dimensional mountains or hills, and have used GRAPHIC DISPLAY OF SENSITIVITY THRESHOLDS MOUNTAIN – Geographical display HILL OF VISION – Perimetric display OCTOPUS REPRESENTATIONS 3D map 3D map No 3D map available on Octopus perimeters Numerical altitude map Numerical sensitivity threshold map Values 27 0m 600m 10 dB 28 28 28 0m 10 dB 1200m 30 1800m 20 dB 31 30 0m 10 dB 1200m 2400m 26 29 600m 1200m 3000m 1800m 20 dB 31 30 27 27 2400m 30 dB 20 dB 33 32 3000m 2400m 3600m 600m 30 dB 28 31 31 31 1800m 1800m 3000m 2400m 10 dB 28 31 31 30 27 25 1800m 10 dB 0m 2400m 1800m 20 dB 27 21 29 33 29 600m 1200m 1200m 29 28 30 28 10 dB 31 29 0m 600m 0m 10 dB 26 29 27 25 27 26 29 26 Color altitude map Color sensitivity threshold map Grayscale of Values Altitude lines map Sensitivity threshold lines map Kinetic Perimetry 10 30 40 50 60 70 80 90 FIGURE 2-7 As in cartography, there are different ways to display the three-dimensional hill of vision in two dimensions. Sampled altitude levels can be displayed numerically, a color code can be used to represent different altitude levels, or altitude lines can show the different altitude levels. Challenges in visual field testing and interpretation 17 two-dimensional maps as a solution. Similar display provide a good representation of a hill on a map. For strategies are used to display the hill of vision in two perimetry, these lines of equal altitude are referred to as dimensions. isopters (lines of equal sensitivity). As in geographical maps (FIG 2-7), the various sensitivity It should be noted that whichever display form is used, thresholds can be displayed numerically (i.e., by sam- there is always some information lost. All three versions pling certain altitudes to give a feel for the overall shape are used to display perimetric results, as each emphasizes of the hill or mountain). Color codes for different altitude different clinical information. For more details of the levels are also often presented on geographical maps. various representations, see Chapters 7, 8, and 11. Last but not least, lines of the same altitude level can CHALLENGES IN VISUAL FIELD TESTING AND INTERPRETATION PERIMETRIC TESTING HAS LOW RESOLUTION So far, this book has presented perimetry as a very accu- From a practical point of view, however, it is nearly rate way of continuously showing the stimuli of increasing impossible to test each location within the visual ield intensity for the patient. It has also been assumed that (spatial resolution) using each possible light intensity thresholding is performed at all locations across the (luminance resolution). This would take too long to be visual ield. useful in a clinical setting. Therefore, referring back to IDEAL VERSUS PRACTICAL PERIMETRIC TESTING SPATIAL RESOLUTION RESOLUTION OF SENSITIVITY THRESHOLDS Ideal Practical Ideal Practical 90 90 180 0 180 0 270 270 FIGURE 2-8 Ideally, the hill of vision would be drawn from an infinite number of test locations and from a continuously changing stimulus luminance. In reality, the time constraints do not allow for this kind of testing, and only sampling at some locations and some luminance levels is possible. 18 Chapter 2 | What is perimetry? the example of the light bulb in a room, the dimmer only ber of light intensity levels are presented. This approach has a set number of discrete levels, such as high, medi- introduces inaccuracies in the perimetric test. In order to um and low, and there are only a few bulbs to illuminate still be able to receive the information necessary for good the room (FIG 2-8). clinical decision-making, a number of elaborate process- es are used in perimetry. This maximizes clinical infor- For perimetry, this means that stimuli are presented at a mation and offers a good trade-off between testing time ixed number of key locations and that only a limited num- and accuracy. These are described in Chapters 4, 5 and 6. NORMAL SENSITIVITIES DEPEND ON AGE AND TEST LOCATION As already illustrated in the section about the hill of For these reasons, sensitivity thresholds are challenging vision, normal sensitivity thresholds depend on the to interpret directly in the clinic, because the representations test location and are higher at the center than in the pe- of normal and abnormal values depend on testing- and riphery. In addition, the normal hill of vision is affected patient-speci ic factors. For correct clinical assessment by age. Normal sensitivity to light in decibels decreases of sensitivity thresholds, a clinician would have to keep approximately linearly with increasing age, beginning at normal reference values in mind for all age groups and the age of 20.1-3 Thus, the hill of vision of a 20-year-old test locations, in order to correctly interpret the results. is typically higher than the hill of vision of an 85-year-old That would be a challenging task. person (FIG 2-9). HILL OF VISION IS AGE- AND LOCATION-DEPENDENT Sensitivity threshold 20-year-old 85-year-old FIGURE 2-9 The normal hill of vision shows the highest sensitivity thresholds at the center, with decreasing sensitivity thresh- olds towards the periphery. Similarly, there is also a decrease in sensitivity thresholds with increasing age at all test locations. Challenges in visual field testing and interpretation 19 Therefore, distinct normative databases have been devel- the respective normative value for someone of that age. oped for most modern perimeters and these databases The calculated comparison to normal is clinically mean- are used to facilitate clinical visual ield interpretation. ingful, as it relates directly to sensitivity loss (FIG 2-10). Normative databases contain normal reference values Alternative expressions that are commonly used for com- for each age group and test location (BOX 2B). They are parisons to normal are deviation from normal or defect. used to compare any measured sensitivity threshold to COMPARISONS SHOW THE DEVIATION FROM NORMAL NORMATIVE VALUES (MEASURED) VALUES COMPARISON (TO NORMAL) Normal sensitivity threshold - Measured sensitivity threshold = Sensitivity loss Sensitivity threshold Comparison Normative Values of 20-year-olds Measured Values of a 20-year-old FIGURE 2-10 The difference between a normal and a measured visual field point is commonly called ‘comparison to normal’ (also referred to as deviation from normal or defect) and its interpretation is independent of a patient’s age or the visual field location. Due to their ease of use, most representations in the and not on the measured sensitivity thresholds. For more Octopus perimeters are based on comparisons to normal information, refer to Chapter 7. 20 Chapter 2 | What is perimetry? BOX 2B NORMATIVE DATABASES IN OCTOPUS PERIMETERS DESIGN OF A NORMATIVE DATABASE By de inition, the normative database of a perimeter consists of a pool of visual ield data from people with normal vision in all age groups. The challenge associated with generating this pool is to ensure that these normal visual ields are truly normal and that there are suf icient visual ields to account for individual differences. The standards to be ful illed for a perimetric normative database are described exhaustively in the ISO norm «Ophthalmic instruments - Perimeters (ISO 12866:1999/Amd1:2008); Amendment A1, Appendix C». All normative databases of Octopus perimeters comply fully with these standards. The typical process to comply with the standards is to perform a clinical study that includes a thorough eye examination and repeated visual ield testing. DISTINCT NORMATIVE DATABASES FOR DIFFERENT DEVICES AND EXAMINATION PARAMETERS Since particular perimeter models vary in design and might use different examination parameters, a stimulus of the same light intensity may be perceived differently on various perimeter models. Therefore, there are distinct normative databases for different perimeter types and settings. PERIMETRY HAS OBJECTIVE AND SUBJECTIVE COMPONENTS In the interest of simplicity, perimetry has been treated as patient does not understand the test, does not pay atten- a purely objective procedure, with exact measurements tion or does not focus continuously on the central target, and distinct sensitivity thresholds at each test location. then the results of the test will be dif icult to interpret. This is true for the equipment and the test conditions. Additionally, some patients may be very conservative in However, there is a subjective element to perimetry, due their judgements, requiring a more intense stimulus for to the subjectivity of the patients undergoing the test. detection, while other patients may be liberal and accept As a result, there is always a certain amount of normal a less intense stimulus for detection. The most important luctuation both among different normal individuals, as person to maximize the performance of the patients is the well as between different measurements of the same visual ield examiner (e.g., a perimetrist or technician). individual over a short period of time. The accuracy of Chapter 3 focuses on potential sources of unreliable and the test results is highly dependent on several factors, thereby highly luctuating visual ields and provides prac- including the cooperation of the patients, their cognitive tical guidance on how to minimize these factors. and physical abilities, and their decision criteria.4-6 If the NORMAL FLUCTUATION DEPENDS ON TEST LOCATIONS AND DISEASE SEVERITY A further complication in visual ield interpretation is the tion is smaller at the center of the visual ield than in the fact that normal luctuation is not uniformly distributed periphery and is also smaller in areas of good vision than across the visual ield (FIG 2-11). Instead, normal luctua- in areas of poor vision.1,7 Challenges in visual field testing and interpretation 21 NORMAL FLUCTUATION IN PERIMETRY Sensitivity threshold Average Hill of Vision Abnormal Normal fluctuation FIGURE 2-11 Since perimetry contains a subjective, patient-related component, there is always normal fluctuation. Its magnitude depends on both the test location and disease severity. These two factors must be kept in mind when making measure luctuation around a sensitivity threshold, the clinical decisions based on visual ield results. To objectively frequency-of-seeing (FOS) curve may be used (BOX 2C). THE FREQUENCY-OF-SEEING (FOS) CURVE BOX 2C Due to luctuation, distinct sensitivity thresholds at a given test location cannot be measured precisely. In reality, the same patient always shows slightly varying responses in repetitive testing. In other words, the likelihood of seeing or not seeing a stimulus is probabilistic. As the luminance (i.e., the light intensity of the stimulus) increases, there is a gradual increase from “unseen” to “seen” responses, so that the probability that a patient will perceive a stimulus changes gradually from 0% to 100%. Because of this, sensitivity thresholds are de ined as the stimulus luminance that is perceived with a probability of 50%. To get a measure of luctuation, one can show a stimulus of a certain luminance to a patient many times at a given test location and determine how often the patient is able to see it. The probability of perceiving a stimulus can be mapped in a graph as a function of stimulus luminance. When doing this for many dif- ferent luminance levels, one can generate a frequency-of-seeing (FOS) curve, which describes the prob- ability that a patient will perceive a target as a function of stimulus luminance. This is a useful tool to illustrate the variability associated with the determination of thresholds.⁸ In areas of normal sensitivity, the FOS curve is typically steep, indicating that there is less variability. In other words, the patient has a high probability of seeing stimuli that are slightly more intense than the luminance at the threshold, and also a high probability of not seeing stimuli that are slightly less intense than those at the threshold. This is illustrated on the left side of the igure by the steep shape of the FOS curve. In areas where defects are present, the FOS curve is typically shallow, indicating that there is greater variability. In other words, there is a gradual change in the probability of detecting stimuli that are higher and lower than the luminance at threshold. This is illustrated on the right side of the igure by the shallow shape of the FOS curve. 22 Chapter 2 | What is perimetry? FREQUENCY-OF-SEEING (FOS) CURVE Fluctuation Fluctuation 100% Probability of seeing the stimulus Normal Abnormal sensitivity sensitivity threshold threshold 50% 0% Dim stimulus Bright stimulus The frequency-of-seeing curve provides the scientiϔic deϔinition of a light sensitivity threshold while taking ϔluctuation into account. It shows the probability of a patient perceiving a certain stimulus luminance. The light sensitivity threshold is deϔined as the stimulus luminance that the patient can see 50% of the time. Fluctuation is quantiϔied as the range of luminance at which the probability of seeing the stimulus is 0% to the luminance at which the probability of seeing the stimulus is 100%. CLINICAL STANDARD FOR VISUAL FUNCTION TESTING Even though perimetry has low resolution and contains daily living, which are the most important factors for the subjective, patient-related components resulting in normal patient. Additionally, slowly progressing diseases such as luctuation, perimetric testing is useful to assess visual glaucoma can be followed accurately through all stages ields in clinical practice. It remains highly important be- of the disease. Perimetry is therefore an indispensable cause visual ield function is most directly related to a tool for every glaucoma specialist. patient’s quality of life and ability to perform activities of References 23 REFERENCES 1. Zulauf M. Normal visual ields measured with Octopus Program G1. I. Differential light sensitivity at individual test locations. Graefes Arch Clin Exp Ophthalmol. 1994;232:509-515. 2. Zulauf M, LeBlanc RP, Flammer J. Normal visual ields measured with Octopus-Program G1. II. Global visual ield indices. Graefes Arch Clin Exp Ophthalmol. 1994;232:516-522. 3. Haas A, Flammer J, Schneider U. In luence of age on the visual ields of normal subjects. Am J Ophthalmol. 1986;101: 199-203. 4. Flammer J, Drance SM, Fankhauser F, Augustiny L. Differential light threshold in automated static perimetry. Factors in luencing short-term luctuation. Arch Ophthalmol. 1984;102:876-879. 5. Flammer J, Niesel P. Reproducibility of perimetric study results. Klin Monbl Augenheilkd. 1984;184:374-376. 6. Stewart WC, Hunt HH. Threshold variation in automated perimetry. Surv Ophthalmol. 1993;37:353-361. 7. Wall M, Woodward KR, Doyle CK, Artes PH. Repeatability of automated perimetry: a comparison between standard automated perimetry with stimulus size III and V, matrix, and motion perimetry. Invest Ophthalmol Vis Sci. 2009;50: 974-979. 8. Chauhan BC, Tompkins JD, LeBlanc RP, McCormick TA. Characteristics of frequency-of-seeing curves in normal subjects, patients with suspected glaucoma, and patients with glaucoma. Invest Ophthalmol Vis Sci. 1993;34:3534-3540. 24 25 CHAPTER 3 HOW TO PERFORM PERIMETRY YOU CAN TRUST INTRODUCTION PERIMETRY – A SUBJECTIVE TEST Perimetry is an elaborate test that depends, to a great In view of the relatively high occurrence of untrust- extent, on subjective factors such as the patient’s cooper- worthy visual ields, it is extremely important to make ation and comfort, as well as on using the correct patient sure that the time invested in perimetry is well spent, information and set-up. Due to this subjective compo- because poorly performed perimetric tests have hardly nent, untrustworthy visual ield tests are common. The any diagnostic value. It therefore pays to take the time extent of untrustworthy results largely depends on how and care necessary to obtain trustworthy results by fol- well perimetry is performed in clinical practice and has lowing certain rules to avoid the most common pitfalls. been reported to range from 3% to 29% of all visual ield tests performed.¹-⁵ 26 Chapter 3 | How to perform perimetry you can trust PERIMETRY – NEED FOR A TEAM APPROACH Three key players are involved in perimetry: the patient, FIG 3-1 shows how each member of the team can contrib- the examiner and the eye doctor. All three should work ute. When this approach is successfully implemented, collaboratively to obtain optimal perimetric test results. perimetry can be performed in a positive atmosphere. PERIMETRY REQUIRES A TEAM APPROACH PATIENT st te e Ex th ure pr of ed As eed es oc e pr t nc sn kq tes Pr ng the Pr o rta ue o vi e en i po st stio break t rm ou etry de P te vi for th ro e im importan estio rfo b im d cle the ns er r of a a Ask qu o pe vid gho ar ce ns ou e th p with e s ut th cou inst dt Emphasiz upe e test rage Be motivate Collaborate ruction rvision ment s EX AM OR CT Pr ov IN ER DO ide ing adequate train All Pr ow to ov ne t i me ide focus cessary etry on perim on rec ogni iati tion and apprec Perf ll orm perimetric test we Ma Accuracy in entering data ance ke n otes ab rform out patient pe FIGURE 3-1 In perimetry, it is essential that doctors, examiners and patients have a positive attitude towards perimetry and that each member of the team contributes to achieving optimum results. THE IMPORTANT ROLE OF THE DOCTOR THE DOCTOR-PATIENT RELATIONSHIP Patients who understand why perimetry is needed and and trust they establish with their patients, doctors are its importance to their eye care are likely to be more moti- in the best position to convey the importance of perimetry vated to undergo a perimetric test. Due to the relationship to their patients. Introduction 27 THE DOCTOR-EXAMINER RELATIONSHIP Eye doctors should also clearly convey the importance of this goal, the doctor must provide training and give feed- perimetry to the visual ield examiners who work with back to the examiners. It is also crucial for the doctor to them in the clinic. For example, the doctor is responsible have reasonable expectations in terms of the time required for ensuring that the visual ield examiners understand to perform trustworthy perimetric tests. Doctors should the importance of trustworthy perimetric results to the arrange for their visual ield examiners to be able to dedi- clinical decision-making process. The visual ield examin- cate time exclusively to performing perimetric tests. This ers should know that the doctor has a genuine interest in means that they should be free of other tasks that might building their perimetric knowledge and skills. Towards reduce the examiner’s focus on the patient. THE IMPORTANT ROLE OF THE VISUAL FIELD EXAMINER The visual ield examiner is in a unique position to have setting up the perimeter, they also directly oversee the an impact on the quality of the perimetric results in two patient during the test. ways. Not only are examiners responsible for correctly ROLE IN CORRECTLY SETTING UP THE PERIMETER The visual ield examiner is responsible for entering in performing this aspect of perimetry can significantly the correct patient information in the perimeter. This is reduce the number of untrustworthy tests and inter- crucial because this information has a direct impact on pretation errors. The examiner is also responsible for whether the results of the test can be trusted. Diligence ensuring that an adequate refractive lens is used. THE EXAMINER-PATIENT RELATIONSHIP A crucial role of the visual ield examiner is to ensure patient. Additionally, the patient should be encouraged that the patients perform perimetry to the very best of to communicate to the examiner any dif iculties or prob- their capacity each time they take a test. To give their best lems encountered, and when a brief rest period would be performance, patients need to be comfortably positioned bene icial. at the perimeter, they need to know what is expected of them, and they need to understand how to perform the There is more, however, to the role of a visual ield exam- test. A competent examiner will ensure that the patient iner. Outstanding examiners will have taken perimetric is not only correctly positioned, but also comfortable. tests themselves and will understand how the patient Similarly, a good examiner will convey what is expected feels during the test. This compassionate approach will of the patient and will give clear instructions on how to go a long way in ensuring patient cooperation and will perform the test. The examiner can also provide brief rest allow the examiner to give genuine encouragement to the periods by pausing the test if this will be helpful to the patient when needed during the test. 28 Chapter 3 | How to perform perimetry you can trust HOW TO PERFORM VISUAL FIELD TESTING SETTING UP THE PERIMETER Perimetry should be performed in a distraction-free room, opaque curtains around the perimeter and earmuffs environment, to enable the patient to concentrate on offer a cost-effective alternative. the perimetric test (FIG 3-2). The room should be quiet, with no activity distracting the patient, and should be at The perimeter is automatically calibrated each time it is a comfortable room temperature. The cupola should be turned on. It is important for the calibration to take place kept clean and free of dust and particles. Additionally, the in the same lighting conditions as those used during peri- room should be dimly lit, to prevent stray light from in- metric testing. Calibration can take up to two minutes and luencing the perimetric result. A dimly-lit environment should be performed prior to testing patients. Thus, the is essential when a cupola perimeter, such as the Octopus perimeter should be turned on prior to the patient visit. 900 is used, but is also helpful for non-cupola perimeters. Ideally, patient data (date of birth, refraction, etc.) are Ideally, perimetry should be performed in a room dedi- entered before the patient enters the room. If an electronic cated solely to this purpose. However, if the layout of the medical record system is in use, it will automatically pop- clinical practice does not offer a stand-alone perimetry ulate the information to the perimeter. PERIMETER SET UP Light Heat Noise Cold FIGURE 3-2 A perimeter should be set up in a distraction-free, dimly-lit environment. How to perform visual field testing 29 PLACING AN ADEQUATE TRIAL LENS The trial lens calculator is helpful in determining the than one trial lens is used, the spherical correction should adequate spherical and cylindrical trial lenses, based on be placed closest to the patient’s eye. Special attention the patient’s current refraction and age. It is vital to en- should be given to the orientation of cylindrical lenses, sure that the patient’s refractive data is up-to-date and it which should be oriented in the angle of the astigmatism is best practice to determine this prior to each test. The (FIG 3-3). correct trial lens should be put into the trial lens holder prior to seating the patient. Trial lenses with a narrow To con irm that adequate refraction is used, the examiner metal rim should be used, to prevent the rim of the should position the patient and ask whether the ixation trial lens from blocking the patient’s ield of view. If more target is clearly visible. PLACEMENT OF CYLINDRICAL TRIAL LENSES 30° 180° 0° 150° 30° 120° 60° 90° FIGURE 3-3 Placement of a cylindrical lens for a patient with a cylindrical correction of 30°, seen from the examiner’s perspective. INSTRUCTING THE PATIENT Due to the subjective components involved in perimetry, 3-4). It can be helpful for examiners to take a perimetry careful patient instruction is fundamental to achieving test themselves, in order to gain a better understanding trustworthy results. Patients will be able to cooperate of what patients are experiencing. more effectively and produce more consistent results if they understand what is expected of them and why the It is fundamental to ensure that the patients know that test is being performed. they are not expected to see all stimuli and that some- times no stimuli are presented. This will help to reduce The visual ield examiner should therefore take the time some of the potential anxiety experienced by patients, to explain the aim of the test, what the patient should ex- who should also know that they can pause the test if they pect to see, and what the patient is expected to do (FIG experience fatigue or have questions. 30 Chapter 3 | How to perform perimetry you can trust STEP-BY-STEP PATIENT INSTRUCTIONS 1. Perimetry tests your central and peripheral vision. 2. Be relatively still once positioned. 3. Always look straight ahead at the fixation target. Do not look around the bowl for stimuli. 4. Press the response button whenever you see the stimulus. a. The stimulus is a flash of light. b. Only one stimulus is presented at a time. c. The stimulus might appear anywhere. d. Some stimuli are very bright, some are very dim, and sometimes no stimulus is presented. e. You are not expected to see all stimuli. f. Do not worry about making mistakes. 5. Blink regularly to avoid discomfort. a. Don’t worry about missing a point, the device does not measure while you blink. 6. If you feel uncomfortable or are getting tired a. Close your eye for a moment, the test will automatically stop. b. The test will resume once you open your eye. 7. If you have a question a. Keep the response button pressed; this will pause the test. FIGURE 3-4 Proper instructions to the patient are essential for the patient to understand their task and consequently to perform perimetry well. The sequence of instructions listed in this Figure can be used. How to perform visual field testing 31 SETTING UP AND POSITIONING THE PATIENT Trustworthy and accurate perimetric results are more tient is correctly positioned and that the non-tested eye likely to be obtained when the patient is comfortable is covered. The optimum ways to ensure patient comfort during the test. It is also important to ensure that the pa- and correct alignment will be discussed in this section. CORRECT EYE PATCH POSITION Before fully positioning the patient, the eye not being test- for the tested eye. If an adhesive eye patch is used, it is ed should be covered with an eye patch that allows the important to make sure that it adheres well all around patient to blink freely (FIG 3-5). If the eye patch is main- the eye. All eye patches should be translucent, to avoid tained in place with a cord, it is important to ensure that adaptation to the dark by the untested eye, which would the cord does not obstruct the patient’s ield of view alter the results of subsequent testing of that eye.⁶ EYE-PATCH POSITION CORRECT INCORRECT Unobstructed view of test eye Cord obstructs view of test eye FIGURE 3-5 An eye patch should cover the eye that is not being tested. It should be positioned so as to not obstruct the patient’s vision in the tested eye. CORRECT PATIENT POSITION The patient should be seated in a comfortable position patient is comfortable. Different Octopus models offer that can be easily maintained throughout the test. A different types of positioning: the Octopus 900 offers a height-adjustable chair with a backrest and, if available, straight-upright patient position and the Octopus 600 armrests should therefore be used. The perimeter should offers a forward-leaning position. be placed on a height-adjustable table to ensure that the 32 Chapter 3 | How to perform perimetry you can trust STRAIGHT-UPRIGHT POSITION FIGURE 3-6 This drawing illustrates the correct straight-upright patient position recommended for the Octopus 900 and older Octopus models. For the Octopus 900 and all older Octopus models, the pa- his or her chin on the chinrest and forehead on the head- tient should sit as close as possible to the device. Then the rest (FIG 3-6). It is important to ensure that the patient height of the table should be adjusted until the patient’s maintains direct contact with the device throughout forehead touches the headrest. The patient should place testing. FORWARD-LEANING POSITION 1. PREPARATION 2. FINAL POSITION Headrest 20cm/8 inches FIGURE 3-7 This drawing illustrates the correct forward-leaning patient position recommended for the Octopus 600. For the Octopus 600, the patient is positioned in a for- headrest, to allow enough space to lean forward. By ward-leaning and downward-gazing position (FIG 3-7). inclining from this position, the patient is automatically The correct position is obtained by irst seating the pa- positioned at the correct height. The patient’s head leans tient in an upright position at a distance of approximately in fully onto the headrest, providing stable ixation. 20cm/8 inches, with the eyes at the upper level of the How to perform visual field testing 33 CORRECT EYE POSITION Once the patient is correctly positioned in the device, it lens. However, the lens should not touch the eyelashes, is important to ensure that the eye is also correctly po- allowing the patient to blink freely and avoiding the lens sitioned. Overall, the eye should be well-aligned with the being smeared with make-up. ixation target and should be relatively close to the trial CORRECT PUPIL POSITION CORRECT INCORRECT Central pupil position Off-center pupil position FIGURE 3-8 The left-hand panel shows an eye in the video monitor that is correctly positioned, with the cross-hair target locat- ed within the boundaries of the pupil. The right-hand panel shows an eye that is incorrectly positioned, with the cross-hair target located outside the boundaries of the pupil. The Octopus perimeters provide a video monitor so that monitor. The patient is correctly positioned when the the examiner can see the patient’s eye. When the patient cross-hair target is within the boundaries of the pupil looks straight at the ixation target, the pupil should be (FIG 3-8). The position of the pupil can be adjusted by aligned with the cross-hair target provided on the video changing the position of the chinrest. CORRECT TRIAL LENS POSITION CORRECT INCORRECT CORRECT INCORRECT Trial lens close to eye Trial lens too far away Central pupil position Off-center pupil position FIGURE 3-9 The patient’s eye should be positioned in the center of the trial lens and as close as possible without touching it. It is important for the patient’s eye to be as close as pos- patient is positioned too far away from the trial lens (FIG sible to the trial lens, in order to avoid the typical “ring” 3-9). The eyelashes should not touch the lens, however. defect (i.e., trial lens rim artifact) that occurs when the 34 Chapter 3 | How to perform perimetry you can trust When a visual ield test assesses both the central and the in order to avoid trial lens rim artifacts. Also, visual itness peripheral visual ields, it will be necessary to remove the to drive is assessed binocularly (both eyes open). In this trial lens for the part of the test that covers the periphery, case, no trial lens should be used. CORRECT FIXATION It is essential for patients to maintain steady ixation target is not recommended for the G, M, N and D patterns throughout the test. The Octopus perimeters offer three (see Chapter 5) and for any pattern where the foveal different ixation targets (FIG 3-10) to promote steady ix- threshold function is turned on. ation in as many patients as possible. Most patients will be able to maintain ixation using the standard cross Finally, some patients with severe visual ield loss in the mark ixation target. If patients have dif iculty under- macula region may not be able to see the standard cross standing where to look when the cross mark ixation mark ixation target. In these patients, the use of the larger target is used, the central point ixation target can be ring target is recommended, to provide an estimate of the used, provided that the test pattern does not test the location of the ixation target. central point. For this reason, the central point ixation FIXATION TARGETS CROSS MARK CENTRAL POINT RING Standard Alternative For patients with severe (do not use on test patterns visual field loss in the with central locations) macula 10° 10° FIGURE 3-10 Octopus perimeters offer 3 different fixation targets. The cross mark target is the default target. The central point target can be used in test patterns that do not test the central point. The ring target is recommended for patients with fixation issues due to severe visual field loss in the macula. How to perform visual field testing 35 MONITORING THE PATIENT DURING THE EXAMINATION To ensure good patient cooperation and trustworthy sponse even if there is no stimulus, or unsteady ixation), results, it is essential to monitor patients throughout the the test should be interrupted and the patient should be examination and not leave them unattended and unmon- reinstructed. If the results seem compromised, it is rec- itored. During the test, it is helpful to encourage patients ommended to start a new test and discard the compro- by telling them that they are doing well and by letting mised one. It is important to note, however, that patients them know how much of the test they have already com- with impaired vision often do not respond due to their pleted. This will help them to remain attentive and may re- condition and not because they answer unreliably. duce anxieties that might negatively in luence the results. If a patient shows inconsistent behavior, the examiner Particular attention should be paid during the irst min- should make a note of this on the examination ile, to com- ute of the test, to ensure that patients have understood municate this information to the clinician. The knowledge what they are expected to do during the test. If a patient that the test has reduced reliability may in luence the shows an unusual response (e.g., no response at all, a re- interpretation of the test. USE OF FIXATION CONTROL Loss of ixation is a primary reason for unreliable visual Control mechanisms active. However, since some patients ield results. Therefore, all current Octopus devices have might not be able to maintain steady ixation for patholog- a built-in Fixation Control for static testing that can track ical reasons (i.e., reduced central vision, unsteady pupil or the patient’s pupil at all times and prevent ixation errors. nystagmus), individual mechanisms within Fixation Con- With Fixation Control, the test is stopped automatically if trol can be turned off individually, to make patient testing the patient loses ixation (due to blinking, searching for possible. If it is necessary to turn off some mechanisms, stimuli or head movements) and automatically restarted careful patient monitoring is key and it is good practice to once proper ixation is regained. Missed stimuli are au- make a note in the patient ile about the patient’s ability to tomatically repeated later during the test. If ixation loss maintain ixation. The clinician should then interpret the occurs for more than just a few seconds, a warning results in the light of this information and should consider message will alert the examiner to properly reposition that the test might have reduced reliability. and reinstruct the patient. FIG 3-11 provides more information about the different Fixation Control consists of several separate control control mechanisms of Octopus Fixation Control. Note mechanisms, as outlined in FIG 3-11, which can be turned that the con iguration depends on the Octopus model. on and off. It is recommended to keep each of the Fixation 36 Chapter 3 | How to perform perimetry you can trust FIXATION CONTROL PREVENTS FIXATION LOSSES BLINK CONTROL Prevents fixation loss due to blinking. RUNNING PAUSED • Detects eye closure due to blinking or falling asleep • Testing occurs only if the patient’s eye is open • Allows the patient to blink normally • Prevents dry eyes • Increases patient comfort • Ensures that no stimuli are missed due to blinking CONTACT CONTROL Prevents loss of contact with the perimeter. RUNNING PAUSED • Detects contact with the headrest or chinrest • Testing occurs only if the head is in contact with the device • Ensures that the head remains close enough to the device to minimize lens rim artifact PUPIL POSITION CONTROL Prevents fixation losses due to incorrect pupil position. RUNNING PAUSED • Detects off-centered pupils due to incorrect fixation or head movement • Testing occurs only if the pupil is correctly centered • Ensures correct gaze direction DART CONTROL Prevents fixation loss due to rapid eye movement. RUNNING PAUSED • Detects rapid eye movement when the patient is searching for stimuli • Testing occurs only if the pupil is steadily fixating • Ensures correct gaze direction AUTOMATED EYE TRACKING (AET) Automatically adjusts the patient’s eye position. RUNNING ADJUSTING POSITION • Moves the headrest and chinrest to keep the eye in the center of the trial lens • Maintains optimum position even if the patient is moving around slightly • Reduces trial lens rim artifacts due to off-centered eye position FIGURE 3-11 Fixation control prevents fixation losses by automatically pausing the test during blinks, loss of contact with the device, off-centered pupils and rapid eye movements. The test is automatically restarted once optimum conditions are achieved. Further, Automated Eye Tracking automatically centers the pupil. Note that not all mechanisms are available on the different Octopus perimeter models. Common pitfalls to avoid 37 COMMON PITFALLS TO AVOID There are many factors that can lead to visual ield tests in the set-up procedure, and external obstructions block- that cannot be trusted. By paying attention to and man- ing the stimuli from reaching the retina, are all commonly aging these factors, a well-trained examiner will have a occurring sources of untrustworthy visual ield results. substantial positive in luence on the quality of the visual Many of these pitfalls can be avoided by paying close at- ield results and on the subsequent clinical decisions. tention to the set-up procedure, by observing the patient Therefore, this section is dedicated to the most common carefully during testing, and by making adjustments or pitfalls in perimetry and provides guidance on how to repeating instructions if necessary, which is the focus of avoid them. this section. Chapters 7 and 8 provide information on how to detect visual ield results that cannot be trusted Patient behavior (i.e., lack of patient cooperation), errors after the test is completed. INCONSISTENT PATIENT BEHAVIOR LEARNING OR PRACTICE EFFECT When taking their irst tests, patients often do not fully visual function more closely. understand the nature of the test and hesitate to press the button when seeing faint stimuli near the sensitivity While learning and practice effects most often occur threshold. This translates into visual ield results that are for patients taking their irst visual ield examination, worse than the patient’s true visual ield, as illustrated in they can also occur when switching from one perimeter FIG 3-12. In subsequent testing, the patients then perform to another, due to small differences in the design (see better and their visual ield results resemble their true Chapter 12). EXAMPLE OF A LEARNING EFFECT LEARNING EFFECT NO LEARNING EFFECT 1st Test 2nd Test 3rd Test 4th Test 5th Test FIGURE 3-12 Example of a patient with normal vision with a strong learning or practice effect from the first to third visual field tests. The fourth and fifth tests represent the true visual field of the patient. 38 Chapter 3 | How to perform perimetry you can trust While learning or practice effects cannot always be also helpful. If a patient does not understand the task of prevented, their frequency can be reduced by careful pa- performing perimetry, the patient will often be hesitant tient instruction and observation. Running a practice test during the irst part of the test, or will not press the re- prior to real testing is a good procedure if time allows. sponse button at all. If this is observed, it is recommended Careful observation during the irst minute of the test is to interrupt the test and reinstruct the patient. FATIGUE EFFECT Visual ield tests require alertness and attention. When vised to blink regularly to avoid dry eyes and discomfort, patients become tired, their attention level may decrease given that Fixation Control is active. Arti icial tear drops and their answers may become less consistent, resulting prior to the test may also reduce fatigue effects due to in a visual ield that is worse than the patient’s true visual dry eyes. Additionally, patients should be encouraged to ield (FIG 3-13).⁷-¹¹ take brief rests, by closing their eyes to relax, if they feel that they are getting tired. Usually, this adds only a few To reduce fatigue effects for patients who have dif iculty seconds to the test duration, but signi icantly improves concentrating for long periods of time, it may be appro- the reliability of the results. Furthermore, using a beep- priate to use tests that are shorter in duration, despite ing sound upon each stimulus presentation may help the the associated loss of accuracy. This may generate more patients to concentrate better on the test. BOX 3A pro- meaningful visual ield results by reducing the unreliabil- vides more information about the advantages and disad- ity due to the fatigue effect. Individual differences exist in vantages of this option. how quickly patients experience fatigue, and this should be considered when selecting a test. Sometimes fatigue is noticeable as drooping eyelids. In such cases, it is best to actively interrupt the test for a while To further reduce fatigue effects, patients should be ad- and to allow the patient to rest before continuing testing. INFLUENCE OF FATIGUE EFFECT ON VISUAL FIELD NO FATIGUE EFFECT FATIGUE EFFECT 9:30 am 9:45 am FIGURE 3-13 Example of a patient tested on the same day and eye within 15 minutes. Note the significant worsening of the visual field in the second test, due to fatigue. Common pitfalls to avoid 39 LOSS OF FIXATION If a patient does not consistently ixate on the central tar- The Octopus Fixation Control should be enabled when- get, the test will lose its reference point and it will not be ever possible, to avoid unreliable visual ields due to ixa- possible to identify the location of abnormal visual ield tion losses. It should only be turned to a lower setting or points (FIG 3-14). This is called ixation loss and is one of completely turned off if a patient is not able to maintain the most common sources of unreliable ields.¹² It occurs steady ixation, for pathological reasons (i.e., reduced especially if the patient is insecure about his or her per- central vision, unsteady pupil or nystagmus). Direct ob- formance and starts looking around, searching for stimuli. servation of the patient’s ixation behavior early in the To avoid ixation losses, it is therefore crucial to explain test can also be helpful in this regard. carefully to the patient that it is perfectly normal not to be able to see all of the stimuli. INFLUENCE OF LOSS OF FIXATION ON VISUAL FIELD CORRECT FIXATION LOSS OF FIXATION Real defect is detected Real defect is missed and/or artifactual defect is identified Do you see? Do you see? Patient Fixation Fixation Target Loss of Fixation Fixation & Patient Target Fixation FIGURE 3-14 If there is a loss of fixation, visual field defects will not be in their exact location, but will either be shifted together with the fixation or masked. In the above example, loss of fixation took place during the entire test. In practice, loss of fixation is typically brief, resulting in more random defect patterns. ADVANTAGES AND DISADVANTAGES OF USING SOUNDS UPON STIMULUS BOX 3A PRESENTATION A beeping sound upon stimulus presentation may be helpful for some patients, to maintain their atten- tion during the perimetric test, because it provides them with a steady rhythm to follow. Additionally, it provides reassurance to the patient that the test is running and everything is working normally. However, the beeping sound may also encourage patients to press the response button even though they cannot see a stimulus. This may increase false answers, resulting in unreliable visual ields. In addition, if more than one perimeter is in a room, the beeping sound of neighboring machines may be distracting. By default, it is thus recommended to turn the beeping sound off and to only use it for selected patients that have dif iculties with maintaining concentration throughout the test. 40 Chapter 3 | How to perform perimetry you can trust LACK OF PATIENT ATTENTION Visual field tests require the patient’s full attention. Soothing and encouraging words from the visual ield Distractions such as noise can negatively in luence the technician can strongly reduce these anxieties and in- patient’s test performance. In addition, some patients crease the reliability of the results. Distractions should experience anxiety when performing visual ield tests, also be reduced to a minimum. If the clinic layout does due to fear that they are not performing well, or anxiety not offer a stand-alone perimetry room, light-dimming about the outcome. curtains around the perimeter and ear muffs can offer a cost-effective alternative. TRIGGER-HAPPY PATIENTS Some patients, consciously or unconsciously, want to answers carefully during the examination. If a patient positively in luence the result of the visual ield test (e.g., responds to more than one false positive stimulus during if their ability to drive is at stake, or if they fear a bad diag- the test, it will be helpful to interrupt the test immediately nosis). These patients may be trigger-happy, pressing and reinstruct the patient, in order to avoid an unreliable the response button even if they do not see a stimulus. result. Note that a beeping sound upon stimulus presen- False positive trials where no stimuli are presented are tation may encourage trigger-happy patients to press the used to detect trigger-happy patients (for more details, response button and it is thus recommended not to use see Chapter 7). It is important to watch for false positive this, except in speci ic situations. MISTAKES IN THE SET-UP PROCEDURE ACCURATE ENTRY OF PATIENT INFORMATION Patient data, such as date of birth and refraction, need to measured sensitivities is compared to the data for an be entered in the perimeter. It is important to ensure that average normal person of the same age, rather than an this information is accurate. For example, if the wrong average normal person who is younger or older. FIG 3-15 date of birth is entered, most representations of the illustrates the in luence of incorrect patient age on the visual ield test will be inaccurate, because each set of patient’s visual ield. INADEQUATE CORRECTION OF REFRACTIVE ERROR Inadequate correction of refractive error can lead to refraction for a patient. To avoid this, it is recommended a blurring of the stimulus. If the patient does not have a to check the refraction on the same day as the perimetric sharp image of the stimulus, the visual ield results will test. Even if the patient’s refraction has been checked be worse than the patient’s true visual ield. Additionally, previously, it is possible that it may have changed since a lens with too much plus power can lead to an arti icially then, especially among older patients. enlarged visual ield, while a lens with too much minus power will have the opposite effect. The second source of error is the incorrect choice of trial lens. It is important to consult the user manual for the The irst source of error is that the patient has been in- respective perimeter, as the choice of trial lens depends correctly refracted, or that the examiner uses the wrong on the perimeter model. The paragraphs below describe Common pitfalls to avoid 41 INFLUENCE OF INCORRECT PATIENT AGE ON VISUAL FIELD RESULTS A CORRECT AGE B INCORRECT AGE C INCORRECT AGE 58 years Too old: 88 years Too young: 18 years FIGURE 3-15 If the date of birth of a 58-year-old patient (A) is incorrectly entered, so that the patient’s age is 88 in the perimeter, the results will be artificially good (B). If the same patient is entered as an 18-year-old patient, the results will be artificially bad (C). the choice of trial lens for the current Octopus models To save time and avoid mistakes, it is recommended to al- 900 and 600. ways use the built-in trial lens calculator to determine the required refractive lens. The trial lens calculator always Patients need their far-distance correction for relaxed uses the patient’s actual best far-distance correction. It vision. Depending on age, an added near-distance correc- then automatically calculates the necessary age-dependent tion for presbyopia is also needed, because perimeters near-distance correction. It determines and recommends test at near distances. It is important to use the adequate the trial lens with the lowest possible power, in order to correction for presbyopia proposed by the perimeter’s minimize the risk of artifacts. BOX 3B presents the under- manufacturer, and not the patient’s reading glass pre- lying assumptions of the trial lens calculator. scription. Special attention should be given to noting the sign (plus or minus) of the correction. If a minus lens is It is best practice to ask each patient prior to starting the employed when a plus lens should have been used, the test whether they can see the ixation target sharply and, patient’s vision may become blurry. if necessary, adjust the refraction so as to avoid inadequate correction of the refractive error. RATIONALE USED IN THE DESIGN OF THE TRIAL LENS CALCULATOR BOX 3B DETERMINATION OF APPROPRIATE SPHERICAL LENS The current Octopus perimeter models 600 and 900 present stimuli at a distance of 30 cm (11.8 inches) from the eye. This corresponds to an approximate refraction of + 3.25 diopters (D), as calculated using the following formula: Power (D) = 1/stimulus distance (m) = 1/0.3 = 3.33 To enable the patient to focus at this distance, the patient’s far-distance refraction values are needed. Depending on the patient’s refraction, different scenarios occur: 42 Chapter 3 | How to perform perimetry you can trust Normal sighted patients: Young emmetropic patients can accommodate at 30cm, so they do not need an additional trial lens. With increasing age, patients gradually lose their ability to accommodate their eyes (i.e., to change their lens power) to objects presented at near distances. To facilitate near optical correction, additional diop- ters (D) of refractive power are needed depending of the age of the patient (see table below). Hyperopic and presbyopic patients: Hyperopic patient may have dif iculty to focus at 30 cm. For these patients, a trial lens is needed, corre- sponding to their refraction (R). As with emmetropic patients who are older, additional diopters (D) are needed to support their near optical correction (presbyopia) (see table below). Myopic patients: Near sighted patients of up to -3 D do not necessarily need corrective lenses, as they can focus at 30 cm. Patients with strong myopia (greater than -3 D) will have dif iculty focusing at 30 cm and need additional correction. For refractive values above -3 D, add 3.25 D to the refractive value (e.g., for R = -4 D; use a -0.75 D lens). As for presbyopic and emmetropic patients, with increasing age, near optical correction is more dif icult and additional diopters are needed. Corrections in the cupola perimeter of Octopus 900 Cupola perimeters allow for full- ield peripheral testing that extends beyond the range of a trial lens. Therefore, all lenses and the lens holder must be removable to allow for peripheral testing. No trial lens should be used for testing beyond 30° eccentricity. The Octopus 900 has a built-in trial lens calculator to determine which trial lens should be used. The following look-up table shows the outputs of the Octopus trial lens calculator. Age Hyperopic Emmetropic Myopic R>0D R=0D R = -0.5 D R = -1 D R = -1.5 D R = -2 D R = -2.5 D R = -3 D R < -3 D < 30 R No lens No lens No lens No lens No lens No lens No lens R + 3.25 30 – 39 R + 1.0 D +1 +0.5 No lens No lens No lens No lens No lens R + 3.25 40 – 44 R + 1.5 D +1.5 +1.0 +0.5 No lens No lens No lens No lens R + 3.25 45 – 49 R + 2.0 D +2.0 +1.5 +1.0 +0.5 No lens No lens No lens R + 3.25 50 – 54 R + 2.5 D +2.5 +2.0 +1.5 +1.0 +0.5 No lens No lens R + 3.25 55 -59 R + 3.0 D +3.0 +2.5 +2.0 +1.5 +1.0 +0.5 No lens R + 3.25 >= 60 R + 3.25 D +3.25 +2.75 +2.25 +1.75 +1.25 +0.75 No lens R + 3.25 Correction in the central field perimeter of the Octopus 600 In order to simplify the clinical work low, the Octopus 600 perimeter has a built-in +3.25 D lens that covers the central 30° of the visual ield. All patients, irrespective of age, therefore receive the maximum correction for presbyopia. Only their actual refraction (R) is needed. If younger patients are over-cor- rected, they are able to compensate by relaxing their lens without negative effect on their visual ield. DETERMINATION OF APPROPRIATE CYLINDER LENS Cylinder Correction (Octopus 900) A cylinder correction can be discarded when the prescription is 0.25 D or less, because it does not alter the result of the visual ield test. For cylinders from 0.5 to 1 D, the spherical equivalent is used and added to the spherical lens needed for each patient. The spherical equivalent is calculated using the following formula: Spherical equivalent = ½ * cylinder correction This formula is an approximation that adequately corrects for small cylinders, but does not suf iciently correct for cylinders larger than 1 D. For cylinders larger than 1 D, a cylindrical correction is needed. Remember to get the cylinder axis oriented to the proper angle on the lens holder. (For the special case of the Octopus 600 refer to the user manual). Common pitfalls to avoid 43 EXTERNAL OBSTRUCTIONS BLOCKING STIMULI FROM REACHING THE RETINA LENS RIM ARTIFACTS If the edge of the trial lens blocks the patient’s view (FIG so that the eye is as close as possible to the trial lens with- 3-16), the visual ield results will be adversely affected out touching it, and aligned in the center of the trial lens and will show absolute defects at the edges. To avoid holder. The Octopus 900 provides a measurement func- trial lens rim artifacts, the patient should be positioned tion to warn if the lens is too far from the eye. INFLUENCE OF LENS RIM ARTIFACTS ON VISUAL FIELD RESULTS A) B) NO ARTIFACT LENS RIM ARTIFACT FIGURE 3-16 If the patient is correctly positioned close to the trial lens (A), rim artifacts do not appear within 30° of the field of view. If the patient is too far away from the trial lens (B), the edge of the visual field shows the rim of the lens. FACIAL STRUCTURE OF THE PATIENT It is important to observe the physiognomy (facial struc- Droopy lids (ptosis) and droopy lid skin (dermatochalasis) ture) of the patient. A prominent nose, a heavy brow or long might also obstruct the patients’ upper ield of view (FIG eyelashes can alter the ield of view, leading to misinter- 3-17). To avoid artifacts caused by ptosis, tape can be used pretation of the visual ield results. If there is a prominent to lift the eyelid. Care should be taken to leave enough facial structure, it is recommended to turn or tilt the freedom to allow blinking. patient’s head to the side slightly, without losing ixation. DIRTY CONTACT LENS Since very high corrections can lead to peripheral dis- lenses are used, they must be inspected before the test. tortions, it is advisable for a patient with very high cor- Dirty contact lenses reduce the amount of light entering rections to wear contact lenses. Patients with moderate the eye, resulting in a diffuse defect. This will also appear myopia may also leave their contact lenses in. If contact in the Defect Curve as a downward shift of the entire curve. 44 Chapter 3 | How to perform perimetry you can trust PUPIL SIZE The amount of light entering the eye is controlled by the observed. These artifacts may simulate glaucomatous diameter of the pupil. As a rule, the pupil must have a visual ield defects. To avoid this, patients with a pupil diameter of at least 3 mm for the results of the test to size of less than 3 mm, as measured in a dimly-lit room, be trustworthy. Small pupils decrease the amount of may be dilated before the perimetric examination. Highly incident light on the retina and result in a uniform de- arti icially dilated pupils may, however, occasionally lead pression of the visual ield (FIG 3-18).¹³,¹⁴ Increasing to mild peripheral visual ield distortions. diffraction around the margin of the pupil may also be INFLUENCE OF FACIAL STRUCTURE ON VISUAL FIELD A) PTOSIS B) PTOSIS WITH LID TAPED UP FIGURE 3-17 Ptosis (droopy lid) results in external superior obstruction of the visual field that is not related to any pathology of the eye (A). Patients with severe ptosis or dermatochalasis should therefore be tested with the lid taped up (B), in order to assess the visual field without the effect of ptosis, as seen in the example below. INFLUENCE OF SMALL PUPIL SIZE ON VISUAL FIELD SMALL PUPIL SIZE NORMAL PUPIL SIZE Light Source FIGURE 3-18 If a patient’s pupil is too small, the overall sensitivity to light will be reduced, resulting in a visual field with diffuse defect. References 45 CLINICAL RELEVANCE OF UNTRUSTWORTHY VISUAL FIELDS Obtaining reliable results is important in order to interpret The second most frequent cause of unreliable visual ields the visual ields correctly. Unreliable visual ields unfor- was false positive errors, which accounted for 18% of all tunately occur relatively frequently in clinical practice. In unreliable visual ields.² Of all the visual ield hemi ields more controlled conditions such as the large Ocular Hy- included in the OHTS, 0.4% had rim artifacts,¹² while pertension Treatment Study (OHTS), ixation losses were superior and inferior depressions due to facial features the most frequently observed cause of unreliable visual accounted for only 0.2% of all hemi ields. In less controlled ields, accounting for 70% of all unreliable visual ields.² conditions, these numbers may be signi icantly higher. REFERENCES 1. Bickler-Bluth M, Trick GL, Kolker AE, Cooper DG. Assessing the utility of reliability indices for automated visual ields. Testing ocular hypertensives. Ophthalmology. 1989;96:616-619. 2. Johnson CA, Keltner JL, Cello KE, et al. Baseline visual ield characteristics in the ocular hypertension treatment study. Ophthalmology. 2002;109:432-437. 3. Johnson CA, Nelson-Quigg JM. A prospective three-year study of response properties of normal subjects and patients during automated perimetry. Ophthalmology. 1993;100:269-274. 4. Katz J, Sommer A. Reliability indexes of automated perimetric tests. Arch Ophthalmol. 1988;106:1252-1254. 5. Katz J, Sommer A, Witt K. Reliability of visual ield results over repeated testing. Ophthalmology. 1991;98:70-75. 6. Fuhr PS, Hershner TA, Daum KM. Ganzfeld blankout occurs in bowl perimetry and is eliminated by translucent occlusion. Arch Ophthalmol. 1990;108:983-988. 7. Gonzalez de la Rosa M, Pareja A. In luence of the "fatigue effect" on the mean deviation measurement in perimetry. Eur J Ophthalmol. 1997;7:29-34. 8. Hudson C, Wild JM, O'Neill EC. Fatigue effects during a single session of automated static threshold perimetry. Invest Ophthalmol Vis Sci. 1994;35:268-280. 9. Johnson CA, Adams CW, Lewis RA. Fatigue effects in automated perimetry. Appl Opt. 1988;27:1030-1037. 10. Marra G, Flammer J. The learning and fatigue effect in automated perimetry. Graefes Arch Clin Exp Ophthalmol. 1991;229:501-504. 11. Wild JM, Searle AE, Dengler-Harles M, O'Neill EC. Long-term follow-up of baseline learning and fatigue effects in the automated perimetry of glaucoma and ocular hypertensive patients. Acta Ophthalmol (Copenh). 1991;69:210-216. 12. Keltner JL, Johnson CA, Cello KE, et al. Classi ication of visual ield abnormalities in the ocular hypertension treatment study. Arch Ophthalmol. 2003;121:643-650. 13. Lindenmuth KA, Skuta GL, Rabbani R, Musch DC. Effects of pupillary constriction on automated perimetry in normal eyes. Ophthalmology. 1989;96:1298-1301. 14. Wood JM, Wild JM, Bullimore MA, Gilmartin B. Factors affecting the normal perimetric pro ile derived by automated static threshold LED perimetry. I. Pupil size. Ophthalmic Physiol Opt. 1988;8:26-31. 46 47 CHAPTER 4 KEY EXAMINATION PARAMETERS FIXED EXAMINATION PARAMETERS Perimetric testing must be as standardized as possible, results. Chapter 12 provides an overview of the most in order to allow comparisons over time and across common differences between devices and provides prac- different eye care providing of ices. Therefore, many tical advice on how to successfully master the transition. examination parameters are ixed by the perimeter used and are not specifically selected by the user of For the sake of completeness, a summary of the most es- the perimeter. These ixed parameters typically include sential ixed examination parameters of current Octopus background color and luminance, maximum stimulus perimeters and the rationales behind them is provided luminance and stimulus duration. in BOX 4A. Note that the settings presented below apply to Standard Automated Perimetry. In special situations, Different perimeter models use different ixed settings. other ixed examination parameters are chosen. They Therefore, when switching from one device to another, are discussed in the respective chapters. it is important to consider their in luence on the perimetric FIXED EXAMINATION PARAMETERS BOX 4A BACKGROUND INTENSITY AND COLOR Background luminance (i.e., the re lected light intensity of the background) determines the contrast between the stimulus presented and the background, and thus has a considerable in luence on stimulus perception. To achieve comparable test results, it must be kept constant. The ideal background luminance of a perimeter should not be too bright, in order to allow display of very dim stimuli for a large dynamic testing range. Neither should it be too dark, to avoid time-consum- ing dark adaptation of the eye. It should stimulate selected cell types. The standard background luminance of current Octopus models consists of white light with a lumi- nance of 31.4 asb, which equals 10 cd/m . This luminance level is at the low end of photopic vision (i.e., the visual system used in normal daylight conditions) and does not require time for dark adaptation, but still provides a high dynamic testing range. White light is used because it is detected by all cell types in the retina and is therefore non-selective. MAXIMUM STIMULUS LUMINANCE As seen in Chapter 2, the maximum stimulus luminance (i.e., the maximum stimulus intensity) of a perimeter de ines the luminance associated with 0 dB on the decibel scale. It is also part of the formula to calculate a decibel value from the stimulus luminance. If the maximum stimulus luminance were to change, then the whole decibel scale would shift, so it must be kept constant for comparable results to be achieved. In order to offer a large dynamic testing range from normal to impaired vision, the maximum stimu- lus intensity value should be as high as possible. However, when the maximum stimulus intensity is 48 Chapter 4 | Key examination parameters too high, a part of it will be re lected from the back of the eye (stray light) and will then be detected by neighboring cells, which will produce inaccurate test results. Empirically, a maximum stimulus luminance of 4,000 asb has been shown to offer a large dynamic range, while minimizing stray light effects. , STIMULUS DURATION In order to reduce ixation losses, the perimetric stimulus duration (i.e., exposure time) is kept below the reaction time of the human re lex of quick eye movements towards rapidly appearing stimulus (i.e., saccadic eye movement). As the reaction time of the saccadic eye movement is around 200 ms, the stimulus duration should be shorter, but still suf iciently long to be seen. For that reason, Octopus perimeters use a standard stimulus duration of 100 ms. PATIENT-SPECIFIC EXAMINATION PARAMETERS As described in Chapter 2, there is always a trade-off 1. Which type of perimetry should be used: static or between testing time and accuracy in perimetric exam- kinetic perimetry? inations. In this respect, it is very important to maximize 2. Which type of stimulus should be used: standard the clinically relevant information, while at the same white-on-white, function-speci ic or low-vision? time minimizing test duration. As perimetry has a wide 3. Which test pattern should be used? range of applications, there is no “one parameter its all” 4. Which test strategy should be used? approach for all situations. Each Octopus perimeter thus contains a library of standardized examination parame- The irst two questions are typically easy to answer. ters from which the optimum set can be chosen for each Indeed, static and standard perimetry are indicated patient. These patient-speci ic examination parameters for the needs of patients in most clinical practices and thus have to be selected for every patient. are by far the most commonly used types of perimetry. With regard to test strategy and test pattern, various In essence, there are four essential questions each clini- selections are commonly employed, and these decisions cian must answer, in the order shown below, prior to must be made individually. ordering a perimetric test: TYPE OF PERIMETRY: STATIC OR KINETIC PERIMETRY STATIC PERIMETRY For reasons of simpli ication, so far this book has concen- (FIG 4-1A). With this type of perimetry, it is possible to trated on static perimetry. In static perimetry, stimuli of detect small changes in sensitivity thresholds with rel- varying luminance levels are used to determine visual sen- atively high accuracy. For this reason, static perimetry sitivity thresholds at a speci ied number of ixed locations is the standard for slowly progressing diseases such as Patient-specific examination parameters 49 glaucoma. Since it is fully automated, it is also easy to use As the majority of visual ield tests are performed for in clinical practice. glaucoma, static perimetry is the most commonly used type of perimetry today. KINETIC PERIMETRY Kinetic perimetry was the irst quantitative method non-seeing to seeing areas. The patient response then of performing visual ield testing and is an alternative de ines the visual ield location of the speci ic light sen- to static perimetry. In kinetic perimetry, moving stim- sitivity threshold (FIG 4-1B). uli of pre-determined light intensities are moved from STATIC AND KINETIC PERIMETRY TESTING METHODS A) STATIC PERIMETRY Dim = Seen Stimulus = Not seen Do you see No the stimulus? No THRESHOLD SENSITIVITY No Fixation Yes Yes Yes Stimulus Yes Bright Stimulus B) KINETIC PERIMETRY 135 = Seen Patient Do you see response the stimulus? 150 = Not seen 165 180 Yes Fixation Yes No 195 Yes 210 Vector (Stimulus trajectory) No 225 FIGURE 4-1 Both static and kinetic perimetry are designed to provide visual sensitivity thresholds that allow mapping the hill of vision of a patient. In static perimetry (A), stimuli of differing light intensity are shown at given locations, to determine the sensitivity threshold at those positions. In kinetic perimetry (B), a stimulus of a given light intensity is moved along the visual field (non-seeing to seeing), to determine the location of that sensitivity threshold. 50 Chapter 4 | Key examination parameters After repeating this process for a speci ic stimulus size for a given stimulus size and intensity and is similar to and intensity across the entire visual ield, the visual sen- an altitude line on a geographical map. Local regions of sitivity thresholds can be connected to form an isopter reduced sensitivity inside the isopter are identi ied in the (line of equal sensitivity). An isopter marks the boundary same way and are called scotomas. FIG 4-2 shows how between seeing and non-seeing around the hill of vision static and kinetic perimetry results are displayed. DISPLAY OF STATIC AND KINETIC VISUAL FIELDS STATIC KINETIC Isopter Scotoma SEEING NON-SEEING (here blind spot) 10 20 30 40 50 60 70 80 90 Scotoma SEEING NON-SEEING (here blind spot) FIGURE 4-2 In static perimetry, each sensitivity threshold is displayed independently, either as a color or as a numerical map (not shown here). In kinetic perimetry, areas of equal sensitivity thresholds form an isopter that provides similar information to static perimetry about the shape of defects. Local areas of depression inside an isopter are called scotomas. Since the patient can report seeing the stimulus at any kinetic perimetry is currently not fully automated, making location along the trajectory of the stimuli, kinetic perime- it more challenging in everyday use. try provides high spatial resolution and fast testing over a large area. It is therefore bene icial for diseases affecting As the majority of visual ield tests are performed to as- the periphery and sharp-edged defects and is frequent- sess glaucoma and due to the ease of use of automation, ly used to evaluate neurological diseases and peripheral static perimetry is by far the most commonly used type retinal diseases. As moving stimuli are easier to see than of perimetry today. For that reason, all of the following non-moving ones in the periphery, kinetic perimetry is paragraphs and chapters focus on static perimetry, while also often used for children and for patients with cog- kinetic perimetry will be discussed in depth in Chapter 11. nitive impairment or severe visual ield loss. However, The key differences between static and kinetic perimetry are summarized in TABLE 4-1. Patient-specific examination parameters 51 COMPARISON BETWEEN STATIC AND KINETIC PERIMETRY TABLE 4-1 STATIC KINETIC ADVANTAGES Clinical gold standard High spatial resolution High precision sensitivity thresholds Fast peripheral testing Fully automated Provides information about other visual functions Highly interactive, lexible and adaptable WHAT IT IS BEST Small changes in sensitivity thresholds Small changes in spatial extent of a defect AT DETECTING Changes in the central area Peripheral changes Remaining vision in advanced diseases COMMON USES Glaucoma Neuro-ophthalmological conditions Macular diseases Peripheral retina diseases Visual ability testing Low vision Children Patient with cognitive impairment STIMULUS TYPE: STANDARD OR NON-CONVENTIONAL STANDARD WHITE-ON-WHITE PERIMETRY The standard perimetric stimulus is white on a white light allows visual ield testing from early to advanced background, and this type of perimetry is commonly re- disease (i.e., it offers a large dynamic testing range). By ferred to as white-on-white perimetry, or Standard Auto- convention, the standard stimulus used is round, with a mated Perimetry (SAP). diameter of 0.43°, which is also the Goldmann stimulus size III, based on the de inition of Professor Hans Gold- The white color stimulus offers the advantage of stim- mann. For more information on Goldmann stimulus ulating all different retinal cell types. As a result, white sizes, refer to BOX 4B. 52 Chapter 4 | Key examination parameters BOX 4B GOLDMANN SIZES I TO V The size conventions used today to describe a perimetric stimulus are derived from the work BLIND SPOT V 1.7° of Professor Hans Goldmann, who developed the Goldmann perimeter in 1946. He de ined standard sizes for perimetric stimuli, and the Goldmann sizes 20 IV 0.8° I to V are still widely used. Each step corresponds to a change in diameter by a factor of 2 and in area by a factor of 4. Size III is several times smaller than III 0.43° the physiological blind spot and was considered to be an accurate measurement size. II 0.2° The Goldmann stimulus sizes I to V are presented in I 0.1° relation to the size of the physiological blind spot. FUNCTION-SPECIFIC PERIMETRY Function-speci ic perimetry uses different stimulus types background (Short-Wavelength Automated Perimetry, to stimulate different visual functions (e.g., motion, or or SWAP); a white lickering stimulus on a white back- color vision), but they all have the same purpose: ground (Flicker Perimetry); or a pulsating stimulus with measuring a subset of the visual system individually, to concentric rings changing in both spatial resolution and get more sensitive responses for early disease detection. contrast (Pulsar Perimetry). They are described in more Different Octopus perimeter models offer different func- detail in Chapter 10. tion-speci ic stimuli (FIG 4-3): a blue stimulus on a yellow FUNCTION-SPECIFIC PERIMETRY Time 1 Time 2 ON OFF SWAP Flicker Pulsar FIGURE 4-3 Stimuli used in function-specific perimetry from left to right: Short Wavelength Automated Perimetry (SWAP), Flicker Perimetry and Pulsar Perimetry. Patient-specific examination parameters 53 PERIMETRY FOR LOW VISION There is a limit to the visibility of the standard size III stimulus size V is typically used, instead of the standard white perimetric stimulus in patients with signi icantly size III. It is 16 times larger in area and is therefore more impaired visual sensitivity. In order to increase the dy- detectable. Chapter 10 provides more information about namic range into the low vision region and to make the stimulus size V. stimulus more visible to these patients, the Goldmann OVERVIEW OF DIFFERENT STIMULUS TYPES TABLE 4-2 STANDARD FUNCTION-SPECIFIC LOW VISION White-on-white, Pulsar, Flicker, SWAP White-on-white, stimulus III stimulus V ADVANTAGES Clinical standard Earlier detection in some Better visibility for patients patients with signi icant visual ield loss Provides information about other visual functions WHAT IT IS BEST Follow-up of a disease Early loss in some patients Advanced visual ield loss AT DETECTING from early to late stage COMMON USES Glaucoma Con irm defects observed on Advanced glaucoma or standard perimetry other ocular or neurological diseases Macular diseases Identify defects in glaucoma suspects who do not show defects on standard perimetry 54 Chapter 4 | Key examination parameters TEST PATTERN In clinical practice, patients can sometimes become tired locations. A very rough grid with 10° degree spacing be- quickly during perimetric testing, which signi icantly tween the stimuli would require approximately 190 test limits the number of test locations that can be reliably locations, but would be highly inaccurate, as there would tested. - A reasonably dense grid of test locations, cov- be only 5 test points in the central 10° of vision, which ering the entire visual ield with 2° degree spacing, would is an important area for visual functions such as reading require around 4,800 size III stimuli, and a grid with 6° and identifying objects (FIG 4-4). degree spacing would require approximately 550 test ILLUSTRATION OF THE LOW SPATIAL RESOLUTION OF PERIMETRIC TESTING 10º SPACING 6º SPACING 2º SPACING ~190 size III targets ~550 size III targets ~4800 size III targets 90 90 90 180 0 180 0 180 0 270 270 270 FIGURE 4-4 Covering the entire visual field with high resolution within a reasonable test duration is not possible. Either the field is only roughly covered, or the test duration is unacceptable, as shown in this example with three different test patterns. In order to maximize perimetric information and mini- offer a large library of testing patterns for common mize test duration, a test pattern should be chosen with perimetric applications. a high density of test locations in the area of high inter- est and a low density of test locations in areas of low The most commonly used test patterns available on the interest (FIG 4-5). For that reason, Octopus perimeters Octopus perimeter and the rationale for which to select are described in depth in Chapter 5. Patient-specific examination parameters 55 EXAMPLES OF DIFFERENT TEST PATTERNS A) G-PATTERN B) M-PATTERN (Glaucoma) (Macula) 90 90 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 270 270 C) ESTERMAN D) PTOSIS (Visual driving ability) 90 90 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 270 270 FIGURE 4-5 Examples of test patterns for various clinical perimetric applications are presented. Each pattern maximizes the relevant information for that clinical situation, while minimizing the test duration by only evaluating the most relevant areas. (A) The G-pattern for glaucoma tests within 30° at locations that follow the retinal nerve fibre bundle patterns. (B) The M-pattern for the macula tests within the central 10°. (C) The Esterman tests binocularly for visual fitness to drive (120° horizontally and 60° vertically). (D) The Ptosis test pattern only evaluates the upper hemifield along common eyelid locations. 56 Chapter 4 | Key examination parameters TEST STRATEGY For the detection and follow-up of a disease, the sensi- stimulus luminance) to 32 dB (approximate foveal sensi- tivity thresholds should be determined with high accu- tivity threshold of a 20-year-old on the Octopus 900), 32 racy. However, in clinical practice, even very cooperative stimuli would have to be presented at one test location. and reliable patients experience fatigue, which limits the Performing the same procedure in 2 dB steps would re- number of stimulus luminance levels that can be pre- quire 16 stimuli, while 4 dB steps would still require the sented during a perimetric test. If we were to sample presentation of 8 stimuli (FIG 4-6). the entire range in steps of 1 dB, from 0 dB (maximum ILLUSTRATION OF THE LOW RESOLUTION OF SENSITIVITY THRESHOLDS IN PERIMETRIC TESTING 4 dB PRECISION 2 dB PRECISION 1 dB PRECISION Up to 8 Up to 16 Up to 32 stimuli/location stimuli/location stimuli/location 32 dB 32 dB 32 dB 0 dB 0 dB 0 dB FIGURE 4-6 Determining a sensitivity threshold with high precision with a sequence of stimuli of increasing intensity is not pos- sible. Either too many stimuli are required, or the step sizes are too large, as the example with three different step sizes demon- strates. Instead of using the strategy of increasing stimulus inten- they only assess whether stimuli are seen or unseen (FIG sity step by step until the sensitivity threshold is reached, 4-8). Qualitative strategies are commonly used in legal vi- an ef icient strategy is therefore needed that maximizes sual ability evaluations, such as in the tests used to assess precision but minimizes test duration. visual itness to drive. Examples of a quantitative and a qualitative test strategy are given in FIG 4-7 and FIG 4-8, Octopus perimeters offer several test strategies with dif- for the sake of illustration. ferent trade-offs between test duration and accuracy for different clinical situations. Some strategies are quanti- The most commonly used strategies available on the tative, which means that they are used to determine a Octopus perimeter and the rationale for which strategy sensitivity threshold (FIG 4-7). Qualitative strategies are to select are described in depth in Chapter 6. also offered in which the testing time is reduced, because Patient-specific examination parameters 57 EXAMPLE OF A QUANTITATIVE STRATEGY = Seen QUANTITATIVE STRATEGY = Not seen Do you see? 30 dB 1 2 Sensitivity Threshold 3 Threshold Zone 5 4 0 dB 1. 2. Sampling in Detailing within large steps threshold zone FIGURE 4-7 Example of a quantitative thresholding strategy: The visual field is first scanned with stimuli with large steps in light intensity, in order to identify a suspected threshold zone. Once that zone has been identified, further testing inside that zone will allow for determination of an accurate threshold with minimal test duration. EXAMPLE OF A QUALITATIVE STRATEGY = Seen QUALITATIVE STRATEGY = Not seen Do you see? Do you see? 30 dB 30 dB Sufficient Sufficient vision to drive vision to drive Patient Patient is not fit is fit Insufficient to drive Insufficient to drive vision to drive vision to drive 0 dB 0 dB FIGURE 4-8 Example of a qualitative strategy: For visual driving ability, one stimulus is shown at the fixed stimulus intensity which is the minimum needed to drive safely. If a person sees that stimulus at a required number of test locations, this means that the person fulfills the visual field criteria to be able to drive. 58 Chapter 4 | Key examination parameters REFERENCES 1. Fankhauser F, Haeberlin H. Dynamic range and stray light. An estimate of the falsifying effects of stray light in perimetry. Doc Ophthalmol. 1980;50:143-167. 2. Anderson RS, Redmond T, McDowell DR, Breslin KM, Zlatkova MB. The robustness of various forms of perimetry to different levels of induced intraocular stray light. Invest Ophthalmol Vis Sci. 2009;50:4022-4028. 3. Johnson CA, Adams CW, Lewis RA. Fatigue effects in automated perimetry. Appl Opt. 1988;27:1030-1037. 4. Wild JM, Searle AE, Dengler-Harles M, O’Neill EC. Long-term follow-up of baseline learning and fatigue effects in the automated perimetry of glaucoma and ocular hypertensive patients. Acta Ophthalmol (Copenh). 1991;69:210-216. 5. Hudson C, Wild JM, O’Neill EC. Fatigue effects during a single session of automated static threshold perimetry. Invest Ophthalmol Vis Sci. 1994;35:268-280. 6. Gonzalez de la Rosa M, Pareja A. In luence of the “fatigue effect” on the mean deviation measurement in perimetry. Eur J Ophthalmol. 1997;7:29-34. 59 CHAPTER 5 SELECTING A TEST PATTERN INTRODUCTION Depending on the pathology or type of ability testing different eye care providing of ices, test patterns are that is to be performed, certain test locations are far standardized. However, various patterns have been more relevant than others. As there is always a trade-off developed and different patterns can be used for the between test duration and accuracy in any perimetric same purpose. Octopus perimeters offer all of the most test, a test pattern should be chosen with locations in commonly used patterns, to allow for testing continuity. the relevant area. The following section focuses on the most commonly used For this reason, all Octopus perimeters offer a library patterns and provides rationales for which to choose in of a variety of test patterns for each application. In or- speci ic situations. TABLE 5-1 provides a summary of the der for test results to be comparable between different most commonly used Octopus test patterns. sessions, between different patients and even between COMMONLY USED TEST PATTERNS FOR VARIOUS INDICATIONS TABLE 5-1 INDICATION RECOMMENDATION COMMON ALTERNATIVES GLAUCOMA/CENTRAL FIELD G (Glaucoma) 32, 30-2, 24-2 MACULA M (Macula) 10-2 FULL FIELD (NEURO, RETINA) 07 Kinetic FOVEA Fovea BLIND SPOT Blind spot Kinetic LOW VISION M, G, 07 depending on pathology Kinetic SCREENING FOR ABNORMAL GST (Glaucoma Screening Test) VISION DRIVING ET (Esterman) FG (Führerscheingutachten), Kinetic BLEPHAROPTOSIS BT (Blepharoptosis) Kinetic BLINDNESS BG (Blindengutachten) 60 Chapter 5 | Selecting a test pattern TEST PATTERNS FOR GLAUCOMA TYPICAL VISUAL FIELD DEFECTS IN GLAUCOMA Glaucoma is a disease resulting in the degeneration of 30° test pattern, which has become the standard for visual retinal nerve iber bundles in the eye. Since the largest ield testing in glaucoma today. Conversely, the periphery proportion of retinal nerve ibers is located within the is rarely affected in isolation in glaucoma,⁶ so that central 30°,¹-⁴ most early to moderate glaucoma visual peripheral testing is less common in cases of glaucoma ield loss occurs within the central 30°. Typical defect for diagnostic reasons and, if used at all, is aimed to patterns follow the distribution of the retinal nerve iber assess a patient’s quality of life. bundles and there is a clear separation along the superi- or and inferior hemi ields at the horizontal meridian. The In very advanced glaucoma, the visual field usually typical patterns of visual ield loss due to glaucoma are constricts to a macular visual ield⁷ and testing outside partial arcuate, paracentral, nasal step, arcuate, temporal the macula does not provide any further diagnostic wedge and altitudinal defects (FIG 5-1).⁵ information. Therefore, it is common to switch to a 10° macular test pattern in advanced glaucoma, in order to Since glaucomatous visual ield defects typically occur track residual vision in that area with higher resolution⁷,⁸ within the central visual ield, the best trade-off between for the same test duration. test duration and accuracy is achieved by using a central Test patterns for glaucoma 61 TYPICAL VISUAL FIELD DEFECTS IN GLAUCOMA PARTIAL ARCUATE PARACENTRAL NASAL STEP DIFFUSE ARCUATE TEMPORAL WEDGE ALTITUDINAL CONSTRICTED (Double arcuate) FIGURE 5-1 The typical visual field defects in glaucoma are the partial arcuate, paracentral, nasal step, diffuse, arcuate, temporal wedge, altitudinal and constricted in advanced glaucoma (ordered according to frequency of occurrence). All manifest themselves within the central 30° visual field, so that central 30° testing in glaucoma care is standard. 62 Chapter 5 | Selecting a test pattern STANDARD TEST PATTERN IN GLAUCOMA CARE The standard perimetric stimulus is white, and is pre- light allows visual ield testing from early to advanced sented on a white background. This type of perimetry disease (i.e., it offers a large dynamic testing range). By is commonly referred to as white-on-white perimetry, convention, the standard stimulus used is round, with a or Standard Automated Perimetry (SAP). diameter of 0.43°, which is also the Goldmann stimulus, size III, based on the de inition by Professor Hans Gold- The white color stimulus offers the advantage of stim- mann. For more information on Goldmann stimulus ulating all different retinal cell types. As a result, white sizes, refer to BOX 4B. G PATTERN The G pattern was designed to serve as a multi-purpose tection of visual loss associated with glaucoma, but also test and offers an excellent trade-off between test dura- neuro-ophthalmological and macular diseases. tion and accuracy.⁹-¹¹ There are 59 different locations within the central 30° of the visual ield and they are To maximize the detection of glaucomatous visual loss, the distributed in a pattern that facilitates not only the de- test locations are distributed along the retinal nerve iber bundles, where visual loss is most likely to occur (FIG 5-2). THE G PATTERN FOR GLAUCOMA FIGURE 5-2 The distribution of the test locations in the G pattern follows the retinal nerve fiber bundles. Test patterns for glaucoma 63 The G pattern (FIG 5-3) offers a high density of points in ant area of visual function for reading and object iden- the paracentral area (down to 2.8° spacing), to facilitate ti ication and allows for additional detection of macular detection of paracentral scotomas, which are common in diseases. Additionally, many recent reports indicate that glaucoma, yet sometimes missed by other patterns.¹²,¹³ there are structural and functional de icits which occur The test grid also accentuates the nasal step and overall in the macula of glaucoma patients.¹⁴ ,¹⁵ To detect has more test points nasally than temporally – partly due common neurological diseases such as hemianopias and to the presence of the blind spot, but also to account for quadrantanopias, there are no points located on the the higher frequency of nasal visual ield loss in glaucoma. vertical and horizontal meridians in the G pattern. Time is saved by not testing in the immediate region of With 5 central points in the fovea and a total of 17 test the blind spot, where unreliable results typically tend locations in the macula, it focuses on the most import- to be observed. PATHOLOGY-BASED G PATTERN FOVEA MACULA 90 90 180 10 20 30 0 180 10 20 30 0 270 30° 270 30° NASAL STEP BORDERS OF QUADRANTS AND HEMIFIELD 90 90 180 10 20 30 0 180 10 20 30 0 270 30° 270 30° FIGURE 5-3 The pathology-based G pattern uses test locations following retinal nerve fiber bundles. It has a high density of test locations (highlighted in red) in the macula and fovea region, to detect foveal and paracentral defects and tests along the horizontal and vertical meridians (i.e., midlines), and to detect nasal step and neurological defects. Valuable testing time is saved with a lower density of test locations towards the periphery and temporal areas. 64 Chapter 5 | Selecting a test pattern ALTERNATIVE TEST PATTERNS FOR THE CENTRAL 30° 32/30-2 AND 24-2 PATTERNS The 32, 30-2 and 24-2 patterns (FIG 5-4) are similar to longer to complete than the G pattern without providing the G pattern in that they cover the central visual ield considerably more meaningful clinical information. and respect the vertical and horizontal meridians. In con- trast however, they are not optimized for speci ic pathol- The 24-2 pattern is based on the 30-2 pattern, but the ogies. Instead, all test locations are equidistant from each most peripheral ring of test locations is removed, except other and separated by 6°. for the two nasal points. With only 54 test points, the test duration of the 24-2 pattern is as short as that of the G Historically, the 32 pattern¹⁶ was initially used in the pattern, but the test pattern is not optimized for typical irst series of Octopus perimeters in 1977, while the 30-2 pathologies. pattern was among the irst central patterns used on the Humphrey Field Analyzer. These patterns are nearly Since it is optimized for pathology and quicker, the G identical to each other. The sole difference is that the pattern is recommended for new patients. However, the 30-2 pattern has 2 extra test locations in the blind spot, 32/30-2 and 24-2 patterns are recommended when a which are omitted in the 32 pattern. With their 74 or 76 large set of existing data taken from one of these patterns test locations respectively, the 32/30-2 patterns take is available for a patient, and the eye care provider wish- es to have continuity in the testing procedure. CENTRAL 30° TEST PATTERNS 32 30-2 24-2 74 test locations 76 test locations 54 test locations 90 90 90 180 10 20 30 0 180 10 20 30 0 180 10 20 30 0 270 30° 270 30° 270 30° FIGURE 5-4 The 30-2 pattern is similar to the 32 pattern, but has 2 additional test locations in the blind spot area. The 24-2 pattern is an abbreviated version of the 30-2 pattern, with most peripheral locations excluded, except for the nasal step region. Test patterns for glaucoma 65 FURTHER TEST PATTERNS FOR GLAUCOMA MACULA TESTING PATTERNS FOR ADVANCED GLAUCOMA In advanced glaucoma, the visual ield is typically con- information, it is common to switch to testing patterns stricted to the macular area. In these situations, testing that solely evaluate the area of the macula.⁷,⁸ the full 30° will not offer a good trade-off between test duration and the clinical information received, because For further information on macula patterns, please see more than 65% of the locations will be in known areas of the section on the macula testing patterns M or 10-2. non-seeing (FIG 5-5). To further maximize useful clinical MACULA PATTERN FOR ADVANCED GLAUCOMA G M 30° 12° 30° 12° FIGURE 5-5 In advanced glaucoma with a severely constricted visual field, the focus of visual field testing is on the remaining vision in the macula. In these situations, a macula pattern like the M pattern provides more clinically relevant information than a central pattern such as the G pattern. KINETIC PERIMETRY FOR ADVANCED GLAUCOMA Since static testing is challenging for patients with alternative to static testing in such cases. For more infor- advanced glaucoma,¹⁷-¹⁹ kinetic perimetry is a good mation, please see Chapter 11 on kinetic perimetry. PERIPHERAL TEST PATTERNS TO ASSESS QUALITY OF LIFE Even though the periphery is rarely affected in isolation be an add-on to the standard G pattern. The G-Periphery in glaucoma,²⁰-²² there may still be a need to assess the pattern places strong emphasis on the nasal step area, peripheral vision, in order to evaluate the patient’s over- as this is the most relevant peripheral location in glau- all quality of life. coma.²⁰,²³ In these situations, the G-Periphery pattern (FIG 5-6) is For an in-depth and ef icient assessment of the periph- a very time-effective peripheral screening pattern, with ery in low-vision patients, the use of kinetic perimetry only 14 test locations in the periphery, and is intended to should be considered. 66 Chapter 5 | Selecting a test pattern ABBREVIATED G PATTERN FOR SCREENING In some instances, a screening visual ield test is a con- test duration and accuracy. It is an abbreviated version venient procedure for every patient, to make sure that of the G pattern with only 28 test locations (FIG 5-6). The visual ield loss is not missed as part of a routine eye locations have been chosen on the basis of their ability to examination. predict glaucoma and other commonly occurring eye dis- eases, such as macula defects and quadrantanopias and For screening of a presumed healthy population, the hemianopias. The G-Screening (GS) pattern is available G-Screening (GS) pattern offers a good trade-off between exclusively with the screening strategy. ADDITIONAL TESTING PATTERNS FOR GLAUCOMA G-PERIPHERY G-SCREENING 14 test locations 28 test locations 90 90 180 10 30 40 50 60 70 80 90 0 180 10 20 30 0 270 90° 270 30° FIGURE 5-6 The G-Periphery pattern is an add-on to the G pattern, to quickly screen the periphery and dominantly the peripheral nasal step in glaucoma. The G-Screening pattern is used for routine screening of dominantly healthy patients. Testing patterns for neurological visual field loss 67 TEST PATTERNS FOR NEUROLOGICAL VISUAL FIELD LOSS TYPICAL VISUAL FIELD DEFECTS IN NEURO-OPHTHALMIC CONDITIONS Neurological conditions lead to a large variety of typical vision) visual ield defects, with congruity (similarity in visual ield defect patterns which are very speci ic, de- location, size and magnitude of the de icit) between the pending on the location at which the visual pathways are two eyes being most common further back in the visual affected (FIG 5-7).²⁴ Lesions of the optic disc and optic pathways at or near the occipital lobe. Large lesions re- nerve lead to unilateral (i.e., only affecting one eye) visual sult in complete hemianopias, although quadrantanopias ield defects. Common optic nerve and nerve head diseas- and wedge-like defects are also common. While large es include disc edema, optic neuropathies, optic neuritis, lesions also affect the central visual ield, small lesions compressive lesions such as those caused by idiopathic may not extend to the central 30°. It should also be noted intracranial hypertension, and a number of congenital that a complete homonymous hemianopia only indicates abnormalities, such as optic nerve head drusen. Typical that the de icit is post-chiasmal and that all visual path- visual ield defect patterns appear in the central 30° and way ibers leading back to the occipital lobe have been include foveal and macular defects, enlarged blind spots, damaged. or patterns similar to those occurring in glaucoma. However, de icits in the far peripheral visual ield beyond To cover all of the aforementioned visual ield defects, the 30° also frequently occur. full visual ield needs to be thoroughly tested, with a high density of test locations in the macula, the blind spot and Chiasmal lesions resulting from diseases such as pitu- the central ield. Therefore, thorough neurological visual itary adenomas and related lesions typically result in ield tests are time-consuming. If the type of disease is bitemporal (i.e., left and right eye defects are mirrored) identi ied, then the focus of testing can be in the area af- hemianopias, which progress from the superior to the in- fected by the disease, in order to reduce the test duration. ferior hemi ield, but always respect the vertical midline. Damage can be more pronounced in one eye than the To maximize perimetric information and minimize test other. Postchiasmal lesions typically lead to homony- duration, kinetic perimetry should also be considered. mous (i.e., left and right eye defects are on same side of 68 Chapter 5 | Selecting a test pattern TYPICAL VISUAL FIELD DEFECTS IN NEUROLOGICAL DISEASES LEFT TEMPORAL FIELD NASAL FIELD RIGHT TEMPORAL FIELD Visual field of the left eye (OS) Visual field of the right eye (OD) 60° 60° 30° 90° Retina Optic disc (blind spot) 1 2 Optic chiasm 3 Optic tract Optic radiations 4 5 (temporal lobe) 6 Lateral geniculate body 7 Optic radiations (parietal lobe) Occipital lobe Visual cortex 8 Optic nerve damage Chiasmal deficit Postchiasmal deficit 1 Caecocentral scotoma 4 Heteronymous 5 Homonymous hemianopia (bitemporal) hemianopia 2 Nerve fiber bundle defect 6 Superior homonymous 4 Heteronymous quadrantanopia (bitemporal) quadrantanopia Temporal lobe lesion 3 Central scotoma 7 Inferior homonymous quadrantanopia Parietal lobe lesion 8 Homonymous hemianopia Occipital lobe lesion FIGURE 5-7 Typical visual field defects in diseases are unilateral if the optic nerve is damaged, heteronymous (the two eyes are mirror images) around the chiasm and homonymous (the two eyes show non-mirror symmetry) beyond the chiasm. Both hemi- anopias (vertical hemisphere defect) and quadrantanopias (quadrant defects) are typical neurological visual field defects. Testing patterns for neurological visual field loss 69 THOROUGH ASSESSMENT OF NEUROLOGICAL VISUAL FIELD DEFECTS N PATTERN The N pattern is speci ically designed for neuro-ophthal- such as nerve compressions, with high accuracy and mic diseases. Given the wide variety of defect patterns, should be used when this kind of pathology is suspected. it consists of several components that can be combined With its 21 test locations extending from 0 to 3°, it pro- independently. The N pattern includes a full ield, fovea vides a rapid assessment of the important foveal region and blind spot testing pattern (FIG 5-8). and can also be useful for other indications in which the fovea is affected. The N-Full Field pattern is designed to test the full visual ield. It is useful to detect any kind of neurological dis- The N-Blind Spot pattern is designed to detect the bound- ease. It extends from 40° nasally to 67° temporally and aries of the blind spot with adequate accuracy to check 40° vertically. With its 54 test locations in the central vi- for blind spot enlargements. It covers the area around the sual ield and an additional 17 locations in the peripheral blind spot from 9 to 19° horizontally, and 9.5° superiorly visual ield, it offers an excellent trade-off between test to 19° inferiorly in 2.5° steps. To account for tilted discs, duration and accuracy in the detection of central and early extra points are added at the superior, temporal corner peripheral neurological defects. and the inferior nasal corner. With its 54 test locations, it takes a relatively long time to complete, while spatial The N-Fovea pattern is designed to detect foveal defects, resolution is limited to 2.5°. DIFFERENT PATTERNS FOR NEURO-OPHTHALMOLOGY N-FULL FIELD N-FOVEA N-BLIND SPOT 71 test locations 21 test locations 54 test locations 90 90 90 180 10 30 40 50 60 70 80 90 0 180 10 0 180 10 20 30 0 270 90° 270 10° 270 30° FIGURE 5-8 The N pattern for neuro-ophthalmic disease consists of three testing patterns: a full field pattern, a fovea pattern and a pattern to test the blind spot. 70 Chapter 5 | Selecting a test pattern 07 PATTERN The 07 pattern is an alternative to the N-Full Field test siderably longer. It is further described in the section on pattern. With its 130 test locations it is more thorough test patterns for retinopathies. than the N-Full Field test pattern, but also takes con- KINETIC PERIMETRY Since static testing is time consuming in the periphery neurological diseases. For more information, see Chapter and provides only a limited spatial resolution, kinetic 11 on kinetic perimetry. perimetry is a very good alternative to static testing in TEST PATTERNS FOR RETINOPATHIES TYPICAL VISUAL FIELD DEFECTS IN RETINOPATHIES Retinal diseases lead to a variety of typical visual ield For many reasons, perimetry is typically not the main defects (FIG 5-9). Diseases such as age-related macular diagnostic tool to detect and follow up retinopathies. degeneration (AMD) or drug-induced maculopathies Firstly, retinal lesions are easily identi ied by fundus lead to macula ield defects and consequently require a examination or imaging. Secondly, perimetry requires macula testing pattern for visual ield testing. the patient to maintain steady ixation, which is chal- lenging for patients with advanced pathologies affecting Other commonly occurring retinopathies often affect the the macula. Many of these patients will also have a far peripheral visual ield. For these conditions, a test non-foveal preferred retinal locus for fixation. And pattern covering the entire visual ield is essential. The thirdly, many retinopathies require peripheral testing visual ield defect patterns observed in retinopathies and a high spatial resolution of the visual ield pattern, are usually irregular. While diabetic retinopathy results making this a challenging test for patients to undergo. in small patchy peripheral visual ield defects, retinal detachments and retinoschises result in one rather large Nevertheless, perimetry is a key test to assess visual cohesive defect, and retinitis pigmentosa shows a ring function in patients with retinopathies and therefore defect in early to moderate disease stages. Due to the continues to play a role in the management of retinal irregularity of these defect patterns, a testing pattern with diseases. Additionally, retinal diseases may occur in a high spatial distribution of test locations is necessary, combination with other common pathologies such as which by de inition is a more time-consuming test. To glaucoma, so a good understanding of retinal visual ield maximize perimetric information and minimize test defects remains essential. duration, kinetic perimetry should also be considered. This may be the most ef icient method of evaluating the far periphery. Testing patterns for retinopathies 71 TYPICAL VISUAL FIELD DEFECTS IN RETINAL DISEASES AMD DRUG-INDUCED MACULOPATHY 10° MACULA DIABETIC RETINOPATHY RETINITIS PIGMENTOSA 90° FULL FIELD FIGURE 5-9 Typical visual field defects in retinal diseases either affect the macula (e.g., AMD, drug-induced maculopathies) or are characterized by patchy, irregular loss affecting the full visual field (e.g., diabetic retinopathy, retinitis pigmentosa). TEST PATTERNS FOR THE MACULA M PATTERN The M pattern is the recommended pattern for macula The M pattern is most commonly used for the testing of visual ield evaluation.²⁵ With its 45 equally spaced test drug-induced maculopathies, to follow up advanced-stage locations, with 1° spacing in the fovea (central 4°), it of- glaucoma patients, and for visual function testing in fers the highest density of test locations in the most es- patients with AMD or other macular dysfunction. sential area for visual function.²⁵ The remaining 36 of the 81 test locations in total are radially arranged outside the fovea (FIG 5-10). 72 Chapter 5 | Selecting a test pattern 10-2 PATTERN The 10-2 pattern is the alternative to the M pattern, but The 10-2 grid is identical to that used on the Humphrey is not physiology-based (i.e., there is no emphasis on the Field Analyzer, thereby allowing for continuity when tran- fovea). Instead, all 76 test locations are equidistant, being sitioning from the Humphrey to the Octopus perimeter. separated from each other by 2° (FIG 5-10). Its test dura- tion is comparable to the M pattern. MACULA TEST PATTERNS M 10-2 81 test locations 76 test locations 90 90 180 10 0 180 10 0 270 10° 270 10° FIGURE 5-10 Both the M pattern and the 10-2 pattern are designed to exclusively test the macula. While the M pattern is physiology-based, with a high density of test locations in the fovea, the test locations are equidistant in the 10-2 pattern. TEST PATTERNS FOR THE FULL FIELD 07 PATTERN The 07 pattern is the recommended static testing pattern It provides reasonably high spatial resolution to be able to identify the patchy peripheral visual ield loss associ- to identify larger retinal lesions with a test duration that ated with a variety of retinal diseases, such as diabetic is acceptable for most patients. Nevertheless, the test retinopathy, retinoschisis, retinal detachment and retinitis duration using a quantitative strategy is long, so that a pigmentosa. qualitative test strategy offers a good trade-off between test duration and accuracy (more details on quanti- It has 130 test locations, extending from 70° temporally to tative and qualitative tests are provided in Chapter 6). 55° nasally, arranged radially with 15° spacing²⁶ (FIG 5-11). Testing patterns for retinopathies 73 FULL FIELD TEST PATTERNS FOR RETINOPATHIES 07 D 130 test locations 58 test locations 90 90 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 270 90° 270 90° FIGURE 5-11 Both the 07 and the D patterns focus on the periphery to detect common retinopathies affecting the periphery. The 07 pattern is recommended because it is more exhaustive, but with 130 test locations it is a long test. D PATTERN FOR DIABETIC RETINOPATHY The D pattern has been designed speci ically for diabetic shorter test duration than the 07 pattern, but it may retinopathy. With only 58 test locations, only extending miss smaller, localized patchy loss if used for diabetic to 50° in the periphery, it has a lower resolution than retinopathy in the early stages. the 07 pattern (FIG 5-11). This allows for a signi icantly KINETIC PERIMETRY Since static peripheral testing is time consuming and peripheral retinopathies. For more information, see provides only limited spatial resolution, kinetic pe- Chapter 11 on kinetic perimetry. rimetry is a very good alternative to static testing in 74 Chapter 5 | Selecting a test pattern TEST PATTERNS FOR VISUAL ABILITY TESTING Visual ability testing is often performed in a legal con- requirements and de ines test conditions in great detail, text, for example to assess a person’s visual ability to other legislation sets broader requirements. It is there- drive, eligibility for a pension or presence of visual fore essential to be familiar with the statutory require- disability. Therefore, highly standardized visual ield ments in one’s own country and to choose a testing tests are prescribed by local law and must be adhered to pattern that adheres to these regulations. strictly. While certain legislation sets very specific TEST PATTERNS FOR VISUAL ABILITY TO DRIVE Safe driving requires a large horizontal ield of view, to By law, in many countries visual ability to drive tests are be able to notice other cars coming from the side, and a mandatory to obtain and maintain a driver's license. fairly intact central ield of view, to be able to notice obsta- While the precise requirements differ according to local cles ahead. As driving is performed with both eyes open, legislation, often a visual ield test is required. While the the binocular ield of view is relevant for safe driving. legislation in some countries is rather vague, in other countries a speci ic test is requested. ESTERMAN TEST The Esterman test was developed by Ben Esterman²⁷ The Esterman test contains 120 test points. It horizon- and has become an accepted standard visual ield test tally spans 160°, and vertically from 30° superior to 60° for driving ability that is available in most modern inferior (FIG 5-12). It is typically a binocular test since perimeters. While this test must be used in countries that driving is undertaken binocularly, but a monocular require it by law, it is also commonly used in countries in version is also available. which there are broader statutory requirements. Testing patterns for visual ability testing 75 ESTERMAN TEST PATTERN 90 180 10 30 40 50 60 70 80 90 0 270 90° FIGURE 5-12 Driving ability tests such as the binocular Esterman test typically extend into the visual field area that can be seen through the front windscreen of a car. As this test has to meet legal requirements, the test pa- that are missed are marked with illed squares. The rameters are clearly outlined and similar for all perim- percentage of seen points relative to all points results eters. Each point is tested using a stimulus intensity of in the Esterman score (FIG 5-13). The Esterman score 1,000 asb on a background intensity of 31.4 asb. Points needed to ful ill driving requirements varies in different that are seen are marked with a plus sign and points legislations. ESTERMAN TEST Demo, John, 1/1/1945 (71yrs) ID 00001 Both eyes / 05/05/2015 / 16:.3:05 Symbols Points seen: 108 / 120 Points missed: 12 / 120 Esterman score: 90 80° Points seen: 108 / 120 Points missed: 12 / 120 Esterman score: 90 OCTOPUS® FIGURE 5-13 Print-out of a binocular Esterman test with the Esterman score. The Esterman score defines the percentage of points seen in relation to all points. In this example, 108 out of 120 points were seen, resulting in an Esterman score of 90%. 76 Chapter 5 | Selecting a test pattern ADDITIONAL DRIVING ABILITY TESTS Octopus perimeters also offer the German driving ability Some legislations also accept driving ability tests per- test FG (Führerscheingutachten). Additional driving ability formed with kinetic perimetry. For more information on test patterns can be created using the custom test function. kinetic perimetry, see Chapter 11. TEST PATTERNS FOR BLEPHAROPTOSIS Visual ield testing for blepharoptosis is performed in To objectively assess the potential bene its of blepha- order to objectively quantify the in luence of ptosis on roplasty for visual function, the affected eye is typically visual function. If it is signi icant, many insurance com- tested twice: once under normal conditions, and once panies accept blepharoplasty as a medically required with the lid taped up to mimic visual function after sur- surgery, instead of a cosmetic surgery, and will cover gery (FIG 5-15). The difference in the superior visual the cost. The acceptance criteria are not standardized ield between the taped and non-taped eye is then used and local legislation, as well as the respective insurance to determine the bene its of blepharoplasty. company, should be consulted. BT PATTERN FOR BLEPHAROPTOSIS The BT pattern is designed speci ically for blepharoptosis ield (FIG 5-14). As there is no vision underneath the lid testing and covers the area of the lid lines in the superior line, qualitative testing (seen, not seen) is suf icient. BLEPHAROPTOSIS TEST PATTERN 90 180 10 30 40 50 60 70 80 90 0 270 90° FIGURE 5-14 The BT pattern for blepharoptosis testing covers the area of the potential lid line. Testing patterns for visual ability testing 77 VISUAL FIELD TESTING FOR BLEPHAROPTOSIS Demo, John, 1/1/1945 (71yrs) Right eye (OD) / 03/21/2014 / 11:36:24 Symbols 90° Points seen: 21 / 87 Points missed: 66 / 87 Score [%]: 24 OCTOPUS® Demo, John, 1/1/1945 (71yrs) Right eye (OD) / 03/21/2014 / 16:53:12 Symbols 90° Points seen: 64 / 87 Points missed: 23 / 87 Score [%]: 74 OCTOPUS® FIGURE 5-15 Visual field testing for blepharoptosis is typically performed twice. Once under normal conditions and once with the lid taped up to mimic post-surgery condition. The difference between the two visual field tests determines the potential benefits of blepharoplasty for visual function. KINETIC PERIMETRY Since static peripheral testing is time-consuming, kinetic visual ield testing in blepharoptosis. For more informa- perimetry is a more time-ef icient method to perform tion, see Chapter 11 on kinetic perimetry. 78 Chapter 5 | Selecting a test pattern TEST PATTERN FOR VISUAL IMPAIRMENT In many countries, there is a pension system to support quality of life. Typically, test patterns for visual impair- visually impaired people. In order to determine a per- ment exhaustively test the central visual function and son’s eligibility for such a pension, an objective visual also extend into the periphery. function test is required that is related to a patient’s BG PATTERN The German examination to assess legal blindness, BG on legal requirements in Germany, but can also be useful (Blindengutachten) tests at 55 locations extending radi- in other countries in which the legislation is less speci ic. ally out to 55° (FIG 5-16). This test was designed based BG 90 180 10 30 40 50 60 70 80 90 0 270 90° FIGURE 5-16 The BG test pattern for visual impairment has 55 test locations and scans the entire visual field up to 55°. References 79 REFERENCES 1. Popovic Z, Sjöstrand J. Resolution, separation of retinal ganglion cells, and cortical magni ication in humans. Vision Res. 2001;41:1313-1319. 2. Curcio CA, Allen KA. Topography of ganglion cells in human retina. J Comp Neurol. 1990;300:5-25. 3. Stone J, Johnston E. The topography of primate retina: a study of the human, bushbaby, and new- and old-world monkeys. J Comp Neurol. 1981;196:205-223. 4. Rovamo J, Virsu V. An estimation and application of the human cortical magni ication factor. Exp Brain Res. 1979;37: 495-510. 5. Keltner JL, Johnson CA, Cello KE, et al. Classi ication of visual ield abnormalities in the ocular hypertension treatment study. Arch Ophthalmol. 2003;121:643-650. 6. Harrington DO, Drake MV. The Visual Fields: Text and Atlas of Clinical Perimetry. St. Louis: CV Mosby; 1990. 7. Weber J, Schultze T, Ulrich H. The visual ield in advanced glaucoma. Int Ophthalmol. 1989;13:47-50. 8. Rao HL, Begum VU, Khadka D, Mandal AK, Senthil S, Garudadri CS. Comparing glaucoma progression on 24-2 and 10-2 visual ield examinations. PLoS One. 2015;10: e0127233. 9. Sugimoto K, Schötzau A, Bergamin O, Zulauf M. Optimizing distribution and number of test locations in perimetry. Graefes Arch Clin Exp Ophthalmol. 1998;236:103-108. 10. Messmer C, Flammer J. Octopus program G1X. Ophthalmologica. 1991;203:184-188. 11. Flammer AJ, Jenni A, Bebie H. Keller B. The Octopus glaucoma G1 program. Glaucoma. 1987;9:67-72. 12. Hood DC, Nguyen M, Ehrlich AC, et al. A test of a model of glaucomatous damage of the macula with high-density perimetry: Implications for the locations of visual ield test points. Transl Vis Sci Technol. 2014;3:5. 13. Ehrlich AC, Raza AS, Ritch R, Hood DC. Modifying the conventional visual ield test pattern to improve the detection of early glaucomatous defects in the central 10°. Transl Vis Sci Technol. 2014;3:6. 14. Asaoka R. Mapping glaucoma patients' 30-2 and 10-2 visual ields reveals clusters of test points damaged in the 10-2 grid that are not sampled in the sparse 30-2 grid. PLoS One. 2014;9:e98525. 15. Asaoka R, Russell RA, Malik R, Crabb DP, Garway-Heath DF. A novel distribution of visual ield test points to improve the correlation between structure-function measurements. Invest Ophthalmol Vis Sci. 2012;53:8396-8404. 16. Gloor B, Gloor E. Detectability of glaucomatous visual ield defects with the Octopus automatic perimeter. A comparison between program G-1 and programs 31 and 32 and their combinations. Klin Monbl Augenheilkd. 1986;188:33-38. 17. Nowomiejska K, Wrobel-Dudzinska D, Ksiazek K, et al. Semi-automated kinetic perimetry provides additional information to static automated perimetry in the assessment of the remaining visual ield in end-stage glaucoma. Ophthalmic Physiol Opt. 2015;35:147-154. 18. Nevalainen J, Paetzold J, Krapp E, Vonthein R, Johnson CA, Schiefer U. The use of semi-automated kinetic perimetry (SKP) to monitor advanced glaucomatous visual ield loss. Graefes Arch Clin Exp Ophthalmol. 2008;246:1331-1339. 19. Scheuerle AF, Schiefer U, Rohrschneider K. Functional diagnostic options for advanced and end stage glaucoma. Ophthalmologe. 2012;109:337-344. 20. Caprioli J, Spaeth GL. Static threshold examination of the peripheral nasal visual ield in glaucoma. Arch Ophthalmol. 1985;103:1150-1154. 21. Drance SM, Susanna R, Fairclough M. Early defects in the visual ield in glaucoma (author's transl). Klin Monbl Augenheilkd. 1978;173:519-523. 22. Schulzer M, Mikelberg FS, Drance SM. A study of the value of the central and peripheral isoptres in assessing visual ield progression in the presence of paracentral scotoma measurements. Br J Ophthalmol. 1987;71:422-427. 23. Stewart WC, Shields MB. The peripheral visual ield in glaucoma: reevaluation in the age of automated perimetry. Surv Ophthalmol. 1991;36:59-69. 24. Rowe F. Visual ields via the visual pathway. Second ed. Boca Raton, FL: CRC Press, Taylor & Francis Group; 2016. 25. Kaiser HJ, Flammer J, Bucher PJ, De Natale R, Stümp ig D, Hendrickson P. High-resolution perimetry of the central visual ield. Ophthalmologica. 1994;208:10-14. 26. Graf M, Meienberg O. Octopus perimetry in neuro-ophthalmologic diseases. A contribution to the problem of optimal program choice based on 427 cases. Klin Monbl Augenheilkd. 1991;198:530-537. 27. Esterman B. Functional scoring of the binocular ield. Ophthalmology. 1982;89:1226-1234. 80 81 CHAPTER 6 SELECTING A TEST STRATEGY INTRODUCTION As illustrated in Chapter 4, determining sensitivity accurate but shorter test may yield more useful visual thresholds by assessing all levels of stimulus intensity ield results. (e.g., stimulus luminance) is not practical because of the time it would require. Several strategies have therefore Another important factor is the reason for which the been developed to minimize test duration while maxi- test is being performed. For example, in order to detect mizing clinically relevant information. Some strategies and follow pathologies such as glaucoma, it is important are quantitative, providing a good estimate of the local to be able to detect small changes in sensitivity thresh- sensitivity thresholds, while others are qualitative and olds with high accuracy. To achieve this, an accurate can only determine whether a stimulus of a given inten- quantitative strategy is needed. On the other hand, sity is seen or not. areas with no clinically meaningful information such as the blind spot or the area under the lid in ptosis test- The optimal strategy for a given test situation depends ing can be identi ied equally well with a qualitative test on a number of factors. The patient’s ability to reliably that simply determines whether stimuli are seen or not complete the test is a crucial factor. A test that is designed seen. Qualitative test strategies are also often suf icient to be very accurate can lead to inaccurate perimetric re- in legal performance ability tests to assess, for example, sults if the patient is only able to perform reliably during whether someone ful ills the visual ield requirements a portion of the test. In such situations, a potentially less to drive. TABLE 6-1 summarizes the differences between qualitative and quantitative testing strategies. 82 Chapter 6 | Selecting a test strategy CHARACTERISTICS OF QUANTITATIVE AND QUALITATIVE STRATEGIES TABLE 6-1 QUANTITATIVE STRATEGIES QUALITATIVE STRATEGIES (THRESHOLDING) (SCREENING) ADVANTAGES Quanti ication of sensitivity Faster compared to most thresholds and visual ield losses quantitative strategies Higher accuracy WHAT IT IS BEST AT DETECTING Small changes in sensitivity Normal versus abnormal vision Compliance with legally required Ocular and neurologic pathologies visual ield criteria (e.g., ability to drive) Absolute defects COMMON USES All pathologies Visual ability tests (e.g., glaucoma, age-related macular (e.g., driving, legal blindness, ptosis) degeneration) Pathologies with predominantly absolute defects (e.g., blind spot enlargement, certain neuro- ophthalmological and retinal pathologies) Screening of eyes for abnormality Finally, it is also important to consider patient comfort. All quantitative and qualitative test strategies available Tests of short duration are easier to perform and may on the Octopus perimeters are described in more detail motivate patients to come regularly for follow-up testing. in the following sections. QUANTITATIVE STRATEGIES Quantitative sensitivity threshold strategies are used to offer higher accuracy, but it also has longer test duration. obtain sensitivity thresholds at various locations within the visual ield. They are commonly used to detect and In the second type of quantitative sensitivity threshold follow pathological visual ield defects. One exception is the strategy, predetermined estimates (e.g., educated guesses) detection and follow-up of pathologies that result in sharp are made about the sensitivity thresholds at each location absolute defects such as blind spot enlargements, which based on information obtained from other neighboring can be equally well detected with a qualitative strategy. visual ield locations. Systematic sampling at each lo- cation is not performed. This approach is used in the Two types of quantitative sensitivity threshold strategies Tendency-Oriented Perimetry (TOP) strategy, a shorter are available. In the irst type, there is a systematic sampling test with reduced accuracy in some situations. of the entire range of light intensities in large steps, with further detailing within the expected sensitivity threshold The characteristics of these strategies are summarized in zone using smaller steps. This approach is designed to TABLE 6-2 and are further detailed in the next paragraphs. Quantitative sensitivity threshold strategies 83 CHARACTERISTICS OF THE TOP, DYNAMIC, LOW VISION AND NORMAL STRATEGIES TABLE 6-2 TOP DYNAMIC LOW VISION NORMAL TEST DURATION* 2-4 minutes 6-8 minutes 6-8 minutes 10-12 minutes WHAT IT IS BEST Contiguous defects Contiguous defects Contiguous defects Contiguous defects AT DETECTING (central 30°) Isolated defects Isolated defects Isolated defects Peripheral defects Peripheral defects Peripheral defects Mild sensitivity Sensitivity Mild sensitivity threshold changes thresholds with low threshold changes sensitivity BEST SUITED FOR Patients struggling Patients with mild Low vision patients Patients with with fatigue changes in sensitivity excellent thresholds cooperation, Busy practices attention and Patients with good minimal fatigue cooperation and attention COMMON USES Glaucoma Glaucoma Low vision Clinical research Macula Macula Periphery (Neuro, Retina) METHODOLOGY Spatial relationship Sampling with Sampling with 4 dB Sampling with 4-2-1 among sensitivity increasing step size step size dB step size thresholds of (2 – 10 dB) Start at 0 dB neighboring zones sensitivity ACCURACY IN dB n/a From ± 1 dB (normal ± 2 dB ± 1 dB vision) to ±5 dB (Low vision) *Test duration estimates are provided for the 30° G pattern with 59 test locations. NORMAL STRATEGY The normal strategy was the irst quantitative testing projects or used to detect early and shallow defects in strategy built into Octopus perimeters. It provides the younger patients who have the endurance necessary to determination of sensitivity thresholds with an accura- perform reliably on longer tests. cy of about 1 dB.¹,² This strategy takes approximately 10 to 12 minutes per eye for the G pattern.³ Due to its rel- The normal strategy uses a 4-2-1 dB sampling procedure atively long test duration and the availability of quicker to determine sensitivity thresholds. In this sampling pro- tests, this strategy is no longer recommended for stan- cedure, stimuli are irst presented in 4 dB steps to ind dard testing. The long test duration can lead to fatigue, the threshold zone. Further detailing occurs in 2 dB steps and many patients show signi icantly reduced reliability and the sensitivity threshold is determined as the in spite of the higher accuracy of this strategy. It is still average between the last seen and not seen stimuli. available, however, and can be useful in clinical research 84 Chapter 6 | Selecting a test strategy NORMAL STRATEGY Do you see? 32 dB 1 20 dB 6 28 dB = Seen 24 dB 2 Threshold 19 dB = Not seen 20 dB 3 Threshold Zone 18 dB 5 16 dB 4 12 dB 8 dB 16 dB 4 dB 0 dB 1. Sampling in 2. Detailing with 2 dB 4 dB steps step in threshold zone FIGURE 6-1 An example of the procedure used in the normal strategy is presented. The 4-2-1 bracketing procedure of the normal strategy proceeds by first presenting stimuli in 4 dB steps (stimuli 1 to 4) to find the threshold zone, then further details in 2 dB steps (stimuli 5 and 6), and finally determines the sensitivity threshold as the average between the last seen and not seen stimuli. Because the widely used dynamic strategy is a variation normal strategy, outlined in BOX 6A, is helpful for grasp- of the normal strategy, understanding the details of the ing the dynamic strategy. BOX 6A THE DETAILS OF THE NORMAL STRATEGY The normal strategy initially tests one anchor point location in each of the four quadrants to determine sensitivity thresholds at one position in each quadrant. Using this information as an initial stimulus for neighboring locations, it then uses a 4-2-1 dB sampling procedure. This sampling procedure is also referred to as bracketing, and is performed using the staircase procedure in which two response reversals ( irst from “not seen” to “seen” and then from “seen” to “not seen”) are required. For example, the test begins by presenting a stimulus at an intensity that corresponds to a given sensitivity threshold (e.g., 28 dB). If this stimulus is not seen, the next stimuli are presented in decreasing 4 dB steps, until the stimulus is seen (e.g., 16 dB; FIG 6-1). At this point, the procedure switches to a second crossing of the threshold, but now in steps of 2 dB. The initial stimulus of that sequence is presented at 18 dB. If it is "seen", the following stimuli are presented in increasing 2 dB steps until "not seen" (second response reversal); however, if the 18 dB stimulus is "not seen", the following stimuli are presented in descend- ing 2 dB steps until "seen". In both cases, the sensitivity threshold is calculated as the mean of the last “seen” and “not seen” stimuli (FIG 6-1). It is expressed in dB with an uncertainty of approximately ±1 dB. Except for the anchor points, the level of the initial stimulus is determined from the results already ob- tained at neighboring test locations, in order to further reduce test duration. It is important to note that even though information from neighboring test locations is used, each sensitivity threshold is deter- mined independently of the neighboring sensitivity thresholds with the sampling procedure described above. Typically, the procedure requires about 4 - 5 stimuli per test location. It is possible for the patient’s sensitivity threshold to be above the level of the initial stimulus. This occurs when the irst stimulus presented is seen. In this situation, the next stimulus is presented in increasing 4 dB steps until "not seen". The rules for the second crossing of the threshold remain the same. Quantitative sensitivity threshold strategies 85 DYNAMIC STRATEGY The dynamic strategy is a widely used procedure because In the dynamic strategy, the determination of the level of it offers an excellent trade-off between test duration and the irst stimulus at a given test location follows the same accuracy.⁴,⁵ It provides detailed information about each rules as the normal strategy (i.e., anchor points and in- visual ield location and has been shown to detect early formation from neighboring locations). Test time is saved visual ield loss and isolated visual ield defects reliably.⁶ mainly because the sensitivity threshold is crossed only It is also a relatively quick test, taking an average of 6 to 8 once (i.e., only one reversal). Depending on whether the minutes per eye when using the G pattern.⁷ Furthermore, irst stimulus is seen or not, the next stimulus is present- this strategy can be used with all test patterns. ed in increasing or decreasing steps until the threshold is crossed. The threshold is determined as the average The dynamic strategy is based on the normal strategy, between the last seen and unseen stimuli. In areas of the but it has been optimized to shorten test duration while visual ield that are near the normal range, an accuracy of missing only a minimal amount of clinically relevant in- approximately ±1 dB is achieved to support early disease formation.⁴,⁸ Similar to the normal strategy, it narrows in detection. In areas of advanced defects, an accuracy of on the sensitivity threshold by using a modi ied staircase approximately ±5 dB is achieved. sampling procedure. While the sensitivity thresholds may not be as accurate In comparison to the normal strategy, the dynamic strat- as those obtained using the normal strategy in more ad- egy step size is smaller in regions of normal sensitivity vanced disease, various clinical studies have shown that and larger in areas where defects are present, ranging the dynamic strategy is adequate for low-vision patients. from 2 dB to 10 dB (FIG 6-2). This saves considerable This is because more accurate testing is not possible due time when signi icant visual ield loss is present. The to an increase in luctuation.⁴,⁸,¹⁰ variable step size is justified, as fluctuation has been shown to increase with increasing defect depth.⁹ Testing can therefore be performed using a step size tailored to the degree of luctuation.⁴ 86 Chapter 6 | Selecting a test strategy DYNAMIC STRATEGY Do you see? 32 dB = Seen 28 dB 1 = Not seen 24 dB 2 20 dB 3 16 dB Threshold Threshold Zone 14.5 dB 12 dB 4 8 dB 4 dB 0 dB Sampling with increasing step size FIGURE 6-2 An example of the procedure used in the dynamic strategy is presented. The dynamic strategy samples with increasing step size (from 2 to 10 dB from normal sensitivity threshold) after the first stimulus is not seen until a stimulus is seen without any further detailing. As a result, the accuracy is between ±1 and ±5 dB, depending on the step size. LOW-VISION STRATEGY The low-vision (LV) strategy is useful for assessing pa- longer because more stimuli would be presented in lo- tients with end-stage diseases, when only limited visual cations where there is no sensitivity. ield function remains. It employs a methodology similar to the normal strategy, but only performs one threshold Besides saving test time, the low-vision strategy is also crossing (4-2 bracketing), which reduces test duration. more patient-friendly for low-vision patients, because While an accuracy of only approximately ±2 dB can be starting the test with the maximum stimulus intensity achieved, this is justi ied by the large luctuation in increases the likelihood that the initial stimulus will be areas of low vision.⁹ In addition, the low-vision strat- seen. This allows patients to feel con ident about their egy starts testing at a sensitivity threshold of 0 dB (FIG performance in the initial stage of the test. 6-3). This means that the initial stimulus used is at the maximum stimulus intensity because of the inverse re- In order to extend the dynamic testing range into the lationship between light intensity and sensitivity thresh- low-vision area and also to make the target easier to see old, as outlined in FIG 2-2. This approach further reduces for low-vision patients, the low-vision strategy is typically test duration when a visual ield contains a large number used in combination with a stimulus size V that is pre- of locations with absolute defects. For such situations, sented for 200 ms (see Chapter 10 for more information testing with the dynamic or normal strategy would take on stimulus size V for low-vision patients). Quantitative sensitivity threshold strategies 87 LOW-VISION STRATEGY Do you see? 32 dB = Seen 28 dB = Not seen 24 dB 20 dB 16 dB 12 dB 8 dB 3 Threshold Threshold Zone 6 dB 4 dB 2 0 dB 1 Sampling in 4 dB steps FIGURE 6-3 The low-vision strategy is optimized for low-vision patients and is a variation of the normal strategy. Significant test time is saved by crossing the threshold only once. Patient confidence is also increased by starting at a sensitivity threshold of 0 dB, the maximum stimulus intensity available on the device. TENDENCY-ORIENTED PERIMETRY (TOP) STRATEGY The Tendency-Oriented Perimetry (TOP) strategy is related to the sensitivity thresholds at nearby locations a widely used and fast procedure. It takes only two to (i.e., there is a spatial correlation among adjacent test lo- four minutes per eye for a complete sensitivity thresh- cations). During the test, answers at any given location old examination with the G pattern.³,¹¹-¹³ Because of its are taken into account to adjust the expected sensitivity short duration, it is especially recommended for patients thresholds at neighboring locations. The test starts by unable to maintain concentration for long periods, such presenting stimuli at 50% of normal sensitivity thresh- as the neurologically impaired or children.¹⁴ For these olds at a quarter of the test locations. If the stimulus at patients, fatigue or lack of concentration in a longer test a certain location is missed, the stimuli at immediately would lead to unreliable results.¹⁵,¹⁶ The TOP strategy is adjacent locations are presented at lower sensitivity also useful as a practical routine method for testing and thresholds. However, if the stimulus is seen, the initial following patients of all age groups, especially in busy stimuli at neighboring locations are presented at higher practices.¹⁶ sensitivity thresholds. This procedure is repeated for all test locations with answers from neighboring test loca- The TOP strategy takes advantage of the fact that sensi- tions leading to an adaptation of all test locations. See tivity thresholds at each location of the visual ield are BOX 6B for more details. 88 Chapter 6 | Selecting a test strategy BOX 6B TOP STRATEGY – STEP-BY-STEP PROCEDURE STEP 0 Normal • Baseline: normal Sensitivity Threshold sensitivities at each test location 27 28 28 28 29 30 31 31 31 28 29 29 29 29 28 • Test pattern divided into 29 30 31 31 31 31 30 30 4 sub-test patterns 30 31 31 32 32 29 30 31 31 32 32 32 31 30 30 29 30 32 32 33 33 32 BS 31 30 30 32 32 33 33 30 30 32 32 33 33 32 BS 31 30 30 30 31 32 32 32 32 31 30 30 30 32 32 33 33 28 30 31 31 31 31 30 30 28 29 30 30 30 29 30 31 32 32 32 29 29 29 29 STEP 1A STEP 1B • Submatrix 1 = ½ of Submatrix 1 Response Matrix 1 • Seen: Add ¼ of normal sensitivity normal sensitivity • Stimulus presentation 15 15 16 16 16 7 8 8 0 -8 • Not seen: Substract ¼ on 1st sub-test pattern of normal sensitivity 15 16 16 16 16 8 8 8 0 -8 STEP 1C 15 16 16 17 17 8 8 8 0 -8 Calculate responses for sub-test patterns 2-4 15 16 16 17 17 8 8 8 4 0 from average of neighboring locations 15 16 16 16 16 8 8 8 8 8 by interpolation STEP 2A STEP 2B • Submatrix 2 = Submatrix 1 Submatrix 2 Response Matrix 2 • Seen: Add 3/16 of + Response Matrix 1 normal sensitivity • Stimulus presentation on 22 23 24 16 8 5 5 3 0 -3 • Not seen: Substract 3/16 2nd sub-test pattern of normal sensitivity 23 24 24 16 8 5 6 6 6 0 STEP 2C 23 24 24 17 9 6 6 6 6 0 Calculate responses for sub-test patterns 1, 3, 4 23 24 24 21 17 6 6 6 6 0 from average of neighboring locations 23 24 24 24 24 6 6 6 6 3 by interpolation STEP 3A STEP 3B • Submatrix 3 = Submatrix 2 Submatrix 3 Response Matrix 3 • Seen: Add 2/16 of + Response Matrix 2 normal sensitivity • Stimulus presentation on 27 28 27 16 5 2 0 0 -2 -4 • Not seen: Substract 2/16 3rd sub-test pattern of normal sensitivity 28 30 30 22 8 4 4 4 0 -4 STEP 3C 29 30 30 23 9 4 4 4 2 0 Calculate responses for sub-test patterns 1, 2, 4 29 30 30 27 17 4 4 4 4 4 from average of neighboring locations 29 30 30 30 27 4 4 4 4 4 by interpolation Quantitative sensitivity threshold strategies 89 STEP 4A STEP 4B • Submatrix 4 = Submatrix 3 Submatrix 4 Response Matrix 4 • Seen: Add 1/16 of + Response Matrix 3 normal sensitivity • Stimulus presentation on 29 28 27 14 1 -2 -2 -2 -2 -2 • Not seen: Substract 1/16 4th sub-test pattern of normal sensitivity 32 34 34 22 4 -2 -2 -1 0 0 STEP 4C 33 34 34 25 9 -2 -2 0 2 2 Calculate responses for sub-test patterns 1 – 3 33 34 34 31 21 -2 -2 0 2 2 from average of neighboring locations 33 34 34 34 31 -2 -2 0 2 2 by interpolation STEP 5 • Sensitivity threshold = Sensitivity Threshold Submatrix 4 + Response 14 3 0 0 Matrix 4 27 26 25 12 0 22 16 3 0 0 0 26 25 23 11 0 0 0 2 30 32 33 22 4 28 29 30 31 21 4 0 0 4 7 29 30 31 34 26 10 4 BS 5 7 31 32 34 27 11 29 30 31 34 33 22 16 BS 13 13 29 30 30 33 35 32 29 26 24 21 31 32 34 33 23 29 29 31 32 32 32 3 30 28 28 29 29 29 28 31 32 34 36 33 27 28 28 28 Because the initial stimuli are presented at a lower sensi- increased accuracy,¹⁵ this strategy also has some short- tivity threshold – and thus higher light intensity (see FIG comings related to accuracy. The TOP strategy can reliably 2-2 for the inverse relationship between light intensity detect large contiguous scotomas such as those present and sensitivity threshold) – in the TOP strategy than in in glaucoma.¹³,¹⁶,¹⁷ However, it smoothes the edges of the dynamic or normal strategies, the likelihood of see- the scotomas¹⁸ (FIG 6-4) and is less sensitive to small, ing most of the initial stimuli is increased. This allows localized defects compared to a systematic sampling pro- patients to feel con ident about their performance for cedure such as the dynamic strategy.⁶,¹¹,¹² These factors the initial stage of the test, resulting in a shorter patient should be kept in mind when making clinical decisions. learning curve, increased reliability for the initial exam- inations, and possibly greater motivation for patients to In addition, the TOP strategy requires a reasonably dense return for follow up testing. test grid to justify the assumption that there is a spatial correlation between the test points. Therefore, it is only While the advantages of the TOP strategy, in terms of available for the central 30° patterns G, 32, 30-2 and 24-2 reduction of test duration and fatigue effects, can lead to and the macula test patterns M and 10-2. 90 Chapter 6 | Selecting a test strategy SPATIAL RESOLUTION OF NORMAL, DYNAMIC AND TOP STRATEGY NORMAL DYNAMIC TOP 12 10-12 minutes 12 6-8 minutes 12 2-4 minutes 9 3 9 3 9 3 6 6 6 FIGURE 6-4 When choosing a test strategy, there is a trade-off between test duration and accuracy as the example above illus- trates. The same patient was tested with the G pattern and the normal (left), dynamic (center) and TOP (right) strategies. Note that while all strategies allow the identification of a double arcuate defect, the visual field measured with the TOP strategy shows the defect as shallower with smoother edges, but it also saves considerable test time. QUALITATIVE STRATEGIES Qualitative strategies are useful and time-ef icient when assessing the vision of patients with certain retinal pa- the quanti ication of a patient’s sensitivity threshold is thologies. Finally, they can be used to quickly screen not necessary. These strategies are used for visual ield patients with assumed normal vision. performance ability testing including driving,¹⁹ legal blindness and ptosis examinations. They are also some- The answers obtained with these strategies are always times used for pathologies that result in absolute defects. qualitative (e.g., seen/not seen or normal visual ield/ For example, qualitative strategies can be used to assess abnormal visual ield). Octopus perimeters offer several the vision of patients with neurological conditions that qualitative strategies for different purposes, as described result in hemianopia, quadrantanopia²⁰ and blind spot below. enlargements. Furthermore, they can also be useful in Qualitative strategies 91 1-LEVEL TEST STRATEGY (TWO-ZONE STRATEGY) The 1-Level Test (1LT) is a fast test strategy most com- used to assess absolute visual ield defects such as blind monly used for legal performance ability tests to assess spot enlargements or the lid margin in blepharoptosis whether someone ful ills the minimal visual ield criteria testing (FIG 6-5). to drive or to perform other daily tasks. In addition, it is EXAMPLES OF THE USE OF THE 1-LEVEL TEST STRATEGY IN DIFFERENT SITUATIONS BLEPHAROPTOSIS BLIND SPOT BINOCULAR ESTERMAN 90 90 90 180 10 30 40 50 60 70 80 90 0 180 0 180 10 30 40 50 60 70 80 90 0 90° 30° 90° 270 270 270 FIGURE 6-5 Three examples of visual field tests performed with a 1LT strategy. For absolute defects such as the lid margin in blepharoptosis testing (left) or blind spot testing (center), the 1LT strategy provides sufficient information to delineate the boundaries. In the Esterman test (right), it is used to determine the percentage of test locations missed and provides informa- tion about a patient’s ability to drive. With the 1LT strategy, only one stimulus is presented at locations with the “+” sign can be interpreted as normal each test location at the predetermined intensity level and those with the “ ”symbol as abnormal. of 6 dB below the normal sensitivity threshold. The pa- tient either sees or misses these stimuli (FIG 6-6). The Note that typically more than one abnormal visual ield visual ield is consequently divided into two zones, seen location showing a disease-speci ic pattern is required (represented by a “+” sign) and not seen (represented to classify a visual ield as abnormal. For more detailed by a “ ”symbol). As a result, this strategy is sometimes information on the distinction between normal and ab- referred to as a two-zone strategy. Clinically, visual ield normal visual ields, see FIG 7-9, 7-10, 8-14 and 8-15.. 92 Chapter 6 | Selecting a test strategy 1-LEVEL TEST STRATEGY Do you see? Normal Abnormal visual field location visual field location Normal Stimulus presented at Abnormal 6dB below normal sensitivity threshold = Seen = Not seen 0 dB 0 dB FIGURE 6-6 With the 1LT strategy, the visual field is divided into areas of likely normal and abnormal visual field locations. SCREENING STRATEGY The screening strategy is used to quickly distinguish be- with a routine eye examination. In that context, the tween people with normal and abnormal visual ields. screening strategy together with the GS pattern offers a It is a very fast strategy and patients with normal visual very good trade-off between testing time and accuracy ields can typically complete it within one minute. It is by being fast while at the same time allowing the identi- designed to be used with the G-Screening (GS) pattern. ication of patients with abnormal visual ields. If a visual ield abnormality is detected during routine screening, The screening strategy is useful and time-effective when further quantitative testing is recommended to assess routine visual ield testing is performed on every patient the extent and depth of visual ield loss (FIG 6-7). to identify pathologies that would otherwise be missed Qualitative strategies 93 EXAMPLE OF THE USE OF THE SCREENING STRATEGY SCREENING TEST QUANTITATIVE TEST Screening strategy Dynamic strategy 12 G-Screening pattern 12 G pattern 9 3 9 3 6 6 + + + + + + + + + + + + + + + + + + + + + + + + + FIGURE 6-7 The routine testing with the screening strategy and the short G-Screening pattern allows the identification of poten- tially unnoticed visual field defects with minimal test duration. This can identify patients who require a more detailed evaluation. In the case presented above, a patient with early glaucoma was detected during routine screening, and consequent quantitative visual field testing confirmed the existence of the previously unnoticed partial arcuate defect. The screening strategy is a modi ied version of the 1-Level with visual ield testing. If the patient sees the stimulus Test strategy (FIG 6-8). The irst stimulus at each location on any of these repetitions, the location is designated as is presented at the intensity that an average subject with normal (represented by a “+” sign), otherwise it is re- a normal visual ield would see 95% of the time. If the corded as abnormal (represented by a “ ” sign). Because point is seen, the location is designated as normal. If it unseen points are tested three times, the likelihood of is not seen however, the same point is repeated twice to obtaining false negative errors is reduced. This approach con irm suspected abnormalities and to avoid false nega- results in better speci icity for the screening test.²¹ tive errors, which are common in patients inexperienced EXAMPLE OF THE USE OF THE SCREENING STRATEGY Do you see? Normal Normal Abnormal visual field location visual field location visual field location Normal Stimulus intensity Abnormal that people with normal visual fields see 95% of time = Seen = Not seen 0 dB 0 dB 0 dB FIGURE 6-8 The screening strategy distinguishes between normal and abnormal visual fields by presenting a stimulus at an intensity that people with normal visual fields would see 95% of the time. 94 Chapter 6 | Selecting a test strategy 2-LEVEL TEST STRATEGY (THREE-ZONE STRATEGY) The 2-Level Test (2LT) strategy is similar to the 1LT, but fatigue. In such situations, the 2LT strategy offers a good presents stimuli at two sensitivity thresholds. It conse- trade-off between test duration and accuracy, often with quently divides the visual ield into three zones and is only minimum loss of clinical information (FIG 6-9). thus also referred to as a three-zone strategy. The 2LT strategy provides only a rough indication of the The 2LT strategy is commonly used as an alternative to status of the visual ield. It is designed to distinguish the quantitative dynamic strategy to assess the full visual between areas of normal visual ield, areas with relative ield in pathologies that result in hemianopia, quadran- defects (i.e., with reduced sensitivity thresholds) and tanopia or certain retinal conditions such as diabetic areas with absolute defects (i.e., with a sensitivity thresh- retinopathies, retinal detachments and retinoschisis. old of 0dB, where the maximum stimulus intensity of Because these pathologies affect a signi icant portion of the perimeter cannot be seen). This information is often the visual ield and often contain a meaningful number of clinically suf icient to identify diseases whose diagnosis absolute defects, quantitative testing with a reasonably is based on the shape of the defect rather than on small dense test pattern can be too time-consuming and conse- changes in sensitivity. quently lead to unreliable results in some patients due to QUALITATIVE TEST WITH 2LT STRATEGY VS QUANTITATIVE TEST QUALITATIVE TEST QUANTITATIVE TEST 2 LT strategy Dynamic strategy 12 07 pattern 12 07 pattern 9 3 9 3 6 6 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + FIGURE 6-9 The qualitative 2LT strategy allows distinction of areas of normal vision and relative and absolute defects with minimal test duration (left) and provides valuable information about the shape of defects. Quantification of the same test pattern with the dynamic strategy reveals the same pattern as well as detailed information about the sensitivity thresholds, but at the expense of a longer test (right). In this example, because the visual field shows an absolute defect, the amount of information provided by the 2LT strategy is comparable to the more detailed information provided by the quantitative dynamic strategy. Qualitative strategies 95 The 2LT strategy starts with the presentation of a stim- seen, it is marked as having an absolute defect (repre- ulus 4dB below the normal sensitivity threshold at each sented by a “ ” sign). location. When this stimulus is seen, it is designated as normal (represented by a “+” sign). When this stimulus is Note that typically more than one abnormal visual ield not seen, a second stimulus is presented at a sensitivity location showing a disease-speci ic pattern is required threshold of 0 dB, which corresponds to the maximum to classify a visual ield as abnormal. For more detailed stimulus intensity that the perimeter can present (FIG information on the distinction between normal and ab- 6-10). If this is seen, the location is marked as having a normal visual ields, see FIG 7-9, 7-10, 8-14 and 8-15. relative defect (represented by a “ ” sign) and if it is not 2-LEVEL TEST STRATEGY Do you see? Normal visual Relative Absolute field location defect defect Normal Stimulus presented at 4dB below Relative defect normal sensitivity threshold = Seen = Not seen Absolute defect 0 dB 0 dB 0 dB FIGURE 6-10 With the 2LT strategy, the visual field is divided into areas of normal visual field results, relative defects and absolute defects. 96 Chapter 6 | Selecting a test strategy RECOMMENDATIONS ON KEY EXAMINATION PARAMETERS There is always a trade-off between test duration and recommended combinations of test patterns and strat- accuracy, but depending on the pathology or visual abil- egies for a variety of visual ield testing situations that ity test performed, certain test parameter combinations maximize clinical information and minimize test duration. offer better trade-offs than others. TABLE 6-3 presents RECOMMENDED TEST PATTERN AND STRATEGY COMBINATIONS TABLE 6-3 TEST PATTERN RECOMMENDED STRATEGIES GLAUCOMA/CENTRAL FIELD G, 32, 30-2, 24-2 Dynamic or TOP MACULA M, 10-2 Dynamic or TOP FULL FIELD (NEURO, RETINA) 07 Dynamic or 2LT FOVEA Fovea Dynamic BLIND SPOT Blind spot 1LT LOW VISION M, G, 07 depending on pathology Low Vision SCREENING FOR G-Screening Screening ABNORMAL VISION DRIVING Esterman Fixed parameters (1LT, stimulus duration 500 ms) German Legal Driving Fixed parameters (Führerscheingutachten FG) (2LT, stimulus duration 200ms) BLEPHAROPTOSIS BT 1LT BLINDNESS BG Fixed parameters (1LT, stimulus duration 500ms) This does not mean that these settings are the best for close to each other, it can only be used with the suf icient- each visual ield test; an expert user may prefer other ly dense central 30° and macula test patterns; 2) legally combinations for certain situations. Therefore, Octopus prescribed ability tests such as the Esterman test are perimeters offer the lexibility to customize examination offered only with their standardized settings to ensure parameters. However, there are two exceptions: 1) because that the legal requirements are met. the TOP strategy requires test locations to be relatively References 97 REFERENCES 1. Bebie H, Fankhauser F, Spahr J. Static perimetry: strategies. Acta Ophthalmol (Copenh). 1976; 54:325-38. 2. Spahr J. Optimization of the presentation pattern in automated static perimetry. Vision Res. 1975;15:1275-1281. 3. Gonzáles de la Rosa M, Morales J, Dannheim F, et al. Multicenter evaluation of tendency-oriented perimetry (TOP) using the G1 grid. Eur J Ophthalmol. 2003;13:32-41. 4. Weber J, Klimaschka T. Test time and ef iciency of the dynamic strategy in glaucoma perimetry. Ger J Ophthalmol. 1995;4:25-31. 5. Weber J. A new strategy for automated static perimetry. Fortschr Ophthalmol. 1990;87:37-40. 6. Maeda H, Nakaura M, Negi A. New perimetric threshold test algorithm with dynamic strategy and tendency oriented perimetry (TOP) in glaucomatous eyes. Eye (Lond). 2000;14:747-751. 7. Zein WM, Bashshur ZF, Jaafar RF, Noureddin BN. The distribution of visual ield defects per quadrant in standard automated perimetry as compared to frequency doubling technology perimetry. Int Ophthalmol. 2010;30:683-689. 8. Zulauf M, Fehlmann P, Flammer J. Perimetry with normal Octopus technique and Weber 'dynamic' technique. Initial results with reference to reproducibility of measurements in glaucoma patients. Ophthalmologe. 1996;93:420-427. 9. Wall M, Woodward KR, Doyle CK, Artes PH. Repeatability of automated perimetry: a comparison between standard automated perimetry with stimulus size III and V, matrix, and motion perimetry. Invest Ophthalmol Vis Sci. 2009;50:974-979. 10. Gardiner SK, Swanson WH, Goren D, Mansberger SL, Demirel S. Assessment of the reliability of standard automated perimetry in regions of glaucomatous damage. Ophthalmology. 2014:121:1359-1369. 11. King AJ, Taguri A, Wadood AC, Azuara-Blanco A. Comparison of two fast strategies, SITA Fast and TOP, for the assessment of visual ields in glaucoma patients. Graefes Arch Clin Exp Ophthalmol. 2002;240:481-487. 12. Morales J, Weitzman ML, Gonzáles de la Rosa M. Comparison between Tendency-Oriented Perimetry (TOP) and octopus threshold perimetry. Ophthalmology. 2000;107:134-142. 13. Wadood AC, Azuara-Blanco A, Aspinall P, Taguri A, King AJ. Sensitivity and speci icity of frequency-doubling technology, tendency-oriented perimetry, and Humphrey Swedish interactive threshold algorithm-fast perimetry in a glaucoma practice. Am J Ophthalmol. 2002;133:327-332. 14. Morales J, Brown SM. The feasibility of short automated static perimetry in children. Ophthalmology. 2001;108:157-162. 15. Gonzalez-Hernandez M, Morales J, Azuara-Blanco A, Sanchez JG, de la Rosa MG. Comparison of diagnostic ability between a fast strategy, tendency-oriented perimetry, and the standard bracketing strategy. Ophthalmologica. 2005;219:373-378. 16. Scherrer M, Fleischhauer JC, Helbig H, Johann Auf der Heide K, Sutter FK. Comparison of tendency-oriented perimetry and dynamic strategy in octopus perimetry as a screening tool in a clinical setting: a prospective study. Klin Monbl Augenheilkd. 2007;224:252-254. 17. Pierre-Filho Pde T, Schimiti RB, de Vasconcellos JP, Costa VP. Sensitivity and speci icity of frequency-doubling technology, tendency-oriented perimetry, SITA Standard and SITA Fast perimetry in perimetrically inexperienced individuals. Acta Ophthalmol Scand. 2006;84:345-350. 18. Anderson AJ. Spatial resolution of the tendency-oriented perimetry algorithm. Invest Ophthalmol Vis Sci. 2003;44: 1962-1968. 19. Esterman B. Functional scoring of the binocular ield. Ophthalmology. 1982;89:1226-1234. 20. Siatkowski RM, Lam BL, Anderson DR, Feuer WJ, Halikman AM. Automated suprathreshold static perimetry screening for detecting neuro-ophthalmologic disease. Ophthalmology. 1996;103:907-917. 21. Turpin A, Myers JS, McKendrick AM. Development of visual ield screening procedures: a case study of the Octopus perimeter. Transl Vis Sci Technol. 2016; doi: 10.1167/tvst.5.3.3. 98 99 CHAPTER 7 OVERVIEW OF VISUAL FIELD REPRESENTATIONS INTRODUCTION Perimetry determines sensitivity thresholds through- For these reasons, Octopus perimeters offer several re- out the visual ield. However, it can be challenging to presentations that are based on the measured sensitivity correctly interpret the raw data in clinical practice be- thresholds, but highlight speci ic aspects of the visual cause 1) normal sensitivity thresholds vary with age ield, in order to support clinical decision-making. In and eccentricity of test location; 2) visual ield testing this chapter, a systematic presentation of all visual ield contains a subjective component due to the patient de- representations and indices is offered, with their de ini- cision processes, which contributes to luctuation; 3) tions, design and relationships. While a detailed under- both visual ield location and disease severity in luence standing of these characteristics is not necessary for fluctuation; and 4) in some patients, more than one correct clinical interpretation, some readers will ind disease may be present. For more information on these this information useful. The clinical meaning and inter- points, see Chapter 2. pretation of these same visual ield representations are subsequently discussed in a clinical step-by-step work- low in Chapter 8, which also includes several examples. RELATIONSHIP AMONG OCTOPUS VISUAL FIELD REPRESENTATIONS Most visual ield representations on Octopus perimeters the in luence of diffuse or widespread defects). FIG 7-1 are based on the following three key representations: provides an overview of these relationships. 1) Values (the sensitivity thresholds); 2) Comparison (the comparison of the sensitivity thresholds with age- In addition, Octopus perimeters determine several indi- matched normative data); and 3) Corrected Comparison cators of visual ield reliability, to assess whether a visual (the comparison of the sensitivity thresholds with age- ield test is trustworthy or not. These are presented at matched normative data, with a correction that eliminates the end of this chapter. 100 Chapter 7 | Overview of visual field representations RELATIONSHIP AMONG DIFFERENT VISUAL FIELD REPRESENTATIONS SENSITIVITY THRESHOLDS Normative Values 26 25 27 26 27 27 26 26 29 29 28 29 29 27 28 31 31 31 31 28 32 32 31 28 27 31 33 33 32 35 32 33 33 32 29 32 33 31 28 27 28 31 32 32 31 29 30 30 29 31 31 28 30 30 29 28 26 29 28 - (Measured )Values Grayscale (Values) 17 17 19 23 12 14 21 1 7 19 15 18 25 5 16 9 15 24 25 24 22 28 16 30 11 1 1 MS 23 24 28 31 30 28 28 30 25 21 (Mean Sensitivity) 18 29 27 29 29 21 27 26 25 26 28 18 22 23 23 23 19 21 24 = SENSITIVITY LOSS Comparison Grayscale (Comparison) Probabilities Cluster Analysis 9 8 8 + 15 13 5 25 15.8 22 10 13 11 + 22 11.3 12 22 16 7 + 10 21 30 9.1 24.9 7 5 17 31 9.7 5 8 + + + 6 + 5 + + 6 4.4 8.3 2.7 10 + 5 + + 8 + + + 5 + 5.8 3.5 10 8 7 6 5 7 8 + Defect Curve Polar Analysis 1 Rank 59 MD -5 (Mean Defect) 0 5% sLV (square root Defect (dB) 5 S 10 95% 30 [dB] 20 10 N I T of Loss Variance) 15 DD 20 (Diffuse Defect) 25 LD (Local Defect) Diffuse Defect - DD = LOCAL SENSITIVITY LOSS Corrected Comparison Corrected Probabilities Corrected Cluster Analysis + + + + 10 8 + 20 10.4 16 5 5.9 8 6 + 17 7 16 10 + 3.7 19.5 + + 15 24 4.3 + + 12 26 + + + + + + + + + + + + 2.9 + 5 + + + + + + + + + + + + 5 + + + + + + + FIGURE 7-1 All visual field representations are based on the measured sensitivity thresholds (i.e., Values) and are mostly compared to age-matched normative data (top), resulting in representations that show sensitivity loss (center). Some representations also only display local sensitivity loss (bottom) because they are additionally corrected to eliminate the influence of diffuse or widespread defects. Representations displaying sensitivity thresholds 101 REPRESENTATIONS DISPLAYING SENSITIVITY THRESHOLDS Some representations display sensitivity thresholds values. The key representations and their relationship as they are measured, without reference to normal are shown in FIG 7-2. VALUES REPRESENTATIONS DISPLAY SENSITIVITY THRESHOLDS HILL OF VISION ~ VALUES ~ GRAYSCALE (Values) 3D map 2D numerical map 2D color map Sensitivity threshold 17 17 19 23 12 14 21 1 Superior Blind spot 7 19 Fixation 15 18 25 5 16 9 15 24 25 22 11 1 24 28 16 1 30 Temporal 24 31 30 28 70˚ 80˚ 90˚ 23 28 28 30 25 21 Nasal 18 29 27 29 29 21 27 26 25 26 28 18 22 23 23 23 19 21 24 Inferior FIGURE 7-2 The Values and Grayscale of Values representations are two-dimensional maps of the hill of vision. They both display sensitivity thresholds as either numerical maps (Values) or color maps (Grayscales). Note that the hill of vision is not available as a representation on Octopus perimeters. VALUES The Values representation shows the sensitivity thres- This representation is of limited diagnostic value, due to holds at each test location and is presented in FIG 7-3. It the dependence of sensitivity thresholds on patient age represents the raw data of visual ield testing and is a and eccentricity of test location, as shown in FIG 2-9. two-dimensional numerical map of a patient’s hill of vision. Sensitivity thresholds are displayed in dB and absolute defects are displayed using a “ ” symbol. 102 Chapter 7 | Overview of visual field representations VALUES 17 17 19 23 12 14 21 1 7 19 15 18 25 5 16 9 15 24 25 22 11 1 7 Sensitivity threshold [dB] 24 28 16 1 30 24 31 30 28 Absolute defect 23 28 28 30 25 21 (i.e., Sensitivity threshold 0 dB) 18 29 27 29 29 21 27 26 25 26 28 18 22 23 23 23 19 21 24 FIGURE 7-3 The Values representation displays sensitivity thresholds in dB. Absolute defects with a sensitivity threshold of 0 dB are displayed using a “ ” symbol. GRAYSCALE OF VALUES The Grayscale of Values representation displays the same Even though a color scale is used, the representation has information as the Values representation (i.e., sensitivity kept its historic name (i.e., Grayscale), which was given at thresholds), but as a two-dimensional color map, as shown a time when no color screens or printers were available. in FIG 7-4. Each color represents sensitivity thresholds within a range of 5 dB. White represents sensitivities of 36 The color representation allows for a more intuitive – 40 dB, while black represents the other end of the scale, assessment of the three-dimensional shape of the hill of showing sensitivity thresholds of 0 dB. Areas between test vision than the numerical Values representation. How- locations are interpolated (i.e., the gaps between test loca- ever, the limitations of the Values representation also tions are illed with “probable” information). apply to the Grayscale of Values representation. GRAYSCALE (VALUES) Sensitivity threshold [dB] 36..40 31..35 26..30 21..25 16..20 11..15 6..10 1..5 0 FIGURE 7-4 The Grayscale of Values representation displays sensitivity thresholds on a color map. Each color represents a range of 5 dB, with white showing the highest sensitivity and black representing absolute defects. Representations based on comparison with normal 103 REPRESENTATIONS BASED ON COMPARISON WITH NORMAL COMPARISON The Comparison representation allows direct assessment test locations. For that reason, it is the most widely used in of the shape and magnitude of disease-related change in clinical practice, and most visual ield representations are sensitivity. In contrast to the Values representation, its in- based on it. terpretation is independent of the age and eccentricity of COMPARISON REPRESENTATION DISPLAYS DEVIATIONS FROM THE NORMAL VISUAL FIELD NORMATIVE VALUES (MEASURED) VALUES COMPARISON (TO NORMAL) Normal sensitivity threshold Measured sensitivity threshold Sensitivity loss 26 25 17 17 9 8 27 26 27 27 26 26 19 23 12 14 21 1 8 + 15 13 5 25 29 29 29 7 19 22 10 28 29 29 27 15 18 25 5 13 11 + 22 28 31 31 31 31 16 9 15 24 12 22 16 7 28 32 32 31 28 27 25 22 11 1 + 10 21 30 31 33 33 32 24 28 16 1 7 5 17 31 32 35 33 33 32 - 24 30 31 30 28 = 8 5 + + + 29 32 33 31 28 27 23 28 28 30 25 21 6 + 5 + + 6 28 31 31 18 29 29 10 + 5 + + 32 32 27 29 29 30 30 29 21 27 26 25 8 + + + 31 31 26 28 5 + 28 30 30 29 28 26 18 22 23 23 23 19 10 8 7 6 5 7 29 28 21 24 8 + Sensitivity threshold Comparison Normative Values of 20-year-olds Measured Values of a 20-year-old FIGURE 7-5 The Comparison representation calculates the deviation of the measured Values (sensitivity thresholds) from the Values of an average normal person of the same age (normal sensitivity threshold at each location obtained from a normative database). 104 Chapter 7 | Overview of visual field representations The Comparison representation is de ined as the individ- Deviations smaller than 5 dB in magnitude are displayed ual deviation from the average normal visual ield (stem- with “+” symbols, because as a rule of thumb, they can be ming from the normative database) of the respective age considered to be approximately within the normal range group. The difference in the normative Value minus the of luctuation within the central 30 degrees of the visual measured Value at each test location is also termed sen- ield. Consequently, these small numerical values are not sitivity loss, loss value or defect value. This principle is meaningful for the interpretation of the visual ield. Test shown in FIG 7-5. More information on normative Values locations with a sensitivity threshold of 0 dB have reached is given in BOX 2B. the loor of perimetric testing and are marked with a “ ” symbol. Deviations from a normative visual ield are displayed for each location in dB. While the Comparisons are calculated Similar representations in non-Octopus devices are alter- at all visual ield locations, their numerical values are not natively called defect map, total deviation (see TABLE 12-5) necessarily presented at all locations, as shown in FIG 7-6. or deviation from normal. COMPARISON 9 8 8 + 15 13 5 25 22 10 13 11 + 22 12 22 16 7 + Sensitivity loss < 5 dB + 10 21 30 7 5 17 31 5 22 Sensitivity loss [dB] 8 + + + 6 + 5 + + 6 Absolute defect 10 + 5 + + (i.e., Sensitivity threshold 0 dB) 8 + + + 5 + 10 8 7 6 5 7 8 + FIGURE 7-6 The Comparison representation allows for a direct assessment of the magnitude and location of a patient’s sensitivity loss in dB. A deviation from a normal sensitivity threshold smaller than 5 dB is marked with a “+” symbol, and an absolute defect with a sensitivity threshold of 0 dB is displayed with a “ ” symbol. Representations based on comparison with normal 105 GRAYSCALE OF COMPARISON The Grayscale of Comparison is used clinically to intui- is based on the Comparison representation, it is indepen- tively assess the magnitude and shape of sensitivity loss. dent of both patient age and eccentricity of test locations. It is also useful for patient education because it is easy to A color scale is used to display sensitivity loss in % in rela- understand. tion to a normal visual ield, with different colors used for different levels of change in sensitivity. For example, a 0% It is a color map in which the areas between test locations to 10% change in sensitivity is displayed in white, 47% to are interpolated (i.e., the gaps in between test locations 58% sensitivity loss is shown in green, and 95% to 100% are illed with “probable” information) (FIG 7-7). Since it change in sensitivity is displayed in black. GRAYSCALE (COMPARISON) Sensitivity loss [% of normal] 0..10 11..22 23..34 35..46 47..58 59..70 71..82 83..94 95..100 FIGURE 7-7 The Grayscale of Comparison representation is a color map showing sensitivity loss in relation to a normal visual field. It allows for assessment of the depth and shape of sensitivity loss and is also useful for patient education. There is an inverse relationship between the sensitivity the Grayscales, similar colors are used to display close to thresholds displayed in the Grayscale of Values and the normal visual ields or fully defective visual ields. Since sensitivity loss displayed in the Grayscale of Comparison. the Values scale is absolute, showing ranges in dB, and the In other words, a high sensitivity threshold means that Comparison scale is relative, showing visual ield loss in there is a small, or no loss of sensitivity (i.e., virtually no percent, the patterns show marginal differences. FIG 7-8 difference from normal). To facilitate the interpretation of demonstrates this relationship. 106 Chapter 7 | Overview of visual field representations INVERSE RELATIONSHIP BETWEEN VALUES AND COMPARISON REPRESENTATIONS High sensitivity Low sensitivity (MEASURED) VALUES COMPARISON (TO NORMAL) threshold [dB] loss [0 dB] 17 17 7 7 19 23 12 14 21 1 7 + 14 12 + 24 7 19 21 9 15 18 25 5 12 10 + 21 16 9 15 24 11 20 14 5 25 22 11 1 + 8 20 20 24 28 16 1 6 + 16 30 30 + 24 31 30 28 6 + + + 23 28 28 30 25 21 5 + + + + 5 18 29 29 9 + + + + 27 29 21 27 26 25 7 + + + 26 28 + + 18 22 23 23 23 19 9 6 5 5 + 6 21 24 6 + Low sensitivity High sensitivity threshold 0 [dB] loss [dB] GRAYSCALE (Values) GRAYSCALE (Comparison) Sensitivity threshold Sensitivity loss [dB] [% of normal] 36..40 0..10 31..35 11..22 26..30 23..34 21..25 35..46 16..20 47..58 11..15 59..70 6..10 71..82 1..5 83..94 0 95..100 FIGURE 7-8 There is an inverse relationship between the Grayscale of Values and the Grayscale of Comparison represen- tations. Yet because an inverse color scale is used, the graphics appear similar, thereby facilitating interpretation (light colors indicate normal areas and darker colors indicate areas with defects). As the Grayscale of Values is an absolute scale (dB) and the Grayscale of Comparison is a relative scale (percent of normal), the shapes and depths are comparable, but not identical. It should be noted that the Grayscale of Comparison pro- Grayscale of Values also shows more severe loss in the vides more clinically meaningful information than the peripheral area compared to the central area, due to Grayscale of Values because it is not affected by patient the effect of eccentricity. The Grayscale of Comparison age or the eccentricity of test location, and thus shows a adjusts for age and eccentricity. For clinical purposes, it patient’s visual ield loss. When comparing Grayscales is therefore recommended to always use the Grayscale of Values for younger and older controls, the Grayscales of of Comparison, which is part of the standard printouts the older person are likely to show more loss, because of Octopus perimeters. normal age effects are not taken into account. The Representations based on comparison with normal 107 PROBABILITIES The Probabilities representation is used clinically to dis- data. More precisely, they show the probability that a given tinguish between normal and abnormal visual ield loca- sensitivity threshold would be obtained at the respective tions. This representation is needed because normal luc- location for a person of the same age as the patient with a tuation is not uniformly distributed across the visual ield; normal visual ield. instead, it is smaller in the center and larger towards the peripheral visual ield. It is therefore not possible to use For example, a person with a normal visual ield has a the same numerical cut-off point (e.g., 6 dB sensitivity loss, high probability of having little to no sensitivity losses. But representing an abnormal visual ield location) for all there is also a small probability that a person with a nor- visual ield locations. mal visual ield would obtain some sensitivity loss. FIG 7-9 illustrates this and also displays examples of patient visual The Probabilities representation uses symbols that are ields in relation to normal visual ields. associated with the statistical distribution of normative DISTRIBUTION OF SENSITIVITY THRESHOLDS OF THE NORMAL POPULATION ADVANCED MODERATE EARLY GLAUCOMA NORMAL GLAUCOMA GLAUCOMA GLAUCOMA SUSPECT Person E Person A Person B Person C Person D Number of people High No 1% 5% 0 dB sensitivity (Average sensitivity loss 0.5% 2% normal value) loss P < 1% P < 2% P < 5% P < 0.5% P > 5% FIGURE 7-9 The distribution displayed in blue indicates the range of possible sensitivity losses at a specific test location and the probability of these being obtained for a person with a normal visual field. It ranges from no sensitivity loss (right) to high sensitivity loss (left), with an average normal value of 0 dB. While it is highly unlikely that a person with a normal visual field would obtain a sensitivity loss at a specific test location similar to those seen on the left of the P < 0.5% mark, a small proportion (0.5%) of the test locations of normal subjects do give these results. The top part of the figure illustrates the results typically obtained for patients at different stages of the disease, at a majority of test locations. Note that for easier readability the distribution is not drawn to scale. 108 Chapter 7 | Overview of visual field representations The Probabilities representation shows the probability (P < 5%): there is a smaller than 5% (and larger than (P) that a normal population shows a given sensitivity 2%) chance that a person with a normal visual ield loss. When the sensitivity loss is high, the likelihood that would show this sensitivity loss. it comes from a person with a normal visual ield is low. From a clinical perspective, one could assume that it is (P < 2%): there is a smaller than 2% (and larger than more likely that the sensitivity loss comes from the pa- 1%) chance that a person with a normal visual ield tient population. would show this sensitivity loss. Increasingly darker symbols are used to show the de- (P < 1%): there is a smaller than 1% (and larger than creasing probability that a person with a normal visual 0.5%) chance that a person with a normal visual ield ield would show a given sensitivity loss at a certain test would show this sensitivity loss. location (FIG 7-10): (P < 0.5%): there is a smaller than 0.5% chance that a (P > 5%): there is a high probability that a person with person with a normal visual ield would show this sen- a normal visual ield would show this sensitivity loss. sitivity loss. PROBABILITIES Probability that a person with a normal visual field shows this result P > 5% P < 5% P < 2% P < 1% P < 0.5% FIGURE 7-10 The various symbols on the Probabilities representation show the likelihood that a given sensitivity loss would be obtained for a person with normal vision. For example, the black square (P < 0.5%) indicates that while it is possible that a person with normal vision could obtain that defect value, the probability of this occurring is very small (less than 0.5%). It should be noted that caution is essential in the clinical for 3 locations). The same is true for a level of signi icance interpretation of the Probabilities representation. This is of P < 2%, which by de inition occurs in 1 out of 50 test due to the fact that a small number of isolated test loca- locations (i.e., on average for one location in the G pattern). tions at a level of signi icance of P < 5% is likely to show up, A level of signi icance of P < 0.5% is even expected to occur even in normal visual ields. For example. in a G pattern, in one out of three normal visual ields. More information which has 59 test locations, by de inition a P value of P < on how to clinically interpret the Probabilities representa- 5% should occur in 1 out of 20 locations (i.e., on average tion is given in FIG 8-15. Representations based on comparison with normal 109 DEFECT CURVE The Defect Curve (also called Bebie Curve¹) is a graphical and makes it possible to distinguish between local and representation that alerts the clinician to the presence of diffuse defects at a glance. For more information on its diffuse defects. It provides a summary of the visual ield design, see BOX 7A. DESIGN OF THE DEFECT CURVE BOX 7A The Defect Curve is based on the Comparison representation (i.e., the sensitivity loss in comparison to the normal visual ield). The Comparisons are irst ranked according to their magnitude, from the small- est to the largest defect. The Defect Curve is drawn by plotting the defects (y-axis) as a function of their rank (x-axis). To give an example, the 28th smallest defect in the igure below is about 7 dB. The y-axis ranges from -5 to 25 dB. It must be noted that negative values indicate that there was no defect com- pared to normal and that the sensitivity is higher than the average normal value. This typically happens randomly at a few locations in every normal visual ield, and therefore the average normal visual ield shows negative values in the irst ranks. This procedure generates the Defect Curve, which by de inition always starts from the top left and moves to the bottom right. Note that spatial information is lost. The average normal Defect Curve is displayed to serve as a reference, lanked by upper and lower limits that show the area in which 90% of normal Defect Curves lie. DEFECT CURVE COMPARISON (TO NORMAL) Y-axis: Defects in dB Average normal 9 8 -5 Defect Curve 8 + 15 13 5 25 0 22 10 + 5% 13 11 + 22 Normal 12 22 16 7 5 band of 7 Defect + 10 21 30 Curve 7 5 17 31 5 10 95% 8 + + + 6 + 5 + + 6 13 10 + 5 + + 15 8 + + + 5 + Defect Curve 10 8 7 6 5 7 20 22 8 + 25 1 59 X-axis: Ranks The Defect Curve is a representation that ranks individual defects according to their size from left to right. Normal visual ϔields have a Defect Curve within the normal band, while the Defect Curve in abnormal visual ϔields lies outside the normal band. The interpretation of the Defect Curve is straightforward. instances, a combination of diffuse (or widespread) loss Parallel downward shifts of the Defect Curve represent and local visual ield loss is present. FIG 7-11 shows these diffuse defects; a drop on the right-hand side of the four main situations, while more examples are provided curve represents local defects and Defect Curves within in FIG 8-10. the normal band are considered to be normal. In many 110 Chapter 7 | Overview of visual field representations DEFECT CURVE NORMAL DIFFUSE DEFECT LOCAL DEFECT LOCAL & DIFFUSE DEFECT Defect Curve within normal Parallel downward shift of Drop of Defect Curve Parallel downward shift on band Defect Curve on the right the left and drop on the right Rank Rank Rank Rank 1 59 1 59 1 59 1 59 -5 -5 -5 -5 0 0 0 0 5% 5% 5% 5% Defect (dB) Defect (dB) Defect (dB) Defect (dB) 5 5 5 5 10 95% 10 95% 10 95% 10 95% 15 15 15 15 20 20 20 20 25 25 25 25 FIGURE 7-11 The Defect Curve is helpful to distinguish intuitively between diffuse and local sensitivity loss. The four main situations: normal, diffuse defect, local defect, and local plus diffuse defect, are shown here. CLUSTER ANALYSIS Cluster Analysis has been designed specifically for For Cluster Analysis, visual ield locations corresponding glaucoma and is very sensitive to detection of subtle to the same retinal nerve iber layer (RNFL) bundle are glaucomatous defects. It capitalizes on the fact that typical grouped and used to calculate a mean cluster defect (Clus- glaucomatous defects consist of a cluster of adjacent de- ter MD). In total, the visual ield is divided into ten clusters, fective visual ield locations that correspond to the path as shown in FIG 7-12. followed by the retinal nerve iber bundles in the retina.² CLUSTER ANALYSIS DISPLAYS TEN CLUSTER MEAN DEFECTS COMPARISON CLUSTER ANALYSIS 9 8 8 + 15 13 5 25 15.8 22 10 11.3 13 11 + 22 12 22 16 7 21 9.1 24.9 + 10 30 9.7 7 5 17 31 5 8 + + + 4.4 6 + 5 + + 6 8.3 2.7 10 + 5 + + 8 + + + 5 + 5.8 3.5 10 8 7 6 5 7 8 + FIGURE 7-12 The Cluster Analysis displays 10 visual field clusters that spatially correlate with retinal nerve fiber bundles. In each Cluster, the average sensitivity loss is calculated and presented as a Cluster Mean Defect (MD). In this example, the superior paracentral cluster is highlighted in red and its corresponding sensitivity losses are written in red font. Representations based on comparison with normal 111 The concept of Probabilities, as presented in the section a signi icance of P < 5% (and P > 1%) and a cluster MD about the Probabilities representation, is also used in in bold font has a signi icance of P <1%. The latter is Cluster Analysis (FIG 7-13). Cluster MDs with a signi i- thus more likely to be abnormal than the former. Addi- cance of P > 5% are displayed with a “+” symbol and tionally, the degree of shading indicates the deviation indicate that for an average person with a normal visual from normal values for the clusters, with lighter shading ield there is a high probability of this cluster MD val- representing lower cluster MDs, and darker shading ue being obtained. A cluster MD in unbolded font has representing higher cluster MDs. CLUSTER ANALYSIS 15.8 11.3 Cluster MD [dB] 9.1 24.9 9.7 + P > 5% 4.4 2.7 P < 5% 8.3 2.7 8.3 P < 1% 5.8 3.5 FIGURE 7-13 The Cluster Analysis indicates the probability (P) of a normal person having a certain Cluster MD by displaying the Cluster MD in bold font for P < 1%, or in normal font for P < 5%. Clusters with P > 5% (not shown here) are displayed with a “+” symbol. A major advantage of Cluster Analysis is that it is more cluster.³ The normal ranges for Cluster MDs are therefore sensitive to detection of signi icant early glaucomatous much smaller, with signi icant change being identi ied change than single point representations such as the Com- earlier.⁴ For more information on the high sensitivity of parison or Probabilities graphs. This is because single test Cluster Analysis for glaucoma detection, see BOX 8B. locations are subject to considerable normal luctuation. The averaging procedure used in the Cluster Analysis sig- Additional information on the design of the Cluster Analy- ni icantly reduces the amount of luctuation within each sis is provided in BOX 7B. 112 Chapter 7 | Overview of visual field representations BOX 7B DESIGN OF CLUSTER ANALYSIS The Cluster Analysis is based on the distribution of retinal nerve ibers in the retina. To design the Cluster Analysis, all test locations of the G pattern were superimposed over the RNFL map described by Hogan et al.³ Next, visual ield locations were grouped into 22 sectors. Test locations whose respective RNFL bundles entered the optic disc in close spatial proximity were grouped into the same cluster. This procedure yielded clusters with 2 to 4 test locations. It was noted that the test locations in each cluster were part of the same 5° sector at the optic disc. Since some of the clusters contained too few test locations to signi icantly reduce variability, these 22 clusters were further grouped to yield the 10 clusters used in the Cluster Analysis. These 10 clusters have been shown to correlate well with structural indings.⁵,⁶ The arithmetic mean of all defects within one cluster results in the Cluster Mean Defect (MD). This number is displayed within each cluster. It should be noted that while the clusters are not strictly sym- metrical, a symmetrical graph is used on the printout, for the sake of simplicity. DESIGN OF CLUSTER ANALYSIS GROUPING OF VISUAL 22 VISUAL 10 VISUAL FIELD LOCATIONS FIELD CLUSTERS FIELD CLUSTERS Based on entry of RNFL Each containing Combining the 22 clusters bundles into optic disc 2-4 locations (at least 4 test locations per cluster) 270 0 180 90 The ten Clusters used in the Cluster Analysis are generated by superimposition of the test pattern onto an RNFL map and segmentation into sectors at the optic nerve head (left). Clustering of the visual ϔield locations included in these sectors leads to 22 initial clusters in the G pattern (center), which are further combined into 10 clusters (right). By using the Cluster boundaries de ined for the G pattern, Cluster Analysis has been designed for the 32/30-2 and the 24-2 patterns. All cluster maps are based on the principle explained above. CLUSTER ANALYSIS FOR DIFFERENT TEST PATTERNS G 32/30-2 24-2 Cluster Analysis is available for the G, 32, 30-2 and 24-2 patterns. Note that the central test locations in the G pattern and the two test locations inside the blind spot in the 30-2/24-2 patterns (test locations shown in gray) are not included in any cluster. Representations based on comparison with normal 113 As with the interpretation of the Probabilities represen- in 50% of normal visual ields), and one cluster at p < 1 % tations, however, some caution is essential in the clinical in one out of 10 normal visual ields (1 out of 10 clusters interpretation of the Cluster Analysis representation. at P < 1% occurs in 10% of normal visual ields). A signif- This is due to the fact that one cluster defect at a signif- icant cluster defect is thus far more clinically meaningful icance of p < 5% is likely to show up in one out of two if it is spatially correlated with another signi icant cluster normal visual ields (1 out of 10 clusters at P < 5% occurs defect, or if it correlates with a structural defect. POLAR ANALYSIS The Polar Analysis has been designed speci ically for Once the visual ield has been lipped across the hor- glaucoma.⁷ It provides information about where struc- izontal axis, each sensitivity loss obtained from the tural defects are to be expected at the optic disc, by dis- Comparison representation is mapped onto the nerve playing visual ield results using structural coordinates. iber that corresponds to it. The nerve iber projects to It is based on the known relationship between structure the optic disc and enters at a speci ic angle around the and function and capitalizes on the fact that each visual optic disc. The angle of entry of each nerve iber is de- ield location corresponds to a speci ic retinal nerve termined and used to place each test location as a radial iber bundle in the retina (i.e., a superior visual ield bar on the Polar Analysis representation. The length of location corresponds to an inferior retinal location and the bar shows the sensitivity loss in dB from the Com- a nasal visual ield location corresponds to a temporal parison representation. Note that if two or more test retinal location).² For more information on the rela- locations map onto the nerve ibers that enter at the tionship between structure and function, see BOX 8C. same angle, the values of the corresponding test point locations are averaged. To facilitate interpretation, a Similar to the Cluster Analysis, the Polar Analysis is gray band ranging from +4 dB to -4 dB provides a rough based on a superimposition of the test pattern onto indication of a normal range. The definitions of the Hogan’s² RNFL map (FIG 7-14). Since the superior visual Polar Analysis are shown in FIG 7-15. ield corresponds to the inferior retina, the visual ield is irst lipped across the horizontal axis. 114 Chapter 7 | Overview of visual field representations DESIGN OF THE POLAR ANALYSIS VISUAL FIELD ORIENTATION COMPARISON S 9 8 8 + 15 13 5 25 22 10 13 11 + 22 12 22 16 7 + 10 21 30 7 5 17 31 T 8 5 + + + N 6 + 5 + + 6 10 + 5 + + 8 + + + 5 + 10 8 7 6 5 7 8 + I STRUCTURAL ORIENTATION 1. Locate nerve fiber 2. Define nerve fiber 3. Draw polar bar at previously entry site at optic disc entry angle at optic disc determined location (Length corresponds to defect size) 4. Repeat for all test locations 270 270 S 30 20 10 N T 0 180 T N 0 180 T [dB] I 90 90 105° FIGURE 7-14 The Polar Analysis orients visual field results (top) like a structural result (bottom), flipping the results across the horizontal meridian. It projects a sensitivity loss from the Comparison chart (e.g., 13 dB, highlighted in red, top) along the corresponding retinal nerve fibers on the retina to the optic disc (red circle, bottom left). At the nerve fiber entry site a red bar is drawn at the angle at which the nerve fiber enters the optic disc (here 105°, bottom middle), with the length of the bar corresponding to the magnitude of the sensitivity loss (i.e., 13 dB, bottom right). By repeating this procedure for all visual field locations, the Polar Analysis is drawn (all red bars, bottom right). (S: Superior; I: Inferior; N: Nasal; T: Temporal) Representations based on comparison with normal, corrected for diffuse defects 115 POLAR ANALYSIS Defect [dB] • Length of bar indicates defect size [dB] • Position along the optic disc represents the entry angle of RNFL fibers associated to each test location S 30 20 10 N T Normal range [dB] I S Superior I Inferior N Nasal T Temporal FIGURE 7-15 The Polar Analysis displays sensitivity losses from the Comparison chart as a projection onto the optic disc, to allow for easy correlation with structural results. The length of the bars indicates the sensitivity loss in dB. The Polar Analysis is a very useful tool to link structural well with structural results⁸ and usefully assists the and functional results because it allows direct side-by- identi ication of the spatially corresponding structural side comparison of the structural and functional results, (RNFL) defects. as can be seen in FIG 8-24. It has been shown to correlate REPRESENTATIONS BASED ON COMPARISON WITH NORMAL, CORRECTED FOR DIFFUSE DEFECTS CORRECTED COMPARISON It is useful to analyze localized visual ield defects inde- The correction applied to the Corrected Comparison is pendently of diffuse defects, which in many cases are based on the DD global index, which represents the caused by cataract. To do so, the Comparison, Probabilities magnitude of diffuse defect. The DD is subtracted from and Cluster Analysis representations are all available in a the sensitivity losses displayed in the Comparison corrected version. This corrected version removes diffuse representation. The DD is explained in detail in the or widespread defects and displays only localized visual section on global indices. FIG 7-16 illustrates how the ield loss. All “corrected” representations are based on the corrected representations are calculated. Corrected Comparison representation. 116 Chapter 7 | Overview of visual field representations CORRECTED COMPARISON AND ITS RELATIONHIP WITH OTHER CORRECTED REPRESENTATIONS COMPARISON PROBABILITIES CLUSTER ANALYSIS 9 8 8 + 15 13 5 25 15.8 22 10 10 11.3 13 11 + 22 12 22 16 7 9.1 24.9 + 10 21 30 7 5 17 31 9.7 5 8 + + + 4.4 6 + 5 + + 6 8.3 2.7 10 + 5 + + 8 + + + 5 + 5.8 3.5 10 8 7 6 5 7 8 + - DD = 4.5 dB = CORRECTED COMPARISON CORRECTED PROBABILITIES CORRECTED CLUSTER ANALYSIS + + + + 10 8 + 20 10.4 16 5 5 5.9 8 6 + 17 7 16 10 + 3.7 19.5 + + 15 24 + + 12 26 4.3 + + + + + + + + + + + + 2.9 + 5 + + + + + + + + + + + + 5 + + + + + + + FIGURE 7-16 The Corrected Comparison representation is calculated by subtracting the magnitude of the diffuse defect ex- pressed in the DD index from each sensitivity loss in the Comparison representation. It allows for the assessment of localized visual field loss without the influence of diffuse defects and is the basis for the calculation of both the Corrected Probabilities and the Corrected Cluster Analysis. Representations based on comparison with normal, corrected for diffuse defects 117 The Corrected Comparison representation is similar symbols to show local sensitivity loss (FIG 7-17). to the Comparison representation and uses the same CORRECTED COMPARISON + + + + 10 8 + 20 16 5 8 6 + 17 7 16 10 + + Local sensitivity loss < 5 dB + + 15 24 + + 12 26 + 16 Local sensitivity loss [dB] + + + + + + + + + + Absolute defect 5 + + + + (i.e., Sensitivity threshold 0 dB) + + + + + + 5 + + + + + + + FIGURE 7-17 The Corrected Comparison representation shows the magnitude of local sensitivity loss once the diffuse defect is removed. It uses the same definitions as the Comparison representation. CORRECTED PROBABILITIES The Corrected Probabilities representation is very similar corrected sensitivity loss at various signi icance levels, as to the Probabilities representation and shows the proba- shown in FIG 7-18. bility that a person with a normal visual ield shows this CORRECTED PROBABILITIES Probability that a person with a normal visual field shows this result P > 5% P < 5% P < 2% P < 1% P < 0.5% FIGURE 7-18 Similar to the Probabilities representation, the Corrected Probabilities representation shows the likelihood that an average person with a normal visual field would have a given sensitivity loss, but is based on the Corrected Comparison representation that only displays local visual field loss. 118 Chapter 7 | Overview of visual field representations CORRECTED CLUSTER ANALYSIS The Corrected Cluster Analysis is very similar to the normal visual ield shows this corrected Cluster Mean Cluster Analysis, but is based on the Corrected Com- Defect in dB at various signi icance levels, as shown in parison. It shows the probability that a person with a FIG 7-19. CORRECTED CLUSTER ANALYSIS 10.4 5.9 Corrected Cluster MD [dB] 3.7 19.5 4.3 + P > 5% + + 2.9 P < 5% 2.9 5.9 P < 1% + + FIGURE 7-19 Similar to the Cluster Analysis, the Corrected Cluster Analysis indicates the probability (P) of an average normal person having a certain Cluster MD, but is based on the Corrected Comparison representation that only displays local visual field loss. GLOBAL INDICES Global indices are useful numerical summaries of the change over time. Some indices summarize the entire entire visual ield, or of an aspect of the visual ield.⁹ They visual ield, while others focus solely on a part of the visual 1) provide a summary of the status of the visual ield, 2) ield. They are presented in detail in the section below. The are useful to objectively assess and classify the severity formula used to calculate each global index is shown in of visual ield loss and 3) are helpful in the assessment of TABLE 7-1. Global indices 119 GLOBAL INDICES AVAILABLE FOR OCTOPUS PERIMETERS TABLE 7-1 INDEX FORMULA VARIABLES MEAN N: Total number of test locations SENSITIVITY (MS) xi: Sensitivity threshold at test location i, or mean of two repeated measurements xi1, xi2 at test location i MEAN DEFECT ni: Normal value at test location i (MD) di: Sensitivity loss at test location i SQUARE ROOT OF LOSS VARIANCE (sLV) CORRECTED ME: Mean Error SQUARE ROOT OF LOSS VARIANCE (CsLV) MEAN SENSITIVITY (MS) The Mean Sensitivity (MS) is the arithmetic mean of based on the Values and its diagnostic value is therefore the sensitivity thresholds displayed in the Values rep- limited by the same factors that affect the Values (e.g., it resentation. It represents the average height of the hill is dependent on patient age and on the spatial distribu- of vision with respect to the locations that are tested, tion of the test locations). and thus a patient’s average sensitivity to light. MS is MEAN DEFECT (MD) The Mean Defect (MD) is the arithmetic mean of the sen- often used to assess visual ield severity.⁹ It is a key index sitivity loss displayed in the Comparison representation. used in the progression analysis available on Octopus It represents the average visual ield loss of a patient perimeters to identify the presence of progression (see derived from the locations that are tested and is thus Chapter 9). 120 Chapter 7 | Overview of visual field representations SQUARE ROOT OF LOSS VARIANCE (sLV) The square root of Loss Variance (sLV) represents the (i.e., diffuse) or localized at some locations. The sLV standard deviation of the individual defects at all visual index thus further summarizes the characteristics of a ield locations and provides a measure of variability visual ield. The sLV index is large in inhomogeneous across the visual ield.⁹ This index is useful because the visual ields (localized defects) and small in homoge- Mean Defect (MD) does not provide any information neous visual ields (diffuse defects), as shown in FIG 7-20. about whether visual ield loss is uniformly distributed SQUARE ROOT OF LOSS VARIANCE (sLV) DIFFUSE DEFECT LOCAL DEFECT COMPARISON COMPARISON 9 9 8 + 5 8 11 10 5 5 8 5 6 6 8 + 8 10 17 10 + 11 + + 15 13 10 + + 6 7 7 6 26 21 19 23 + + 5 7 7 + + 19 22 22 21 18 + 5 7 7 + 21 15 13 17 7 12 9 6 6 7 + 5 6 + 8 6 6 6 6 5 + + + + + + 5 9 5 + + + 7 + + + 9 15 5 5 + + + + 10 6 + + 7 8 8 6 5 5 + + + + + + 5 5 + + MD 6.3 dB MD 6.5 dB sLV 2.5 dB sLV 8.5 dB 26 23 22 22 15 21 21 21 19 19 18 11 11 17 17 15 15 10 10 10 13 sLV sLV 12 8.5 dB 9 9 99 9 2.5 dB 11 10 10 8 8 8 8 8 7 7 7 7 7 7 7 7 7 7 MD 8 8 8 MD 6 6 6 6 66 6 66 6 6 6.3 dB 6 6 6 6.5 dB 5 55 5 5 55 5 5 5 5 5 5 5 5 + + + + + + ++ ++ + + + ++ + ++ + + + ++ + + + + + + + + ++ + FIGURE 7-20 The sLV provides a measure of the inhomogeneity of a visual field. This is illustrated in this figure, which shows a homogeneous visual field with diffuse defect (left) and a heterogeneous visual field with localized defect (right). If the visual field is homogeneous, the sensitivity losses at specific test locations (shown on the y-axis in the bottom part of the figure) do not deviate strongly from MD, and sLV is small (left). If the visual field is heterogeneous, some locations deviate strongly from MD, and therefore sLV is large (right). Note that sLV is the standard deviation of the local defects and thus does not span the full range of determined sensitivity losses. Global indices 121 CORRECTED SQUARE ROOT OF LOSS VARIANCE (CsLV) The Corrected square root of Loss Variance (CsLV) is The reliability index used for CsLV is Short-term Fluc- similar to the sLV, with an added correction factor to tuation (SF), which is explained in detail in the section account for the variability of patient responses that occurs about reliability indices. Note that CsLV is only displayed during a perimetric test. It is a useful index to distin- if SF is actively determined during the visual ield test by guish between a truly heterogeneous visual ield and repeated testing at all test locations. a visual ield that is heterogeneous due to Short-term Fluctuation.⁹ DIFFUSE DEFECT (DD) DEFINITION OF DIFFUSE DEFECT (DD) BOX 7C As shown in the section about the Defect Curve, diffuse defects result in a parallel downward shift of the Defect Curve. The magnitude of that shift is measured by assessing the distance between the Defect Curve and the average normal Defect Curve at a representative location along the curve. This generates the index DD. As the Defect Curve may not be fully parallel with the average normal Defect Curve, it is essential to measure at a location that represents diffuse visual ield loss. DD is calculated from the 20th to the 27th percentile of the ranks. For the G pattern, which includes 59 test locations, this translates into the range from the 12th to the 16th rank from the left. This area is neither too close from the left to be meaningfully affected by random abnormally high sensitivity responses, nor too close to the right to be meaningfully affected by local defects. To be less in luenced by variability, an average of the deviations of the respective ranks from the median Defect Curve is used. DIFFUSE DEFECT (DD) Y-axis: Defects in dB Average normal -5 Defect Curve 0 5% 5 Diffuse Defect 10 95% 15 20 25 1 12 16 59 X-axis: Ranks In the Defect Curve, all individual defects are ranked from 1 to the total number of test locations (e.g., the 59 locations of the G pattern are shown here). The DD is calculated from the magnitude of the downward shift of the Defect Curve at the ranks from the 20th to the 27th percentile (for the G pattern, ranks 12 to 16). 122 Chapter 7 | Overview of visual field representations The index DD allows quanti ication of diffuse defect in section of this chapter. It is also used in the progression dB and is derived from the Defect Curve, as explained in analysis available on Octopus perimeters to identify the BOX 7C. It is mainly used to calculate the Corrected Com- presence of diffuse progression (see Chapter 9). parison representation, which is discussed in the previous LOCAL DEFECT (LD) The index LD allows quantification of the mean local available on Octopus perimeters to identify the presence defect in dB and is also derived from the Defect Curve, as of local progression. explained in BOX 7D. It is used in the progression analysis BOX 7D DEFINITION OF LOCAL DEFECT (LD) Any point on the Defect Curve outside normal limits represents an abnormal visual ield point. Shifting down the average normal Defect Curve by the amount of the diffuse defect DD yields a curve represent- ing the diffuse defect. Any further deviation of the individual Defect Curve downwards indicates local defects. The local defect index LD is de ined as the average of these deviations measured between the 14th and 59th ranks for the G pattern. In more general terms and also applicable to other test patterns, the LD index is de ined as the average of these deviations measured between the 23rd percentile of ranks and the last rank. LOCAL DEFECT (LD) Y-axis: Defects in dB Average normal -5 Defect Curve 0 5% 5 Diffuse Defect 10 95% 15 Local Defect 20 25 1 59 X-axis: Ranks The LD index represents the magnitude of the average local defect and is derived from the Defect Curve. It is calculated from the deviation between the Diffuse Defect and the Defect Curve, as indicated by the red area. Reliability indices 123 RELIABILITY INDICES Due to the subjective component of perimetric testing, These are presented below with their respective formula, unreliable visual ield results occur and it is essential shown in TABLE 7-2 at the end of this section. For more to identify them in clinical practice. Octopus perime- information on how to clinically interpret reliability ters provide several indicators of visual ield reliability. indices, see the section on reliability in Chapter 8. FALSE POSITIVE (FP) ANSWERS False positive (FP) answers are used to detect trigger- in the natural rhythm of perimetric testing in which happy patients. These are patients who respond even no stimulus is presented. If a patient responds, this is when no stimulus is presented. This type of patient marked as a false positive answer (FIG 7-21). behavior occurs if patients do not understand the nature of the test, or if they wish to positively in luence the result. The false positive rate is calculated as the ratio of false positive answers to the total amount of positive catch Positive catch trials are used to identify false positive an- trials presented. swers. Positive catch trials consist of a gap introduced FALSE POSITIVE (FP) ANSWERS Yes Do you see the stimulus? Fixation Target FIGURE 7-21 False positive answers occur when patients respond even though no stimulus is presented. They are useful to detect unreliable visual fields. 124 Chapter 7 | Overview of visual field representations FALSE NEGATIVE (FN) ANSWERS False negative answers are used to detect fatigue, loss of should be able to see these bright stimuli, and when they attention and potential ixation loss during perimetric are missed, this is marked as a false negative answer testing. (FIG 7-22). Negative catch trials are used to identify false negative The false negative rate is calculated as the ratio of false answers. Negative catch trials consist of stimuli that negative answers to the total amount of negative catch are presented at a higher intensity than the patient has trials presented. previously seen. Patients who perform the test reliably FALSE NEGATIVE (FN) ANSWERS Yes No Do you see Do you see the stimulus? the stimulus? Fixation Fixation Target Target FIGURE 7-22 False negative answers occur when patients do not respond to a stimulus of higher intensity (right) than a stimulus they had previously seen at that location (left). A high false negative response rate can indicate an unreliable field and may be an indicator of fatigue. RELIABILITY FACTOR (RF) The Reliability Factor (RF) summarizes the false positive both false positive and false negative answers to the sum and false negative answers. RF is calculated as the ratio of of positive and negative catch trials presented. SHORT-TERM FLUCTUATION (SF) The Short-term Fluctuation (SF) index provides a measure the deviations between the irst and second sensitivity of the variability of patient responses that occurs during a thresholds are determined. SF is de ined as the standard perimetric test.⁹ In order to determine SF, the sensitivity deviation of the distribution of the results of repeated thresholds are measured again at the end of the test, and measurements of the same threshold.¹⁰ Reliability indices 125 RELIABILITY INDICES AVAILABLE ON OCTOPUS PERIMETERS TABLE 7-2 INDEX FORMULA VARIABLES FALSE POSITIVE (FP) ANSWERS [%] nf+ : Number of false positive answers ntot+ : Total number of positive catch trials presented FALSE NEGATIVE (FN) ANSWERS [%] nf- : Number of false negative answers ntot- : Total number of negative catch trials presented RELIABILITY FACTOR (RF) [%] SHORT-TERM FLUCTUATION (SF) xi1 : Sensitivity threshold at test location i determined in 1st of two repeated measurements xi2 : Sensitivity threshold at test location i determined in 2nd of two repeated measurements N : Total number of test locations 126 Chapter 7 | Overview of visual field representations REFERENCES 1. Bebie H, Flammer J, Bebie T. The cumulative defect curve: separation of local and diffuse components of visual ield damage. Graefes Arch Clin Exp Ophthalmol. 1989;227:9-12. 2. Hogan MJ, Alvarado JA, Weddel JE. Histology of the human eye: an atlas and textbook. Philadelphia: WB Saunders; 1971. 3. Mandava S, Zulauf M, Zeyen T, Caprioli J. An evaluation of clusters in the glaucomatous visual ield. Am J Ophthalmol. 1993;116:684-691. 4. Naghizadeh F, Holló G. Detection of early glaucomatous progression with octopus cluster trend analysis. J Glaucoma. 2014;23:269-275. 5. Holló G. Comparison of structure-function relationship between corresponding retinal nerve ibre layer thickness and Octopus visual ield cluster defect values determined by normal and tendency-oriented strategies. Br J Ophthalmol. 2016; doi: 10.1136/bjophthalmol-2015-307759. 6. Naghizadeh F, Garas A, Vargha P, Holló G. Structure-function relationship between the Octopus perimeter cluster mean sensitivity and sector retinal nerve iber layer thickness measured with the RTVue optical coherence tomography and scanning laser polarimetry. J Glaucoma. 2014;23:11-18. 7. Buerki E. Update Octopus Perimetrie 1. Teil: Die Polardarstellung der Gesichtsfelddaten. Ophta. 2006;7:9-12. 8. Holló G, Naghizadeh F. Evaluation of Octopus Polar Trend Analysis for detection of glaucomatous progression. Eur J Ophthalmol. 2014;24:862-868. 9. Flammer J. The concept of visual ield indices. Graefes Arch Clin Exp Ophthalmol. 1986;224:389-392. 10. Bebie H, Fankhauser F, Spahr J. Static perimetry: accuracy and luctuations. Acta Ophthalmol (Copenh). 1976;54:339-348. 127 CHAPTER 8 CLINICAL INTERPRETATION OF A VISUAL FIELD INTRODUCTION Octopus perimeters offer a variety of visual ield repre- a normal and a borderline visual ield, as well as visual sentations that are based on the raw data (i.e., the sen- ields with localized loss, diffuse loss, and both local and sitivity thresholds). Each of them focuses on different diffuse loss, and a visual ield with advanced loss. These clinically relevant aspects of visual ield interpretation, examples provide an excellent starting point to become to facilitate clinical decision-making. While there is often familiar with the various representations and their be- overlap in the information provided by the different rep- havior in standard clinical situations and are referenced resentations, there is typically one representation that is throughout the book. A removable poster of these exam- best suited to provide information about a certain clinical ples is also included in the back cover of this book. aspect of a visual ield. Thereafter, this chapter presents the various representa- This chapter provides a systematic approach on how to tions in a step-by-step work low. Because this chapter interpret visual ields in a clinically meaningful way and focuses on how to interpret visual fields for clinical highlights particular representations to answer speci ic purposes, only an introduction to the de initions, design clinical questions. To illustrate how the various repre- and relationships between the representations is present- sentations can be used in clinical situations, this chapter ed. Detailed information about each representation is starts by presenting six typical visual ields at different provided in Chapter 7 and should be consulted as required. stages of disease severity (FIG 8-1). The examples include 128 Chapter 8 | Clinical interpretation of a visual field EXAMPLES OF SIX TYPICAL VISUAL FIELDS NORMAL BORDERLINE EARLY TO MODERATE Diffuse defect 1 Correct patient & examination parameters? 2 Reliable, free of artifacts and trustworthy? 3 Diffuse loss? Rank Rank Rank 1 59 1 59 1 59 -5 -5 -5 0 0 0 DEFECT CURVE 5% 5% 5% 5 5 5 Defect (dB) 10 95% 10 10 95% 95% 15 15 15 20 20 20 25 25 25 DD 0.1 dB 1.3 dB 6.2 dB LD 0.2 dB 0.2 dB 0.6 dB 4 Significant local loss? PROBABILITIES CORRECTED PROBABILITIES FIGURE 8-1 A systematic approach to visual field interpretation is recommended and this workflow can be used as a guide (this figure is also included as a poster in the back cover of this book). Introduction 129 EARLY TO MODERATE ADVANCED Local defect Local & diffuse defect Correct patient & examination parameters? 1 Reliable, free of artifacts and trustworthy? 2 Diffuse loss? 3 Rank Rank Rank 1 59 1 59 1 59 -5 -5 -5 0 0 0 DEFECT CURVE 5% 5% 5% 5 5 5 10 95% 10 95% 10 95% 15 15 15 20 20 20 25 25 25 DD 1.3 dB 5.9 dB 19.3 dB LD 7.0 dB 6.1 dB 4.7 dB Significant local loss? 4 PROBABILITIES CORRECTED PROBABILITIES 130 Chapter 8 | Clinical interpretation of a visual field EXAMPLES OF SIX TYPICAL VISUAL FIELDS (CONTINUED) NORMAL BORDERLINE EARLY TO MODERATE Diffuse defect 5 Assess shape & depth of defect. GRAYSCALE (COMPARISON) + + + + 9 9 + + + + + + + 5 + + + + 5 8 11 10 5 5 + + + 8 8 10 COMPARISON + + + + + + + 7 + 11 + + + + + + + + + + + + + 6 7 7 6 + + + + + + + + + + + + + 5 7 7 + + + + + + + + + + 5 7 7 + + + 7 + + + + + + + + 9 6 6 7 + + + + + + + + + + + + 8 6 6 6 6 5 + + + + + + 5 9 5 + + + + 7 + + + + + + + + + 9 15 5 5 + + + + 10 6 + + + + + + + + + + 6 + 7 8 8 6 5 5 + + + + 5 5 + + + + + + CORRECTED COMPARISON + + + + + + + + + + + + + + 5 + + + + + + 7 + + + + + + + + + 5 + 5 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 8 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + Introduction 131 EARLY TO MODERATE ADVANCED Local defect Local & diffuse defect Assess shape & depth of defect. 5 GRAYSCALE (COMPARISON) 8 + 19 8 5 6 6 8 + + 14 22 17 10 17 17 15 13 10 + + 6 23 24 COMPARISON 26 21 19 23 + + 9 12 12 12 19 26 15 17 27 19 22 22 21 18 + + 14 13 24 25 21 24 9 18 21 15 13 17 10 10 11 12 14 8 9 17 12 5 9 + 5 6 + + + 5 + 17 22 11 15 + + + + + + 6 + + 7 5 + 19 10 + + + 5 5 5 12 19 22 27 + + + + + + + + 7 5 12 5 24 + + 7 7 22 + + + + + + 8 9 8 9 10 5 25 + + 12 + 20 7 + 13 CORRECTED COMPARISON 6 + 5 + 7 + + 8 16 16 9 11 12 14 11 9 + + + 17 5 25 20 18 21 + + + 7 6 6 + 6 + + 7 17 20 21 20 16 + + 8 7 18 19 15 5 + + 20 14 12 15 + + 5 6 + + + + 10 + + + + 5 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 8 + + + + + + + + + + + + 6 + 5 + + + + + + + + + + + + + + + + + 6 + + 6 + + 132 Chapter 8 | Clinical interpretation of a visual field EXAMPLES OF SIX TYPICAL VISUAL FIELDS (CONTINUED) NORMAL BORDERLINE EARLY TO MODERATE Diffuse defect 6 For glaucoma: Significant cluster defects? + 2.5 6.0 CLUSTER ANALYSIS + + 8.5 + + + 3.4 + 5.1 + + 6.1 + + 6.6 + + + + 7.2 6.1 + + 6.9 + + 6.4 CORRECTED CLUSTER ANALYSIS + + + + + 2.3 + + + 2.1 + + 2.0 + + + + + + + + + + + + + + + + + 7 For glaucoma: Where to look for structural defects. POLAR ANALYSIS S S S 30 20 10 N T T N 10 20 30 30 20 10 N T [dB] [dB] [dB] I I I 8 Severity? MD -0.2 dB 1.0 dB 6.3 dB sLV 1.5 dB 1.9 dB 2.5 dB Introduction 133 EARLY TO MODERATE ADVANCED Local defect Local & diffuse defect For glaucoma: Significant cluster defects? 6 17.4 17.1 26.1 CLUSTER ANALYSIS 6.6 19.2 23.7 20.7 + + 17.0 22.5 20.6 15.5 11.1 10.1 3.0 3.2 17.0 + + 6.4 4.2 25.8 20.3 + 7.5 24.4 + 7.8 25.4 CORRECTED CLUSTER ANALYSIS 11.2 6.8 16.1 5.3 13.3 4.4 19.4 + + 11.1 3.2 1.3 14.2 5.2 + 1.7 + + + + + + 6.5 1.0 + 1.6 5.1 + 1.9 6.1 For glaucoma: Where to look for structural defects. 7 POLAR ANALYSIS S S S 10 20 30 30 20 10 30 20 10 T N N T N T [dB] [dB] [dB] I I I Severity? 8 MD 6.5 dB 10.1 dB 21.7 dB sLV 8.3 dB 7.2 dB 5.6 dB 134 Chapter 8 | Clinical interpretation of a visual field STEP-BY-STEP INTERPRETATION OF A VISUAL FIELD OVERVIEW OF STEP-BY-STEP WORKFLOW VISUAL FIELD INTERPRETATION WORKFLOW Patient name & age 1 Refraction Correct patient & examination parameters? Pattern/strategy Yes False positives 2 False negatives Reliable, free of No Retest if clinically Repetitions artifacts & trustworthy? relevant Duration Yes Defect Curve 3 DD, LD Diffuse loss? Yes Caused by Potentially unreliable, No pathology? retest if clinically relevant No Consider pathology Yes Yes leading to diffuse defect Probabilities 4 Corrected Probabilities Borderline or significant local loss? No Normal visual field OR Diffuse defect only Yes Grayscale (Comparison) 5 Comparison Assess shape & depth of defect. Consider non- Corrected Comparison No glaucomatous field defects Typical for glaucoma? Yes Cluster Analysis 6 Corrected Cluster Analysis Glaucoma only: Consider non- No glaucomatous field defects Significant cluster defects? Yes Consider glaucoma Polar Analysis 7 Glaucoma only: No Consider non- Where to look for structural defects. glaucomatous field defects Is there a relationship? Yes Consider glaucoma MD, sLV 8 Severity? FIGURE 8-2 A systematic approach to visual field interpretation is recommended and this workflow can be used as a guide. Step-by-step interpretation of a visual field 135 This chapter provides a systematic step-by-step approach excellent starting point to interpret visual ield results. on how to interpret visual ields in a clinically meaningful Different sequences may, however, be equally valid or way and highlights particular representations to answer even more adequate in speci ic cases and should be used speci ic clinical questions. This suggested sequence has accordingly. An overview of that work low is presented been validated by many experts and can serve as an in FIG 8-2. STEP 1 – CONFIRM PATIENT AND EXAMINATION PARAMETERS IMPORTANCE OF CONFIRMING PATIENT AND EXAMINATION PARAMETERS It is essential to con irm that the correct information is used for each test, in order to make accurate clinical decisions. STEP 1 – CONFIRM PATIENT AND EXAMINATION PARAMETERS 1 Correct patient & examination parameters? FIGURE 8-3 Before interpreting visual field results, it is important to confirm that the correct patient data has been entered and that the correct examination parameters have been used during the test. Octopus perimeters display key patient and examination • Patient’s name and identi ication number parameters for all visual ields (FIG 8-4). Special attention • Patient’s date of birth and age should be paid to patient age and refraction. If these are • Tested eye incorrect, this can lead to non-pathological diffuse visual • Date and time of examination ield loss. The following parameters are displayed: • Test pattern and strategy • Stimulus type • Maximum stimulus intensity and background luminance • Refraction entered or trial lens used • Pupil size 136 Chapter 8 | Clinical interpretation of a visual field OVERVIEW OF PATIENT AND EXAMINATION PARAMETERS Demo, John, 1/5/1942 (63yrs) Demo, John, 1/5/1942 (63yrs) Left eye (OS) / 01/24/2005 / 16:25:23 Seven-in-One Grayscale (CO) Values [dB] MD [dB] MS [dB] 16.8 17 17 11.3 12.2 17.6 [%] 19 23 12 14 21 1 0..10 Left eye (OS) / 01/24/2005 / 16:25:23 11..22 23..34 35..46 16 25 15 18 7 22 19 9 15 11 24 25 1 5 47..58 24 28 16 1 30 24 31 30 28 59..70 23 28 28 30 25 21 71..82 18 29 29 27 29 83..94 21 27 26 25 26 28 95..100 18 22 23 23 23 19 6.1 3.9 24.1 25.9 21 24 Comparison [dB] Corrected comparisons [dB] Defect curve 1 Rank 59 9 8 + + 8 + 15 13 5 25 + + 10 8 + 20 -5 22 10 16 5 0 5% 13 11 + 22 8 6 + 17 Defect (dB) 12 22 16 7 7 16 10 + 5 + 10 21 30 + + 15 24 7 5 17 31 + + 12 26 10 95% 5 + 8 + + + + + + + 6 + 5 + + 6 + + + + + + 15 10 + + 5 + + 5 + + + 8 + + + + + + + 20 5 + + + 25 10 8 7 6 5 7 5 + + + + + 8 + + + Diffuse defect [dB]: 5.4 Probabilities Corrected probabilities Programs: G Standard White/White / Normal Questions / repetitions: 356 / 23 [%] P>5 Parameters: 4 / 1000 asb III 100 ms Duration: 15:32 P<5 P<2 P<1 Catch trials: 1/18 (6%) +, 1/18 (6%) - RF: 5.5 P < 0,5 Refraction S/C/A: -3.5 / 1.25 / 35 VA [m]: Programs: G Standard White/White / Normal Questions / repetitions: 356 / 23 MS [dB]: 30° 19.7 Parameters: 4 / 1000 asb III 100 ms Duration: 15:32 MD [< 2.0 dB]: 9.9 Pupil [mm]: 5.6 IOP [mmHg]: Catch trials: Refraction S/C/A: Pupil [mm]: 1/18 (6%) +, 1/18 (6%) - -3.5 / 1.25 / 35 5.6 RF: VA [m]: IOP [mmHg]: 5.5 sLV [< 2.5 dB]: 8.1 NV: T21 V2.1 NV: T21 V2.1 Comment: Good fixation OCTOPUS® EyeSuite™ Static perimetry, V3.5.0 OCTOPUS 101 Comment: Good fixation FIGURE 8-4 All patient and examination parameters are displayed for every perimetric result. STEP 2 - DETERMINE WHETHER THE VISUAL FIELD CAN BE TRUSTED IMPORTANCE OF ASSESSING WHETHER THE VISUAL FIELD CAN BE TRUSTED Due to the subjective, patient-related component of reasons can occur frequently, must be identi ied and perimetric testing, unreliable visual field tests, tests should not be clinically interpreted. with artifacts or tests that cannot be trusted for other STEP 2 – ASSESS WHETHER THE VISUAL FIELD CAN BE TRUSTED 2 Reliable, free of artifacts & trustworthy? FIGURE 8-5 Before interpreting visual field results, it is important to confirm that the visual field can be trusted. Visual fields that are not reliable, contain artifacts or cannot be trusted for other reasons should be retested if this is clinically relevant. Visual ield results that cannot be trusted may occur for ing of the task to perform, or a desire to in luence the a number of reasons, as shown in Chapter 3. They can be results. Untrustworthy tests can also occur following set- caused by inconsistent patient behavior resulting from up errors, for example when incorrect test parameters fatigue, learning effects, distraction, lack of understand- or inadequate refraction are used, or when the incorrect Step-by-step interpretation of a visual field 137 UNTRUSTWORTHY VISUAL FIELD TESTS CAN SHOW SIGNIFICANT DEFECTS 1st Test 2nd Test 3rd Test 4th Test 5th Test Reliable normal Less reliable normal Normal who experiences difficulties with perimetry Normal with learning effects (tests 1 to 3) Normal with artifactual defects (here: lens rim artifact on 1st, 2nd and 4th test) FIGURE 8-6 The examples above show several visual field series from different individuals with clinically confirmed normal visual fields and no pathology. Note that while some individuals perform perimetric testing consistently, some show improve- ment over time due to learning effects, and some perform variably from one examination to the next. This results in untrust- worthy visual field results, which may be misinterpreted. 138 Chapter 8 | Clinical interpretation of a visual field date of birth is entered. In addition, artifacts stemming the visual ield results of several individuals with clinically from incorrect positioning of the patient, droopy eye- con irmed normal visual ields and no pathology. lid or incorrectly centered correction lenses can also lead to untrustworthy results. While a well-trained and Since visual ields that cannot be trusted may not repre- observant visual ield examiner can signi icantly reduce sent the true status of a patient’s visual ield, they may be the amount of untrustworthy visual ields in a clinical clinically meaningless. It is thus essential to identify them practice, some patients are simply unable to perform as a irst step in visual ield interpretation. The reliabili- perimetric testing consistently. ty indicators provided by Octopus perimeters, as well as further indicators of whether a visual ield can be trusted, FIGURE 8-6 shows the impact of unreliable visual ield should be used. These are presented in this section. tests, inconsistent patient behavior and set-up errors on FALSE POSITIVE AND FALSE NEGATIVE ANSWERS Octopus perimeters offer several indicators to detect interpreted with caution and the test should ideally be unreliable visual ields (see TABLE 7-2 for the de initions repeated if it is essential for clinical decision-making. of each of these indicators). The two most important In most clinical studies however, false positive rates of indicators of unreliability are the false positive (see FIG up to 20 to 33% are accepted.²-⁶ 7-21) and false negative answers (see FIG 7-22). Note that if only a few positive catch trials are present- False positive answers occur when the patient presses ed (e.g., the default setting of the G TOP test contains the response button when no stimulus is presented. only six positive catch trials), one accidentally missed Patients who respond in the absence of a stimulus are positive catch trial has a great impact on the false pos- referred to as trigger-happy, and may have visual ield itive rate. In this situation, more lenient acceptance results that are better than their true visual ield sta- criteria may be appropriate. tus, as shown in FIG 8-7. When the false positive an- swer rate exceeds 10 to 15%,¹ the results should be IMPACT OF FALSE POSITIVE ANSWERS ON VISUAL FIELD RESULT HIGH FALSE POSITIVES NO FALSE POSITIVES Real defect is missed Real defect is visible 1st Test 2nd Test FIGURE 8-7 The example above shows the impact of a high rate of false positive answers on the visual field. The field on the left is unreliable because the patient responded in the absence of a stimulus. As a result, the visual field appears better than the true visual field of the patient, which is shown on the right. Step-by-step interpretation of a visual field 139 False negative answers occur when patients do not re- in luctuation with increasing visual ield loss. This can spond to stimuli that they should be able to see. Patients result in false negative rates above 50%, even though the who do not respond to stimuli they should be able to see visual ield test is performed without any subjective mis- may experience fatigue or a loss of attention, and may takes.⁷ False negative answers should thus be interpreted have results that are worse than their true visual ield with care in more advanced vision loss. status, as shown in FIG 8-8. For most patients, clinical studies often exclude results with false negative rates Note that if only a few negative catch trials are presented, above 20 or 30%.⁴ In patients with severe vision loss, more lenient acceptance criteria may be appropriate, as however, false negative errors are not a meaningful explained in the section on false positive answers. indicator of reliability because there is a large increase IMPACT OF FALSE NEGATIVE ANSWERS ON VISUAL FIELD RESULT HIGH FALSE NEGATIVES NO FALSE NEGATIVES Defect is deeper Real defect shape & depth 1st Test 2nd Test FIGURE 8-8 The example above shows the impact of a high rate of false negative answers on the visual field. The field on the left is unreliable because the patient did not respond to stimuli that should have been seen. As a result, the visual field appears worse than the true status of the patient’s visual field, which is shown on the right. CONSISTENCY OF RESULTS WITH FURTHER DIAGNOSTIC TESTS Any drastic inconsistency in the location of a visual ield ield defect.⁸ These visual ield tests can be used in the defect in repeated testing can suggest that some of the future to evaluate progression or stability. visual ield tests may not be trusted. This is because pathologies lead to characteristic visual ield defect Furthermore, if a visual ield defect corresponds to the patterns in speci ic locations. While these defects may results of another diagnostic test, this strongly supports deepen, expand or in some instances also improve over the decision that the visual ield result can be trusted. time, they are usually consistently located at the same For example, if a patient shows a visual ield defect position in repeated visual ield testing. If defect pat- characteristic of glaucoma and shows a related RNFL terns shift to different locations on repeated testing, as thinning or rim thinning at the related optic disc loca- can be seen in some of the examples shown in FIGURE tion, as well as high IOP, it will be highly likely that the 8-6, this is typically a sign of an untrustworthy visual patient has glaucoma and that the visual ield result ield test. Therefore it is good clinical practice to base is thus trustworthy. The results of visual ield tests a clinical decision on two to three visual ield tests, in should therefore always be interpreted in light of the order to con irm or discard an initially observed visual full clinical pro ile. 140 Chapter 8 | Clinical interpretation of a visual field OTHER INDICATORS TO DETERMINE WHETHER VISUAL FIELD TESTS CAN BE TRUSTED In addition to the false positive and false negative detect trigger-happy behavior using the Defect Curve. answers, other indicators are also useful to determine whether visual ield test results can be trusted. One of Test duration can be a further indicator of whether the most powerful indicators remains the visual ield visual ield results can be trusted. Abnormally long test examiner’s direct observation of the patient during the durations can indicate that a patient is struggling with test. Examiners should note their observations in the pa- the task of performing perimetry. tient’s chart. Finally, if a patient can sustain prolonged testing, one can In addition, besides the false positive answers, the Defect also retest the determined visual sensitivity thresholds Curve can also be helpful to identify trigger-happy patient to determine Short-term Fluctuation (SF), a further index behavior. See FIG 8-10 for more information on how to de ined in TABLE 7-2. STEP 3 – IDENTIFY DIFFUSE VISUAL FIELD DEFECTS NEED FOR THE DETECTION OF DIFFUSE DEFECTS It is helpful to be alerted to the presence of diffuse defects cataracts, glaucoma, retinal and neurological diseases), early in the process of visual ield interpretation, because they may also indicate the presence of untrustworthy although they are commonly caused by pathology (e.g., visual ield results. STEP 3 – IDENTIFY DIFFUSE VISUAL FIELD LOSS 3 Diffuse loss? FIGURE 8-9 Diffuse visual field loss should ideally be identified early on, as it can be a sign of both a pathology leading to diffuse defects or an untrustworthy visual field. Diffuse defects are present when most visual ield defect. The etiology of diffuse and local visual ield defects locations show defects of approximately the same is presented in TABLE 8-1. magnitude (i.e., there is no apparent visual ield loss pattern). Conversely, a visual ield with a local defect In clinical decision-making it is essential to clarify the is characterized by a speci ic defect pattern in which cause of diffuse defects. If pathology can be ruled out, certain visual ield points are affected more than others. the visual ield should be treated as potentially untrust- Diffuse loss can also occur in the presence of a local worthy and should be retaken, if clinically relevant. Step-by-step interpretation of a visual field 141 THE ETIOLOGY OF DIFFUSE AND LOCAL VISUAL FIELD DEFECTS TABLE 8-1 EXAMPLES OF PATHOLOGIES EXAMPLES OF UNTRUSTWORTHY RESULTS DIFFUSE • Lens opacity (e.g., cataract) • Incorrect refraction (WIDESPREAD) • Cornea opacity (e.g., Fuchs dystrophy) • Incorrect patient age DEFECT • Dense vitreous opacity • Small pupil size • Any advanced pathology resulting • Learning effect in severe visual ield loss • Distraction (e.g., advanced glaucoma) • Fixation loss • Fatigue LOCAL DEFECT • Glaucoma • Lens rim artifact • Age-related macular degeneration • Lid artifact • Hemianopia • Quadrantanopia • Vitreous opacity It is important to note that when advanced visual ield loss To quickly identify the presence of diffuse defects, the is present (e.g., MD > 20 dB), most visual ield locations Defect Curve is useful. are affected. As a result, diffuse loss is always present. DEFECT CURVE The Defect Curve is a graphical representation that pro- representations, and also provides other clinically valu- vides a summary of the visual ield and distinguishes able information, as shown in FIG 8-10. For more details between local and diffuse defects.⁹ In clinical practice it of the design and de initions of the Defect Curve, see is very helpful in alerting the clinician to the presence of BOX 7A. diffuse defects that may be missed by looking at other 142 Chapter 8 | Clinical interpretation of a visual field DEFECT CURVE – INTERPRETATION AID NORMAL BORDERLINE DIFFUSE DEFECT Defect Curve within normal Limited diagnostic value Parallel downward shift of band Defect Curve 1 Rank 59 1 Rank 59 1 Rank 59 -5 -5 -5 0 0 0 5% 5% 5% 5 5 5 Defect (dB) 10 95% 10 10 95% 95% 15 15 15 20 20 20 25 25 25 LOCAL DEFECT LOCAL & DIFFUSE DEFECT ADVANCED Drop of Defect Curve Parallel downward shift on Limited diagnostic value on the right the left and drop on the right 1 Rank 59 1 Rank 59 1 Rank 59 -5 -5 -5 0 0 0 5% 5% 5% Defect (dB) 5 5 5 10 95% 10 95% 10 95% 15 15 15 20 20 20 25 25 25 TRIGGER-HAPPY HEMISPHERE DEFECTS QUADRANT DEFECTS Steep rise of Defect Curve Vertical drop of Defect Vertical drop of Defect on the left Curve in the center Curve towards the right 1 Rank 59 1 Rank 59 1 Rank 59 -5 -5 -5 0 0 0 5% 5% 5% Defect (dB) 5 5 5 10 95% 10 95% 10 95% 15 15 15 20 20 20 25 25 25 FIGURE 8-10 The Defect Curve alerts the clinician to the presence of diffuse defects and allows a rapid distinction to be made between local and diffuse defects in early to moderate disease. It furthermore allows the identification of trigger-happy patients and has a characteristic shape for localized hemisphere and quadrant defects. Note that it is of limited diagnostic value in borderline (i.e., suspect) situations or in advanced pathology. The interpretation of the Defect Curve is based on its local defects are present when there is a drop on the graphical representation and is straightforward. A right-hand side of the Defect Curve (steepening of the visual ield is normal when the entire Defect Curve lies downward slope), while the left side remains within the within the normal band (i.e., the 90% con idence inter- normal band. If both local and diffuse defects are present, val). Diffuse defects are present when there is a parallel there is both a parallel downward shift on the left and a downward shift of the Defect Curve. Alternatively, only drop on the right. Step-by-step interpretation of a visual field 143 The Defect Curve can also identify trigger-happy re- nearly vertical drop at a given location along the curve. sponse behavior, which results in a steep slope above FIG 8-11 illustrates the usefulness of the Defect Curve in a the normal band on the left. Hemisphere and quadrant clinical situation. defects, on the other hand, usually show a characteristic EXAMPLE OF THE CLINICAL USEFULNESS OF THE DEFECT CURVE LOCAL DEFECT DIFFUSE DEFECTS OF VARIOUS MAGNITUDE 1st Test 2nd Test 3rd Test 4th Test 5th Test Grayscale (Comparison) 1 Rank 59 1 Rank 59 1 Rank 59 1 Rank 59 1 Rank 59 -5 -5 -5 -5 -5 Defect Curve 0 0 0 0 0 5% 5% 5% 5% 5% Defect (dB) 5 5 5 5 5 10 95% 10 95% 10 95% 10 95% 10 95% 15 15 15 15 15 20 20 20 20 20 25 25 25 25 25 FIGURE 8-11 This example shows a series of five visual field tests of a patient with glaucoma with a local superior nasal de- fect that deepens from the 1st to the 5th test. In addition, visual fields 2 to 5 show diffuse defects of various magnitudes. The diffuse defect is most pronounced on the 3rd test, as can be seen from the large parallel downward shift of the Defect Curve. An inspection of the Defect Curve thus immediately alerts the clinician to the presence of the fluctuating diffuse defect. In this example, the near-absence of diffuse defect on the 4th and 5th test indicates that the diffuse loss observed on the 3rd test was due to fluctuation and not pathology. While the Defect Curve is very helpful and straightfor- cally remain within the normal band. In severe pathology, ward to interpret in early to moderate disease, it has lim- most visual ield points are affected to some extent and ited clinical usefulness in suspect situations or advanced absolute defects are not drawn on the Defect Curve. As a disease. In suspect situations, all visual ield points typi- result, the Defect Curve lies in the lower left-hand corner. CORRECTING FOR DIFFUSE DEFECTS Local and diffuse defects may occur together, for exam- which the unspeci ic, diffuse defect is removed, as shown ple in glaucoma patients who also have cataracts. In such in FIG 7-16. cases, the diffuse defects may mask localized defects. It is therefore desirable to distinguish between the local The corrected representations provide very helpful and diffuse visual ield components, in order to analyze clinical information to determine whether there is local the local visual ield loss independently. To achieve this, loss when diffuse loss is also present, as illustrated in Octopus perimeters offer corrected representations, in FIG 8-12. 144 Chapter 8 | Clinical interpretation of a visual field EXAMPLE OF THE CLINICAL USEFULNESS OF THE CORRECTED REPRESENTATIONS LOCAL DEFECT DIFFUSE DEFECTS OF VARIOUS MAGNITUDE 1st Test 2nd Test 3rd Test 4th Test 5th Test Grayscale (Comparison) Probabilities Corrected Probabilities FIGURE 8-12 Example of the glaucoma patient with a local superior nasal defect presented in Figure 8-11. Due to the presence of fluctuating diffuse defects of various magnitudes, the extent of the local defect is difficult to judge. This is the purpose of the Corrected Probabilities representation, which eliminates the influence of diffuse defect and allows the identification of local defects. The corrected representations are very helpful when dif- When there is advanced visual ield loss (e.g., MD > 20 dB), fuse loss is present or suspected, as can be seen in the correcting the visual ield for diffuse loss does not pro- borderline, diffuse loss and diffuse and local loss exam- vide clinically useful information, because most visual ples in FIG 8-1. However, when mainly local defects are ield locations are relatively severely affected. Local present, the corrected representations are very similar to defects no longer exist in this situation, because the the uncorrected representations and thus provide only entire visual ield is affected. This can be seen in the limited additional information, as is visible in the normal advanced example of a constricted glaucoma visual and local loss examples shown in FIG 8-1. ield in FIG 8-1. Step-by-step interpretation of a visual field 145 STEP 4 – DISTINGUISH BETWEEN NORMAL AND ABNORMAL VISUAL FIELDS NEED TO DISTINGUISH BETWEEN NORMAL AND ABNORMAL VISUAL FIELDS Distinguishing between normal and abnormal visual ields smaller than normal luctuation. In sum, the challenge is is challenging because 1) there is luctuation in healthy to detect faint signals within noise. For example, there are eyes, 2) this luctuation is not uniformly distributed across borderline ields which may remain stable and normal, the visual ield, as shown in FIG 2-11, and 3) subtle visu- while others, although appearing the same, have already al ield defects, as they occur in early glaucoma, are often undergone the irst steps towards pathology. STEP 4 – DISTINGUISH BETWEEN NORMAL AND ABNORMAL VISUAL FIELDS 4 Borderline or significant local loss? FIGURE 8-13 Before analyzing a visual field in detail, statistical analysis is used to assess whether a visual field is within nor- mal limits, or is abnormal. The Probabilities and Corrected Probabilities are used to achieve this essential step, which results from the normal fluctuation present in perimetry. In view of the challenges mentioned above, there is a need the direct assessment of the visual ield sensitivity thresh- for representations that allow for the distinction between olds can be very challenging. normal and abnormal visual ield locations. This is the purpose of the Probabilities and Corrected Probabilities It is thus worth looking at these representations prior to representations, which employ statistical analysis to performing an in-depth analysis of the visual field in distinguish between normal and abnormal visual ields. order to 1) avoid spending unnecessary time on analysis These representations are especially useful in borderline of a normal visual ield and 2) avoid confusion between situations or to detect subtle visual ield change in which normal luctuation and truly abnormal visual ields. PROBABILITIES AND CORRECTED PROBABILITIES The Probabilities and Corrected Probabilities repre- given test location. Increasingly darker symbols are used sentations serve the purpose of distinguishing between to show the decreasing probability that a given visual normal and abnormal visual ields. They show the prob- ield result would be obtained for a person with an aver- ability (P) that a person of the same age with an average age normal visual ield (FIG 8-14). For more details on the normal visual ield (or one with a visual ield corrected de initions used in the Probabilities representations, see for diffuse loss in the case of the Corrected Probabilities FIG 7-9, 7-10 and 7-18. representation) has a certain visual field result at a 146 Chapter 8 | Clinical interpretation of a visual field PROBABILITIES AND CORRECTED PROBABILITIES – INTERPRETATION AID PROBABILITIES CORRECTED PROBABILITIES DEFINITION INTERPRETATION Probability that a person with a normal visual field shows this result P > 5% Likely normal location P < 5% Potentially abnormal location P < 2% P < 1% P < 0.5% Highly likely abnormal location Corrected for diffuse defect FIGURE 8-14 The various symbols on the Probabilities representations show the likelihood that a person with a normal visual field would show a given sensitivity loss. For example, the black square (P < 0.5) indicates that while it is possible that a person with an average normal visual field could obtain that defect value, the probability of this occurring is very small. Note that the Corrected Probabilities representation shows the same information, but is adjusted to remove diffuse visual field defects and is based on the Corrected Comparison representation. The clinical interpretation of the Probabilities repre- determine a visual ield as abnormal, these guidelines sentation is straightforward in that it is easy to see the typically require the presence of one or more clusters of pattern of visual ield loss marked by dark symbols. abnormal visual ield locations that are consistent with However, there are some factors to be aware of in clinical the expected visual ield loss pattern of a disease. This decision-making. Firstly, there are no criteria allowing is because it is highly unlikely that such clusters would for an unambiguous distinction between normal and form in normal visual ields. If, however, the distribution of abnormal visual ields. Secondly, it is common to have a a few likely abnormal test locations is random and does few random test locations that show a P value lower than not correspond with a disease pattern, this can often be 5% in normal visual ields. For further details concerning attributed to normal luctuation. FIG 8-15 illustrates how these points, see FIG 7-9 and 7-10. to clinically interpret the Probabilities plots of several visual ields with potential early glaucomatous visual Due to these factors, the Probabilities representation ield loss, in which the magnitude of visual ield loss, as must be clinically interpreted with care. Depending on illustrated in the Grayscale of Comparison representa- the pathology, different clinical guidelines are available tion, is similar. to de ine visual ield abnormality and severity.¹⁰,¹¹ To Step-by-step interpretation of a visual field 147 CLINICAL INTERPRETATION OF PROBABILITIES IN BORDERLINE SITUATIONS GRAYSCALE (Comparison) PROBABILITIES DESCRIPTION INTERPRETATION Number of locations at Likely normal P < 5% 2 P < 2% 2 P < 1% 1 Random distribution of likely abnormal locations Number of locations at Likely normal p < 5% 2 Two adjacent likely abnormal test locations, no cluster Number of locations at Likely abnormal p < 5% 2 Investigate further P < 2% 1 P < 1% 1 P < 0.5% 2 Five likely abnormal locations clustered in an inferior partial arcuate defect pattern One likely abnormal location at random position Number of locations at Likely abnormal p < 5% 7 Investigate further P < 2% 3 P < 0.5% 1 Six likely abnormal locations clustered in a superior partial arcuate defect pattern Three likely abnormal locations clustered in an inferior, paracentral defect pattern FIGURE 8-15 The visual field results obtained from four potential early glaucoma cases are presented. They are challenging to interpret by simply looking at the relative sensitivity loss, which is marked with yellow in the Grayscale of Comparison repre- sentation. In the two examples at the top, the few randomly distributed test locations with a probability smaller than 5% also occur frequently in normal visual fields. The absence of clusters of likely abnormal visual field locations suggests that these two examples can be interpreted as likely normal. In the two examples at the bottom, the few test locations with a probability smaller than 5% are organized in clusters and may be interpreted as likely abnormal. 148 Chapter 8 | Clinical interpretation of a visual field The Probabilities representation is the key graph to look In more advanced disease, however, the Probabilities at in borderline situations because it is better suited than representation loses diagnostic value because once the other representations to distinguish between normal disease has progressed to a certain level, most visual and abnormal visual ields, as illustrated in FIG 8-15. In ield points are highly unlikely to be normal at a signi i- early to moderate disease, it is mainly helpful to detect cance of P < 0.5%. Even though there might still be visual subtle change, as sensitivity loss is also apparent from ield worsening, it may no longer be apparent from the the Comparison representations, as can be seen in the Probabilities representation, as illustrated in FIG 8-16. examples shown in FIG 8-1. Methods offered to detect and measure progression are given in Chapter 9. LIMITATIONS OF THE PROBABILITIES REPRESENTATION IN ADVANCED DISEASE PROGRESSIVE ADVANCED GLAUCOMA 1st Test 2nd Test 3rd Test 4th Test 5th Test Grayscale (Comparison) Probabilities FIGURE 8-16 Example of a series of visual fields from a patient with progressing advanced glaucoma. Even though the visual field is worsening over time, the change is not apparent in the Probabilities representation because most visual field locations already show a probability of P < 0.5% in the 1st of the 5 tests. In case of diffuse loss, the Corrected Probabilities repre- localized loss independently of the diffuse defect, as is sentation should also be consulted to assess abnormal shown in FIG 8-12. Step-by-step interpretation of a visual field 149 STEP 5 – ASSESS SHAPE AND DEPTH OF DEFECT NEED FOR ASSESSING SHAPE AND DEPTH OF DEFECT Once it has been determined that a visual ield is trust- severity of the visual ield defect, and to indicate potential worthy and abnormal, the shape of the defect area and further diagnostic tests. Typical visual ield defects for the depth of the defect should be assessed. Since different glaucoma, neuro-ophthalmic and retinal diseases are pathologies show different disease patterns, these char- presented in FIG 5-1, 5-7 and 5-9. acteristics are helpful to determine the possible cause and STEP 5 – ASSESS SHAPE AND DEPTH OF DEFECT 5 Assess shape & depth of defect. Typical for glaucoma? FIGURE 8-17 The shape and depth of a defect provide valuable clues to identify and characterize pathology. They can be analyzed from a graphical (Grayscale of Comparison) or numerical (Comparison and Corrected Comparison) map. GRAYSCALE OF COMPARISON, COMPARISON AND CORRECTED COMPARISON The Comparison representations are key in that they Three representations are available. The Grayscale of provide a thorough analysis of both the depth and shape Comparison is a color map of a patient’s visual ield of defects, thus providing information about the possible loss. The Comparison and Corrected Comparison repre- causes of the visual ield loss. They do so by comparing sentations show the same information using numerical the measured sensitivity thresholds to a normal visual maps. An overview of how to clinically interpret them ield, as shown in FIG 7-5. is provided in FIG 8-18. For further details, see FIG 7-6, 7-7 and 7-17. 150 Chapter 8 | Clinical interpretation of a visual field GRAYSCALE OF COMPARISON, COMPARISON AND CORRECTED COMPARISON – INTERPRETATION AID GRAYSCALE (Comparison) DEFINITION INTERPRETATION Sensitivity loss [% of normal] 0..10 Normal 11..22 23..34 35..46 47..58 Visual field loss 59..70 (the darker the worse) 71..82 83..94 95..100 COMPARISON CORRECTED COMPARISON DEFINITION INTERPRETATION 9 8 + + 8 + 15 13 5 25 + + 10 8 + 20 22 10 16 5 + Sensitivity loss < 5 dB Normal 13 11 + 22 8 6 + 17 12 2216 7 7 1610 + + 10 21 30 + + 15 24 22 Sensitivity loss [dB] Visual field loss 75 17 31 + + 12 26 8 5 + + + + + + + + (the larger the worse) 6 + 5 + + 6 + + + + + + 10 + 5 + + 5 + + + + Absolute defect Maximum visual field loss 8 + + + + + + + (i.e., Sensitivity threshold 0 dB) 5 + + + 10 8 7 6 5 7 5 + + + + + 8 + + + Corrected for diffuse defect FIGURE 8-18 The Grayscale of Comparison is a color map that is especially useful to determine the shape of the sensitivity loss, whereas the Comparison and Corrected Comparison representations are numerical maps showing sensitivity loss in dB. The Corrected Comparison representation shows localized loss only. All representations are key to identifying possible causes of disease. The Grayscale of Comparison representation is ideally ity of test locations (see FIG 2-9 for more information), suited to assess defect shapes and to gain a quick irst it represents a patient’s true sensitivity loss. However, impression of a patient’s overall visual ield loss. Since caution is essential when interpreting the precise it is intuitive to understand, it is also very useful for boundaries of the Grayscale of Comparison representa- patient education. tion, as its high spatial resolution might give the impres- sion that the detailed boundaries of a defect are known, Since it is based on the Comparison representation, which is not true, as explained in BOX 8A. which eliminates the effect of patient age and eccentric- Step-by-step interpretation of a visual field 151 BOUNDARIES OF GRAYSCALE OF COMPARISON CAN BE MISLEADING BOX 8A It is important to remember that in perimetric testing only a discrete number of locations are tested, as illustrated in FIG 4-4. As a result, there are large gaps between test points for which no information is available. These gaps are illed with interpolated (i.e., probable or likely) information in the Grayscale of Comparison representation. The boundaries of a visual ield defect shown in the Grayscale of Compari- son representation are thus only estimated and may not re lect the exact boundaries. The resolution of a test is only as good as that of the test pattern. This is important to remember when interpreting Grayscale of Comparison maps, to avoid incorrect interpretation of a slightly changing defect pattern. GRAYSCALE OF COMPARISON IS AN INTERPOLATED COLOR MAP OCTOPUS GRAPHIC REALISTIC GRAPHIC Gaps between test points Poor spatial resolution are interpolated in most perimetric tests Sensitivity loss [% of normal] 0..10 11..22 23..34 35..46 47..58 59..70 71..82 83..94 95..100 It is essential to be aware that the Grayscale of Comparison is an interpolated visual field map, where gaps between visual field points are filled by interpolation (left). Its true spatial resolution is much poorer, as illustrated in the panel on the right. Conversely, both the Comparison and Corrected Com- The Comparison representations should be looked at in parison representations are better suited to assess all clinical situations, as the shape and depth of defect precise defect depth than the Grayscale of Comparison are key information sources in any clinical situation, from representation, because they show visual ield loss in early to advanced disease, as shown in the examples in 1 dB steps. Even small sensitivity loss can be seen in FIGURE 8-1. An exception may be borderline visual ields these representations. While the Comparison represen- in which defect depth is small and thereby challenging to tation shows the actual local visual ield loss (deviation distinguish from normal luctuation. In those situations, of measured sensitivity threshold from normal), the the Probabilities representations are better suited to Corrected Comparison representation shows localized identify the shape and depth of a potential defect. visual ield loss only, as explained in FIG 7-16 and 7-17. 152 Chapter 8 | Clinical interpretation of a visual field STEP 6 - ASSESS CLUSTER DEFECTS IN GLAUCOMA NEED TO ASSESS CLUSTER DEFECTS IN GLAUCOMA Typical glaucomatous defects (just like other neurolog- the retinal nerve iber bundles in the retina. Thus, in the ical defects caused by localized retinal nerve iber dam- assessment of glaucomatous visual ield defects, one is age) consist of a cluster of adjacent defective visual ield looking for a cluster of affected visual ield locations both locations (FIG 5-1) that correspond to the path followed by in the Probabilities and Comparison representations. STEP 6 – ASSESS CLUSTER DEFECTS IN GLAUCOMA 6 Glaucoma only: Significant cluster defects? FIGURE 8-19 Assessment of visual field defects in clusters is helpful for the detection of subtle glaucomatous changes. This is the purpose of the Cluster and Corrected Cluster Analysis. Many glaucomatous visual ield changes, however, are and speci ic to analyze individual test locations to identify smaller than the normal range of luctuation and are not clusters of visual ield defects. marked as abnormal. In those cases, the Probabilities rep- resentation may not be sensitive enough to detect very Therefore, further representations are offered to facilitate subtle glaucomatous visual ield loss. In addition, it is the interpretation of localized glaucomatous visual ield time consuming, subjective and not suf iciently sensitive loss. The Cluster Analysis and the Corrected Cluster Anal- ysis were developed for this purpose. CLUSTER ANALYSIS AND CORRECTED CLUSTER ANALYSIS The Cluster Analysis and the Corrected Cluster Analysis Similar to the Probabilities representation, they show have been designed speci ically for glaucoma and are the probability (P) that a person with a normal visual very sensitive to detect subtle glaucomatous visual ield ield (or one with a visual ield corrected for diffuse loss defects. In Cluster Analysis, visual ield locations corre- in the case of the Corrected Cluster Analysis) would sponding to the same retinal nerve iber layer (RNFL) have a given Cluster value. They thus provide clinical bundle are grouped in 10 visual ield clusters and used information as to whether a visual ield cluster is likely to calculate the respective average Cluster Mean Defects to be normal or not. This is summarized in FIG 8-20. For (Cluster MDs). further details of the design and the de initions of both Cluster and Corrected Cluster Analysis, see FIG 7-12, 7-13 and 7-19, and BOX 7B. Step-by-step interpretation of a visual field 153 CLUSTER ANALYSIS AND CORRECTED CLUSTER ANALYSIS – INTERPRETATION AID CLUSTER ANALYSIS CORRECTED CLUSTER ANALYSIS DEFINITION INTERPRETATION Cluster Probability that a 15.8 10.4 MD [dB] person with a 11.3 5.9 normal visual field shows this result 9.1 24.9 3.7 19.5 9.7 4.3 4.4 + + P > 5% Likely normal cluster 8.3 2.7 2.9 + 2.7 P < 5% Potentially abnormal cluster 5.8 + 3.5 + 8.3 P < 1% Highly likely abnormal cluster Corrected for diffuse defect FIGURE 8-20 The Cluster Analysis representations group defects into ten clusters according to the paths followed by the nerve fiber bundles in the retina. Highly likely normal clusters (P > 5%) are marked with a “+” symbol, and likely abnormal Cluster Mean defects are displayed in normal font (P < 5%) or bold font (P < 1%). The Corrected Cluster Analysis representa- tion is similar, but eliminates diffuse visual field loss and solely considers local loss. Clustering visual ield defects according to the paths fol- Probabilities representations.¹² This is due to the fact lowed by the nerve fiber bundles in the retina is more that the clustering and averaging procedure signi icantly sensitive to detect glaucoma and some other optic reduces the in luence of normal luctuation.¹³ This is neuropathies than using individual test locations in the further explained in BOX 8B. CLUSTER ANALYSIS IS HIGHLY SENSITIVE TO DETECT GLAUCOMA BOX 8B Cluster Analysis has been shown to be more sensitive to detect subtle glaucomatous change¹² than look- ing at individual test locations, due to the reduction of the in luence of normal luctuation. For example, in the clinical situation shown in the igure included in this box, most test locations in the supero-nasal cluster show a small numerical sensitivity loss (as shown in the adapted Comparison representation, which is not available on Octopus perimeters). This sensitivity loss is on average larger than the one in the infero-nasal cluster. However, when looking at the sensitivity losses at a speci ic test location in the supero-nasal segment, most of these sensitivity losses are too small to manifest as a likely abnormal vi- sual ield location in the Probabilities representation. As a result, such a visual ield would be considered as normal, as shown in FIG 8-15. However, it is highly unlikely that all test locations within the same cluster show such a degree of sensitiv- ity loss. By averaging the sensitivity losses of all test locations within the cluster, this cluster is very likely not to be normal at a signi icance of P < 1%. As a consequence, it can be concluded that the visual ield is likely to be abnormal. Note that the Cluster Analysis uses an idealized graphical display. Consult BOX 7B for the real boundaries of the Cluster Analysis. 154 Chapter 8 | Clinical interpretation of a visual field ILLUSTRATION OF THE CLINICAL USEFULNESS OF CLUSTER ANALYSIS SENSITIVITY LOSS PROBABILITIES CLUSTER ANALYSIS (Adapted Comparison representation) 3 1 3 4 2 3 3 5 + 2 1 3.2 2 6 2 0 0 2 2 2 1 + + 0 0 3 1 3 1 + 0 3 1 2 2 4 1 0 0 + 1 1 1 1 3 1 + + 3 1 1 2 2 2 2 1 0 1 1 + + 0 3 0 0 2 0 2 4 This example highlights the high sensitivity of Cluster Analysis for the detection of subtle glau- comatous visual field defects. When looking at the sensitivity loss of the individual test locations (left) in the superior arcuate cluster (red shading), only one location is marked as abnormal in the Probabilities representation (center). However, most locations are slightly, but not significantly elevated, which results in a significantly abnormal (P < 1 %) Cluster MD in the Cluster Analysis. Besides being more sensitive than the Probabilities rep- potentially abnormal locations to detect clusters of ab- resentation to detect early glaucomatous visual ield loss normal visual ield locations. This makes the Cluster (FIG 8-21), the Cluster Analysis is also easier to read and Analysis a fast and useful tool in clinical practice. avoids having to spend time identifying and counting ILLUSTRATION OF THE HIGH SENSITIVITY OF CLUSTER ANALYSIS TO DETECT GLAUCOMA VISUAL FIELD ORIENTATION GRAYSCALE (Comparison) PROBABILITIES CLUSTER ANALYSIS Two superior paracentral locations at p < 5% Supero-nasal cluster at p < 1% 2.3 + + + + + + + + + FIGURE 8-21 Example of a borderline visual field. By just looking at the Grayscale of Comparison (left) and Probabilities (middle) representations, one may interpret this visual field as likely to be normal, as there is no pattern of contiguous ab- normal locations. However, examination of the Cluster Analysis (right) shows a small, but significant superior arcuate defect pattern, which calls for further investigation. Step-by-step interpretation of a visual field 155 As with the interpretation of the Probabilities represen- Thus, clinicians can be more con ident that a cluster at tations, some caution is essential in the clinical interpre- P < 5 % is truly abnormal when a contiguous cluster is tation of the Cluster representation. This is because one also abnormal, or when there is a spatially corresponding random cluster showing a P value smaller than 5% is structural defect. expected to occur frequently, also in normal visual ields. STEP 7– WHERE TO LOOK FOR STRUCTURAL DEFECTS NEED TO IDENTIFY RELATIONSHIP BETWEEN FUNCTIONAL AND STRUCTURAL DAMAGE IN GLAUCOMA When an eye is investigated for glaucoma, both functional ticularly dif icult in eyes with early stages of the disease. alterations and structural damage (neuroretinal rim A mild alteration in the visual ield has more clinical value tissue loss; decrease of retinal nerve iber layer thickness, for decision-making if a spatially corresponding structur- RNFLT) should be considered. al alteration is also detected, and vice versa. However, it is not quite straightforward to understand the geometric In clinical practice, spatially corresponding structural and relation between the usual presentation of the visual visual ield alterations are necessary to detect glaucoma ield (perimetry) and the structural results (i.e., fundus and to separate glaucoma from other diseases. This is par- photography or optical coherence tomography OCT). STEP 7 – WHERE TO LOOK FOR STRUCTURAL DEFECTS 7 Glaucoma only: Where to look for structural defects. Is there a relationship? FIGURE 8-22 Knowing where to look for structural defects to identify a spatial relationship between structural and functional results is helpful for the detection of subtle glaucomatous changes. This is the purpose of the Polar Analysis. Glaucomatous structural damage occurs at the optic disc While there is a correspondence between the structural and results in a degeneration of the nerve ibers that con- and functional defect locations, the reference coordinates nect the damaged disc location to the retina. Perimetric are different. Different conventions are therefore used to testing presents stimuli at various retinal locations along display structural and functional results. See BOX 8C for the defective layer and is able to identify the defect. more information on the spatial relationship between structural and functional results. 156 Chapter 8 | Clinical interpretation of a visual field BOX 8C ANATOMICAL RELATIONSHIP BETWEEN STRUCTURAL AND FUNCTIONAL RESULTS Glaucomatous structural damage can be observed at the level of the optic disc and results in a degen- eration of the nerve ibers that connect from the damaged disc location to the retina. As a result, light entering the retina anywhere along the defective nerve iber bundle cannot be processed and this results in visual ield defect at the respective retinal location. Furthermore, while visual ield results are oriented like a real-world image associated with post-process- ing in the brain, the real world image is lipped both horizontally and vertically when passing through the lens and entering the retina, and thus the structural and visual ield results are also lipped horizontally and vertically. This means that a superior visual defect is produced by inferior optic nerve head damage and a nasal visual ield defect is produced by temporal optic nerve head damage. In addition, while visual ields are oriented from a patient’s view, structural results are oriented from a doctor’s view, looking onto a patient’s retina. Due to these different viewpoints, the graphical represen- tations of structural and functional results appear like mirror images lipped at the horizontal axis, as is illustrated in the graphic below. SPATIAL RELATIONSHIP BETWEEN VISUAL FIELDS AND STRUCTURAL RESULTS VISUAL FIELD ORIENTATION STRUCTURAL ORIENTATION COMPARISON RETINA WITH OPTIC DISC S 9 8 8 + 15 13 5 25 22 10 13 11 + 22 12 22 16 7 + 10 21 30 270 7 5 17 31 T 5 N N 0 180 T 8 + + + 6 + 5 + + 6 90 10 + 5 + + 8 + + + 5 + 10 8 7 6 5 7 113dB 13 13d 3dB 3d 3dB dB 8 + I I Structural damage and visual field results are flipped across the horizontal midline (i.e., a superior visual fi eld defect corresponds to an inferior structural defect at the corresponding location at the optic disc). Note that even though structural and functional results are also flipped across the vertical midline, the defects are displayed on the same side because of the different viewing directions of the patient (visual field) and the observing clinician (structure). Due to the different coordinates used to display structural and functional representations in an intuitive way, to save and functional results it is useful to have an analysis tool valuable time. This is the purpose of the Polar Analysis. that facilitates inding the relationship between structural Step-by-step interpretation of a visual field 157 POLAR ANALYSIS The Polar Analysis representation is designed to facili- The Polar Analysis displays individual visual ield defects tate the identi ication of the spatial relationship between as red bars along the perimeter of the optic disc. The lo- structural and functional results by mapping visual ield cation of the bar indicates the corresponding structural defects onto the optic disc in the same orientation as a area, and the length of the bar shows the amount of sen- structural result. This allows intuitive side-by-side sitivity loss in dB, with longer bars indicating greater comparison between structural and functional results. magnitude of defect, as shown in FIG 8-23. For more infor- mation on the design of the Polar Analysis, see FIG 7-14. POLAR ANALYSIS - INTERPRETATION AID POLAR ANALYSIS DEFINITION INTERPRETATION Defect [dB] Location of potential structural damage on • Length of bar indicates defect size [dB] optic disc • Position along the optic disc represents the entry angle of RNFL fibers associated to each test location Short bar Normal location (Within gray normal range) S 30 20 10 N T Long bar Abnormal location [dB] Normal range I S Superior I Inferior N Nasal T Temporal FIGURE 8-23 The Polar Analysis maps functional results onto the optic disc, to appear like a structural result. This assists in assessing the spatial relationship between visual field defects and possibly associated structural defects. Clinical use of the Polar Analysis is straightforward. After losses have occurred. Using this graphical representation, placing it next to a structural result taken during the same the visual ield results can be related to structural results, time period, a clinician should look for locations in the thereby making detailed and accurate comparison of Polar Analysis with a cluster of red bars that are outside of damaged segments much easier (see FIG 8-24 for an ex- normal range. This allows clinicians to see the signi icantly ample). The results of the Polar Analysis have been shown deviating visual ield test locations that may correspond to correlate well with structural OCT results.¹⁴ to structural regions of the optic nerve head rim where 158 Chapter 8 | Clinical interpretation of a visual field ILLUSTRATION OF THE CLINICAL USEFULNESS OF THE POLAR ANALYSIS VISUAL FIELD ORIENTATION GRAYSCALE (Comparison) PROBABILITIES CLUSTER ANALYSIS Two superior paracentral locations at p < 5% Supero-nasal cluster at p < 1% 2.3 + + + + + + + + + STRUCTURAL ORIENTATION POLAR ANALYSIS FUNDUS IMAGE OCT MACULA MAP Subtle visual field loss Splinter hemorrhage and subtle RNFL loss Retinal ganglion cell loss at 7 o’clock position at 7 o’clock position at 7 o’clock position S 30 20 10 T N [dB] I FIGURE 8-24 Patient with suspected very early glaucoma. While the Probabilities representation is not sensitive enough to show significant visual field loss, the Cluster Analysis shows that the supero-nasal cluster is likely abnormal at P < 1%. The Polar Analysis shows a potential defect at the 7 o’clock position of the optic disc, where a very subtle disc hemorrhage is also found in the fundus photo (darker area within the blue circle). The Macula map picks up the loss of retinal ganglion cells at a comparable location. Due to the spatial relationship between the subtle defect in the visual field (Polar Analysis) and structur- al measurements (Fundus Image and Macula Map), glaucoma is confirmed. Step-by-step interpretation of a visual field 159 STEP 8 – ASSESS SEVERITY NEED TO ASSESS SEVERITY OF VISUAL FIELD LOSS A key element prior to clinical decision-making is to as- It is desirable to have summarizing quantitative measures sess the severity of visual ield loss in an objective manner, (i.e., global indices) that allow for a characterization of a in order to decide on an adequate clinical intervention. visual ield in a few words. Summarizing global indices¹⁵ This is challenging to perform from the representations are needed for visual ield severity staging systems, but discussed so far because there is a wide variety of visual they are also very useful when patients are referred, and ield defect patterns and depths. they also ind use in clinical studies or guidelines. An overview of the design and de initions of available global indices is provided in TABLE 7-1. STEP 8 – ASSESS VISUAL FIELD SEVERITY 8 Severity? FIGURE 8-25 Global indices provide useful information to quickly characterize a visual field and to assess disease severity. MEAN DEFECT (MD) The Mean Defect (MD) provides a summary of the over- defects of all test locations, expressed in decibels. Its all severity of visual ield loss, which is useful to assess calculation formula is shown in TABLE 7-1 and its clinical overall disease severity and essential to judge disease relevance is illustrated in FIG 8-26. progression.¹⁵ If a visual ield defect worsens, indepen- dent of whether it is a local or a diffuse defect, MD will The MD is an essential index used in both the Brusini worsen too. As a general interpretation rule, it can thus and Hodapp-Parrish-Anderson glaucoma staging sys- be said that the higher the MD, the greater the visual tems.¹⁰,¹¹,¹⁶,¹⁷ In the Hodapp-Parrish-Anderson system, ield damage. early visual ield defects are characterized by an MD of up to 6 dB, moderate visual ield defects are character- As its name suggests, the MD is a mathematical repre- ized by an MD ranging from 6 to 12 dB, and severe visual sentation of the average of the individual visual ield ield defects have an MD worse than 12 dB. 160 Chapter 8 | Clinical interpretation of a visual field ILLUSTRATION OF THE USEFULNESS OF MD NORMAL SUSPECT EARLY TO MODERATE ADVANCED Diffuse defect Local defect Local & diffuse defect MD -0.2 dB 1 dB 6.3 dB 6.5 dB 10.1 dB 21.7 dB FIGURE 8-26 The Mean Defect (MD) summarizes the severity of visual field loss in one number, for comparison with other patients and to quickly communicate the severity of visual field loss. The examples above show different visual fields with increasingly severe visual field loss. SQUARE ROOT OF LOSS VARIANCE (sLV) In clinical practice, local and diffuse defects typically Clinical interpretation is straightforward. If sLV is small, have very different causes, as shown in TABLE 8-1, and a visual ield is homogeneous (i.e., all defects have ap- therefore require different clinical management. How- proximately the same size), suggesting that the visual ever, the global index MD does not distinguish between ield is normal, or that the deterioration is predom- them, because it is based on an average visual field inantly diffuse, or that the visual ield has advanced, defect. For example, two visual ields with similar MD severe visual ield loss. However, if sLV is larger, then (FIG 8-27) can look completely different, depending on the visual ield is heterogeneous, which means that the whether there is diffuse or local loss. individual defects vary substantially. The larger the sLV, the greater the variability among the different defects. It is thus useful to use an additional global index to It is noteworthy to mention that in early to advanced distinguish between local and diffuse loss. This is the glaucoma, sLV becomes increasingly higher; but in very purpose of the square root of Loss Variance (sLV) which advanced glaucoma, sLV is low, since in this stage most provides a measure of variability of local loss across all visual ield locations are very severely damaged and test locations. The formula used to calculate it is shown the defect pattern is therefore diffuse. in TABLE 7-1. Note that sLV is merely the standard deviation of the local defect values. Step-by-step interpretation of a visual field 161 ILLUSTRATION OF THE USEFULNESS OF sLV DIFFUSE DEFECT LOCAL DEFECT COMPARISON COMPARISON 9 9 8 + 5 8 11 10 5 5 8 5 6 6 8 + 8 10 17 10 + 11 + + 15 13 10 + + 6 7 7 6 26 21 19 23 + + 5 7 7 + + 19 22 22 21 18 + 5 7 7 + 21 15 13 17 7 12 9 6 6 7 + 5 6 + 8 6 6 6 6 5 + + + + + + 5 9 7 + 5 + + + + + 9 15 5 5 + + + + 10 6 + + 7 8 8 6 5 5 + + + + + + 5 5 + + MD 6.3 dB MD 6.5 dB sLV 2.5 dB sLV 8.5 dB 26 23 22 22 15 21 21 21 19 19 18 11 11 17 17 15 15 10 10 10 13 sLV sLV 12 9 9 9 9 9 11 8.5 dB 2.5 dB 10 10 8 8 8 8 8 7 7 7 77 7 7 7 7 7 MD 8 8 8 MD 6 6 6 6 66 6 66 6 6 6.3 dB 6 6 6 6.5 dB 5 5 5 5 5 55 5 5 5 55 5 5 5 + + + + + + ++ + ++ + + + ++ + ++ + + ++ + + + + +++ +++ + FIGURE 8-27 Visual fields with either diffuse defects (left) or local defects (right) appear fundamentally different, but can have similar MD values, as this example illustrates. The square root of Loss Variance (sLV) is then useful to distinguish between the two situations, as sLV is smaller in the case of homogeneous or diffuse visual field defects and larger in the case of heterogeneous or local visual field defects. In short, sLV is a measure of how much the defects at different test locations differ from the mean defect, as illustrated in the graphic at the bottom. 162 Chapter 8 | Clinical interpretation of a visual field sLV is an essential index used in the Brusini Glaucoma used to judge local disease progression in glaucoma. For Staging System¹¹,¹⁶,¹⁷ in combination with MD to divide more information on how to judge disease progression, visual field loss into 5 stages, and is also commonly see Chapter 9. References 163 REFERENCES 1. Lee M, Zulauf M, Caprioli J. The in luence of patient reliability on visual ield outcome. Am J Ophthalmol. 1994; 117:756-761. 2. Advanced Glaucoma Intervention Study. 2. Visual ield test scoring and reliability. Ophthalmology. 1994;101:1445-1455. 3. Gillespie BW, Musch DC, Guire KE, et al. The collaborative initial glaucoma treatment study: baseline visual ield and test-retest variability. Invest Ophthalmol Vis Sci. 2003;44:2613-2620. 4. Johnson CA, Keltner JL, Cello KE, et al. Baseline visual ield characteristics in the ocular hypertension treatment study. Ophthalmology. 2002;109:432-437. 5. Keltner JL, Johnson CA, Cello KE, Wall M, NORDIC Idiopathic Intracranial Hypertension Study Group. Baseline visual ield indings in the Idiopathic Intracranial Hypertension Treatment Trial (IIHTT). Invest Ophthalmol Vis Sci. 2014;55: 3200-3207. 6. Miglior S, Zeyen T, Pfeiffer N, et al. The European glaucoma prevention study design and baseline description of the participants. Ophthalmology. 2002;109:1612-1621. 7. Bengtsson B, Heijl A. False-negative responses in glaucoma perimetry: indicators of patient performance or test reliability? Invest Ophthalmol Vis Sci. 2000;41:2201-2204. 8. European Glaucoma Society. Terminology and Guidelines for Glaucoma. 4th ed. Savona: PubliComm; 2014. 9. Bebie H, Flammer J, Bebie T. The cumulative defect curve: separation of local and diffuse components of visual ield damage. Graefe's Arch Clin Exp Ophthalmol. 1989;227:9-12. 10. Hodapp E, Parrish RK, Anderson DR. Clinical decisions in glaucoma. 1st ed. St. Louis: Mosby; 1993. 11. Susanna R Jr., Vessani RM. Staging glaucoma patient: why and how? Open Ophthalmol J. 2009;3:59-64. 12. Naghizadeh F, Holló G. Detection of early glaucomatous progression with octopus cluster trend analysis. J Glaucoma. 2014;23:269-275. 13. Mandava S, Zulauf M, Zeyen T, Caprioli J. An evaluation of clusters in the glaucomatous visual ield. Am J Ophthalmol. 1993;116:684-691. 14. Holló G, Naghizadeh F. Evaluation of Octopus Polar Trend Analysis for detection of glaucomatous progression. Eur J Ophthalmol. 2014;24:862-868. 15. Flammer J. The concept of visual ield indices. Graefe's Arch Clin Exp Ophthalmol. 1986;224:389-392. 16. Koçak I, Zulauf M, Hendrickson P, Stümp ig D. Evaluation of the Brusini glaucoma staging system for follow-up in glaucoma. Eur J Ophthalmol. 1997;7:345-350. 17. Koçak I, Zulauf M, Bergamin O. Evaluation of the Brusini glaucoma staging system for typing and staging of perimetric results. Ophthalmologica. 1998;212:221-227. 164 165 CHAPTER 9 INTERPRETATION OF VISUAL FIELD PROGRESSION INTRODUCTION Vision-related quality of life is severely diminished both Because progression in diseases such as glaucoma is when diffuse deterioration within the central 30-degrees typically slow, the magnitude of luctuation can be larger of the visual ield (increase of MD) reaches a critical than the annual rate of progression. Identifying disease level and when localized progression prevents the per- progression from a series of visual ields is therefore a formance of normal daily activities (e.g., due to severe challenging and time-consuming task in clinical prac- progression of a localized inferior paracentral scotoma). tice (FIG 9-1). As a result, expert agreement is moderate In clinical practice, it is essential to detect progression at best (45% to 65%).¹-⁵ Statistical progression analy- and to measure its speed (i.e., rate of progression ses greatly support the assessment of progression that expressed as change per year in dB) as early as possible is needed for clinical decision-making. The use of pro- to make decisions about potential interventions before gression software options was shown to improve expert signi icant visual ield loss develops. agreement,¹-⁵ but to also reduce overall visual field analysis time.⁴ 166 Chapter 9 | Interpretation of visual field progression CHALLENGES ASSOCIATED WITH ASSESSING VISUAL FIELD PROGRESSION Stable? 1st Test 2nd Test 3rd Test 4th Test 5th Test 6th Test Or progressing? Glaucoma Patient 1 Glaucoma Patient 2 Glaucoma Patient 3 FIGURE 9-1 Determination of whether visual fields are stable over time or whether they are progressing can be challenging, especially when the change is small and there is considerable fluctuation. This is illustrated with the visual field series of three different patients. The EyeSuite Progression Analysis function of the Octo- The Cluster Trend Analysis and Polar Trend Analysis pus perimeters has been designed to assess visual ield have been speci ically designed to detect subtle glau- progression in an effective and ef icient way. It includes comatous change. The Cluster Trend Analysis assesses the following three types of progression analysis: Global cluster-specific progression within ten nerve fiber Trend Analysis (GTA), (Corrected) Cluster Trend Analy- bundle regions separately, which is particularly useful sis (CTA and CCTA), and Polar Trend Analysis (PTA) in glaucoma in which localized (cluster) progression are shown in FIG 9-2. and stability occur at different locations independently from each other in the same eye. Furthermore, the Polar The Global Progression Analysis measures and statisti- Trend Analysis facilitates the detection of spatially cor- cally classi ies long-term change in the global indices, responding structural and visual ield changes. namely Mean Defect (MD), Diffuse Defect (DD), Local Defect (LD) and square Root of Loss Variance (sLV). The different types of progression analyses make a It not only assesses whether a series of visual ields is statement about whether a visual ield series is stable stable or shows significant change, but also provides or not and also show the location of progression. How- information about the rate of change in dB/year and on ever, it is also important to know the shape, location and the local, diffuse or combined nature of progression. depth of a defect. For example, a deep defect approaching Introduction 167 the fovea solicits a much more aggressive treatment of all visual ield tests are also displayed as a default than a shallow defect in the periphery. To provide this and may be changed to any other single ield represen- information, the Grayscale of Comparison representations tation such as the Cluster Analysis. PROGRESSION ANALYSIS TOOLS AVAILABLE IN OCTOPUS PERIMETERS EYESUITE PROGRESSION ANALYSIS GLOBAL TREND ANALYSIS (CORRECTED) CLUSTER TREND ANALYSIS MD sLV Cluster MD progression? Corrected Cluster MD Stable or overall Increasing/decreasing progression? progression? inhomogeneity? 0 0 4.1 3.5 2.1 1.5 -0.8 6.7 -1.4 6.0 2.6 2.0 0.4 -0.2 1.0 0.4 -0.3 -0.2 15 0.7 0.0 0.6 -0.1 25 15 2000 2001 2002 2003 2004 2005 2000 2001 2002 2003 2004 2005 Slope: 1.9 dB / Yr (p < 0.5%) Slope: 2.1 dB / Yr (p < 0.5%) POLAR TREND ANALYSIS DD LD Diffuse progression? Local progression? Where to look for 0 structural progression? 0 S 30 20 10 N T [dB] I 25 15 2000 2001 2002 2003 2004 2005 2000 2001 2002 2003 2004 2005 Slope: 0.6 dB / Yr Slope: 1.8 dB / Yr (p < 0.5%) SERIES OF VISUAL FIELDS Defect location & depth? FIGURE 9-2 Octopus perimeters offer 3 types of progression analysis to assess visual field change over time. A Global Trend Analysis based on the four global indices MD, sLV, DD and LD, and, for glaucoma, both Cluster (and Corrected Cluster) Trend Analysis and Polar Trend Analysis. In contrast to simply looking at a series of visual fields, most of these analyses employ statistical methods to determine progression. To provide orientation about both defect location, shape and defect depth, the series of Grayscale representations is also provided. 168 Chapter 9 | Interpretation of visual field progression ASSESSMENT OF GLOBAL VISUAL FIELD CHANGE CHANGE OF MEAN DEFECT (MD) AS A MEASURE OF GLOBAL CHANGE To judge whether a current treatment strategy is effec- the area of the visual ield that was tested. It is the aver- tive as well as to make a clinical decision about future age of all individual sensitivity losses and is expressed in interventions, it is essential to know whether, overall, a dB. Consequently, if any visual ield location worsens or visual ield series is stable, worsening or improving. This improves, MD will also change accordingly, even though can be achieved by analyzing the change of the global in- the change may be small. This makes MD a good index to dex Mean Defect (MD) over time. track overall visual ield change. For more information on the de inition of the index MD see TABLE 7-1, and for more The global index MD summarizes the sensitivity loss over information on its clinical interpretation see FIG 8-26. TREND ANALYSIS FOR THE VISUALIZATION OF CHANGE The simplest way to assess MD change is to plot the MD If there is no luctuation and the change in MD over time of each visual ield test in a two-dimensional graph. The is suf iciently large, it is simple to graphically determine MD is plotted on the y-axis and test date is plotted on the whether a series of visual ields is stable, worsening or x-axis. This allows graphical assessment of visual ield improving by drawing a trend line. Intuitively, the trend change over time as shown in FIG 9-3. Because an in- line corresponds to the line that provides the best linear creasing MD represents visual ield worsening, it is most it for all the MD points. If this line is lat, then the visual intuitive to use a scale showing the smallest MD at the ield series is stable, if it is sloping upwards, then the top and the largest at the bottom. series is improving and if it is sloping downwards, then the series is worsening (FIG 9-4). Assessment of global visual field change 169 TREND ANALYSIS FOR DISPLAYING OVERALL CHANGE MD [dB] 0.9 2.2 1.7 4.5 3.3 6.0 Test date 10/2002 01/2003 06/2003 12/2003 06/2004 01/2005 0 5 10 MD [dB] 15 20 25 2002 2003 2004 2005 Test date [years] FIGURE 9-3 A simple way to assess visual field change over time is to draw a two-dimensional graph with the test date of each visual field test on the x-axis and the corresponding MD on the y-axis. By drawing a trend line that provides the best linear fit for the individual MD points (red line), it is easy to see that this visual field series is worsening (downward slope). GRAPHICAL INTERPRETATION OF A TREND LINE STABLE WORSENING IMPROVING 0 0 0 5 5 5 MD [dB] 10 MD [dB] 10 MD [dB] 10 15 15 15 20 20 20 25 25 25 2002 2003 2004 2005 2002 2003 2004 2005 2002 2003 2004 2005 FIGURE 9-4 If visual field change is sufficiently large, just looking at the red trend line allows one to intuitively assess whether a visual field series is stable (flat line, left) worsening (downward sloping line, middle) or improving (upward sloping line, right) over time. 170 Chapter 9 | Interpretation of visual field progression This approach is referred to as trend analysis and is used The steepness of the line is referred to as the slope and for all representations that are part of the EyeSuite Pro- is used to assess the rate of change in MD over time. The gression Analysis. To best it the trend line to the mea- rate of change is expressed in dB per year and is derived sured MD values, linear regression analysis with the by determining the amount of change in MD (y-axis) that ordinary least squares it is used. For more details on this occurs over the selected period of time (x-axis). In FIG approach as well as key characteristics of trend analysis, 9-5, the rate of change for MD is 1.9 dB/year. refer to BOX 9A. SLOPE OF TREND ANALYSIS PROVIDES RATE OF CHANGE 0 Slope 1.9 dB/year 5 1.9 dB 10 MD [dB] 15 20 1 year 25 2002 2003 2004 2005 Test date [years] FIGURE 9-5 To determine the rate of overall visual field change, the best-fit line is drawn through the MD data points in the Global Trend Analysis. Once this trend line is drawn, the actual data points can be discounted and the rate of change can be determined using the slope of the trend line. The rate of change is automatically expressed in dB per year. In this example, the slope or rate of change is 1.9 dB/year. Assessment of global visual field change 171 USING PROBABILITIES TO DISTINGUISH BETWEEN STABLE AND CHANGING VISUAL FIELD SERIES A key challenge in the assessment of visual ield progres- Worsening at P < 5%: this visual ield series shows sion is the distinction between a series of visual ields overall worsening. There is a smaller than 5% (and larg- that is truly changing and one that is stable but shows er than 1%) chance that a stable visual ield series would luctuation. This challenge is greater in cases in which look like the series in question, which means there is a the magnitude of the change is small and the amount of high likelihood that the visual ield series is worsening. luctuation is large, which is a common situation when assessing glaucomatous progression. Worsening at P < 1%: this visual ield series shows overall worsening. There is a smaller than 1% chance In clinical practice, the trend line alone is not suf icient that a stable visual ield series would look like the series to distinguish between stable and changing visual ields. in question, which means there is very high likelihood This is because most visual ield series will show at least that the visual ield series is worsening. a small trend upwards or downwards. The challenge is to determine whether this trend is signi icantly different Improvement at P < 5%: this visual ield series shows from a lat line (i.e., one with a slope of zero). overall improvement. There is a smaller than 5% (and larger than 1%) chance that a stable visual ield series To distinguish between a stable and a truly changing se- would look like the series in question, which means there ries of visual ields, a t-test is used. The t-test is a statis- is a high chance that the visual ield series is improving. tical test of hypothesis that allows the determination of whether two sets of data are signi icantly different from Improvement at P < 1%: this visual ield series shows each other. For trend analysis, the t-test is applied to the overall improvement. There is a smaller than 1% chance observed slope to determine whether it is signi icantly that a stable visual ield series would look like the series different from a slope of zero (e.g., lat line showing no in question, which means there is a very high chance that change over time, which represents the typical situation the visual ield series is improving. of a stable visual ield series). The concept of probability is then used to determine the probability (P) that a stable Floor effect: There is more than 20 dB sensitivity loss visual ield series with an assumed slope of zero would in the visual ield series and no signi icant change, which show a given slope (see BOX 9A). Its interpretation is means that the determination of progression or stability similar to the P values used in the Probabilities plot (see is not possible due to the advancement of the disease. FIG 7-9 and 7-10). If there is a low probability that a stable visual ield series would look like the series in question, If there is no symbol, then there is a probability of P > 5% then that series is unlikely to be stable and consequently that a stable visual ield series would look like the series it is likely that the visual ield series is changing. in question or in other words that the data do not show change at the levels mentioned above. This either means To facilitate interpretation, the EyeSuite Progression that the visual ield is stable, or that the data available Analysis uses red downward arrows to show signi icant are not suf icient to capture change. This is often the case worsening and green upward arrows to show signi icant when only a few visual ield tests are available and pro- improvement at two probability levels and also marks gression is slow or when luctuation is large as explained loor effects using the following symbols: in BOX 9A. 172 Chapter 9 | Interpretation of visual field progression BOX 9A GENERAL CHARACTERISTICS OF TREND ANALYSIS LINEAR REGRESSION ANALYSIS WITH ORDINARY LEAST SQUARES ESTIMATES TO DETERMINE THE TREND LINE To determine the trend line, EyeSuite Progression Analysis uses linear regression analysis with least squares estimates. Linear regression analysis is a statistical approach for modeling the relationship between two variables using a straight line. An excellent it for the trend line is obtained using the least squares method, which is a commonly used approach to it the regression line. The best it of the trend line is achieved by minimizing the sums of squared residuals (i.e., the vertical distance between each data point and the itted regression line). OCTOPUS PERIMETERS USE LEAST SQUARES LINEAR REGRESSION TO BEST FIT A TREND LINE 0 5 10 MD [dB] 15 Residuals – Distance between MD data and trend line Best fitted trend line 20 25 2002 2003 2004 2005 Test date [years] The least squares linear regression approach determines a best ϔitted trend line by minimizing the vertical distance between the individual test points and the trend line. This vertical distance is called the residual and is depicted by the red lines in this example. If the ϔit were perfect, all individual test points (black dots) would fall exactly on the trend line (gray line). SIGNIFICANCE IS INFLUENCED BY THE AMOUNT OF FLUCTUATION The trend line describes the data and allows for the determination of the slope. However, this is not suf icient to distinguish between a stable and changing visual ield series because there is typically at least some positive or negative slope (even if it is very small) due to the luctuation of the variable (e.g., MD) over time. Therefore, it is necessary to determine whether the observed slope corresponds to a true change, or whether it may be explained by luctuations in the data. The t-test is used to determine whether the observed slope is signi icantly different from zero (as would be expected if the series of visual ields was stable) using two levels of signi icance (P < 5% and P < 1%). The amount of luctuation is taken into account by the t-test. This is necessary because the same slope may indicate a signi icant trend when the luctuations are small, but may not be signi icant when luctu- ations are large. In other words, a larger slope is needed to detect true change for the same number of tests and the same follow-up length when large luctuations are present. Assessment of global visual field change 173 ILLUSTRATION OF THE INFLUENCE OF FLUCTUATION ON SIGNIFICANCE TREND ANALYSIS DESCRIPTION INTERPRETATION MD Mean defect No symbol indicating change NO PROGRESSION 0 Slope Slope 0.9 dB/year Large slope 0.9 dB/year Outliers in 3rd & 4th test Considerable fluctuation 15 Stable series Outliers Considerable fluctuation 25 may prevent a change to 2010 2011 2012 be declared significant MD Mean defect Significant worsening at P < 5% WORSENING 0 Slope 0.6 dB/year Smaller slope Test data close to trend line Consistent results Slope 15 0.6 dB/year Progressing series 25 2010 2011 2012 In this ϔigure, the visual ϔield series from two different patients are shown over a comparable time period with approximately the same amount of test data. In the example on top, ϔluctuation is large because the 3rd and 4th test are outliers. As a result, even the relatively large slope (0.9 dB/year) is insufϔicient to indicate signiϔicant change and the series appears to be stable (no symbol). More visual ϔield tests may be needed to identify whether the series is truly stable or progressing. However, when there is less ϔluctuation in the visual ϔield data (bottom), even a small slope (0.6 dB/year) sufϔices to detect signiϔicant change (red downward arrow) and the series is conϔirmed as progressing. SIGNIFICANCE IS INFLUENCED BY THE NUMBER OF VISUAL FIELD TESTS The number of visual ield examinations (n) included in a trend analysis is important because it in lu- ences the outcome of the t-test. The EyeSuite Progression Analysis can be performed with a minimum of three visual ield tests. However, if there are only three or four visual ield tests included in the analysis, the slope must be quite steep to be able to separate true change from luctuations. On the other hand, if there are many visual ield tests included, even a visual ield series with a shallow slope can identify signi icant change. For typical progressing visual ields, trends will not become signi icant before ive or six examinations are included in the analysis. Guidelines on glaucoma treatment⁶ typical- ly recommend a minimum of 6 visual ields in the irst two years to reliably detect glaucomatous visual ield progression. However, if luctuation is large and the slope is small, an even larger number of visual , ield tests are required to detect progression.⁷ ⁸ 174 Chapter 9 | Interpretation of visual field progression THE INFLUENCE OF THE NUMBER OF VISUAL FIELD TESTS ON SIGNIFICANCE TREND ANALYSIS DESCRIPTION INTERPRETATION MD Mean defect No symbol indicating change NO PROGRESSION 0 Slope 0.9 dB/year Large slope 6 Tests Outliers in 3rd & 4th test Considerable fluctuation 15 Outliers 6 Visual field tests Series declared stable because of fluctuation Slope 0.9 dB/year 25 More test data is needed 2010 2011 2012 2013 2014 MD Mean defect Significant worsening at P < 1% WORSENING 0 Slope 0.7 dB/year 10 Tests Test data close to trend line Overall consistent results 15 10 Visual field tests Progressing series Sufficient number of tests Slope 0.7 dB/year 25 2010 2011 2012 2013 2014 SERIES OF VISUAL FIELDS Defect location & depth? Tests 1-6 All tests The number of visual ϔield tests signiϔicantly inϔluences whether a visual ϔield series is considered stable or not. More visual ϔields are required when the slope is shallow or when ϔluctuation is large. In this example, due to ϔluctuation in tests 3 and 4, the visual ϔield series doesn’t show signiϔicant change after the initial 6 tests (top) even though the slope of 0.9 dB/year is relatively large. Change is only detected by the trend analysis upon inclusion of more tests (bottom). MD TREND ANALYSIS Interpretation of MD Trend Analysis in clinical practice is improving can be made solely by looking at the red down- a fast and straightforward process (if adequate visual ields ward (signi icant worsening) or the green upward (sig- are selected, which is described in more detail later in ni icant improvement) arrows displayed. To assess rate this chapter). The decision about whether a visual ield of change, the slope is numerically displayed as change in series is stable, signi icantly worsening or signi icantly dB/year at the bottom of the graph (FIG 9-6). Assessment of global visual field change 175 MD TREND ANALYSIS - INTERPRETATION AID MD TREND DEFINITION INTERPRETATION Probability of stable visual field series Y-axis showing this result MD [dB] Magnitude of visual field loss No symbol P > 5% Stable Worsening at P < 5% Likely worsening 0 Worsening at P < 1% Highly likely worsening Improvement at P < 5% Likely improving 15 Improvement at P < 1% Highly likely improving Floor effect (MD > 20 dB, Maximum loss Slope: 1.9 dB/year no significant progression) (No further measurement of progression possible) 25 2000 2001 2002 2003 2004 2005 MD trend line X-axis Time (years) 1.9 dB/year Slope [dB/year] MD change in dB/year MD of visual field test (taken at specific date) Normal range of MD 15 MD of 15 dB Seriously impaired visual field FIGURE 9-6 MD Trend Analysis allows for a quick identification of worsening (red downward arrow) or improvement (green upward arrow) of a visual field series. In addition, it displays the rate of change (slope in dB/year) and shows the trend graphic including slope and individual test points to graphically assess severity of visual field loss, rate of progression, test interval, amount of fluctuation and number of tests included in the analysis. While the detailed graphical presentation of the trend line Visual ields included in the analysis are marked in a and the test data is not necessary for deciding about the different color which supports the visual ield selection presence and rate of MD change, it provides valuable infor- process. Lastly, different symbols are used for each pe- mation. It allows for a quick assessment of disease severity rimeter model to draw attention to a possible perimeter as well as rate of disease progression. The lower the level model-related bias. This can for example occur when a of the curve, the more the disease has progressed and patient is tested for the irst time on a new perimeter the steeper the curve, the more rapid the change. model and shows a strong learning effect. For more information on transitioning from one perimeter model The graph also allows for a quick determination of the to another, please refer to Chapter 12. frequency of the visual ield tests performed. In addition, it allows one to see at a glance if there is a signi icant out- Further orientation is provided by a gray band at the top lier, which calls for more careful evaluation to make sure which indicates the normal range of MD (i.e., the 95% that this visual ield is reliable and whether it should be con idence interval) and a red line at 15 dB which rep- included in the analysis. For more information, consult resents seriously impaired visual ields. The graph stops the next section in this chapter on adequate selection of at 25 dB because in many countries, an MD of 20 to 25 dB visual ield tests. is considered legal blindness. 176 Chapter 9 | Interpretation of visual field progression INTERPRETATION OF MD TREND ANALYSIS MD Trend Analysis provides information about the pres- at the end of their respective lifespan, the 80-year-old pa- ence and rate of progression as well as the magnitude of tient would have an MD of 7 dB whereas the 50-year-old the sensitivity loss (i.e., magnitude of MD) of a patient. patient would have an MD of 19 dB. However, if this same However, this data is not suf icient to make a clinical 80-year-old patient showed a progression rate of 2 dB decision as these factors have a very different meaning per year, at age 90 this patient would have an MD of 23 depending on their relation to each other as well as the dB, which represents near total visual ield loss. patient’s age and life expectancy. It therefore goes without saying that these factors as well For example, an MD of 3 dB in a patient progressing at as a patient’s lifestyle, adherence to and persistence with a rate of 0.4 dB/year has a very different meaning in a medications, other clinical issues and the practitioner’s 50-year-old patient compared to an 80-year-old patient. overall clinical assessment have to be taken into account Assuming a life expectancy of 90 years for both patients to make a clinical decision. and projecting the current slope linearly into the future, SELECTION OF ADEQUATE VISUAL FIELDS FOR ANALYSIS IMPORTANCE OF SELECTING ADEQUATE VISUAL FIELD TESTS FOR ANALYSIS A trend analysis is only clinically meaningful if adequate included in the progression analysis should be reliable, visual ields are selected for analysis. To facilitate the be part of a relevant time period, and be tested using selection process, the EyeSuite Progression Analysis the same test parameters. Each of these requirements is allows examiners to choose the visual ields to be in- described in this section. cluded in the analysis with a simple click. Visual ields EXCLUSION OF UNTRUSTWORTHY VISUAL FIELD TESTS It is important that only trustworthy visual ields, re- luctuation in a visual ield series and may change the liable and free of artifacts, be included in the analysis. outcome of visual field trend analysis as illustrated Untrustworthy visual fields increase the amount of in FIG 9-7. Assessment of global visual field change 177 IMPORTANCE OF EXCLUSION OF UNTRUSTWORTHY VISUAL FIELD TESTS TREND ANALYSIS DESCRIPTION INTERPRETATION MD Mean defect High false Filled green upward arrow IMPROVING positives (Improvement at P < 1%) 0 Slope – 0.6 dB/year Unreliable visual fields included Considerable fluctuation 15 Artifact Slope -0.6 dB/year 25 2010 2011 2012 2013 2014 MD Mean defect High false No symbol STABLE positives 0 Slope 0.0 dB/year 1st test with ptosis artifact excluded Considerable fluctuation 15 Artifact Slope 0.0 dB/year 25 2010 2011 2012 2013 2014 MD Mean defect High false Filled red downward arrow WORSENING positives (Worsening at P < 1%) 0 Slope 0.9 dB/year 1st test with ptosis artifact and Consistent results 5th & 6th test with high false positive 15 rates excluded Artifact Slope 0.9 dB/year 25 2010 2011 2012 2013 2014 SERIES OF VISUAL FIELDS Defect location & depth? False False Ptosis positives: positives: 71% 57% FIGURE 9-7 Visual field tests that are not trustworthy can significantly alter the trend analysis result as the example above illustrates. In this example, the first test is not trustworthy due to a ptosis lid artifact and tests five and six are unreliable due to high false positive rates. If all seven visual field tests are included in the analysis, the series seems to be improving (top), if the lid artifact is excluded (middle), the series appears to be stable and if all three untrustworthy visual fields are excluded from analysis, a significant visual field worsening becomes apparent (bottom). 178 Chapter 9 | Interpretation of visual field progression ADEQUATE TIME PERIOD FOR ANALYSIS When choosing a time period for visual ield progression Another example is the situation in which a switch to analysis, it is important to keep in mind that changes in more aggressive glaucoma treatment is made. This treatment as well as surgical interventions can signi i- switch can change the rate of progression. In that situ- cantly change both visual ield severity and progression ation, it would be much harder to detect the change in rates. For example, a patient with both cataract and rate if pre-treatment data are included. However, it should glaucoma typically shows a significant improvement be noted that the impact of the switch in treatment on rate of the MD after cataract surgery. This improvement of progression may only be assessed once a suf icient makes it challenging to assess glaucomatous progres- number of visual ield tests become available after the sion rates after surgery, if pre-surgery visual ield data switch. Thus the new rate cannot be assessed immedi- are included in the progression analysis. In those cases ately following the change in treatment. only post-surgery data should be analyzed. COMPARABLE TEST PARAMETERS All visual ields included in a given progression analysis only one type of test strategy is used, the EyeSuite Pro- must have the same test parameters in order to obtain gression Analysis allows inclusion of visual ield results meaningful information about visual ield progression. obtained using different quantitative testing strategies. Therefore, the EyeSuite Progression Analysis offers the The rationale for this is that even though the levels of trend calculations only on visual ields tests that have accuracy between the TOP and the other strategies been done with the same test pattern and stimulus and slightly differ, these effects are minimized at the level of background characteristics. However, although ideally the global indices.⁹,¹⁰ DISTINCTION BETWEEN LOCAL AND DIFFUSE CHANGE IMPORTANCE OF DISTINCTION BETWEEN LOCAL AND DIFFUSE CHANGE When both local and diffuse defects are present, it is by both local and diffuse change, it is impossible to deter- not only desirable to know whether there is change but mine the nature of the change by looking at MD alone. also whether the detected change is local or diffuse. This is important because local and diffuse change can For example, a patient may have both a local defect due be caused by different clinical situations that may call to glaucoma and a diffuse defect due to a cataract. If for different types of intervention (see TABLE 8-1 on the the MD is worsening in this patient, it is essential for a etiology of local and diffuse loss). Because MD is affected clinician to know whether the cataract, the glaucoma or Distinction between local and diffuse change 179 both are worsening. Examples of the presence of both tions in which MD is not suf iciently sensitive to detect local and diffuse change are presented in FIG 9-9. subtle local changes. This can for example be the case if there is subtle local glaucomatous change, but the visual In addition, the distinction between local and diffuse ield series also shows increased diffuse luctuation. change is not only helpful in the presence of both a local An example of this is given in FIG 9-10. and diffuse pathology, it is also very useful in all situa- USE OF DIFFUSE DEFECT (DD) INDEX TO IDENTIFY DIFFUSE CHANGE To determine whether there is diffuse visual field of diffuse change. No symbol is displayed if there is no change independent from the presence or absence of diffuse change, signi icant diffuse worsening is indicated local change, Octopus perimeters use the global index by red downward arrows and significant diffuse im- DD. This index represents the magnitude of the diffuse provement is shown by green upward arrows, similarly defect and is calculated from the Defect Curve. For more to that described for the MD slope. information on its design and de inition see BOX 7C. Four typical situations (stable, local progression, diffuse The DD Trend Analysis uses comparable de initions as progression, local and diffuse progression) and the the MD Trend Analysis but displays DD values on the respective behavior of the DD Trend Analysis are shown y-axis instead of MD values and thus allows assessment in FIG 9-8. TYPICAL BEHAVIOR OF GLOBAL TREND ANALYSES FROM EARLY TO MODERATE DISEASE MD sLV DD LD Stable Diffuse progression Local progression Diffuse & local progression FIGURE 9-8 This figure illustrates the typical behavior of the four Global Trend Analyses in potentially worsening visual field series from early to moderate disease. A quick visual inspection of the four global indices provides a straightforward assess- ment of whether a visual field series is worsening (MD worsening) and of whether the change is caused by diffuse worsening (MD and DD worsening), local worsening (MD, LD, and sLV worsening) or both diffuse and local worsening (MD, DD, LD and sLV worsening). Note that in more advanced disease (e.g., MD > 20 dB), with most visual field locations showing some degree of sensitivity loss, MD and also DD shows worsening while LD and sLV show improvement. 180 Chapter 9 | Interpretation of visual field progression USE OF LOCAL DEFECT (LD) INDEX TO IDENTIFY LOCAL CHANGE To determine whether there is local visual ield change of localized change. No symbol is displayed if there is independent from the presence or absence of diffuse no local change, red downward arrows indicate signi i- change, Octopus perimeters use the global index LD. cant local worsening and signi icant local improvement This index represents the magnitude of the local defect is shown by green upward arrows, similarly to that de- and is calculated from the Defect Curve. For more infor- scribed for the MD slope. mation on its design and de inition see BOX 7D. The typical behavior of the LD Trend Analysis in pro- The LD Trend Analysis uses comparable de initions as gressing from early to moderate disease (e.g., worsening the MD Trend Analysis but displays LD values on the glaucoma) is shown in FIG 9-8. y-axis instead of MD values and thus allows assessment USE OF SQUARE ROOT OF LOSS VARIANCE (sLV) TO IDENTIFY LOCAL CHANGE While the combined evaluation of the DD and LD Trend challenging to understand the visual ield change in Analysis is suf icient to distinguish between local and case of simultaneous local and diffuse change. For more diffuse change, some users are more familiar with the information on the design and de inition of sLV see FIG square root of Loss Variance(sLV) index. Octopus perim- 7-20 and TABLE 7-1. For more information on its clinical eters therefore also provide a trend graphic of the index interpretation, see FIG 8-27. sLV as an alternative to using the DD and LD Trend Anal- ysis. This allows clinicians to choose the analysis they The sLV Trend Analysis uses comparable de initions as prefer to assess progression. the MD Trend Analysis but displays sLV values on the y-axis instead of MD values and thus allows distinction The sLV global index provides a measure for the inho- between homogenous and inhomogenous change. No mogeneity of the visual ield. If a visual ield is normal, symbol is displayed if there is no change; increasing shows a diffuse defect or shows severe pathology (e.g., inhomogeneity is indicated by red downward arrows MD > 20 dB), it is very homogenous and sLV is low. On and increasing homogeneity is shown by green upward the other hand, if a visual ield shows one or more local arrows, similarly to that described for the MD Trend defects, it is more inhomogenous and sLV is larger. sLV Analysis. therefore increases if a local defect is increasing, and it remains stable if a diffuse defect is increasing. While The typical behavior of the sLV Trend Analysis in pro- this provides comparable information in a situation in gressing from early to moderate disease (e.g., worsening which there is only local or only diffuse change, it becomes glaucoma) is shown in FIG 9-8. Distinction between local and diffuse change 181 CLINICAL INTERPRETATION OF GLOBAL TREND ANALYSES In clinical practice it is helpful to jointly consider the of a patient’s series of visual ields. Further analysis of information from the four indices presented in the the individual graphs can then be performed given the Global Trend Analysis. It is useful to assess the symbols clinical situation. Two clinical examples are shown in marking signi icant change irst to get a quick overview FIG 9-9 and 9-10. CASE EXAMPLE 1: PATIENT WITH BOTH PROGRESSING CATARACT AND GLAUCOMA GLOBAL TREND ANALYSIS DESCRIPTION INTERPRETATION MD sLV MD Worsening Fast (1.5 dB/year) worsening Stable or overall Increasing/decreasing progression? inhomogeneity? of visual field series 0 DD Diffuse 0 worsening Both local and diffuse worsening sLV Stable Progression of cataract and 15 LD Local glaucoma worsening 25 15 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 Slope: 1.5 dB / Yr (p < 0.5%) Slope: 0.2 dB / Yr DD LD Diffuse progression? Local progression? 0 0 25 15 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 Slope 1.2 dB / Yr (p < 0.5%) Slope: 0.6 dB / Yr (p < 0.5%) SERIES OF VISUAL FIELDS Defect location & depth? FIGURE 9-9 This figure illustrates the usefulness of looking at the four global indices in combination. In this example, a patient has both confirmed glaucoma and cataract. While the visual field shows overall significant worsening (MD worsening at p < 0.5% ), the MD Trend Analysis does not show which disease is progressing. An analysis of the Diffuse (DD) and Local (LD) Trend Analyses shows both significant local and diffuse progression, suggesting that both glaucoma and the cataract are progressing. 182 Chapter 9 | Interpretation of visual field progression CASE EXAMPLE 2: SUBTLE LOCAL GLAUCOMATOUS CHANGE AND MAINLY DIFFUSE FLUCTUATION GLOBAL TREND ANALYSIS DESCRIPTION INTERPRETATION MD sLV Clinically confirmed Slow (0.3 dB/year) local Stable or overall Increasing/decreasing pathology: glaucoma progression? inhomogeneity? worsening of visual field series 0 0 MD Stable Some fluctuation DD Stable Slow local glaucoma progression 15 sLV Increasing inhomogeneity 25 15 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 LD Local Slope: 0.4 dB / Yr Slope: 0.5 dB / Yr (p < 0.5%) worsening DD LD Diffuse progression? Local progression? 0 0 15 25 15 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 Slope: 0.1 dB / Yr Slope: 0.3 dB / Yr (p < 2%) SERIES OF VISUAL FIELDS Defect location & depth? FIGURE 9-10 This glaucoma patient shows a marked nasal step and some diffuse visual field loss in visual fields 3 and 4 in the series of Grayscale (Comparison) representation. Looking solely at MD change, the visual field series appears to be stable (no symbol). Nevertheless, the MD Trend Analysis also shows an outlier on the 3rd test, which is also present in the DD Trend Analysis suggesting this is caused by diffuse fluctuation. Assessment of the DD Trend Analysis (no change), sLV Trend Analysis (significant worsening at P < 1%) and LD Trend Analysis (significant worsening at P < 5%) reveals no diffuse change but significant local change. In conclusion, in this situation MD is too affected by diffuse fluctuation to show the significant but local worsening of the nasal step defect. Thus, the additional assessment of local and diffuse change in this situation is more sensitive in detecting subtle local change than the assessment of only the MD Trend Analysis. Cluster Trend and Corrected Cluster Trend Analysis 183 CLUSTER TREND AND CORRECTED CLUSTER TREND ANALYSIS IMPORTANCE OF ASSESSING CLUSTER PROGRESSION IN GLAUCOMA Typical glaucomatous defects caused by localized retinal For example, to determine whether there is a corre- nerve iber damage, as well as some visual ield defects sponding structural change in glaucoma to con irm a caused by optic nerve damage, consist of a cluster of suspected glaucomatous change, it is helpful to know in adjacent defective visual ield locations (FIG 5-1) that which area of the visual ield the change is happening. correspond to the path followed by the retinal nerve iber In addition, in a constricted glaucomatous visual ield with bundles in the retina (see step 5 in Chapter 8). Localized many visual ield locations showing absolute defects visual ield progression therefore typically occurs in a (i.e., sensitivity thresholds of 0 dB), progression in the re- cluster of visual ield locations. maining central visual ield of a patient is of key impor- tance for quality of life but may not be apparent from However, if localized glaucomatous progression is small the MD Trend Analysis, due to its relative insensitivity and there is additional luctuation, the global index MD in detecting localized change. may not be sensitive enough to detect that subtle clus- ter change because MD is an average of the sensitivity It is thus helpful to assess Cluster MD progression in loss of the whole visual ield. While in some instances addition to the global indices to detect subtle localized looking at local change using the LD or sLV indices can visual field change in glaucoma as well as to receive lead to the detection of such change, spatial information additional spatial information about where the change about where the change occurs is missing. is happening. This is the purpose of both the Cluster and Corrected Cluster Trend Analysis. CLUSTER AND CORRECTED CLUSTER TREND ANALYSIS Cluster Trend Analysis (CTA) is a trend analysis based Both types of Cluster Trend Analysis employ the same on the single ield Cluster Analysis whose design and statistical analysis also used in the global MD Trend de initions have already been explained in FIG 7-12 and Analysis and use comparable symbols to indicate sig- 7-13 and BOX 7B and whose clinical interpretation and ni icance of change. However, instead of looking for usefulness have already been shown in FIG 8-20 and signi icant MD change over time, they are looking for 8-21 and BOX 8B. The Corrected Cluster Trend Analysis signi icant Cluster or Corrected Cluster Mean Defect (CCTA) is very similar to CTA, but is based on the Correct- (MD) change over time. ed Cluster Analysis (see FIG 7-19) which eliminates the in luence of diffuse defect. 184 Chapter 9 | Interpretation of visual field progression CLUSTER TREND AND CORRECTED CLUSTER TREND – INTERPRETATION AID CLUSTER TREND DEFINITION INTERPRETATION Probability of stable visual field series showing this result 4.1 2.1 No symbol P > 5% Stable -0.8 6.7 2.6 1.0 0.4 0.4 Worsening at P < 5% Likely worsening cluster 0.7 Worsening at P < 1% Highly likely worsening cluster 0.6 Improvement at P < 5% Likely improving cluster Improvement at P < 1% Highly likely improving cluster CORRECTED CLUSTER TREND Floor effect Maximum cluster loss (Cluster MD > 20 dB, (No further measurement of progression possible no significant progression) in this cluster) 3.5 1.5 6.7 Cluster MD Slope Cluster MD change in dB/year -1.4 6.0 [dB/year] 2.0 -0.2 -0.3 -0.2 6.0 Corrected Cluster Corrected Cluster MD change MD Slope [dB/year] in dB/year 0.0 -0.1 FIGURE 9-11 The Cluster Trend representations display 10 visual field clusters that spatially correlate with retinal nerve fiber bundles. In each cluster, a Cluster MD change in dB/year is indicated. Significant Cluster MD worsening is marked with a red downward arrow, whereas significant Cluster MD improvement is marked with green upward arrows. Stable clusters do not have a symbol and clusters which show a floor effect are marked with a black symbol. CTA and CCTA also use the red downward arrows and more clinically meaningful if it is spatially correlated with green upward arrows to show signi icant cluster wors- another meaningful cluster defect or if it correlates with ening or improvement. However, the graphical display a signi icant structural change. is different from the MD Trend Analysis. The individual Cluster MDs are not shown in a two-dimensional trend Similar to Cluster Analysis (see BOX 8B), CTA has been graph. Instead, both the Cluster MD change in dB/year shown to be highly sensitive in detecting subtle, early and a symbol indicating the signi icance of this change glaucomatous change and has been shown to be more are displayed in each of the 10 clusters as shown in sensitive in detecting change than MD Trend Analysis FIG 9-11. and local event analysis¹¹ (not available as a statisti- cal tool in the EyeSuite Progression Analysis). Similar to the interpretation of Cluster Analysis, some caution is essential in the clinical interpretation of CTA These indings can be explained with the same rationale and CCTA. This is because one random cluster showing a used to explain why Cluster Analysis is highly sensitive in P value smaller than 5% is expected to occur even in sta- detecting early glaucomatous defects. Because glauco- ble visual ields. Thus, a signi icant cluster defect is much matous change is mostly local, the averaging used to de- Cluster Trend and Corrected Cluster Trend Analysis 185 rive the MD global index reduces the chances of detecting hand, single point event analysis is too in luenced by early localized change (FIG 9-12 and 9-13). On the other luctuation to detect signi icant change early. ILLUSTRATION OF THE USEFULNESS OF CLUSTER TREND ANALYSIS EYESUITE PROGRESSION ANALYSIS GLOBAL TREND ANALYSIS CLUSTER TREND ANALYSIS MD sLV CLUSTER TREND Stable or overall Increasing/decreasing Cluster MD progression? progression? inhomogeneity? 0 0 Significant superior paracentral, superior and infero-temporal progression 15 1.6 25 15 0.5 2011 2012 2013 2014 2011 2012 2013 2014 Slope: 0.5 dB / Yr Slope: 0.3 dB / Yr 0.6 0.0 DD LD 1.4 Diffuse progression? Local progression? 0 -0.6 0 0.5 1.9 -0.1 15 -0.7 25 15 2011 2012 2013 2014 2011 2012 2013 2014 Slope: 0.0 dB / Yr Slope: 0.6 dB / Yr (p < 0.5%) SERIES OF VISUAL FIELDS Defect location & depth? FIGURE 9-12 The usefulness of Cluster Trend Analysis (CTA) in a case which shows a considerable amount of fluctuation is visible in the data. This visual field series of a glaucoma patient appears to be stable (no symbol indicating change) on the global index MD, but shows local worsening on the LD index. Using CTA, significant worsening (red downward arrow) is apparent in the superior paracentral, superior and infero-temporal clusters indicating clear local worsening. In this situation, CTA is more sensitive in detecting progression than MD and provides additional information about the location of progression compared to the LD index. 186 Chapter 9 | Interpretation of visual field progression ILLUSTRATION OF THE USEFULNESS OF CLUSTER TREND ANALYSIS IN ADVANCED DISEASE EYESUITE PROGRESSION ANALYSIS GLOBAL TREND ANALYSIS CLUSTER TREND ANALYSIS MD sLV CLUSTER TREND Stable or overall Increasing/decreasing Cluster MD progression? progression? inhomogeneity? 0 0 Floor effect 15 -0.1 25 15 0.1 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 Slope: 0.2 dB / Yr Slope: 0.0 dB / Yr 0.9 0.7 DD LD -0.6 Diffuse progression? Local progression? 0 -0.3 0 0.1 -0.6 1.6 15 0.2 25 15 Significant inferior 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 progression Slope: 0.0 dB / Yr Slope: 0.2 dB / Yr SERIES OF VISUAL FIELDS Defect location & depth? FIGURE 9-13 This example presents the visual fields of a glaucoma patient with a severe superior altitudinal defect and no remaining sensitivity in most of the upper visual field (floor effect, no further progression can be detected). All four global indices are stable with no symbol indicating change. However, using the Cluster Trend Analysis, significant localized worsening (red downward arrow) is apparent in the inferior cluster. In such advanced situations, Cluster Trend Analysis can assist in the detection of progression in areas with remaining sensitivity, which is important for the management of the patient. Polar Trend Analysis 187 POLAR TREND ANALYSIS IMPORTANCE OF ESTABLISHING A RELATIONSHIP BETWEEN STRUCTURAL AND FUNCTIONAL PROGRESSION In eyes with early glaucomatous damage or only subtle Because visual ield damage is often detected in a repre- progression, detection of pathological changes is chal- sentation of a retinal location while structural damage lenging. Therefore, it is often useful to consider both is evident at the optic disc, there is a need to use a repre- functional and structural change (i.e., neuroretinal sentation that links the structural to the functional visual rim tissue loss; decrease of retinal nerve fiber layer ield progression. This is the purpose of Polar Trend thickness. Analysis. USE OF POLAR TREND ANALYSIS TO ASSIST IN THE DETECTION OF GLAUCOMATOUS STRUCTURAL PROGRESSION Polar Trend Analysis is based on Polar Analysis, whose bundles arrive at the margin of the disc. It does so by design and de initions have already been shown in FIG employing the same trend analysis approach also used 7-14 and 7-15 and whose clinical interpretation and use- in the global MD Trend Analysis (see FIG 9-3 and 9-5 and fulness have been presented in FIG 8-23 and 8-24. BOX 9A), but applies it to sensitivity loss at each test lo- cation (pointwise trend analysis). For more information It graphically represents change at each visual ield test on the design of Polar Trend Analysis, refer to BOX 9B. location where the corresponding retinal nerve fiber THE DESIGN OF POLAR TREND ANALYSIS BOX 9B Polar Trend Analysis performs pointwise trend analysis on sensitivity loss data to determine the trend line but not the signi icance of the slope for each visual ield location individually. This is illustrated in the graphic in this box, which uses the example of one superior nasal test location circled in red in the Grayscale representation. However, the graphical display of Polar Trend Analysis is fundamentally different from the other representations discussed previously. Instead of using the slope to determine a rate of change, the trend line is used to determine a best itted sensitivity loss for the irst (blue point in the graphic in this box) and the last (yellow point) of the visual ield tests. It should be noted that these two data points are based on the trend line at the respective test dates, not on the individual visual ield test result at a given test date. These two itted sensitivity loss values are then marked in the same Polar grid also used for Polar Analysis and connected by a straight line at the position where the corresponding nerve iber bundles of the test location arrive at the margin of the disc. If there is worsening between that irst and last itted sensitivity loss, then the bar is drawn in red, while it is drawn in green if there is improvement. 188 Chapter 9 | Interpretation of visual field progression THE DESIGN OF POLAR TREND ANALYSIS SENSITIVITY LOSS POLAR TREND ANALYSIS AT ONE TEST LOCATION AT ONE TEST LOCATION 0 Best fitted first sensitivity loss one test location (dB) Sensitivity loss at S 30 20 10 [dB] N T I 15 Best fitted last sensitivity loss 25 2002 2003 2004 2005 Test date (years) SERIES OF VISUAL FIELDS Polar Trend Analysis performs point-wise trend analysis to determine the trend line at each visual ϔield location. The individual sensitivity losses from one test location over time, shown as red circles in the series of visual ϔields (bottom) are used to determine the trend line of the sensitivity losses at that test location (top left). The trend line, and not on the actual test data (gray squares), is used to determine the initial (blue) and last sensitivity loss (yellow). These sensitivity losses are used as the start and end location of the progression bar in the Polar Trend Analysis (top right). Overall worsening is illustrated with a red bar (shown in this example), and overall improvement is illustrated with a green bar (not shown in this example). The length of the bar indicates the magnitude of change. Progression (worsening) is represented by a red bar, the change is not given numerically, the approximate change length of which corresponds to the best- itted change of each defect can be identi ied on the graph in dB. A in the sensitivity loss in dB. Improvement is similarly gray band in the center indicates approximate normal represented using a green bar. Though the quantity of ranges for those bars (FIG 9-14). Polar Trend Analysis 189 POLAR TREND ANALYSIS – INTERPRETATION AID POLAR TREND DEFINITION INTERPRETATION Indication of change at each test location projected onto optic disc Sensitivity loss [dB] Worsening location • Length of bar indicates total • Short bar – small worsening S change from first to last • Long bar – large worsening 30 20 10 [dB] N T selected test I Sensitivity gain [dB] Improving location • Length of bar indicates total • Short bar – small improvement change from first to last • Long bar – large improvement selected test Normal range FIGURE 9-14 Polar Trend Analysis representation projects local progression per test location onto the optic disc to allow for easy linking with structural results. Red bars indicate worsening while green bars indicate improvement. The starting and end location point (i.e. the length) of each bar is based on the loss indicated by the local trend line between the first and the last examination. Clinical interpretation of Polar Trend Analysis is straight- potential structural progression. However, it is import- forward and based solely on the graphical representation. ant to note, no rates of progression or signi icance of The longer the bar, the more absolute change has occurred progression are provided by Polar Trend Analysis. For during the time period of interest and the further away an exact evaluation of these parameters, one can refer the bar is located from the center, the more damage was to Cluster Trend and Corrected Cluster Trend Analyses. already present at a given test location at the time of It is important to remember that those representations the irst test. are oriented as visual ields and not as structural data. This means that related defects will be positioned at the If there are many red bars indicating worsening clus- location lipped vertically across the horizontal midline. tered at one optic disc location, this indicates a visual ield worsening at that position. One can determine Polar Trend Analysis has been shown to correlate well whether a corresponding structural change at that same with structural progression data¹² and is therefore a very position is present. Defect progression on the Polar Trend useful and quick tool for assistance with the combined Analysis report can be considered as a warning message evaluation of both structural and functional progres- for localized visual ield progression, which may draw sion. A clinical case is illustrated in FIG 9-15. the clinicians’ attention to the spatially corresponding 190 Chapter 9 | Interpretation of visual field progression ILLUSTRATION OF THE USEFULNESS OF POLAR TREND ANALYSIS GLOBAL TREND ANALYSIS CORRECTED CLUSTER POLAR TREND TREND ANALYSIS ANALYSIS Fast superior & paracentral MD Mean defect sLV Loss variance progression 0 2.6 0 2.5 0.4 -0.5 1.7 S 30 20 10 N T 0.7 [dB] I -0.1 0.2 15 0.0 -0.0 MD change 25 1.0 dB/year 15 2013 2013 Slope: 1.0 dB / Yr (p < 0.5%) Slope: 0.4 dB / Yr (p < 0.5%) OCT DD Diffuse defect LD Local defect Significant further 0 RNFLT decrease 0 inferotemporally 2008 2013 and superotemporally 25 15 2013 2013 Slope: 0.2 dB / Yr Slope: 1.1 dB / Yr (p < 0.5%) SERIES OF VISUAL FIELDS 2008 2013 FIGURE 9-15 This glaucoma patient shows significant local visual field worsening (MD, LD and sLV worsening at P < 1%) over a period of 5 years starting from the superior paracentral and superior nasal step areas and expanding to the inferior paracentral area, while deepening at the original defect locations (significant corrected cluster worsening in these areas). Polar Trend Analysis displays strong supero- and infero-temporal worsening. Looking at the change on the OCT retinal nerve fiber layer thickness between 2008 and 2013 (supero-and infero-temporal structural progression), there is a clear spatial relationship between structural and functional change, thus confirming that these changes stem from glaucoma. References 191 REFERENCES 1. Tanna AP, Bandi JR, Budenz DL, et al. Interobserver agreement and intraobserver reproducibility of the subjective determination of glaucomatous visual ield progression. Ophthalmology. 2011;118:60-65. 2. Tanna AP, Budenz DL, Bandi J, et al. Glaucoma Progression Analysis software compared with expert consensus opinion in the detection of visual ield progression in glaucoma. Ophthalmology. 2012;119:468-473. 3. Viswanathan AC, Crabb DP, McNaught AI, et al. Interobserver agreement on visual ield progression in glaucoma: a comparison of methods. Br J Ophthalmol. 2003;87:726-730. 4. Lin AP, Katz LJ, Spaeth GL, et al. Agreement of visual ield interpretation among glaucoma specialists and comprehensive ophthalmologists: comparison of time and methods. Br J Ophthalmol. 2011;95:828-831. 5. Iester M, Capris E, De Feo F, et al. Agreement to detect glaucomatous visual ield progression by using three different methods: a multicentre study. Br J Ophthalmol. 2011;95:1276-1283. 6. European Glaucoma Society. Terminology and Guidelines for Glaucoma. 4th ed. Savona: PubliComm; 2014. 7. Taketani Y, Murata H, Fujino Y, Mayama C, Asaoka R. How Many Visual Fields Are Required to Precisely Predict Future Test Results in Glaucoma Patients When Using Different Trend Analyses? Invest Ophthalmol Vis Sci. 2015;56:4076-4082. 8. Chauhan BC, Garway-Heath DF, Goñi FJ, et al. Practical recommendations for measuring rates of visual ield change in glaucoma. Br J Ophthalmol. 2008;92:569-573. 9. Anderson AJ. Spatial resolution of the tendency-oriented perimetry algorithm. Invest Ophthalmol Vis Sci. 2003;44: 1962-1968. 10. Maeda H, Nakaura M, Negi A. New perimetric threshold test algorithm with dynamic strategy and tendency oriented perimetry (TOP) in glaucomatous eyes. Eye (Lond). 2000;14:747-751. 11. Naghizadeh F, Holló G. Detection of early glaucomatous progression with octopus cluster trend analysis. J Glaucoma. 2014;23:269-275. 12. Holló G, Naghizadeh F. Evaluation of Octopus Polar Trend Analysis for detection of glaucomatous progression. Eur J Ophthalmol. 2014;24:862-868. 192 193 CHAPTER 10 NON-CONVENTIONAL PERIMETRY INTRODUCTION Static Standard Automated Perimetry (SAP, alternative- variability in patient responses in areas of significant ly called white-on-white perimetry), which uses a white vision impairment or low vision and 2) there is a marked Goldmann size III stimulus presented on a white back- loor effect in areas of signi icant vision impairment or ground, is by far the most commonly used type of peri- low vision. metric test today. It is the standard of care to detect and follow glaucoma. The white stimulus stimulates nearly Other forms of perimetry have been developed to allow all types of retinal ganglion cells and as a result the test has for earlier detection and to overcome the shortcomings a large dynamic range. Nevertheless, it would be desirable of SAP. Non-conventional perimetry includes function- to have a more sensitive test than SAP for early detection speci ic perimetric tests that use stimuli which target of irreversible vision loss in diseases such as glaucoma. speci ic pathways and visual functions (e.g., licker) and also white-on-white perimetry performed with the larger Furthermore, the following shortcomings are associat- size V stimulus, which provides a useful alternative for ed with SAP using a size III stimulus: 1) there is large testing in areas of vision impairment or low vision. 194 Chapter 10 | Non-conventional perimetry FUNCTION-SPECIFIC PERIMETRY RATIONALE FOR USING FUNCTION-SPECIFIC PERIMETRY Different Octopus perimeter models offer different types While the stimuli used in SWAP, Flicker perimetry and of function-speci ic stimuli. Pulsar perimetry uses a lick- Pulsar perimetry differ substantially from each other, the ering stimulus with concentric rings changing in both same rationale was used to develop them. These tests are spatial resolution and contrast that resembles a bullseye. designed to overcome the redundancy of the visual sys- Flicker perimetry uses a white lickering stimulus present- tem by selectively stimulating a subset of retinal cells and ed on a white background. Short-Wavelength Automated as a result get a more sensitive response to early changes Perimetry (SWAP - alternatively called blue-on-yellow (FIG 10-1). This rationale is based on the hypothesis that perimetry) uses a blue (short wavelength) stimulus pre- different types of retinal ganglion cells process different sented on a yellow background. Similar to SAP, all these visual functions, but nearly all retinal ganglion cells can tests are based on functional decline due to retinal detect the white stimulus used in SAP. While some cells ganglion cell loss in glaucoma. are adversely affected by pathology such as glaucoma, ILLUSTRATION OF THE RATIONALE BEHIND FUNCTION-SPECIFIC PERIMETRY STIMULUS TYPE OF RETINAL NORMAL EARLY PATHOLOGY GANGLION CELLS SAP Parvocellular Koniocellular Magnocellular Pulsar Magnocellular FIGURE 10-1 Function-specific perimetry has been developed to reduce the redundancy within the visual system with the goal of detecting visual field loss earlier. The idea is based on the hypothesis that white light universally stimulates nearly all retinal ganglion cell types. The loss of a few retinal cells should therefore be easily compensated by the remaining cells, as the example with the SAP stimulus (top) illustrates. The white stimulus stimulates many retinal cells and even when several are dysfunctional, the white stimulus (white circle) is still seen. In function-specific perimetry, only one cell type is predominantly stimulated. In the example with the Pulsar stimulus (bottom), there is no remaining functional magnocellular cell that can be stimulated by the Pulsar stimulus. As a result, the stimulus is not seen. Function-specific perimetry 195 other neighboring cells may still detect the SAP stimulus. pathology such as glaucoma, there are a smaller number This presumably makes the SAP test less sensitive to early of cells that are able to detect the function-speci ic stim- visual ield loss. To give a simple analogy, it is as though ulus, making the test more sensitive to early visual ield one person out of the 20 who promised to help you move loss. Using the previous analogy, this would translate into calls in sick on moving day. The other 19 helpers can having one person out of only two cancel on moving day. effectively carry on the task and the impact of the one There is only one person to help with the move and the missing person is not felt too strongly. task becomes much more dif icult. In contrast, function-specific perimetry targets only The function-speci ic stimuli currently available have all a subset of retinal ganglion cells. It is assumed that been developed for early glaucoma detection, but have if a few cells in this subset are adversely affected by also been used for other diseases. USE OF FUNCTION-SPECIFIC PERIMETRY IN CLINICAL PRACTICE While many studies have reported that function-spe- While there are distinct normative databases for each ci ic perimetry detects glaucomatous vision loss earlier function-speci ic stimulus as well as for SAP, it is es- than SAP,¹,² other studies have found no such effect.³,⁴ sential to consider that function-speci ic perimetry has As a result, experts have not yet reached a consensus on a smaller dynamic range than SAP. Therefore, while whether function-specific perimetry provides added normal subjects may show comparable responses on value in comparison to SAP. all tests, patients with more advanced disease are likely to show visual ield defects that appear more severe on When making a decision about whether or not to use function-speci ic perimetry due to the smaller dynamic function-speci ic perimetry, it is essential to keep in mind range. that the quantitative results cannot be directly compared with white-on-white perimetry. While SAP is the recom- Consequently, function-speci ic perimetry cannot be mended standard, one may choose either SAP or one of used through all disease stages. If there is advanced the function-speci ic perimetry tests as a default test for disease, one should use SAP. If function-speci ic perime- disease detection. If time allows, one might choose to try is chosen as a default for disease detection, switching perform an additional test, particularly in situations of to SAP is recommended for follow-up at some point. In uncertainty (i.e., to con irm suspected but uncon irmed order to avoid a lack of historic reference data, it may be visual ield loss as shown in the example in FIG 10-2). best to switch to SAP early in the follow-up process. 196 Chapter 10 | Non-conventional perimetry EXAMPLE OF SAP AND FUNCTION-SPECIFIC PERIMETRY IN THE SAME EYE GRAYSCALE (CO) COMPARISON PROBABILITIES + + + + + + 7 + 8 + + + + + + + + + + + + + + + + SAP + + + + + + + + + + + + + + + + + + + + + + + 7 + + + + + + + + + + FUNCTION-SPECIFIC PERIMETRY 7 7 + + 7 9 6 7 7 11 + + 11 7 + + 5 7 8 + + + + + + + 5 8 11 + + + + + + + 5 10 + + + + + + 10 + + + 7 + + + + + + + + + + FIGURE 10-2 The same patient with an early glaucomatous defect is tested twice, once with the SAP test (top) and once with the function-specific Pulsar stimulus (bottom). While SAP does not show a statistically significant defect in this patient, there is a clear defect visible when using function-specific Pulsar perimetry. Note that the locations with p < 5% for SAP are within the area in which the defect is present for function-specific perimetry. PULSAR PERIMETRY The Pulsar stimulus is a function-speci ic stimulus that two images alternate at a frequency of 10 Hz over 500 ms. tests both licker sensitivity and contrast sensitivity. It If licker sensitivity is reduced, the visual system cannot has been developed speci ically for early glaucoma de- detect the change between the phase and counter-phase tection and has been shown to be both sensitive and images. As a result, the phase and counter-phase images speci ic in the detection of early glaucoma.¹,² It is a very are perceived as a single image. Because the average patient-friendly perimetric test. intensity of the rings of the phase and counter-phase images are equal to the mean intensity of the back- The stimulus used in Pulsar perimetry consists of a ring ground, the Pulsar stimulus blends with the background pattern with a diameter of 5° of visual angle, which is and is not visible anymore (FIG 10-3). However, if licker- more than 10 times larger in radius and 100 times larger sensitivity is not affected, the visual system distinguishes in area than the white size III stimulus used in SAP. The between the phase and counter-phase images and the Pulsar stimulus consists of phase and counter-phase im- Pulsar stimulus is perceived like a pulsating ring pattern, ages. This means that light rings on the phase image are similar to the ripple pattern generated if a water drop displayed as dark rings on the counter-phase image. The enters a smooth water surface.¹ Function-specific perimetry 197 DESIGN OF THE PULSAR STIMULUS PHASE COUNTER-PHASE HEALTHY DISEASED flicker-sensitive cells flicker-sensitive cells Seen Not seen FIGURE 10-3 The Pulsar stimulus consists of a flickering phase and counter-phase image. If the function of the flicker-sen- sitive cells is intact, the stimulus can be seen (bottom left). If it is decreased, then the phase and counter-phase images are perceived as one image that equals the background and is invisible (bottom right). The Pulsar test uses a very patient-friendly stimulus. It is In addition, sensitivity thresholds can also be deter- easy to instruct the patients on how to perform the test mined. Pulsar perimetry employs its own unit scale, the (seen or not seen) and patients have more con idence src scale, consisting of 36 distinct steps, with increased about seeing the stimulus both because of its large size spatial resolution (sr) and contrast (c) with each step and perceived motion. As a result, Pulsar perimetry (FIG 10-4). The results of this threshold test are then has low test-retest variability and a minimal learning displayed as any SAP result and all the visual ield repre- effect.⁵,⁶ These features make it very suitable for screening sentations presented in Chapters 7-9 are available. Pulsar purposes. perimetry uses all representations available for SAP. SENSITIVITY THRESHOLDS WITH PULSAR PERIMETRY LESS VISIBLE MORE VISIBLE CONTRAST RESOLUTION SPATIAL FIGURE 10-4 Pulsar perimetry allows the determination of sensitivity thresholds by showing stimuli of both increasing spatial resolution (sr) and contrast (c). Sensitivity thresholds are expressed in src. 198 Chapter 10 | Non-conventional perimetry FLICKER PERIMETRY Flicker perimetry is similar to Pulsar perimetry in that it stimulus that the perimeter can display) lickers over a stimulates licker sensitive cells and has been created for period of 1 second and the patient is instructed to press early glaucoma detection. However, the stimulus design the response button only when the stimulus seems to is fundamentally different from that of Pulsar perimetry. licker (FIG 10-5). The licker frequency ranges from Flicker perimetry determines the critical fusion fre- very fast (approximately 50 cycles per second) to slow quency (CFF), or in other words, the frequency at which (i.e., 1-5 cycles per second). The CFF represents the sen- the licker appears to fuse into continuous steady light. sitivity threshold of Flicker perimetry (FIG 10-6) and is In this test, a white stimulus of Goldmann size III with expressed in Hertz (Hz). a stimulus intensity of 4,000 asb (i.e., the most intense DESIGN OF THE FLICKER STIMULUS Do you see the Stimulus stimulus flicker? Time 1 Time 2 Fixation ON OFF Flickering 4,000 asb t Time (t) = 1 s Frequency = 4 stimuli/s = 4 Hz FIGURE 10-5 Flicker perimetry uses a flickering white stimulus (size III) of 4,000 asb on a white background that flickers at different temporal frequencies. The frequency is expressed in Hertz, a unit that defines how many times the stimulus is flickering per second. In the example above, the stimulus has a frequency of 4 Hz. Flicker perimetry was shown to be both sensitive and are minimally in luenced by media opacities stemming speci ic in the detection of early glaucoma.⁷-⁹ One of its from pathologies such as cataracts or refractive errors, major additional advantages is that sensitivity thresholds for example.⁹,¹⁰ Function-specific perimetry 199 Flicker perimetry is more demanding of patients com- mended only for patients who perform very well on pared to Pulsar perimetry, because they must pay atten- perimetry. In these patients, it is a useful perimetric test. tion to both the presence of a stimulus and whether it is lickering or not. Thus, careful patient instruction and BOX 10A provides practical guidance on how to best observation are even more essential in licker perimetry perform licker perimetry. than in other perimetry forms. Its use is therefore recom- SENSITIVITY THRESHOLDS WITH FLICKER PERIMETRY LESS VISIBLE MORE VISIBLE High frequency Low frequency FREQUENCY (HZ) t t t FIGURE 10-6 In flicker perimetry, stimuli flicker from high frequencies (50Hz, flicker is more difficult to see) to low frequencies (1-5 Hz, flicker is easier to see) to determine the Critical Fusion Frequency (i.e., the frequency in Hertz (Hz) at which a fl icker- ing stimulus appears to fuse into continuous steady light). The CFF defines the sensitivity threshold at a given location. HOW TO PERFORM RELIABLE FLICKER PERIMETRY BOX 10A Most points highlighted in Chapter 3 on how to run a reliable visual ield test also apply to licker perimetry. However, there are some speci ic points to which particular attention should be given. First, patient instructions need to be slightly adapted and should include a description of a lickering stimulus. An example referring to old television sets or to a candle in the wind might prove helpful. It might also be useful to describe that the test examines one’s ability to recognize when lights go on and off when they are switched rapidly. It also needs to be stressed that all stimuli are visible for a full second, but that the patient should only respond when a lickering motion is perceived and not upon the mere presence of a stimulus. It might be worth starting with a practice test to make sure that the patient understands the task. It is also recommended that the examiner pays very close attention to ixation losses, because patients are more likely to search for stimuli in licker perimetry than in other forms of perimetry because of its inherent challenges. 200 Chapter 10 | Non-conventional perimetry SHORT WAVELENGTH AUTOMATED PERIMETRY (SWAP) Short Wavelength Automated Perimetry (SWAP) is com- receive input from blue-sensitive retinal ganglion cells) monly referred to as blue-on-yellow perimetry, because while the intense yellow background is used to suppress it displays a large blue (short wavelength) stimulus of (i.e., adapt or fatigue) the relative sensitivity of both the Goldmann size V on a bright yellow background with a green (M-cones for “middle” wavelength cones) and the luminance of 315 asb (100 cd/m²).¹¹ The patient is asked red cones (L-cones for “long” wavelength cones). Sensitivity to respond whenever a blue stimulus is visible. thresholds are determined by increasing the luminance (i.e., light intensity) of the blue stimuli from less visible SWAP is designed to elicit a response from the blue sen- to more visible and are expressed in dB (FIG 10-7). Never- sitive pathway (S-cones for “short” wavelength cones and theless, the numerical dB values are not directly compa- the koniocellular cells in the lateral geniculate body that rable to those obtained with SAP. DESIGN OF SHORT WAVELENGTH AUTOMATED PERIMETRY (SWAP) LESS VISIBLE MORE VISIBLE LUMINANCE STIMULUS FIGURE 10-7 SWAP allows the determination of sensitivity thresholds by showing blue stimuli of increasing light intensity on an intense yellow background. Sensitivity thresholds are expressed in dB but are not directly comparable to results from SAP. Like other types of function-specific perimetry, SWAP addition, the patient’s eye needs to adapt to the very in- has also been shown to be useful for early glaucoma tense background for several minutes before starting the detection.¹²,¹³ Unlike licker perimetry, it is in luenced test in order to avoid false results. This light adaptation by media opacities and blur.¹⁴ is time-consuming and makes SWAP an overall longer test to perform than SAP. The task of performing SWAP is easy to understand for the patients (seen or unseen). Nevertheless, this test is However, given a patient who is able to perform the challenging for patients because the intensity of the yellow test reliably, SWAP is a useful perimetric test. BOX 10B background makes it dif icult to perceive the blue stimuli. provides practical guidance on how to best administer a This results in increased test-retest variability.¹⁵,¹⁶ In SWAP test. Stimulus V for patients with low vision 201 HOW TO ADMINISTER A RELIABLE SWAP TEST BOX 10B Most points highlighted in Chapter 3 on how to run a reliable visual ield test also apply to SWAP perimetry, However, particular attention needs to be given to some speci ic points. For SWAP, allow the patient’s eye to adapt to the very intense background for several minutes before starting the test in order to avoid untrustworthy results. Patients should be instructed to press the response button when they see a blue light presented anywhere in the bowl. The examiner should let the patient know that the color of the stimulus may appear to be slightly different from blue, as some patients report seeing the stimulus as bluish or purplish. SWAP is a more challenging test to perform than SAP. The examiner should closely monitor the patients as they are taking the test, to identify any need to rest. Particular attention should also be paid to reliability indices to ensure that patients are performing the test to the best of their ability. It is often helpful to provide a brief demonstration test to familiarize the patient with the test procedure. STIMULUS V FOR PATIENTS WITH LOW VISION There is a limit to the visibility of the standard size III dynamic range in regions of poor vision, the Goldmann white perimetric stimulus in patients with signi icantly stimulus V can be used. When this stimulus, which is 16 impaired sensitivity. This is because there are no longer times larger in area than the size III stimulus (FIG 10-8), enough intact cells to elicit a response to a stimulus even is displayed for a longer period of time (i.e., 200 ms), though the patient has some vision remaining (FIG 10-9). it provides a useful alternative perimetric stimulus for In order to overcome this loor effect and to increase the patients with severe visual ield loss. GOLDMANN STIMULUS SIZE III VS V FOR LOW VISION III 0.43° V 1.7° FIGURE 10-8 The Goldmann stimulus size V used for patients with severe vision loss is 16 times larger in area than the stan- dard Goldmann size III. Both are displayed with the same intensities on the same white background, but due to its greater size, the size V stimulus is more visible for low-vision patients than the size III. Because the larger stimulus V reaches more intact cells, longer can,¹⁷ as illustrated in the example shown in it can elicit a response when the smaller stimulus III no FIGURE 10-9. 202 Chapter 10 | Non-conventional perimetry ILLUSTRATION OF THE PRINCIPLE OF USING STIMULUS V FOR LOW VISION PATIENTS STIMULUS NORMAL ADVANCED PATHOLOGY SAP stimulus III 0.43° SAP stimulus V 1.7° Receptive retinal ganglion cells Parvocellular Koniocellular Magnocellular FIGURE 10-9 The standard Goldmann size III white stimulus is too small to reach sufficient cells to elicit a response in this example (top). The larger Goldmann stimulus size V can still trigger cells, offering an increased dynamic testing range for patients with severe vision loss. In addition to the increased dynamic range, the larger and intense stimulus available (as illustrated in FIG 6-3), is also thus more visible stimulus size V has also been shown to recommended. This approach saves valuable testing time have signi icantly lower test-retest variability compared to and is easier for patients to complete. For more informa- stimulus size III.¹⁸-²⁰ This is thought to be due to a larger tion on the low-vision strategy, see Chapter 6. stimulus being easier to see, which is essential in low- vision patients who struggle much more with perimetric Because stimulus sizes III and V are not directly compa- testing than patients with normal visual ields. rable, switching to stimulus V is only recommended for patients for whom testing with stimulus III no longer Besides using stimulus size V for low-vision patients, use renders useful clinical results, either due to the loor effect of the low-vision strategy, which starts with the most or the large variability of stimulus III. References 203 REFERENCES 1. Gonzalez de la Rosa M, Gonzalez-Hernandez M. Pulsar perimetry. A review and new results. Ophthalmologe. 2013;110:107-115. 2. Zeppieri M, Brusini P, Parisi L, Johnson CA, Sampaolesi R, Salvetat ML. Pulsar perimetry in the diagnosis of early glaucoma. Am J Ophthalmol. 2010;149:102-112. 3. van der Schoot J, Reus NJ, Colen TP, Lemij HG. The ability of short-wavelength automated perimetry to predict conversion to glaucoma. Ophthalmology. 2010;117:30-34. 4. Sample PA, Medeiros FA, Racette L, et al. Identifying glaucomatous vision loss with visual-function-speci ic perimetry in the diagnostic innovations in glaucoma study. Invest Ophthalmol Vis Sci. 2006;47:3381-3389. 5. Gonzalez-Hernandez M, de la Rosa MG, de la Vega RR, Hernandez-Vidal A. Long-term luctuation of standard automatic perimetry, pulsar perimetry and frequency-doubling technology in early glaucoma diagnosis. Ophthalmic Res. 2007;39:338-343. 6. Salvetat ML, Zeppieri M, Parisi L, Johnson CA, Sampaolesi R, Brusini P. Learning effect and test-retest variability of pulsar perimetry. J Glaucoma. 2013;22:230-237. 7. Matsumoto C, Takada S, Okuyama S, Arimura E, Hashimoto S, Shimomura Y. Automated licker perimetry in glaucoma using Octopus 311: a comparative study with the Humphrey Matrix. Acta Ophthalmol Scand. 2006;84:210-215. 8. Nomoto H, Matsumoto C, Takada S, et al. Detectability of glaucomatous changes using SAP, FDT, licker perimetry, and OCT. J Glaucoma. 2009;18:165-171. 9. Rota-Bartelink A. The diagnostic value of automated licker threshold perimetry. Curr Opin Ophthalmol. 1999;10: 135-139. 10. Lachenmayr BJ, Gleissner M. Flicker perimetry resists retinal image degradation. Invest Ophthalmol Vis Sci. 1992;33:3539-3542. 11. Sample PA, Johnson CA, Haegerstrom-Portnoy G, Adams AJ. Optimum parameters for short-wavelength automated perimetry. J Glaucoma. 1996;5:375-383. 12. Horn FK, Brenning A, Jünemann AG, Lausen B. Glaucoma detection with frequency doubling perimetry and short- wavelength perimetry. J Glaucoma. 2007;16:363-371. 13. Johnson CA, Brandt JD, Khong AM, Adams AJ. Short-wavelength automated perimetry in low-, medium-, and high- risk ocular hypertensive eyes. Initial baseline results. Arch Ophthalmol. 1995;113:70-76. 14. Delgado MF, Nguyen NT, Cox TA, et al. Automated perimetry: a report by the American Academy of Ophthalmology. Ophthalmology. 2002;109:2362-2374. 15. Mojon DS, Zulauf M. Normal values of short-wavelength automated perimetry. Ophthalmologica. 2003;217:260-264. 16. Kwon YH, Park HJ, Jap A, Ugurlu S, Caprioli J. Test-retest variability of blue-on-yellow perimetry is greater than white-on- white perimetry in normal subjects. Am J Ophthalmol. 1998;126:29-36. 17. Wall M, Woodward KR, Doyle CK, Zamba G. The effective dynamic ranges of standard automated perimetry sizes III and V and motion and matrix perimetry. Arch Ophthalmol. 2010;128:570-576. 18. Wall M, Doyle CK, Eden T, Zamba KD, Johnson CA. Size threshold perimetry performs as well as conventional automated perimetry with stimulus sizes III, V, and VI for glaucomatous loss. Invest Ophthalmol Vis Sci. 2013;54:3975-3983. 19. Wall M, Doyle CK, Zamba KD, Artes P, Johnson CA. The repeatability of mean defect with size III and size V standard automated perimetry. Invest Ophthalmol Vis Sci. 2013;54:1345-1351. 20. Wall M, Woodward KR, Doyle CK, Artes PH. Repeatability of automated perimetry: a comparison between standard automated perimetry with stimulus size III and V, matrix, and motion perimetry. Invest Ophthalmol Vis Sci. 2009;50:974-979. 204 205 CHAPTER 11 KINETIC PERIMETRY WHAT IS KINETIC PERIMETRY? LIMITATIONS OF STATIC PERIMETRY LOW SPATIAL RESOLUTION Static perimetry is currently the most commonly used The major drawback of static perimetry is that the most type of perimetry. With static perimetry, sensitivity common static test patterns have low spatial resolution. thresholds are determined at a speci ied number of test Because testing the entire visual ield with a densely locations. These thresholds are then compared to the spaced test grid would be very time-consuming, only a sensitivity thresholds of normal controls of the same representative sampling of potential visual ield locations age as the patient. Small changes in sensitivity can be is tested. As a result, static perimetry provides very lim- detected with high accuracy. Because this is essential ited information about small-sized scotomas such as the for detecting glaucoma and monitoring its progression, blind spot, as shown in FIG 11-1. Additionally, de ining the static perimetry is well suited for glaucoma care and boundaries of scotomas can also be compromised by the management. low spatial resolution of static perimetry. LOW SPATIAL RESOLUTION WITH STATIC PERIMETRY STATIC, 6° SPACING STATIC, 2° SPACING KINETIC (30-2) 10 20 10 20 10 20 FIGURE 11-1 Static perimetry has relatively low spatial resolution as demonstrated in this example in which the blind spot is tested. Using a 30-2 pattern with 6°spacing, only one or two locations are tested within the blind spot, providing no details about its size. Using a customized test pattern with 2°spacing provides higher, but not optimal resolution, while increasing test duration. Kinetic perimetry in this situation provides much higher spatial resolution with similar or lower test duration. 206 Chapter 11 | Kinetic perimetry SLOW PERIPHERAL TESTING Static perimetric testing is typically limited to the central FIG 5-13) or with widely spaced test grids such as in the 30° visual ield because this is the most crucial area of G-Periphery pattern (FIG 5-6) for glaucoma to save test visual function and the region in which most early and time. More detailed full threshold tests like the 07 pat- moderate glaucomatous scotomas occur. When static tern (FIG 5-11) require considerable test time and are too perimetry is performed in the periphery, it is often used long for some patients to complete reliably. In addition, in a qualitative way such as in legal documentation or vi- their accuracy is still limited due to the large extent of the sual disability tests (e.g., visual ield driving examinations, peripheral visual ield as illustrated in FIG 11-2. SLOW PERIPHERAL TESTING WITH STATIC PERIMETRY STATIC STATIC Quantitative dynamic Qualitative 2LT strategy strategy KINETIC 12 12 12 07 pattern 07 pattern 9 3 9 3 9 3 6 6 6 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 10 20 30 40 50 60 70 80 90 + + + + + + + + + + + + + + + + + + + + + + + + + + + + FIGURE 11-2 Peripheral testing with static perimetry is time-consuming under both quantitative and qualitative strategies, as this example of a postchiasmal lesion resulting in hemianopia with macular sparing demonstrates. Note that a kinetic test can be up to three times faster than a quantitative static test. What is kinetic perimetry 207 DESCRIPTION OF KINETIC PERIMETRY Kinetic perimetry is an alternative method to static that it is used to map a patient’s hill of vision in order perimetry. Its major advantages are that it provides high- to identify regions of normal and abnormal sensitivity to er spatial resolution, is faster for peripheral testing and light. However, the procedure used to achieve this goal is involves greater interaction between the examiner and fundamentally different. the patient. It has the same goal as static perimetry, in MOVING STIMULI ALONG VECTORS With kinetic perimetry, sensitivity thresholds are deter- visual ield location at which that response occurs has a mined by moving stimuli of various sizes and light inten- sensitivity threshold equal to the speci ic light intensity sities from a region of non-seeing to a region of seeing. used along the vector. The process continues so that all The trajectory of the stimulus is called a vector. regions of the visual ield are evaluated with this light intensity and stimulus size. This procedure is then repeat- As in static perimetry, the patient is asked to press the ed with stimuli of different intensities and size so that a response button once the stimulus is seen. The speci ic map of visual ield sensitivity can be generated (FIG 11-3). ISOPTERS When a suf iciently large number of vectors are tested light intensity. In pathological situations this does not throughout the visual ield with the same stimulus, the always apply because within the isopters there may be response points of each vector can be connected to form smaller areas of non-seeing (scotomas) that will be dis- a boundary of equal sensitivity. This boundary is called cussed in the next section. Several isopters can be drawn an isopter and is comparable to the contour line on a by varying the size and intensity of the stimuli from more topographical map. If a person has normal vision, then all visible (larger and more intense) to less visible (smaller points inside the isopter are areas of seeing and all points and dimmer) targets. outside the isopter are areas of non-seeing for a given SCOTOMAS Not all locations within a given isopter are areas of seeing. evaluations are called spot checks. Once located, radi- There may also be areas of non-seeing (i.e., scotomas). al vectors can be drawn moving again from the area of Using the analogy of the hill, these areas of non-seeing non-seeing (here the location of the center of the scoto- are like lakes or local depressions on the hill of vision, ma) towards an area of seeing (i.e., outwards). which are not identi iable using the procedure described above. Instead, static points of the same intensity as the Using this approach and combining all isopters and outer isopter already drawn have to be evaluated at differ- scotomas, the hill of vision can be drawn as illustrated ent locations inside the isopter to locate scotomas. These in FIG 11-4. 208 Chapter 11 | Kinetic perimetry KINETIC TESTING METHODOLOGY Patient Do you see Threshold along first vector response the stimulus? Sensitivity threshold Fixation Vector (Stimulus trajectory) Do you see Isopter Patient the stimulus? (Thresholds of one stimulus type) responses Sensitivity threshold Fixation Vector (Stimulus trajectory) Do you see Hill of vision Patient the stimulus? (Thresholds of several stimulus types) responses Sensitivity Fixation threshold Fixation Vector (Stimulus trajectory) FIGURE 11-3 In kinetic perimetry, sensitivity thresholds are determined by moving a stimulus of fixed intensity and size along a vector from an area of non-seeing to an area of seeing (top). In a normal visual field, the area of non-seeing to seeing is typically in the direction from the periphery towards fixation. The hill of vision can be drawn by connecting several thresholds of equal sensitivity (middle) thus forming an isopter and by drawing several isopters (bottom). An isopter can be thought of as a contour line of the hill of vision. What is kinetic perimetry 209 IDENTIFICATION OF LOCAL SCOTOMAS WITH KINETIC PERIMETRY Patient Do you see Static points inside isopter responses the stimulus? (Identification of areas of local depression) Sensitivity threshold Fixation = Seen = Not seen Static points Do you see Isopter of local depression Patient the stimulus? (Identification of boundaries of local depression) responses Sensitivity threshold Fixation = Seen = Not seen Vector (Stimulus trajectory) Do you see Hill of vision Patient the stimulus? (Thresholds of several stimulus types responses including local depressions) Sensitivity Fixation threshold Fixation Vector (Stimulus trajectory) FIGURE 11-4 Static points (spot checks) are used to identify areas of local depression. Once identified, radial vectors originating from the location of the local depression allow drawing the isopter representing the boundary of the local depression. The hill of vision can be drawn by connecting several thresholds of equal sensitivity thus forming an isopter and by drawing several isopters. 210 Chapter 11 | Kinetic perimetry HILL OF VISION AS A TOPOGRAPHICAL MAP Sensitivity threshold Fixation Fixation Isopter 4 Isopter 3 Isopter 2 Isopter 1 FIGURE 11-5 Kinetic results are displayed similarly to a topographical map. Lines of equal stimulus intensity and size are called isopters and are used to display the hill of vision in a two-dimensional map, similar to contour lines on a topographical map. Localized areas of non-seeing, such as that shown by the filled light blue circle, represent scotomas or areas of non- seeing for that target. THE HILL OF VISION AS A TOPOGRAPHICAL MAP Kinetic results are displayed as a topographical map. the hill of vision largely depends on its expected shape Similar to contour lines on a topographical map, isopters (i.e., the pattern of a speci ic pathology). In addition to are used to display the hill of vision with its outline, its the outline of the hill of vision, crevices, ridges and lo- crevices, ridges and even local depressions as shown in cal depressions have to be identi ied individually, and the FIG 11-5. In this manner the three-dimensional hill of slope of sensitivity transitions should be noted. Because vision can be represented in a two-dimensional drawing. of this, kinetic perimetry today is not fully automated and requires an interaction between the examiner and the The procedure used to create the topographical map of patient. WHY PERFORM KINETIC PERIMETRY? BENEFITS OF KINETIC PERIMETRY In contrast to static perimetry, in which thresholds are carefully determined at a number of pre-determined will be at threshold (scanning through a large area and locations (assessing a wide range of light intensities to identifying a speci ic location). This leads to a number of determine thresholds at each location), kinetic perimetry very distinct advantages of kinetic perimetry over static searches for the location at which a given light intensity perimetry. Why perform kinetic perimetry? 211 STATIC VERSUS KINETIC PERIMETRY STATIC Dynamic strategy 12 07 pattern 12 KINETIC 9 3 9 3 6 6 10 20 30 40 50 60 70 80 90 FIGURE 11-6 A patient with a ring scotoma due to retinitis pigmentosa tested both with static (left) and kinetic (right) perimetry. Note that kinetic perimetry provides a much higher spatial resolution that allows detection of even small defects. Static perimetry, in contrast, provides much less information during equal testing time. HIGH SPATIAL RESOLUTION Kinetic perimetry is better at de ining the pattern and present in quadrantanopia and hemianopia¹ or a con- shape of visual ield loss than static perimetry, as illus- stricted visual ield in end-stage glaucoma.² It is also very trated in FIG 11-6. Because the patient can report seeing bene icial if small scotomas need to be mapped reliably, the stimulus at any location along the entire trajectory of such as the blind spot or a scotoma due to a retinal hem- a vector, many possible response locations can be mapped orrhage. with a small number of vectors and the sequence of kinetic scanning can be different for each eye rather than However, while stimulus intensities may be varied, typ- using the same test pattern for all tests. This is especially ically only a small number of light intensities are used, bene icial if one is interested in identifying sharp-edged making it challenging to detect small threshold changes scotomas or steep isopter boundaries such as the de icits throughout the hill of vision. FAST PERIPHERAL TESTING Kinetic perimetry is a very ef icient method of evaluat- central visual field; thus kinetic perimetry has many ing the periphery (beyond 30 degrees of eccentricity), advantages for these conditions.¹,²,⁴-⁶ because a large area can be covered in a relatively short time due to the moving stimuli,³ as shown in FIG 11-6. Driving ability testing, legal blindness examinations or pto- sis testing³,⁷ also require peripheral visual ield evaluation. Several neurological and retinal diseases affect the pe- Thus, in some countries (e.g., Germany), kinetic perimetry ripheral visual ield earlier or more signi icantly than the is a legally accepted method to perform these tests. 212 Chapter 11 | Kinetic perimetry EASIER FOR PATIENTS Kinetic perimetry is highly lexible and interactive, and factors, kinetic perimetry is often used for low vision hence can be adjusted to the reliability and capabilities patients⁴,⁵ or patients who experience challenges in of the patient. Additionally, a moving stimulus is easier performing perimetry, including children.⁹ to see than a non-moving stimulus.⁸ Because of these LIMITATIONS OF KINETIC PERIMETRY REQUIRES HIGH SKILL OF THE EXAMINER Even though kinetic perimetry is highly versatile, one on the other hand, is a heuristic procedure that is highly of its drawbacks is that it cannot be fully automated for interactive between the patient and the examiner. Every all clinical situations, as the shape and height of an indi- stimulus manipulation by the examiner affects how the vidual hill of vision depends on pathology. Thus, kinetic patient will respond, and these responses will in turn in- perimetry requires much more interaction between the luence the next maneuver of the examiner. In this sense, examiner and patient than static perimetry. kinetic perimetry is similar to chess in that it incorpo- rates a lexible and adaptive strategy. Conceptually, the difference between static and kinetic perimetry is similar to the difference between checkers Being able to correctly map all possible clinical situations and chess. Static perimetry uses a pattern of visual ield requires great skill. Depending on prior knowledge, it locations (placed along either a Cartesian coordinate grid may take a training period of three months or more for or a polar coordinate system) that are ixed for each test, the examiner to become fully familiar and comfortable and uses the same strategy to determine the sensitivity with the test procedure in any situation. In this view, it threshold for an increment of light on the uniform back- is a very challenging procedure to implement on an auto- ground. It is similar to checkers in that the procedure is mated device. With a skilled and experienced examiner, essentially the same for each eye tested, which limits the however, it is possible to obtain the highest quality infor- amount of information one can obtain. Kinetic perimetry, mation concerning the peripheral visual ield. VARIABILITY AMONG EXAMINERS There is no consensus or standard method of conducting within one clinical center, the quality and ef iciency of kinetic perimetry, making it more challenging to com- kinetic perimetry can vary considerably from one exam- pare results from one clinical center with the indings iner to the next. from another than it is with static perimetry. And even Why perform kinetic perimetry? 213 COMPARISON BETWEEN STATIC AND KINETIC PERIMETRY TABLE 11-1 STATIC KINETIC LOCATIONS Fixed number of pre-determined Individually adjustable moving locations targets AUTOMATION Fully automated Semiautomated, needs involvement of examiner SPATIAL RESOLUTION Low High ACCURACY OF VISUAL Higher Lower SENSITIVITY THRESHOLDS WHAT IT IS BEST AT Small changes in sensitivity Small changes in spatial extent DETECTING (e.g., sharp-edged scotomas) Changes in central 30° Changes in periphery Remaining vision in advanced disease Defects in children COMMON USES Glaucoma Neuro-ophthalmological conditions Macular diseases Peripheral retinal diseases Visual ability testing Low vision Children CHALLENGING IDENTIFICATION OF SMALL SENSITIVITY CHANGES AND DIFFUSE LOSS While kinetic perimetry is better at identifying the pat- loss are more dif icult to identify with kinetic perimetry. terns and shapes of visual loss compared to static perime- A direct comparison between static and kinetic perimetry try, small sensitivity changes²,⁶ and widespread or diffuse is provided in TABLE 11-1. 214 Chapter 11 | Kinetic perimetry HOW TO PERFORM KINETIC PERIMETRY THE GOLDMANN PERIMETER: KINETIC VISUAL FIELD TESTING Quantitative kinetic perimetry was developed in 1946 Because of the lexible and adaptive properties of kinet- by Hans Goldmann and Haag-Streit¹⁰ and was the stan- ic perimetry, the manual Goldmann perimeter (FIG 11-7) dard of visual ield testing prior to the invention of the is still widely used and remains the reference for kinetic irst automated perimeter, the Octopus 201, in 1974.¹¹,¹² perimetry today. THE GOLDMANN PERIMETER AND ITS SUCCESSOR, THE OCTOPUS 900 FIGURE 11-7 The Octopus perimeters (right) retain all the characteristics of the manual Goldmann perimeter (left). To allow for continuity, the Octopus kinetic perimeter TABLE 11-2 summarizes the major differences and similar- retains all the characteristics of the manual Goldmann ities between Octopus and Goldmann kinetic perimetries. perimeter including the same lexible and adaptive prop- erties. It has been shown to be fully comparable to a It is helpful to keep the legacy of manual Goldmann manual Goldmann perimeter.¹³-¹⁶ In addition, it provides perimetry in mind because many de initions and uses standardized test conditions and semiautomation of ki- stem from the time when the Goldmann perimeter netic perimetry to optimize clinical work low and increase was invented, and they are easier to understand when consistency of results among examiners and centers. one is familiar with the manual Goldmann perimeter. How to perform kinetic perimetry 215 COMPARISON BETWEEN OCTOPUS KINETIC PERIMETRY AND GOLDMANN TABLE 11-2 KINETIC PERIMETRY OCTOPUS KINETIC PERIMETRY GOLDMANN KINETIC PERIMETRY METHODOLOGY Computer controlled stimulus Manual stimulus presentation presentation DESIGN Goldmann bowl (radius = 30cm) Goldmann bowl (radius = 30cm) Background illumination 31.4 asb Background illumination 31.4 asb (10 cd/m²) (10 cd/m²) STIMULUS TYPES Goldmann sizes I to V Goldmann sizes 0 to V Intensities 1a to 4e Intensities 1a to 4e STIMULUS SPEED Fixed (1 – 10°/s) Manually guided Manually guided VECTOR TYPES Guided vector Straight Free-hand vector Curvilinear Static points Static points INDIVIDUALIZATION & Full individualization Full individualization AUTOMATION Automation with added individualization Full automation ADDITIONAL FEATURES Reaction time compensation Normal isopter ranges KEY DECISIONS IN KINETIC PERIMETRY As with static perimetry, a number of key questions need • Which stimulus type should be used? to be asked before starting a kinetic test and the answers • Which stimulus size? will largely determine the results that one is able to • Which stimulus intensity? achieve. These questions are similar to those asked for • Which stimulus speed? static perimetry, but are answered differently. These • Which testing methodology should be used? questions are: • What is the trajectory of the vector? • Can some of the testing be automated? 216 Chapter 11 | Kinetic perimetry STIMULUS TYPES Similarly to the questions asked in static perimetry, the isopters and scotomas. Stimuli can be made more visible irst question about stimulus type in kinetic perimetry by changing the stimulus size or intensity or by varying has no clearly right or wrong answer. One can de ine both together. For a normal visual ield, the most visible standard testing methodologies for certain situations stimuli lead to the largest isopters and the least visible and follow them through for each patient. stimuli lead to the smallest isopters. In FIG 11-8, common stimuli are shown that allow a thorough assessment of In order to scan a patient’s entire hill of vision, one needs the full visual ield. more and less visible stimuli to be able to identify different NORMAL ISOPTERS FOR DIFFERENT STIMULUS TYPES 120 105 90 75 60 135 45 150 V4e 30 III4e 165 I4e 15 I2e I1e 180 10 30 40 50 60 70 80 90 0 195 345 210 330 225 315 240 255 270 285 300 FIGURE 11-8 By using stimuli of different size and intensity, the hill of vision of a person with normal vision can be drawn. The III4e stimulus is larger and more intense and leads to a larger isopter than the smaller and dimmer I1e stimulus. STIMULUS SIZE Octopus kinetic perimetry uses ive distinct stimulus siz- larger stimuli III to V are detected outside of the central es, Goldmann I to V, with Goldmann I being the smallest visual ield in people with normal vision. Goldmann size and each subsequent size being four times larger in area I is also often used to map small or shallow scotomas than the previous one as shown in TABLE 11-3. The sizes that require high spatial resolution (e.g., the blind spot). and naming scheme stem from the convention used by Although size 0 is available on the Goldmann perimeter, the manual Goldmann perimeter and were kept exactly it has not been included on the Octopus perimeter. This the same to provide direct continuity. is because the size 0 stimulus is dif icult to perceive through the optics of the eye, which can lead to unre- While there is no standardized procedure for kinetic liable and artefactual test results. The size O stimulus perimetry, and stimulus selection depends on the exam- also has a limited dynamic range. iner and the patient, Goldmann sizes I to V at the highest intensity are commonly used to test the far and inter- Goldmann V is the largest and most visible stimulus and is mediate peripheral visual ield. Goldmann sizes I and II often used for low vision patients who cannot see smaller combined with lower intensities are then used for the stimuli. highly sensitive central area because the isopters of the How to perform kinetic perimetry 217 GOLDMANN STIMULUS SIZES I TO V TABLE 11-3 SIZE DIAMETER AREA [MM2] RECOMMENDED FOR V 64 Low vision (end stage disease) Far periphery (determination of anatomical visual ield borders) 1.7° IV 16 0.8° III 4 Periphery Standard for static testing 0.43° II 1 0.2° I 0.25 Peripheral and central testing 0.1° Small area and high resolution (e.g., blind spot, small or shallow scotomas) STIMULUS INTENSITY Stimulus intensities in Octopus kinetic perimetry range for stimulus intensity stems from the manual Goldmann from 1a to 4e, with 1a being the dimmest and 4e being perimeter (Box 11A). Because this scale is the accepted the brightest. A total of 20 distinct stimulus intensities are standard in kinetic perimetry, it is also incorporated available, as shown in FIG 11-9. The naming convention into Octopus kinetic perimetry. STIMULUS INTENSITIES IN KINETIC PERIMETRY 1a: Darkest stimulus 4e: Brightest stimulus 19 dB 0 dB 1 2 3 4 a b c d e a b c d e a b c d e a b c d e FIGURE 11-9 The intensities of the Goldmann stimuli used in kinetic perimetry are presented in 1 dB steps from the darkest 1a to the brightest 4e intensity. 218 Chapter 11 | Kinetic perimetry As a rule, higher intensity stimuli such as the 4e are used dB) are usually chosen. When mapping absolute defects for peripheral testing and dimmer stimuli such as the 1e (i.e., areas of blindness), none of the stimuli are visible to are used for central testing. Using stimuli with very simi- the patient. Then, the brightest 4e stimulus can be select- lar intensities adds little diagnostic information because ed, as it is the easiest for the patient to see and possibly their isopters are very close to each other and would respond to at the borders of the defect. When there is a clutter the picture and represent a generally poor trade- wide separation between contour lines (isopters or sco- off between test duration and information gained. Thus, tomas), intermediate stimulus intensities can be selected stimuli with several dB differences in intensity (3 to 5 to test the region between the isopters. BOX 11A THE ORIGIN OF THE STIMULUS INTENSITY SCALE The manual Goldmann perimeter only contains one bright light source. In order to generate dimmer stimuli, ilters are placed in front of the light source, making the stimulus dimmer. There are two sets of ilters. Filters a, b, c, d and e dim the stimulus by 1 dB, and ilters 1, 2, 3 and 4 dim it by 5 dB. In combination, 20 different stimuli can be produced, with the brightest, 4e, representing a maximum stimulus brightness of 1,000 asb (315 cd/m²). STIMULUS SPEED Each stimulus for Octopus kinetic perimetry moves As a rule, stimulus velocities of 3-5°/s have been shown at a constant speed to allow for reproducible results. to optimize the trade-offs among accuracy, reliability and The stimulus speed should be selected to optimize the ef iciency¹³,¹⁷ and are recommended as a standard set- trade-off between accuracy and test duration. While the ting. For small scotomas such as the blind spot, slower in luence of patient reaction time is smaller for a slower stimuli of 2-3°/s are recommended as the clinically rel- stimulus, the longer testing time can result in fatigue. In evant spatial changes are small and are more accurately such cases, using a stimulus that moves faster leads to mapped with a slower stimulus. more reproducible results. GENERAL TESTING METHODOLOGIES Finding the adequate testing methodology for any pa- struction and advice from a colleague highly experienced tient is a process that requires an experienced examiner in performing this procedure is highly recommended. who can adapt to the patient’s responses. Consulting a textbook focusing speci ically on kinetic perimetry¹⁸-²⁰ The next sections will illustrate key concepts of kinetic is recommended for guidance. In addition, obtaining in- perimetry as a starting point for beginners, but are insuf i- cient to attain high pro iciency in kinetic perimetry. How to perform kinetic perimetry 219 IDENTIFICATION OF NORMAL ISOPTER LOCATION AND SHAPE For each stimulus size and intensity, Octopus kinetic pe- to age-matched normative data will allow correct in- rimetry automatically provides the age-matched normal terpretation of the results. As the hill of vision is rather isopter location as a reference. The inner dark central steep towards the far periphery, large age-related sen- band represents 25–75% of age-matched normals; the sitivity changes have only a small in luence on isopter outer light band denotes 5–95% of age-matched healthy location.²⁰-²² normals, as shown in FIG 11-10. In practical terms, the normal isopter location provides These zones support at-a-glance identi ication of devia- guidance on where to start placing vectors. Placing vec- tions from normal and are especially helpful in interpret- tors far outside of a normal isopter would only waste ing central visual ield defects and generalized diffuse or time, as the patient cannot see the stimuli in these areas. widespread loss. As the hill of vision is rather lat from Conversely, starting too near the anticipated location of the mid-periphery to the macula, those isopter locations detection can make the patient unprepared to respond are signi icantly in luenced by age and only comparison and can produce untrustworthy results. NORMAL ISOPTERS 120 105 90 75 60 135 45 5 – 95% 25 – 75% 150 30 3 165 1 15 180 10 00 30 40 50 60 70 80 90 195 45 345 210 30 330 225 315 240 255 270 285 300 FIGURE 11-10 The normal isopters provide guidance on where to start a vector of a given intensity. They also serve as a guide in judging whether an isopter is normal. The dark red band represents 25–75% of healthy normals; the outer light red band represents 5–95% of healthy normals of the same age. Note that the isopters are not round, but egg-shaped. They extend farthest in the inferior temporal visual field and least in the superior nasal visual field. 220 Chapter 11 | Kinetic perimetry MAPPING THE OUTLINE OF THE HILL OF VISION The overall outline of the hill of vision provides valuable affected area of the visual ield. As a general rule, stimuli information about a patient’s visual ield because devi- should not move directly along the horizontal or vertical ations from normal isopter shapes indicate abnormal meridians, because inconsistent results will be obtained. visual fields. Thus, mapping the outline of the hill of This is because the boundaries of quadrantanopia and vision is usually the irst step in kinetic perimetric testing. hemianopia are typically positioned along the horizon- To map the outline of the hill of vision, stimuli are moved tal and vertical meridians and a stimulus moving along from the peripheral end of the normal band towards the these meridians cannot map them clearly. Glaucomatous center ( ixation) along a given radial meridian. By repeating de icits along the nasal horizontal meridian (e.g,. nasal this procedure with different stimulus types, the outline steps and arcuate scotomas) represent another example of the hill of vision can be drawn in detail, as shown in where the stimulus should not be moved along the hor- FIG 11-11. izontal meridian. Thus, for these conditions, the radial vectors are best placed with an offset of a few degrees and This procedure is a fast and easy way to identify quadran- possibly parallel to the horizontal and vertical meridians. tanopia and hemianopia, as the isopter will dip in the MAPPING THE OUTLINE OF THE HILL OF VISION 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-11 Superior-nasal quadrantanopia identified with radial vectors along meridians. Note that the vectors along the horizontal and vertical midlines are placed parallel to them to allow for better detection of the boundaries of the visual loss in that quadrant. There are no responses in the superior nasal quadrant of this right eye, indicating the quadrantanopia. DETAILING THE BOUNDARIES OF AN ISOPTER As with any contour or topographic map, the hill of vision which either manifests as inconsistent with adjacent vec- may have crevices or depressions, which represent tors or outside of the expected normal sensitivity, which relative or absolute scotomas. As shown in FIG 11-11, these requires further investigation. defects may not be identi ied with standard vectors moving from the periphery to the center. This is where custom- Conceptually, the process is always the same. When ized individual assessment is needed. The examiner has alerted to a potential abnormal isopter shape, the operator to identify where there is a lack of normal response, should estimate where the isopter is likely to be. To verify How to perform kinetic perimetry 221 that this isopter is correct, additional vectors are drawn to confirm that it is outside of the normal expected perpendicular to the anticipated boundary of the isopter, responses. as shown in FIG 11-12. The perpendicular vectors optimize the likelihood that the hill of vision will be met “head-on”, If the patient response is as expected on the imagined which will reduce variability and provide more clinically isopter, the isopter shape is con irmed and can be drawn. meaningful information. Before initiating this process, it If not, the procedure has to be repeated, taking into account is important to recheck the abnormal isopter shape the new information until the isopter location is con irmed. DETAILING THE BOUNDARIES OF AN ISOPTER 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-12 Procedure for detailing the boundaries of abnormal isopters on a superior-nasal quadrantanopia. The lack of normal responses allows the examiner to estimate the location of the isopter (dotted gray line), and then test using perpendicular vectors (bold red) crossing that line to confirm the shape of the true isopter. IDENTIFICATION OF ISOLATED SCOTOMAS While the procedure shown in FIG 11-12 allows identi i- possible areas of sensitivity loss (areas of non-seeing or cation of the outline of the hill of vision, it usually misses scotomas). This allows for quick identi ication of scoto- isolated absolute defects or local depressions located in- mas as shown in FIG 11-13. side of an isopter or between isopters. In keeping with the analogy of a hill, isolated defects can be thought of as If areas of defects are identi ied, their boundaries can be lakes or depressions of different shapes and depths. In mapped by moving radial stimuli from inside of the de- order to identify these defects, spot-checking inside the fects from the center towards its edges. This procedure hill of vision must be performed. Spot-checking quickly can be repeated with stimuli of different visibility to examines locations between isopters using static points de ine the slope and depth of the defect. of the same size and intensity as the outer isopter, to ind 222 Chapter 11 | Kinetic perimetry IDENTIFICATION OF ISOLATED SCOTOMAS 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 Defect 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 Defect 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-13 By placing a static point of the same intensity inside of an isopter or between isopters (spot checking, red circles), one can identify local defects that would otherwise be missed (no response, gray circle). Using radial vectors (bold red lines) from the center of the area of non-seeing (from the inside) to the area of seeing (to the outside) allows drawing the boundaries (gray bold line) of the defect in detail. For ease of reading, the defect should be filled with the appropriate color. MAPPING THE HILL OF VISION USING SEVERAL STIMULUS TYPES By repeating the procedures described in the previous identifying an unnatural isopter shape without having to sections using different stimulus types with different sizes use extra vectors. and intensities, several isopters can be drawn to charac- terize the patient’s entire hill of vision. There are many When spot checking to identify local areas of depres- tips and tricks to make this procedure ef icient. A few of sion, the size and intensity of the outer isopter should them are presented here. be used between the outer and the inner isopters (FIG 11- 14). Then, only the size and intensity of the inner isopter When drawing a second isopter, placing the vectors of the should be used farther towards the center. second isopter with a radial offset to the ones used in the irst isopter is recommended, as seen in FIG 11-14. In other It is also important to remember that there may be more words, the vectors used to determine the second isopter than one isopter for the same stimulus size and intensity. should be placed at different locations than those used to There may be a region of detecting the target in the far determine the irst isopter. This increases the chance of periphery, with an area of non-seeing closer to ixation, How to perform kinetic perimetry 223 followed by a second area that can detect the target. This the visual pathways. Because of this, it is important to can occur in some cases of retinal disease, moderate to make good use of spot checking and evaluate the entire advanced glaucoma, and neurologic disorders affecting visual ield. PLACEMENT OF VECTORS AND STATIC POINTS USING DIFFERENT STIMULUS TYPES 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 III4e III4e I2e I2e FIGURE 11-14 Vectors of different stimulus sizes and intensities are best placed with an offset to increase the chance of identification of abnormal isopter shapes. When placing static points between two isopters, always use the intensity of the more visible outer isopter. Local scotomas can be absolute defects with sharp-edged FIG 11-15. For easy interpretation, these local depressions boundaries such as the blind spot or relative defects with are typically illed with color to indicate that the corre- a gentle slope on the edge of the defect as in glaucoma. To sponding stimulus cannot be seen within that visual ield distinguish between the two, more than one stimulus is area. needed to characterize a local scotoma as can be seen in DISTINCTION BETWEEN ABSOLUTE AND RELATIVE SCOTOMAS 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 III4e I2e FIGURE 11-15 More than one isopter is needed to distinguish between absolute and relative scotomas. This example shows a nasal step for a glaucoma patient. 224 Chapter 11 | Kinetic perimetry CHECKING FOR VISUAL FIELD RELIABILITY Like static visual ield testing, kinetic perimetry has a certain vectors to check for consistency of responses, patient-related subjective component and the reliability as shown in FIG 11-16. To do this, two vectors should be of the results largely depends on good patient coopera- placed as close together as possible (or repeated) and tion and minimizing variability due to learning or fatigue then compared for consistency. If the responses are effects.²¹,²³,²⁴ Therefore, it is also essential to check reliable, the two patient responses should be very close for patient reliability in kinetic perimetry. While static pe- together, as shown in the igure below to the left which rimetry uses global indices such as false positive and false means there is low test-retest variability. If they are sep- negative catch trials and short-term luctuation, kinetic arated, as in the example below to the right, it indicates perimetry employs other methodologies to test for similar an unreliable result with high test-retest variability. This reliability indicators. procedure provides a good indicator for the quality of the results. Similarly, spot checking can be repeated at To assess short-term luctuation, it is worth duplicating various locations to assess response consistency. CHECKING FOR SHORT-TERM FLUCTUATION LOW SHORT-TERM FLUCTUATION HIGH SHORT-TERM FLUCTUATION The two gray dots on each vector The two gray dots on each vector are close to each other are far from each other 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-16 By repeating some vectors, short-term fluctuation and thus test-retest variability can be assessed. If the responses are close together (left), it indicates good patient cooperation, good repeatability and high reliability. If the responses largely differ (right), it indicates an unreliable visual field. In legal driving and blindness examinations performed positive and false negative answers even though the pro- with kinetic perimetry, it is worth checking for false an- cedure is different. Checking for false positive answers swers to identify patients who may simulate responses can be easily done by presenting stimuli outside of the or a lack of response (functional changes or visual mea- normal isopter area (FIG 11-17). By de inition, the patient sures that are non-physiologic and non-pathologic). This is not supposed to see these stimuli. If there are many can produce visual ield results that are either better or positive responses, this is a strong indicator of a patient worse than the actual visual ield sensitivity pro ile. As who is malingering. in static perimetry, it is possible to check for both false How to perform kinetic perimetry 225 CHECKING FOR FALSE POSITIVES NO FALSE POSITIVES FALSE POSITIVES No response to stimuli 3 responses to stimuli outside normal isopter outside normal isopter 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-17 Checking for false positive responses can be done by placing vectors or static points outside of a normal isopter. If a patient responds, then these are false positives, as the patient cannot see them. To detect false negative answers one places a more intense the patient to observe (FIG 11-18). Failure to see a more or larger stimulus at a location where the stimulus was intense or larger stimulus than the one that was detected previously detected. This stimulus should be easy for at threshold is considered to be a false negative response. CHECKING FOR FALSE NEGATIVES NO FALSE NEGATIVES FALSE NEGATIVES Immediate response to larger or more intense Random response to larger or more intense stimulus at a location previously seen stimulus at a location previously seen 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 Larger or more 345 195 Larger or more 345 intense stimulus intense stimulus 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 Smaller or less Smaller or less intense stimulus intense stimulus FIGURE 11-18 Checking for false negative responses can be done by placing larger or more intense vectors or static points at a location where a smaller or less intense stimulus was previously detected. If a patient does not respond, then these are false negatives, as the patient should be able to see them. 226 Chapter 11 | Kinetic perimetry PATIENT REACTION TIME COMPENSATION Patient reaction time in luences the size of an isopter as For this reason, Octopus kinetic perimetry offers the the patient’s response is produced some time after the possibility of adjusting for patient reaction time by mea- stimulus is actually seen.²¹,²²,²⁵ This also adds signi icant suring its magnitude in the patient’s intact visual ield variability to the test procedure.²³ If a patient’s re- and applying a reaction time correction for it, as illus- sponses were always instantaneous, outlines of the hill trated in FIG 11-19. In order to do so, the examiner should of vision would be larger and isolated defects would be choose a reaction time vector of the same stimulus type smaller than they appear on the printout. This makes as the isopter and place it into the patient’s seeing area. The the interpretation of results challenging, especially in pa- patient should be able to see the stimulus immediately tients with long or inconsistent reaction times. as it is presented. Thus, the time between stimulus pre- sentation and when the patient presses the response button represents the patient’s reaction time. PATIENT REACTION TIME COMPENSATION Reaction time (RT) vector Patient sees Patient responds Standard vector Standard vector with RT compensation 0 ms 523 ms FIGURE 11-19 There is always a lag between the moment the patient sees a stimulus and the moment a patient presses the response button. This constitutes the patient’s reaction time. By placing reaction time (RT) vectors into the patient’s seeing area, one can account for this lag. For a precise measurement of patient reaction time, us- placing the reaction time vectors close to the correspond- ing the average reaction time obtained from two or three ing isopter. FIG 11-20 provides an example of the clinical different vectors for each stimulus type is recommended, usefulness of reaction time compensation. How to perform kinetic perimetry 227 EXAMPLE OF THE CLINICAL USEFULNESS OF REACTION TIME COMPENSATION REACTION-TIME REACTION-TIME compensation compensation turned OFF turned ON Reaction time vectors 10 30 40 50 60 70 80 10 30 40 50 60 70 80 FIGURE 11-20 Without reaction time compensation, local depressions look uncharacteristically large (left). By using reaction time vectors (bold red, double arrows) to determine the patient’s reaction time and by turning reaction time compensation on (right), the patient’s adjusted defect size is revealed. STEP-BY-STEP EXAMPLE OF KINETIC PERIMETRY A real-life example of a complete kinetic test as performed in clinical practices is provided in FIGURE 11-21. STEP-BY-STEP EXAMPLE OF A KINETIC TEST WITH SEVERAL ISOPTERS (STEPS 1-2) 1. Mapping outline of hill of vision 2. Detailing boundaries of isopter I4e, 5°/s I4e, 5°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-21 This example above shows a full kinetic perimetric test of a quadrantanopia with 4 isopters (shown here in blue, red, gray and green), static points and reaction time compensation. Checks for consistent results and false positives are not shown in this example. 228 Chapter 11 | Kinetic perimetry STEP-BY-STEP EXAMPLE OF A KINETIC TEST WITH SEVERAL ISOPTERS (STEPS 3-8) 3. Drawing isopter 4. Mapping the next outline of hill of I4e, 5°/s vision & detailing boundaries of isopter in abnormal response region V4e, 5°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 5. Drawing isopter 6. Spot-checking between isopters V4e, 5°/s Use stimulus type from outer isopter V4e, 0°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 7. Mapping the next outline of hill of 8. Drawing isopter vision & detailing boundaries of isopter I2e, 5°/s I2e, 5°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 How to perform kinetic perimetry 229 STEP-BY-STEP EXAMPLE OF A KINETIC TEST WITH SEVERAL ISOPTERS (STEPS 9-14) 9. Spot-checking between isopters 10. Mapping the next outline of hill of vision Use stimulus type from outer isopter & detailing boundaries & drawing isopter I4e, 0°/s I1e, 2°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 11. Spot-checking between isopters 12. Mapping of isolated defect Use stimulus type from outer isopter (blind spot) I2e, 0°/s, I1e, 0°/s I4e, 2°/s 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 13. Draw reaction time vectors 14. Reaction-time compensation in visible area RT on RT vectors, same intensity, size and speed as respective standard vector 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 230 Chapter 11 | Kinetic perimetry AUTOMATION OF KINETIC PERIMETRY MANUAL KINETIC PERIMETRY – FULL FLEXIBILITY In manual kinetic perimetry, the operator draws each uation. A drawback of manual kinetic perimetry is the vector individually for each patient. This procedure, lack of consensus for a standard way to conduct it. As a which is used on manual Goldmann perimeters, is fully result, there is limited comparability between the results implemented on the Octopus perimeters. Therefore, a obtained from different examiners and clinics. Another Goldmann manual perimetric test can be performed on drawback is that manual kinetic perimetry requires the Octopus perimeter. The example presented above intensive training and there is a certain operator bias. illustrates the lexibility of manual kinetic perimetry. Simpler procedures are therefore desirable for more consistent and effective clinical work lows. Manual kinetic perimetry is still widely used today be- cause it allows full lexibility to adapt to any patient sit- AUTOMATED KINETIC PERIMETRY– STANDARDIZATION While kinetic perimetry testing often needs to be individ- responses are already known. An example is visual ield ualized, there are certain indications where the expected testing for ptosis, as illustrated in FIG 11-22. EXAMPLE OF FULLY AUTOMATED KINETIC PERIMETRY TO TEST FOR PTOSIS PTOSIS TEMPLATE PTOSIS TEST Full automation possible Test performed twice with taped and untaped lid 120 105 90 75 60 120 105 90 75 60 Taped 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 0 90 0 0 60 70 80 Untaped 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-22 In ptosis testing, one is trying to identify the exact position of the lid, which always curves upwards from the nasal to temporal side. Therefore, a standardized testing procedure of a few vertical vectors is all that is needed and a very visible and adequately fast III4e to V4e at 3–5°/s is a good stimulus choice. This procedure can be fully automated and performed both on taped and untaped lids. How to perform kinetic perimetry 231 For any such indication with a clearly known defect pat- Full automation not only standardizes kinetic testing and tern, Octopus kinetic perimetry allows storage of fully makes it much more comparable across examiners and automated templates that can, once programmed, be clinics, it also makes the procedure as easy to learn and run in the same way as Standard Automated Perimetry perform as static perimetry. As there is currently no con- by simply pressing the start button. Only the isopters sensus on how a certain indication should be tested, each remain to be drawn manually. clinic can de ine the automated templates according to its current testing methodologies. SEMIAUTOMATED KINETIC PERIMETRY – STANDARDIZATION AND FULL FLEXIBILITY Semiautomated kinetic perimetry offers the bene its of mode. In contrast to automated kinetic perimetry, vec- both automated and manual kinetic perimetry with much tors can be individually added, but responses can also be less of their respective shortcomings, and is a part of Octo- repeated or deleted if the examiner deems it necessary. pus kinetic perimetry. Because of the full lexibility offered by semiautomated kinetic perimetry, it can provide results that are as pre- In semiautomated kinetic perimetry, the examination is cise as manual kinetic perimetry while greatly improving started using a given prede ined template in an automated the standardization within a clinic, as all examiners use EXAMPLE OF CUSTOMIZED TEMPLATES FOR NEURO-OPHTHALMIC CONDITIONS GENERAL ASSESSMENT PITUITARY ADENOMA 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 HEMIANOPIA BLIND SPOT 120 105 90 75 60 120 105 90 75 60 135 45 135 45 150 30 150 30 165 15 165 15 180 10 30 40 50 60 70 80 90 0 180 10 30 40 50 60 70 80 90 0 195 345 195 345 210 330 210 330 225 315 225 315 240 255 270 285 300 240 255 270 285 300 FIGURE 11-23 Kinetic templates allow testing standardization, as the same methodology is always used. Full flexibility of adap- tation to a patient’s specific situation is also enabled. Above are four examples of templates regularly used in a neuro-ophthalmic - clinic.²⁴ ²⁶ For simplicity, only one stimulus type is displayed, but templates with more than one stimulus type are also possible. 232 Chapter 11 | Kinetic perimetry the same underlying technique and only make adapta- commonly occurring indications, based on each clinic’s tions if the patient requires it. This greatly improves needs. FIG 11-23 shows a number of templates that can be consistency amongexaminers and facilitates clinical used in a neuro-ophthalmic clinic. These templates are not result interpretation. considered the only possible templates for such condi- tions, but rather examples of performing effective kinetic Many different templates can be created for the most perimetry in these situations. References 233 REFERENCES 1. Rowe FJ, Noonan C, Manuel M. Comparison of octopus semi-automated kinetic perimetry and humphrey peripheral static perimetry in neuro-ophthalmic cases. ISRN Ophthalmol. 2013;doi: 10.1155/2013/753202. 2. Scheuerle AF, Schiefer U, Rohrschneider K. Functional diagnostic options for advanced and end stage glaucoma. Ophthalmologe. 2012;109: 337-344. 3. Alniemi ST, Pang NK, Woog JJ, Bradley EA. Comparison of automated and manual perimetry in patients with blepharoptosis. Ophthal Plast Reconstr Surg. 2013;29:361-363. 4. Nowomiejska K, Brzozowska A, Koss MJ, et al. Quanti ication of the visual ield loss in retinitis pigmentosa using semi- automated kinetic perimetry. Curr Eye Res. 2016;doi: 10.3109/02713683.2015.1079328. 5. Nowomiejska K, Wrobel-Dudzinska D, Ksiazek K, et al. Semi-automated kinetic perimetry provides additional information to static automated perimetry in the assessment of the remaining visual ield in end-stage glaucoma. Ophthalmic Physiol Opt. 2015;35:147-154. 6. Agarwal HC, Gulati V, Sihota R. Visual ield assessment in glaucoma: comparative evaluation of manual kinetic Goldmann perimetry and automated static perimetry. Indian J Ophthalmol. 2000;48:301-306. 7. Riemann CD, Hanson S, Foster JA. A comparison of manual kinetic and automated static perimetry in obtaining ptosis ields. Arch Ophthalmol. 2000;118:65-69. 8. Nevalainen J, Paetzold J, Krapp E, Vonthein R, Johnson CA, Schiefer U. The use of semi-automated kinetic perimetry (SKP) to monitor advanced glaucomatous visual ield loss. Graefes Arch Clin Exp Ophthalmol. 2008;246:1331-1339. 9. Patel DE, Cumberland PM, Walters BC, Russell-Eggitt I, Rahi JS, OPTIC study group. Study of Optimal Perimetric Testing in Children (OPTIC): Feasibility, Reliability and Repeatability of Perimetry in Children. PLoS One. 2015; doi: 10.1371/ journal.pone.0130895. 10. Haag-Streit AG, (Hrsg.). 1858 - 2008: 150 Jahre Haag-Streit/150 Years of Haag-Streit. Bern: Stämp li Publikationen AG; 2008. 11. Johnson CA, Wall M, Thompson HS. A history of perimetry and visual ield testing. Optom Vis Sci. 2011;88:E8-15. 12. Fankhauser F. Remembrance of Hans Goldmann, 1899-1991. Surv Ophthalmol. 1992;37:137-142. 13. Rowe FJ, Rowlands A. Comparison of diagnostic accuracy between Octopus 900 and Goldmann kinetic visual ields. Biomed Res Int. 2014;doi: 10.1155/2014/214829. 14. Nowomiejska K, Vonthein R, Paetzold J, Zagorski Z, Kardon R, Schiefer U. Comparison between semiautomated kinetic perimetry and conventional Goldmann manual kinetic perimetry in advanced visual ield loss. Ophthalmology. 2005;112:1343-1354. 15. Ramirez AM, Chaya CJ, Gordon LK, Giaconi JA. A comparison of semiautomated versus manual Goldmann kinetic perimetry in patients with visually signi icant glaucoma. J Glaucoma. 2008;17:111-117. 16. Rowe FJ, Hanif S. Uniocular and binocular ields of rotation measures: Octopus versus Goldmann. Graefes Arch Clini Exp Ophthalmol. 2011;249:909-919. 17. Johnson CA, Keltner JL. Optimal rates of movement for kinetic perimetry. Arch Ophthalmol. 1987;105:73-75. 18. Anderson DR. Testing the ield of vision. St.Louis: CV Mosby; 1982. 19. Anderson DR. Perimetry - With and without automation. 2nd ed. St.Louis: CV Mosby; 1987. 20. Walsh TJ. Visual Fields: Examination and interpretation. 3rd ed. American Academy of Ophthalmology Monograph Series; Oxford University Press; 2010. 21. Nowomiejska K, Brzozowska A, Zarnowski T, Rejdak R, Weleber RG, Schiefer U. Variability in isopter position and fatigue during semi-automated kinetic perimetry. Ophthalmologica. 2012;227:166-172. 22. Grobbel J, Dietzsch J, Johnson CA, et al. Normal values for the full visual ield, corrected for age- and reaction time, using semiautomated kinetic testing on the Octopus 900 perimeter. Transl Vis Sci Technol. 2016;5:doi:10.1167/tvst.5.2.5. 23. Hirasawa K, Shoji N. Learning effect and repeatability of automated kinetic perimetry in healthy participants. Curr Eye Res. 2014;39:928-937. 24. Rowe FJ, Sarkies NJ. Assessment of visual function in idiopathic intracranial hypertension: a prospective study. Eye (Lond). 1998;12:111-118. 25. Rowe FJ, Cheyne CP, García-Fiñana M, et al. Detection of visual ield loss in pituitary disease: Peripheral kinetic versus central static. Neuro-Ophthalmology. 2015;39:116-124. 26. Rowe FJ, Wright D, Brand D, et al. A prospective pro ile of visual ield loss following stroke: prevalence, type, rehabilitation, and outcome. Biomed Res Int. 2013; doi: 10.1155/2013/719096. 234 235 CHAPTER 12 TRANSITIONING TO A DIFFERENT PERIMETER MODEL INTRODUCTION At the end of the life span of a perimeter or in order to are obtained on different perimeter models is irst pre- bene it from technologies only available on a different sented. Then, this chapter highlights that while sensi- perimeter model or brand, transitioning to a new perimeter tivity thresholds are not directly comparable between with distinct characteristics may be necessary. Due to different models, sensitivity losses (i.e., deviations from differences in the design and test parameters between normal sensitivity thresholds) are comparable to a large perimeter models, the measured sensitivity thresholds extent because of the use of device-speci ic normative are not directly comparable. As a result, the variability databases. This chapter also provides practical guidance introduced by a transition must be acknowledged and on how to minimize patient-related luctuation that may addressed. arise during the transition and subside as patients become familiar with the new device. Octopus perimeters offer several features that make it possible to transition smoothly between perimeter In addition, when transitioning from an HFA to an Octopus models, regardless of whether the transition is from perimeter, it is important to recognize that each perimeter one Octopus model to another Octopus model or from a uses its own, sometimes proprietary, test parameters Humphrey Field Analyzer (HFA) to any Octopus model. and result displays. As a result, the transition may appear These features minimize, to a large extent, the impact of challenging. Practical recommendations for the selection the different parameters used in the various perimeter of test patterns and strategies are presented to facilitate models and are systematically presented in this chapter. the transition. Furthermore, information is provided on how to interpret the perimetric result after the transition An explanation of why different sensitivity thresholds from an HFA to an Octopus perimeter. 236 Chapter 12 | Transitioning to a different perimeter model GENERAL ASPECTS OF TRANSITIONING MEASURED SENSITIVITY THRESHOLDS CANNOT BE COMPARED ACROSS DIFFERENT PERIMETER MODELS Since different perimeter models vary in design and sensitivity thresholds cannot be directly compared. BOX sometimes use different test parameters, patients may 12A presents an overview of major causes of variability perceive perimetric stimuli differently. As a result, the between different Octopus perimeter models as well as measured sensitivity thresholds can vary¹ and measured the HFA perimeter and Octopus perimeter models. BOX 12A MAJOR DIFFERENCES BETWEEN VARIOUS OCTOPUS PERIMETER MODELS Since the various Octopus perimeter models vary in design and sometimes use different test parameters, measured sensitivity thresholds also vary.¹ Firstly, design differences can lead to a different perception of perimetric stimuli. For example, there are two fundamentally different designs used in recent Octopus perimeter models. Cupola perimeters (e.g., Octopus 101 and 900) allow for testing of the full ield (e.g., 90° radius) and use a moving projector to present the perimetric stimuli onto the whitish surface of a cupola. On the other hand, screen-based perimeters (e.g., Octopus 600) allow for testing of the central ield only (e.g., 30° radius) and generate the stimulus on a computer display. Because of the different stimulus presentation technologies used, patients may perceive stimuli differently. In addition, the full ield cupola perimeters are open and thus need to operate under dim room lighting conditions to avoid stray light in luencing the result, whereas screen-based perimeters are closed, not in luenced by stray light and thus can be operated under daylight conditions. Further, while the mechanical projector of the cupola perimeters makes some noise upon stimulus presentation, screen- based perimeters are silent during stimulus presentations. As a result, even if completely identical test conditions are used (i.e., same stimulus size, same stimulus luminance and same background luminance), patients may respond differently. They can be in luenced by these differences and, as a result, determined sensitivity thresholds may vary. Secondly, different test parameters may also lead to different perimetric results. For this reason, all recent Octopus models (e.g., Octopus 900, Octopus 600, Octopus 300 and Octopus 123) use the same ixed test parameters, which are described in Box 4A. An exception is the Octopus 101, which uses a background luminance of 4 asb (instead of 31.4 asb), operating under mesopic illumination (i.e., midway between daylight and night vision) instead of photopic illumination (i.e., daylight vision), which may in luence the perception of the perimetric stimulus. To reduce this bias when transitioning from an Octopus 101 to an Octopus 900, the Octopus 900 can be optionally operated using a background luminance of 4 asb. MAJOR DIFFERENCES BETWEEN THE HFA PERIMETER AND OCTOPUS PERIMETER MODELS As already explained in the section above, design differences between the HFA perimeter (which is a cupola perimeter) and other Octopus perimeter models may lead to different perception of perimetric stimuli even if the same test conditions were used. However, the HFA perimeter and the various Octopus perimeter models also use different ixed test parameters. The most marked difference between the determined sensitivity thresholds of an HFA perimeter and recent Octopus perimeter models (e.g., Octopus 900, 600, 300 and 123) stems from the different maximum stimulus luminances used (4,000 asb in Octopus perimeters compared to 10,000 General aspects of transitioning 237 asb in HFA perimeters). This difference leads to an offset of 4 dB in the default decibel scale used to display sensitivity thresholds. This is due to the fact that both instruments take the maximum stimulus luminance as the origin of their dB scale (0 dB), as explained in BOX 2A. A stimulus of 1,000 asb intensity therefore corresponds to a sensitivity threshold of 10 dB on an HFA II perimeter and to 6 dB on an Octopus 900 perimeter. DEVICE-SPECIFIC NORMATIVE DATABASES ALLOW COMPARISON OF SENSITIVITY LOSSES BETWEEN DEVICES SENSITIVITY LOSSES CAN BE COMPARED BETWEEN DIFFERENT OCTOPUS PERIMETER MODELS Whenever an Octopus perimeter model is developed, an Octopus 300) are imported into another model (e.g., data are collected from people with healthy eyes and of Octopus 900), the user can be sure that the imported sen- different ages on that model in order to develop a nor- sitivity thresholds are compared with the Octopus 300 mative database for it (see BOX 2B for more detail on device-speci ic normative database to calculate the sen- normative databases). As a result, each Octopus model sitivity losses. has its respective normative database. Furthermore, all Octopus models contain the normative databases of all Using device-speci ic normative databases largely elimi- other models in order to allow for smooth transitions nates device-speci ic differences in sensitivity losses. As between models. When transitioning from one Octopus a result, sensitivity losses and all related representations, model to another, the existing data of one device can be with the exception of the Values and Grayscale (Values), imported into the other device and the data compared to are largely comparable across perimeter models as shown the appropriate normative database. For example, when in FIG 12-1. the visual ield tests taken on a given Octopus model (e.g., SENSITIVITY LOSSES CAN BE COMPARED BETWEEN HFA PERIMETERS AND DIFFERENT OCTOPUS PERIMETER MODELS HFA perimeters use an HFA-speci ic normative database measured sensitivity thresholds of an HFA II and an to calculate the sensitivity losses presented in the Total Octopus 900 show an offset of 4 dB as explained in BOX Deviation representation, while each Octopus model 12A, the respective normative databases show the same uses its own normative database. As a result, the use of offset, and as a result the sensitivity losses are comparable. these device-speci ic normative databases largely elimi- This means that all representations with the exception of nates any model-related bias between perimetric results the Values and Grayscale (Values) representations are when looking at sensitivity losses. For example, while the comparable. 238 Chapter 12 | Transitioning to a different perimeter model SENSITIVITIY LOSSES BETWEEN DIFFERENT DEVICES ARE LARGELY COMPARABLE MEASURED SENSITIVITY NORMATIVE VALUES SENSITIVITY LOSSES THRESHOLDS Values Normative Database Corrected Probabilities 6 16 20 18 OCTOPUS 900 13 16 23 20 17 19 18 19 23 23 23 21 19 19 17 19 18 24 29 24 19 19 18 19 17 18 20 24 26 28 22 16 17 13 16 24 21 28 27 23 9 11 18 19 20 23 22 21 11 1 5 13 18 20 18 8 5 2 0 15 3 2 5 3 NOT COMPARABLE COMPARABLE Values Normative Database Corrected Probabilities 5 14 18 17 OCTOPUS 600 10 16 20 21 14 21 17 19 22 22 23 22 21 20 16 21 25 20 25 21 24 21 23 17 18 20 23 24 25 26 23 16 18 18 23 26 23 27 29 22 1 6 21 23 26 26 22 21 11 1 0 23 23 18 8 5 2 15 4 2 5 3 NOT COMPARABLE COMPARABLE Threshold Values Normative Database Pattern Deviation Probability Map HFA II FIGURE 12-1 This example illustrates the benefits of using device-specific normative databases (i.e., an individual normative database for each device). In this example, sensitivity thresholds of a patient with retinal detachment were determined on an Octopus 900, Octopus 600 and on an HFA II perimeter on the same day (left). These sensitivity thresholds cannot be compared to each other due to the different characteristics of the three perimeter models. However, because distinct normative databases are used for the Octopus 900, Octopus 600 and the HFA II perimeter (middle), the sensitivity losses are comparable. Sensitivity losses are calculated as the deviation of the measured sensitivity thresholds of each model from its respective normative database and are the basis of most visual field representations such as the Corrected Probabilities or Pattern Deviation Probability Map shown in this figure. Note that comparability applies to all representations with the exception of the Values and Grayscale (Values) representations. General aspects of transitioning 239 IMPORT OF EXISTING DATA TO ENSURE CONTINUITY IMPORT OF EXISTING DATA FROM ONE OCTOPUS PERIMETER MODEL TO ANOTHER As presented in Chapter 9, a series of visual ield tests data. All current Octopus perimeters therefore allow for over time is necessary to adequately assess visual ield the import of electronically stored visual ield results progression in diseases such as glaucoma. When tran- from the Octopus models 500, 101, 123, 300, 900 and sitioning from one perimeter to another, it is therefore 600. Data can be transferred either in a single session or essential to be able to use a patient’s existing visual ield on a continuous basis if the other perimeter is still in use. RAW DATA AFTER IMPORT CAN BE DISPLAYED IN ANY OCTOPUS FORMAT RAW DATA FROM DIFFERENT DISPLAY IN ANY OCTOPUS FORMAT PERIMETER MODELS Values Normative Database OCTOPUS 900 6 16 20 18 13 16 23 20 17 19 18 19 23 23 23 21 19 19 17 19 18 24 29 24 19 19 18 19 Octopus HFA-style Octopus 7-in-1 17 18 20 24 26 28 22 16 17 13 16 24 21 28 27 23 9 11 Single field analysis Right eye (OD) 18 19 20 23 22 21 11 1 5 13 Name: Demo John ID: GS_103 Date of birth: 1977/01/01 Demo John, 1977/01/01 (39yrs) 18 20 18 8 5 2 0 32 Right eye (OD) / 08/08/2016 / 09:05:48 15 3 2 Fixation monitor: Min Stimulus: III / 4000 asb / White Pupil diameter: Date: 08/08/2016 Seven-in-One Fixation target: Cross marks Background: 31 asb / White Visual acuity: null Time: 09:05:48 5 3 Fixation losses: 0/0 Strategy: TOP RX: Age: 39 Grayscale (CO) Values [dB] False pos errors: 0 % MD [dB] 6 16 20 18 MS [dB] False neg errors: 0 % 7.5 6.9 19.4 19.8 [%] Test duration: 02:04 13 16 23 20 17 19 95..100 Fovea: Off 18 19 23 23 23 21 19 19 6 16 20 18 83..94 71..82 17 19 18 24 29 24 19 19 18 19 13 16 23 20 17 19 59..70 NOT COMPARABLE 18 19 23 23 23 21 19 19 47..58 35..46 17 13 18 16 20 24 24 21 26 28 28 27 22 23 16 9 17 11 17 19 18 24 29 24 19 19 18 19 23..34 18 19 20 23 22 21 11 1 5 13 11..22 17 18 20 24 26 28 22 16 17 18 20 18 8 5 2 0 30 30 0..10 13 16 24 21 28 27 23 9 11 15 3 2 Values Normative 18 19 20 23 22 21 11 1 5 13 11.8 20.6 15.6 5 3 7.1 Comparison [dB] Corrected comparisons [dB] Defect curve OCTOPUS 600 Database 18 20 15 18 3 8 2 5 <0 2 <0 <0 <0 0 <0 <0 <0 <0 18 8 + 5 11 + + + 1 Rank 74 12 10 + 6 9 6 5 + + + + + -5 5 14 18 17 5 3 <0 <0 <0 <0 8 8 5 5 5 6 8 7 + + + + + + + + 0 5% 10 16 20 21 14 21 9 8 10 + + 5 10 9 9 7 + + + + + + + + + + Defect (dB) 5 17 19 22 22 23 22 21 20 9 9 9 5 5 + 7 11 9 + + + + + + + + + 95% -18 -8 -4 -5 -11 -1 3 2 10 13 11 5 9 + + 6 18 16 6 + + + + + + 11 9 16 21 25 20 25 21 24 21 23 17 -12 -10 -3 -6 -9 -6 -5 -3 4 1 -2 1 15 8 8 8 5 7 8 18 27 22 14 + + + + + + 11 20 15 7 -8 -8 -5 -5 -5 -6 -8 -7 -1 -1 3 2 2 1 -1 0 20 18 20 23 24 25 26 23 16 18 8 7 10 21 23 26 27 + + + 14 16 19 20 -9 -8 -10 -4 0 -5 -10 -9 -9 -7 -2 -1 -3 3 7 2 -3 -2 -2 0 25 18 23 26 23 27 29 22 1 6 11 24 25 + 17 18 -9 -9 -9 -5 -5 -3 -7 -11 -9 -2 -2 -2 2 2 4 0 -4 -2 Diffuse defect [dB]: 7.0 21 23 14 16 21 23 26 26 22 21 11 1 0 -13 -11 -5 -9 -3 -4 -6 -18 -16 -6 -4 2 -2 4 3 1 -11 -9 -8 -8 -8 -5 -7 -8 -18 -27 -22 -14 -1 -1 -1 2 0 -1 -11 -20 -15 -7 MD -10.6 Probabilities Corrected probabilities 23 23 18 8 5 2 -8 -7 -10 -21 -23 -26 -27 -1 0 -3 -14 -16 -19 -20 PSD 7.8 15 4 2 -11 -24 -25 -4 -17 -18 5 3 -21 -23 -14 -16 Total deviation Pattern deviation [%] P>5 P<5 P<2 P<1 NOT COMPARABLE P < 0,5 30° Programs: 32 Standard White/White / TOP Questions / repetitions: 74 / 0 MS [dB]: 15.5 Parameters: 31.4 / 4000 asb III 100 ms Duration: 02:04 Catch trials: 0/4 (0%) +, 0/4 (0%) - RF: 0.0 MD [< 2.0 dB]: 11.6 sLV [< 2.5 dB]: 7.8 Threshold Normative Refraction S/C/A: VA [m]: [%] Pupil [mm]: IOP [mmHg]: <5 NV: T12 V2.1 Values Database <2 <1 Comment: < 0.5 OCTOPUS 900 SN3704 OCTOPUS® EyeSuite™ Static perimetry, V3.5.0 OCTOPUS 900, SN 3704, V 2.3.1 / 3.6.0 HFA II FIGURE 12-2 All recent Octopus perimeter models can import data from other Octopus models and from the HFA II perim- eter. Because the raw data is imported (i.e., the sensitivity thresholds, reliability indices and general test parameters) and the Octopus models that allow data import contain device-specific normative databases for all other models, the existing data is treated as a new measurement. Consequently, all representations and printouts available on an Octopus perimeter are avail- able, including the Octopus HFA-style (middle), the Octopus 7-in-1 printout (right), the Cluster Analysis and the Polar Analysis (not shown in this example of a retinal detachment case) and any trend analysis (not shown). 240 Chapter 12 | Transitioning to a different perimeter model To ensure a seamless transition, Octopus perimeters can treat the existing data like any new measurement import the measured sensitivity thresholds, reliability and display it in exactly the same format as shown in indices and general test parameters, including infor- FIG 12-2. Potential differences in de initions of represen- mation as to which perimeter model the data is coming tations and indices used are thus eliminated and pro- from. The imported measured sensitivity thresholds gression of visual ield data can be assessed as shown in are then compared to the relevant normative database FIG 12-3. In addition, this approach offers the advantage as described in the previous section (e.g., if importing that data taken years ago can be viewed with the latest existing data from an Octopus 300 into an Octopus 900, analysis tools (e.g., Cluster Trend Analysis). the measured sensitivity thresholds are compared to the Octopus 300 normative database). Because all Octo- For full transparency, the device from which a mea- pus representations are calculated from the measured surement stems is clearly marked on each visual ield sensitivity thresholds (SEE FIG 7-1), by comparing them test and assigned a distinct symbol in the global trend to device-speci ic normative databases, the new device analysis. IMPORT OF EXISTING DATA FROM AN HFA TO AN OCTOPUS PERIMETER To ensure that existing data collected on an HFA can thresholds and the HFA normative database, thus largely be used after a transition to an Octopus perimeter, all eliminating device-speci ic differences. Because raw data recent Octopus perimeter models allow import of elec- (i.e., sensitivity thresholds) are imported, the Octopus tronically stored data from an HFA II. This includes the perimeter can treat the existing data like any new mea- sensitivity thresholds, general test parameters, perim- surement and display it in exactly the same format as eter model from which the data stems (HFA II), as well shown in FIG 12-2. as reliability indices. To largely eliminate the in luence of any device-related differences at the level of sensitivity To assess visual ield progression, it is important to be losses, each Octopus perimeter also contains a normative able to use the existing HFA data imported into an Octo- database for the HFA II perimeter. The sensitivity losses pus perimeter and the new measurements in the same (i.e., Total Deviation on the HFA-style printout and the trend analysis. This is possible as long as comparable test Comparisons on the Octopus-style printout) of the HFA parameters (i.e., same stimulus type, same test pattern) data are then calculated from the imported sensitivity are used. FIG 12-4 provides an example. General aspects of transitioning 241 OCTOPUS PERIMETERS CAN JOINTLY DISPLAY DATA FROM ANY OCTOPUS PERIMETER IN A TREND ANALYSIS MD TREND ANALYSIS MD Mean defect O’123 0 O’300 15 Octopus 123 Octopus 300 25 2006 2007 2008 2009 2010 2011 2012 2013 2014 SERIES OF VISUAL FIELDS OCTOPUS 123 OCTOPUS 300 FIGURE 12-3 All Octopus perimeters allow import of existing patient data to ensure data continuity. The measured sensitivity thresholds are imported and compared to the appropriate device-specific normative database. The data can then be displayed in any Octopus format. In the example above, a glaucoma patient with an inferior arcuate defect has been tested on an Octopus 123 perimeter (unfilled triangle) from 2006 to 2009 using Standard Automated Perimetry (SAP) with a G test pattern. In 2010, the clinic transitioned to an Octopus 300 (filled triangle) and continued testing the patient with the same test parameters. The data of both devices can be used in the same Global Trend Analysis to monitor progression. Note that this patient shows typical levels of fluctuation both before and after the transition. 242 Chapter 12 | Transitioning to a different perimeter model OCTOPUS PERIMETERS CAN JOINTLY DISPLAY HFA AND OCTOPUS DATA IN A TREND ANALYSIS MD Mean defect O’900 0 HFA II 15 HFA II Octopus 900 25 2006 2007 2008 2009 2010 2011 2012 2013 2014 SERIES OF VISUAL FIELDS HFA II OCTOPUS 900 FIGURE 12-4 In this example, a glaucoma patient with a superior arcuate defect has been tested on an HFA II perimeter from 2006 to 2009 using SAP with a 24-2 test pattern. In 2010, the clinic transitioned to an Octopus 900 and continued testing the patient with the same test parameters. The HFA II data can be imported into the Octopus 900 perimeter and the data of both devices can be used in the same Global Trend Analysis because of the device-specific normative databases used by the Octopus perimeters. Specific aspects related to transitioning from the Humphrey Field Analyzer 243 MANAGING PATIENT-RELATED FLUCTUATION As described in the sections above, Octopus perimeters learning effects, it is thus good practice for technicians to offer several features that minimize the impact of the take some examinations with the new device themselves transition between different perimeter models. Never- and to make sure to include noticeable differences in the theless, one may still observe differences in some but not patient instructions. Furthermore, running a practice all patients after switching devices. test with a patient on a new device is also helpful. This can be explained in part by the fact that patient-related In addition to learning effects, some patients may show luctuation is always present in perimetry and should personal preferences for one perimeter model over the therefore also be expected during the transition from one other. For example, while cupola perimeters such as perimeter to another. During the transition, patient- the Octopus 900 or the HFA II need to be operated under related luctuation can be associated with the transition dim room light conditions, closed 30° perimeters like the itself or it may be independent of it. Chapter 3 provides Octopus 600 or 300 may also be operated at daylight many practical tips on how to minimize patient-related levels. Different ambient light conditions may in luence luctuation. The transition between perimeter models the patient’s performance during the perimetric test, itself may increase the amount of patient-related luc- with dim light conditions enhancing concentration in tuation in some but not all patients. Because the design some, while making others sleepy and less alert. While and working conditions of different perimeter models personal preferences cannot be eliminated, typically the vary, some patients may show learning effects during the impact on the visual ield test results is within expected initial tests on the new device (for more information on levels of luctuation. learning effects, see FIG 3-12). To minimize the impact of SPECIFIC ASPECTS RELATED TO TRANSITIONING FROM THE HUMPHREY FIELD ANALYZER SELECTION OF TEST PARAMETERS As shown in FIGURES 12-1 and 12-4, visual ield tests tak- TABLE 12-1 provides an overview of the most common en on either the HFA or the Octopus perimeter result in choices of Octopus test patterns and strategies following comparable test results that can be used equally well a transition from an HFA perimeter. More detailed infor- for clinical decision-making.²-⁴ However, because both mation on all available Octopus test patterns is presented perimeter brands use their own test patterns and strate- in Chapter 5 and more details on the available test strate- gies to perform visual ield testing, one may not intuitively gies are presented in Chapter 6. know which ones to choose on an Octopus perimeter after a transition from an HFA perimeter. 244 Chapter 12 | Transitioning to a different perimeter model COMMON CHOICES OF TEST PATTERNS AND STRATEGIES IN HFA TABLE 12-1 AND OCTOPUS PERIMETERS TEST PATTERN TEST STRATEGY HFA OCTOPUS HFA OCTOPUS 30° 24-2, 30-2 24-2, 30-2, G SITA Fast TOP (Glaucoma/General) 24-2, 30-2 24-2, 30-2, G SITA Standard Dynamic 10° 10-2 10-2, M SITA Fast TOP (Macula, constricted field, hydroxy-chloroquine retinopathy) 10-2 10-2, M SITA Standard Dynamic FULL FIELD (Threshold) 60-4 07 SITA Standard Dynamic FULL FIELD (Screening) 81FF, 120FF, 07 3-zone 2LT 135 FF DRIVING ABILITY Esterman test Esterman test While EyeSuite Progression Analysis (see Chapter 9) can in the patient’s existing visual ield tests. For this reason, be performed on tests that use different test strategies, it Octopus perimeters provide the most commonly used requires the same test pattern and the same overall test HFA test patterns, namely the 24-2, 30-2 (FIG 5-4) and conditions to be used for all tests included in the visu- 10-2 (FIG 5-10). If any other HFA pattern not available on al ield series. If progression analyses are needed when an Octopus perimeter is needed, it is possible to create transitioning from an HFA perimeter to an Octopus perim- that test pattern using the Custom Test function available eter, it is thus best to select the same test pattern used on some Octopus models. INTERPRETATION OF A SINGLE VISUAL FIELD Both Octopus and HFA perimeters have developed their imported on an Octopus perimeter can be displayed using own brand-speci ic visual ield representations. While both the Octopus-style as well as the HFA-style printout. the underlying reasoning and de initions are comparable, they have different names, a different graphical style and HFA and corresponding Octopus representations are the formulas used in their calculation can vary.⁵,⁶ very similar. Once the Octopus-speci ic terminology of each representation becomes familiar, those familiar with To facilitate the transition from an HFA to an Octopus the HFA terminology can easily interpret the results. FIG perimeter with minimal training in visual ield interpre- 12-5 presents a side-by-side comparison of all available tation, all Octopus perimeters offer an HFA mode. In this HFA and Octopus representations and also highlights mode, an HFA-style printout is available in which the differences relevant for clinical interpretation. Guidance single ield representations and indices are named and on how to transition from the Glaucoma Hemi ield Test calculated based on the de initions used in the original HFA (GHT) to the Defect Curve is provided in BOX 12B. printout. FIG 12-2 shows that any visual ield test taken or Specific aspects related to transitioning from the Humphrey Field Analyzer 245 SIMILARITIES AND DIFFERENCES BETWEEN HFA AND CORRESPONDING OCTOPUS REPRESENTATIONS HFA REPRESENTATION CORRESPONDING COMMENTS OCTOPUS REPRESENTATION THRESHOLD VALUES VALUES Both perimeters display the measured sensitivity thresholds. 17 20 20 18 Octopus perimeters display absolute 15 21 22 24 23 27 defects (i.e., sensitivity thresholds 0 5 22 20 22 26 25 27 28 dB) with a “@” sign (see FIG 7-2 and 7-3). 8 9 16 14 29 27 27 30 29 27 15 13 15 12 30 31 31 28 28 26 26 23 27 28 28 30 31 30 29 19 23 28 28 28 29 29 30 29 26 21 24 28 29 30 30 28 28 23 24 25 29 29 28 25 21 26 28 GRAYSCALE GRAYSCALE (COMPARISON) Both perimeters use an interpolated graphical map to assess magnitude and shape of defects. The HFA Grayscale is based on sensitivity thresholds (Threshold Values) in dB, thus it is influenced by both patient-age and eccentricity of test location. The Octopus Grayscale (Comparison) is based on sensitivity loss in %, thus its interpretation is independent of patient-age and eccentricity of test locations (see FIG 7-7 and 8-18). TOTAL DEVIATION COMPARISON Both perimeters display sensitivity NUMERICAL MAP loss (i.e., deviation from age-corrected 6 + + 6 normal values), but they use opposite signs. 10 5 5 + + + 21 6 9 7 + + + + Octopus perimeters display sensitivity 18 19 14 17 + + + + + + loss < 5 dB with a “+” sign (see FIG 7-6 and 8-18). 12 16 16 20 + + + + + + 6 + + 5 + + + + 7 6 + + + + + + + + 7 5 + + + + + + 5 5 + + + + + 6 + + FIGURE 12-5 Side-by-side comparison of the HFA Single Field Analysis and the Octopus 7-in-1 printout of the same visual field test that was taken on an HFA II perimeter and then imported into an Octopus perimeter. Many representations in the two printouts are based on the same principles, but use different names. It should be noted that while differences between the results of the two perimeters are present, they are typically very small and do not alter the clinical interpretation of the case. Small differences in the definitions used between the perimeters are highlighted in the comment column. 246 Chapter 12 | Transitioning to a different perimeter model PATTERN DEVIATION CORRECTED COMPARISON Both perimeters display local NUMERICAL MAP sensitivity loss (i.e., deviation from + + + + age-corrected normal values with a correction applied to eliminate any 8 + + + + + influence of diffuse loss). 19 + 7 5 + + + + 16 17 12 15 + + + + + + Octopus and HFA perimeters use opposite signs. 10 14 14 18 + + + + + + + + + 5 + + + + Octopus perimeters display local 6 + + + + + + + + + sensitivity loss < 5 dB with a “+” sign (see FIG 7-16, 7-17 and 8-18). 5 + + + + + + + + + + + + + + 5 + + TOTAL DEVIATION PROBABILITIES Octopus and HFA perimeters show PROBABILITY MAP the same levels of probabilities using similar symbols. Octopus perimeters use the following symbols (see FIG 7-10, 8-14 and 8-15). P > 5% P < 5% P < 2% P < 1% P < 0.5 % PATTERN DEVIATION CORRECTED PROBABILITIES Octopus and HFA perimeters show PROBABILITY MAP the same levels of probabilities using similar symbols. Octopus perimeters use the following symbols (see FIG 7-18, 8-14 and 8-15). P > 5% P < 5% P < 2% P < 1% P < 0.5 % Specific aspects related to transitioning from the Humphrey Field Analyzer 247 MD MD HFA and Octopus perimeters use opposite signs. MEAN DEVIATION MEAN DEFECT HFA perimeters put extra weight on central visual field locations. -4.66 dB 4.4 dB Octopus perimeters weigh each location equally,5,6 as the standard G pattern has higher density of central test locations (see TABLE 7-1 and FIG 8-26). PSD sLV HFA perimeters put extra weight on central visual field locations. PATTERN STANDARD SQUARE ROOT OF LOSS DEVIATION VARIANCE Octopus perimeters weigh each location equally, as the standard G pattern has higher density of central 6.11 dB 5.3 dB test locations (see TABLE 7-1 and FIG 8-27). GHT DEFECT CURVE Both GHT and Defect Curve provide 1 Rank 74 information on the overall status of GLAUCOMA HEMIFIELD TEST -5 the visual field, though the methods differ. 0 5% 5 For more details, see BOX 12B. Defect (dB) Outside normal limits 10 95% 15 20 25 VFI MD Both VFI and MD are measures of the overall visual field loss, and give VISUAL FIELD INDEX MEAN DEFECT comparable results in patients with MD values larger than ±5 dB. 90% 4.4 dB VFI is expressed as a percentage of normal function, ranges from 100% to 0 % and is not influenced by diffuse visual field loss. MD is expressed in dB, ranges from 0 up to 25 dB and is affected by diffuse visual field loss but is also more sensitive in detecting early visual field loss. 7 FALSE POS ERRORS FALSE POSITIVE ANSWERS Both HFA and Octopus perimeters display the percentage of false positive errors (see FIG 7-21). 12% 1/8 (12%) + Octopus perimeters additionally present the absolute numbers of false positive answers and the total number of positive catch trials. 248 Chapter 12 | Transitioning to a different perimeter model FALSE NEG ERRORS FALSE NEGATIVE ANSWERS Both HFA and Octopus perimeters display the percentage of false negative errors (see FIG 7-22). 12% 1/8 (12%) - Octopus perimeters additionally present the absolute numbers of false negative answers and the total number of negative catch trials. FIXATION LOSSES NOT AVAILABLE HFA perimeters use the Heijl-Krakau method to determine the percentage of fixation losses. 0/12 Octopus perimeters prevent fixation losses by using Fixation Control, in which the test is interrupted when adequate fixation is not maintained (see FIG 3-11). GAZE TRACKER NOT AVAILABLE HFA perimeters record eye movements using the gaze tracker. Octopus perimeters prevent fixation losses by using Fixation Control, in which the test is interrupted when adequate fixation is not maintained (see FIG 3-11). BOX 12B RELATIONSHIP BETWEEN THE GLAUCOMA HEMIFIELD TEST (GHT) AND THE DEFECT CURVE The Glaucoma Hemi ield Test (GHT) is an intuitive text-based index that provides information about the overall status of the visual ield and classi ies the visual ield results as “Within normal limits”, “Border- line”, “Outside normal limits”, “General reduction of sensitivity” and “Abnormally high sensitivity”. Its design is based on the asymmetry of sensitivity thresholds for the superior and inferior arcuate nerve iber bundle regions. It therefore determines statistically signi icant differences between two corre- sponding visual ield clusters divided by the horizontal midline. In Octopus perimeters, the Defect Curve is used to determine overall visual ield status. And while it is based on different principles, it provides similar information about whether visual ields are normal or whether local or diffuse defects are present. The table below summarizes some rules of thumb on how to read the Defect Curve to obtain information that is comparable to the GHT. For more details on the Defect Curve, refer to FIG 7-11 and 8-10. Specific aspects related to transitioning from the Humphrey Field Analyzer 249 GHT DEFECT CURVE DEFECT CURVE INTERPRETATION Rank 1 59 WITHIN NORMAL LIMITS NORMAL -5 0 5% Defect Curve within normal 5 band Defect (dB) 10 95% 15 20 25 Rank 1 52 BORDERLINE BORDERLINE -5 0 5% Defect Curve along/slightly 5 below normal band Defect (dB) 10 95% OR 15 Defect Curve within normal 20 band, but with characteristic drop on the right (not shown) 25 Rank 1 59 OUTSIDE NORMAL LIMITS LOCAL DEFECT -5 0 5% Drop of Defect Curve on 5 the right Defect (dB) 10 95% 15 20 25 Rank 1 59 GENERAL REDUCTION DIFFUSE DEFECT -5 OF SENSITIVITY 0 5% Parallel downward shift of 5 Defect Curve Defect (dB) 10 95% 15 20 25 1 Rank 59 ABNORMALLY HIGH TRIGGER-HAPPY -5 SENSITIVITY 0 5% Steep rise of Defect Curve 5 on the left Defect (dB) 10 95% 15 20 25 250 Chapter 12 | Transitioning to a different perimeter model INTERPRETATION OF VISUAL FIELD PROGRESSION Both HFA and Octopus perimeters offer methods for change and the rate of change. In addition, both HFA assessing visual field progression. FIG 12-6 presents a and Octopus perimeters provide tools to determine side-by-side comparison of all available HFA and Octopus whether there is local progression beyond what is ap- progression analyses and also highlights differences parent in the MD trend analysis and where the change relevant for clinical interpretation. For more detailed happens. The Octopus also offers a method for identify- information on EyeSuite Progression Analysis, refer to ing diffuse progression independently. Furthermore, Chapter 9. to facilitate the combined evaluation of structural and functional progression in glaucoma, Octopus perimeters To judge whether a visual ield series is stable or pro- offer a trend procedure, the Polar Trend Analysis, which gressing, both HFA and Octopus perimeters use a trend facilitates inding a relationship between structural and analysis approach and determine both signi icance of functional losses. OVERVIEW OF PROGRESSION TOOLS AVAILABLE ON HFA AND OCTOPUS PERIMETERS HFA REPRESENTATION OCTOPUS REPRESENTATION COMMENTS GUIDED PROGRESSION EYESUITE PROGRESSION GPA uses both trend analysis ANALYSIS (GPA) ANALYSIS and point-wise event analysis, which requires two reliable baseline tests. EyeSuite Progression Analysis uses trend analysis. GPA TREND ANALYSIS MD TREND ANALYSIS Both perimeters use trend analysis to determine signi- ficance and rate of change. OVERALL PROGRESSION VFI MD Mean defect 100% HFA uses the Visual Field 0 80% Index (VFI), which typically 60% ranges from 100% to 0%. Significant change is shown 40% in text. 20% 15 Octopus uses the Mean 0% 74 AGE 84 94 Defect (MD), which typically ranges from 0 to 25 dB. Rate of progression: -2.6 ± 1.8% Significant worsening is (95% confidence) 25 2015 shown with red downward Slope significant at P < 5% Slope: 0.5 dB / Yr (p<0.5%) arrows (see FIG 9-6). FIGURE 12-6 Side-by-side comparison of the HFA and the Octopus progression analyses of the same visual field series that was taken on an HFA II perimeter and then imported into an Octopus perimeter. Some analyses identify similar aspects of progression, such as whether there is progression and where localized progression occurs, but use a different approach. Further, the Octopus perimeter offers analyses for identifying diffuse progression and providing guidance on where to look for structural progression. Differences in the methods used between the perimeters are presented in the comment column. Specific aspects related to transitioning from the Humphrey Field Analyzer 251 GPA EVENT ANALYSIS (CORRECTED) CLUSTER HFA uses point-wise event TREND ANALYSIS analysis to assess local Deviation from Baseline progression and requires two trustworthy baseline tests. GPA Alert displays likelihood 0.8 of progression as text. 0.8 Octopus uses trend analysis 0.2 1.0 0.1 to determine significance 0.1 and rate of change of the 0.2 0.2 variables (see FIG 9-8) LOCAL PROGRESSION 0.2 • (Corrected) Cluster MD 0.7 (average defects of 10 clusters following RNFL distribution on the retina, see FIG 9-11) Progression Analysis LD TREND ANALYSIS • LD (Local Defect, see BOX 7D) LD Local defect LOCAL PROGRESSION 0 • sLV (Octopus equivalent of PSD, see FIG 8-27) Trend analysis is relatively robust to outliers. Furthermore, individual visual fields can be easily excluded from the trend analysis in the Octopus perimeter by simply clicking 15 on a given test. 2015 Slope: 0.5 dB / Yr (p<0.5%) POSSIBLE PROGRESSION sLV TREND ANALYSIS sLV Loss variance LOCAL PROGRESSION 0 5 2015 Slope: 0.4 dB / Yr (p<5%) 252 Chapter 12 | Transitioning to a different perimeter model NOT AVAILABLE DD TREND ANALYSIS Octopus uses trend analysis to determine significance DD Diffuse defect and rate of change of the DIFFUSE PROGRESSION variable DD (Diffuse Defect, 0 see BOX 7C). 25 2015 Slope: 0.1 dB / Yr NOT AVAILABLE POLAR TREND ANALYSIS Octopus uses Polar Trend STRUCTURAL PROGRESSION Analysis to show point-wise progression per visual field WHERE TO LOOK FOR location projected onto the optic disc as guidance on where to look for structural progression (see FIG 9-14). S 30 20 10 N T [dB] I References 253 REFERENCES 1. Anderson DR, Feuer WJ, Alward WL, Skuta GL. Threshold equivalence between perimeters. Am J Ophthalmol. 1989;107:493-505. 2. King AJ, Taguri A, Wadood AC, Azuara-Blanco A. Comparison of two fast strategies, SITA Fast and TOP, for the assessment of visual ields in glaucoma patients. Graefes Arch Clin Exp Ophthalmol. 2002;240:481-487. 3. Pierre-Filho Pde T, Schimiti RB, de Vasconcellos JP, Costa VP. Sensitivity and speci icity of frequency-doubling technology, tendency-oriented perimetry, SITA Standard and SITA Fast perimetry in perimetrically inexperienced individuals. Acta Ophthalmol Scand. 2006;84:345-350. 4. Wadood AC, Azuara-Blanco A, Aspinall P, Taguri A, King AJ. Sensitivity and speci icity of frequency-doubling technology, tendency-oriented perimetry, and Humphrey Swedish interactive threshold algorithm-fast perimetry in a glaucoma practice. Am J Ophthalmol. 2002;133:327-332. 5. Funkhouser A, Fankhauser F. The effects of weighting the "mean defect" visual ield index according to threshold variability in the central and midperipheral visual ield. Graefes Arch Clin Exp Ophthalmol. 1991;229:228-231. 6. Funkhouser AT, Fankhauser F. A comparison of the mean defect and mean deviation indices within the central 28 degrees of the glaucomatous visual ield. Jpn J Ophthalmol. 1990;34:414-420. 7. Artes PH, O'Leary N, Hutchison DM, et al. Properties of the statpac visual ield index. Invest Ophthalmol Vis Sci. 2011;52:4030-4038. 254 255 CHAPTER 13 CLINICAL CASES INTRODUCTION The previous chapters of this book have systematically relevant clinical information. Background information on presented various aspects of visual field testing and the patient’s history as well as other diagnostic results interpretation. To conclude, visual ield interpretation such as visual acuity, IOP, fundus images, OCT scans and is now put into a clinical context. In this chapter, 23 clin- MRIs which are relevant for clinical decision making, are ical cases are presented that show visual ields or visual shown. In all examples, visual acuity is expressed in dec- ield series of patients with glaucoma, neuro-ophthal- imal units for uniformity, but the Octopus allows users to mic disorder and retinal disease. The selected cases are select different units when performing the test. In each model cases. They present typical defect patterns of the case, key diagnostic indings leading to disease diagnosis disease rather than unusual cases and are reliable, free are presented and summarized. of artifacts and can be fully trusted. An overview of all available cases is presented on the To link visual ield interpretation to the clinical situation, next page. the visual ield results are presented in addition to other 256 Chapter 13 | Clinical cases GLAUCOMA – SINGLE FIELD NEUROLOGICAL DISEASES 1. Very early stage glaucoma 16. Cerebral infarction (normal tension glaucoma) (bilateral) 2. Early stage glaucoma 17. Leber hereditary optic neuropathy (normal tension glaucoma) (bilateral) 3. Early stage glaucoma 18. Bilateral optic neuritis (primary open-angle glaucoma) (multiple sclerosis) 4. Early stage glaucoma 19. Tuberculum sellae meningioma (with cataract) (bilateral) 5. Early stage glaucoma (normal tension glaucoma) 6. Early stage glaucoma RETINAL DISEASES (primary open-angle glaucoma) 20. Age-related macular degeneration 7. Moderate glaucoma 21. Branch central retinal artery occlusion (normal tension glaucoma) 22. Macular hole 8. Moderate glaucoma 23. Branch central retinal vein occlusion (primary open-angle glaucoma) 9. Late stage glaucoma (normal tension glaucoma) GLAUCOMA – TREND 10. Early to moderate glaucoma (normal tension glaucoma) 11. Early to moderate glaucoma (primary open-angle glaucoma) 12. Early to moderate glaucoma (primary open-angle glaucoma) 13. Early to moderate glaucoma (normal tension glaucoma) 14. Early to moderate glaucoma (primary open-angle glaucoma) 15. End-stage glaucoma (exfolitative glaucoma) Glaucoma | Single field 257 1 VERY EARLY STAGE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 57-year-old female, no family history • Patient reported decreased visual acuity in both eyes and discomfort in left eye IOP/VA corr • 15 mmHg/ 1.2 – 5.25 (sph) FUNDUS • C/D = 0.9 • Rim thinning at 6 to 11 o'clock position • Optic disc hemorrhage and narrow slit-like RNFL defect at 11 o'clock position • Temporal alpha zone and beta zone peripapillary chorioretinal atrophy (PPA) Demo Jane, 1947/01/01 (57yrs) Right eye (OD) / 2004/11/18 / 12:01:27 Seven-in-One Grayscale (CO) Values 22 25 23 26 -1.6 -1.9 27.4 27.9 95%..100% 28 28 27 27 28 29 83%...94% 71%...82% 28 28 30 28 28 29 28 29 59%...70% 26 29 30 29 30 31 29 29 26 28 47%...58% 21 26 29 30 30 31 32 28 27 35%...46% 23%...34% 26 28 28 26 33 30 31 29 29 11%...22% 26 28 29 29 29 29 30 31 29 31 0%...10% 28 29 27 31 31 27 27 30 27 29 28 29 28 28 -1.1 -1.3 27.9 2 8 .8 26 28 27 28 Comparison Corrected comparisons Defect curve + + + + + + + + Rank 1 74 + + + + + + + + + + + + -5 + + + + + + + + + + + + + + + + 0 5% + + + + + + + + + + + + + + + + + + + + 5 Defect Curve Defect (dB) + + + + + + + + + 5 + + + + + + + + 95% 10 in normal + + + + + + + + + + + + 5 + + + + + 15 range + + + + + + + + + + + + + + + + + + + + 20 + + + + + + + + + + + + + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + + + + + Probabilities Corrected probabilities All test locations at P > 5% P>5 P<5 P<2 P<1 P < 0,5 Programs: 32 Standard White/White / Normal Questions / repetitions: 520 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 16:41 MS [dB]: 28.0 Catch trials: 0/26 (0%) +, 1/26 (4%) - RF: 1.9 MD [< 2.0 dB]: -1.5 Refraction S/C/A: -3.25// VA: 1.0 sLV [< 2.5 dB]: 1.8 Pupil [mm]: IOP [mmHg]: CsLV [dB]: 1.5 SF [dB]: 1.6 Comment: Classification: OCTOPUS® OCTOPUS 101 • No visual ield loss • Fundus indings show changes indicative of very early glaucoma including neuroretinal rim loss, optic disc hemorrhage, and RNFL loss 258 Chapter 13 | Clinical cases 2 EARLY STAGE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 53-year-old female, no family history • Optic nerve cupping observed during unrelated emergency eye surgery IOP/VA corr • 12 mmHg/ 1.2 + 0.25 (sph) FUNDUS • C/D = 0.8 • Rim thinning and RNFL loss at 5 to 6 o'clock position Demo Jane, 1942/01/01 (53yrs) Left eye (OS) / 1996/06/21 / 14:24:40 Seven-in-One Grayscale (CO) Values 26.2 23 21 20.0 0.7 6.9 95%..100% 23 24 21 24 24 14 83%...94% 71%...82% 26 27 28 27 18 21 59%...70% 28 25 27 23 24 47%...58% 28 26 16 24 17 10 35%...46% 32 32 23 17 34 30 34 33 32 23%...34% 29 30 33 31 28 25 11%...22% 27 30 30 31 32 0%...10% 28 25 31 27 30 27 27 27 28 28 27 23 -0.3 -0.9 28.5 28.6 27 27 Comparison Corrected comparisons Defect curve Rank + + + + 1 74 + + + + + 11 + + + + + 10 -5 + + + + 0 + + 9 + + + 9 + 5% + + + 6 5 + + + 6 5 5 Defect (dB) + + 14 5 10 15 + + 14 5 10 15 + + 9 13 + + 9 13 10 95% + + + + + + + + + + + + + + + + + + + + + + 15 + + + + + + + + + + + + + + + + + + 20 + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities Cluster of abnormal locations with P < 0.5 % P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 481 / 1 30° Parameters: 4 / 1000 asb III 100 ms Duration: 16:49 MS [dB]: 25.9 Catch trials: 0/24 (0%) +, 1/24 (4%) - RF: 2.0 MD [< 2.0 dB]: 1.6 Refraction S/C/A: // VA: 1.2 sLV [< 2.5 dB]: 4.2 Pupil [mm]: 5.3 IOP [mmHg]: 12 CsLV [dB]: 3.9 SF [dB]: 1.7 Comment: NTG Classification: OCTOPUS® OCTOPUS 101 • Mild superior nasal step and mild superior paracentral scotoma • Spatial relationship between visual ield loss and both rim thinning and RNFL loss in fundus photo Glaucoma | Single field 259 3 EARLY STAGE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 56-year-old female, her brother has POAG • Patient visited clinic to rule out glaucoma because of her family history IOP/VA corr • 24 mmHg /1.0 - 3.25 (sph) FUNDUS • Inferior RNFL defects OCT • RNFL and ganglion cell loss inferotemporally at 7 to 8 o’clock position Demo Jane, 1958/01/01 (56yrs) Right eye (OD) / 2015/01/09 / 01:31:08 Four-in-One Grayscale (CO) Cluster analysis [dB] MD [dB] 2.2 2.5 Nasal step, 3.1 superior arcuate and superior + paracentral defect 5.4 + 7.4 + + + + + -1.0 -0.5 Defect curve Polar analysis Rank 1 59 -5 0 5% 5 Defect (dB) S 10 95% T N 10 20 30 RNFL loss [dB] I inferortemporally 15 20 Local defect (glaucoma) 25 Diffuse defect [dB]: 2.3 Structural damage suggested at 7 to 8 o'clock position 30° Programs: G Standard White/White / TOP Questions / repetitions: 72 / 0 MS [dB]: 26.4 Parameters: 31.4 / 4000 asb III 100 ms Duration: 02:18 Catch trials: 0/4 (0%) +, 0/4 (0%) - RF: 0.0 MD [< 2.0 dB]: 0.9 Refraction S/C/A: VA [m]: sLV [< 2.5 dB]: 4.5 Pupil [mm]: 3.0 IOP [mmHg]: NV: T31 V1.0 Comment: GC thinning inferortemporally OCTOPUS ® • Nasal step, superior arcuate and superior paracentral defect apparent in Cluster Analysis • Spatial relationship between visual ield loss (Polar Analysis suggests structural damage at 7 to 8 o'clock position) and inferotemporal structural loss (fundus photo, RNFL & GC thickness map) 260 Chapter 13 | Clinical cases 4 EARLY STAGE GLAUCOMA (WITH CATARACT) PATIENT • 71-year-old male, no family history • Patient reported defective vision in both eyes over the last 6 months and glare at night while crossing roads IOP/VA corr • 24 mmHg/ 0.7 + 1.75 (sph), - 1.25 (cyl) x 80° FUNDUS • Fundus image hazy due to cataract OCT • RNFL loss and ganglion cell loss at 5 to 6 o’clock position Demo John, 1944/01/01 (71yrs) Left eye (OS) / 2015/04/29 / 14:57:18 Four-in-One Grayscale (CO) Corrected cluster analysis [dB] Hazy due to MD [dB] 4.2 6.4 cataract 2.2 Nasal step + (local defect) + 7.4 + + + + + + 0.1 2.5 Defect curve Polar analysis 1 Diffuse defect Rank 59 -5 (cataract) 0 5% 5 Defect (dB) S 10 95% 30 20 10 Structural N T [dB] I damage 15 suggested at 5 to 6 o'clock Local defect position 20 (glaucoma) 25 Diffuse defect [dB]: 2.3 30° Programs: G Standard White/White / TOP Questions / repetitions: 74 / 0 MS [dB]: 23.0 Parameters: 31.4 / 4000 asb III 100 ms Duration: 02:25 Catch trials: 2/4 (50%) +, 0/4 (0%) - RF: 25.0 MD [< 2.0 dB]: 3.3 Refraction S/C/A: VA [m]: sLV [< 2.5 dB]: 3.4 Pupil [mm]: 3.7 IOP [mmHg]: NV: T31 V1.0 Comment: OCTOPUS® EyeSuite™ Static perimetry, V3.5.0 OCTOPUS 300 • Both diffuse defect (due to cataract) and local defect (due to glaucoma) in Defect Curve • Corrected Cluster Analysis (removing diffuse defect) shows superior nasal step • Spatial relationship between visual ield loss (Polar Analysis suggests structural damage at 5 to 6 o'clock position) and inferotemporal structural loss (fundus photo, RNFL & GC thickness map) Glaucoma | Single field 261 5 EARLY STAGE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 58-year-old female, father had glaucoma • Optic nerve cupping detected during routine medical visit IOP/VA corr • 16 mmHg/ 1.2 – 1.0 (sph), - 0.75 (cyl) x 80° FUNDUS • C/D = 0.9 • Rim thinning and wide RNFL loss at 5 to 6 o'clock position Demo Jane, 1944/01/01 (58yrs) Left eye (OS) / 2003/03/14 / 16:43:49 Seven-in-One Grayscale (CO) Values 19.8 20 21 23.2 6.9 3.4 95%..100% Absolute defect 21 21 22 23 24 23 83%...94% (sensitivity threshold 0 dB) 71%...82% 27 27 25 23 24 23 59%...70% 27 23 28 47%...58% 24 21 15 21 26 18 35%...46% 21 27 24 31 32 31 31 33 31 23%...34% 29 36 32 30 29 26 11%...22% 28 29 29 32 30 0%...10% 27 27 28 26 29 30 26 26 28 27 24 23 -0.7 -0.5 28.6 28.0 26 26 Comparison Corrected comparisons Defect curve + + + + 1 Rank 74 + + + + + + + + + + + + -5 + + + + 0 + + + + + + + + 5% + 5 + + 5 + 5 Defect (dB) + 8 15 7 + 7 + 8 14 7 + 6 8 + 7 + 8 + 7 + 10 95% + + + + + + + + + + + + + + + + + + + + + + 15 + + + + + + + + + + + + + + + + + + 20 + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 459 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 15:23 MS [dB]: 25.0 Catch trials: 3/23 (13%) +, 0/23 (0%) - RF: 6.5 MD [< 2.0 dB]: 2.2 Refraction S/C/A: +0.5/-0.75/80 VA: 1.2 sLV [< 2.5 dB]: 5.8 Pupil [mm]: 5.0 IOP [mmHg]: 13 CsLV [dB]: 5.5 SF [dB]: 2.8 Comment: Gla Classification: OCTOPUS® OCTOPUS 101 • Dense paracentral scotoma • Spatial relationship between visual ield loss and both rim thinning and RNFL loss in fundus photo 262 Chapter 13 | Clinical cases 6 EARLY STAGE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 55-year-old male, no family history • Patient reported decreased visual acuity and blurred vision IOP/VA corr • 23 mmHg/ 1.2 – 4.25 (sph), - 1.0 (cyl) x 180° FUNDUS • C/D = 0.8 • Small disc • Rim thinning at 5 to 6 o'clock position Demo John, 1951/01/01 (55yrs) Left eye (OS) / 2007/05/11 / 10:02:54 Seven-in-One Grayscale (CO) Values 25.6 24 23 11.7 1.2 15.1 95%..100% 27 26 25 23 19 15 83%...94% 71%...82% 17 18 29 21 6 6 59%...70% 27 21 23 3 9 47%...58% 28 31 10 3 5 4 35%...46% 30 33 32 3 34 23%...34% 31 33 31 31 30 31 32 30 30 29 11%...22% 29 31 30 31 30 0%...10% 29 30 28 27 31 31 29 31 30 30 24 28 -2.0 -1.4 30.1 29.0 30 29 Comparison Corrected comparisons Defect curve Rank + + + + 1 74 -5 + + + + 5 9 + + + + 6 10 10 10 10 10 0 5% + 6 21 19 + 7 22 20 + 7 6 26 20 + 8 7 26 21 5 Defect (dB) + + 20 26 22 21 + + 21 26 22 22 95% + + + 27 + + + 28 10 + + + + + + + + + + + + + + + + + + + + + + 15 + + + + + + + + + + Local 20 defect + + + + + + + + + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 447 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 13:43 MS [dB]: 24.2 Catch trials: 0/22 (0%) +, 0/23 (0%) - RF: 0.0 MD [< 2.0 dB]: 3.3 Refraction S/C/A: -2.75/-1.0/180 VA: 1.0 sLV [< 2.5 dB]: 8.7 Pupil [mm]: 6.7 IOP [mmHg]: CsLV [dB]: 8.6 Very reliable test SF [dB]: 1.5 Comment: Classification: OCTOPUS® OCTOPUS 101 • Visual ield shows superior arcuate defect • Spatial relationship between visual ield loss and rim thinning in fundus photo indicative of glaucoma Glaucoma | Single field 263 7 MODERATE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 57-year-old female, no family history • Patient reported decreased visual acuity in both eyes and discomfort in left eye IOP/VA corr • 16 mmHg/ 1.0 – 5.5 (sph) FUNDUS • C/D = 0.95 • Rim thinning at 12 to 6 o'clock position • Vein angulation and bayoneting at 12 and 6 o'clock position Demo Jane, 1947/01/01 (57yrs) Left eye (OS) / 2004/11/18 / 12:25:31 Seven-in-One Grayscale (CO) Values 24 17 20 23 1.9 14.1 24.1 11.7 95%..100% 29 27 25 23 22 15 83%...94% 71%...82% 27 27 25 25 22 20 13 11 59%...70% 28 28 27 16 2 6 3 2 47%...58% 28 28 24 32 29 1 4 9 4 35%...46% 23%...34% 28 28 27 31 32 26 11%...22% 28 27 29 11 24 17 1 0%...10% 27 28 28 18 3 17 14 26 29 29 22 27 26 1.6 14.5 25.9 12.3 27 27 26 26 Comparison Corrected comparisons Defect curve + 6 + + + 5 + + Rank 1 74 + + + + + 9 + + + + + 8 -5 + + + + 5 6 13 13 + + + + 5 6 12 13 0 5% + + + 11 26 22 23 22 + + + 11 26 22 22 22 5 Defect (dB) + + 5 + + 29 24 18 21 + + + + + 28 24 17 20 95% 10 + + + + + + + + + + + + 15 + + + 18 6 12 24 + + + 17 5 11 23 20 + + + 10 24 10 11 + + + 10 24 9 10 25 + + + 5 + + + + + + + + Diffuse defect [dB]: -1.1 + + + + + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Large sLV shows Programs: Parameters: 32 Standard White/White / Normal 4 / 1000 asb III 100 ms Questions / repetitions: Duration: 642 / 0 21:13 severe 30° local defect MS [dB]: 18.4 Catch trials: 0/32 (0%) +, 2/33 (6%) - RF: 3.0 MD [< 2.0 dB]: 8.2 Refraction S/C/A: -3.25// VA: 1.0 sLV [< 2.5 dB]: 10.9 Pupil [mm]: 5.8 IOP [mmHg]: CsLV [dB]: 11.0 SF [dB]: 2.2 Comment: Classification: OCTOPUS® OCTOPUS 101 • Dense partial double arcuate visual ield defect • Spatial relationship between visual ield loss and both rim thinning and vein bending in fundus photo 264 Chapter 13 | Clinical cases 8 MODERATE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 52-year-old female, no family history • Patient diagnosed with glaucoma during medical check-up IOP/VA corr • 20 mmHg/ 1.2 – 4.0 (sph), - 0.25 (cyl) x 180° FUNDUS • C/D = 0.9 • Rim thinning at 6 to 8 o'clock position and notching at 11 o’clock position • Large RNFL loss at 6 to 8 o'clock position and small RNFL loss at 11 o’clock position • Angulation of lower vein and undermining due to optic disc cupping • Temporal alpha zone and beta zone peripapillary chorioretinal atrophy (PPA) Demo Jane, 1954/01/01 (52yrs) Right eye (OD) / 2007/02/06 / 09:27:13 Seven-in-One Grayscale (CO) Values 3.1 5 18.4 23.9 8.7 95%..100% 11 9 21 15 83%...94% 71%...82% 12 21 1 24 23 59%...70% 11 12 15 24 47%...58% 28 35%...46% 6 7 33 30 27 23%...34% 28 29 33 31 7 14 16 29 29 29 11%...22% 21 31 25 24 27 0%...10% 9 18 28 26 25 26 11 22 26 27 26 27 7.9 0.8 20.0 27.5 25 24 Comparison Corrected comparisons Defect curve Rank 19 16 1 74 15 16 5 10 12 13 + 7 -5 16 6 13 + 0 5% 27 + + 24 + + 18 17 13 + 15 14 10 + 5 Defect (dB) + + 24 24 + + 21 21 + + 10 95% 6 + + + + + + + + + 19 14 13 + + + 16 11 11 + + + 15 Absolute 8defects 6 + + + 5 + + + + 20 18 11 + + 15 8 + + (sensitivity threshold + +0 dB) + + 25 13 + + + + + 10 + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 448 / 4 30° Parameters: 4 / 1000 asb III 100 ms Duration: 14:18 MS [dB]: 17.2 Catch trials: 0/22 (0%) +, 5/23 (22%) - RF: 11.1 MD [< 2.0 dB]: 10.4 Refraction S/C/A: -3.0/-0.25/180 VA: 1.2 sLV [< 2.5 dB]: 10.4 Pupil [mm]: IOP [mmHg]: 13 CsLV [dB]: 10.3 SF [dB]: 2.1 Comment: Classification: OCTOPUS® OCTOPUS 101 • Dense visual ield loss in superior nasal quadrant with many locations showing absolute defects and little remaining sensitivity near ixation corresponding with RNFL loss at 6 to 8 o’clock position • Mild sensitivity loss on lower nasal ield relating to RNFL loss at 11 o’clock position Glaucoma | Single field 265 9 LATE STAGE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 52-year-old male, no family history • Patient reported decreased visual acuity in both eyes IOP/VA corr • 15 mmHg/ 1.2 + 1.25 (sph), - 0.5 (cyl) x 80° FUNDUS • C/D = 1.0 • Rim disappearance at 12 and 6 to 8 o’clock position • Narrowing of retinal artery Demo John, 1954/01/01 (52yrs) Right eye (OD) / 2006/11/24 / 16:24:03 Seven-in-One Grayscale (CO) Values 0.8 2 4 13.0 26.2 14.0 95%..100% 9 6 7 15 83%...94% 71%...82% 5 18 59%...70% 1 21 47%...58% 27 24 35%...46% 3 28 29 30 23%...34% 29 25 28 28 26 20 11%...22% 1 21 0%...10% 15 20 1 5 14 14 17 22.9 13.3 5.0 15.0 12 13 Comparison Corrected comparisons Defect curve 1 Rank 74 22 20 12 11 -5 17 19 19 10 7 10 9 + 0 5% 23 9 13 + 28 5 18 + 5 Defect (dB) + + + + 95% 29 + + 19 + + 10 + + + 7 + + + + + 5 8 + + + 15 29 6 19 + 13 7 + + 20 25 25 22 14 14 10 15 13 5 + + Diffuse defect [dB]: -1.1 14 14 + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 120 105 90 75 60 P < 0,5 135 45 150 30 30° 165 15 Programs: G Standard White/White / Normal Questions / repetitions: 438 / 0 Parameters: 4 / 1000 asb III 100 ms Duration: 13:53 MS [dB]: 8.6 Catch trials: 0/22 (0%) +, 11/22 (50%) - RF: 25.0 MD [< 2.0 dB]: 19.0 180 10 30 40 50 60 70 80 90 0 Refraction S/C/A: +2.75/-0.5/80 VA: 1.0 sLV [< 2.5 dB]: 10.1 Pupil [mm]: IOP [mmHg]: 16 CsLV [dB]: 10.0 SF [dB]: 2.4 195 345 Comment: Classification: 210 330 OCTOPUS ® OCTOPUS 101 225 240 255 270 285 300 315 • Dense double arcuate defect with many locations showing absolute defects • No sensitivity loss at ixation • Kinetic perimetry shows intact temporal and central visual ield • Late stage glaucoma with preserved ixation and peripheral temporal visual ield 266 Chapter 13 | Clinical cases 10 EARLY TO MODERATE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 40-year-old male, no family history • Glaucoma was suspected after routine medical check-up IOP/VA corr • 16 mmHg/1.2 - 2.5 (sph), - 1.5 (cyl) x 110° FUNDUS • 1998 Rim thinning RNFL loss at 7 o’clock position • 2007 Rim thinning & RNFL loss at 6 to 8’clock position indicating progression Cluster MD change MD Mean defect sLV Loss variance 1.1 – 2.4 dB/year 0 0 2.4 1.9 1.9 -0.0 1.1 S 30 20 10 N T 0.0 [dB] I 0.1 -0.2 15 MD change -0.1 0.8 dB/year -0.0 25 15 Structural progression 2007 2007 Slope: 0.8dB / Yr (p<0.5%) Slope: 0.7dB / Yr (p<0.5%) suggested at 6 to 8 o'clock position DD Diffuse defect LD Local defect 0 0 1998 2007 25 15 2007 2007 Slope: 0.0dB / Yr Slope: 1.0dB / Yr (p<0.5%) 1998 2007 • Grayscale series shows expansion of superior nasal defect to a superior arcuate defect from 1998 to 2007 • Signi icant (P < 1%) MD worsening at 0.8 dB/year due to fast progression in affected superior clusters (Cluster MD change 1.1 to 2.4 dB/year) • Large (up to 30 dB) progression at 6 to 8 o’clock position in Polar Trend Analysis • Rim thinning and RNFL loss spreading from 7 o’clock position towards 6 and 8 o’clock position • Clear relationship between fundus and visual ield progression con irming glaucomatous progression Glaucoma | Trend 267 11 EARLY TO MODERATE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 68-year-old female, no family history • High IOP identi ied during visit initiated due to eye pain IOP/VA corr • 22 mmHg/1.5 + 0.75 (sph), - 0.25 (cyl) x 10° FUNDUS • 2001 Mild, slit-like RNFL loss at 7 o’clock position. No rim thinning or notching. • 2008 RNFL loss & additional rim thinning with undermining at 6 to 8 o’clock position indicating progression; laser scar at 1 to 3 o’clock position due to treated BRVO, which developed in 2002 during follow up MD Mean defect sLV Loss variance 0 0 1.8 1.1 2.1 0.3 1.9 S 30 20 10 N T -0.2 [dB] I -0.3 0.5 15 -0.4 0.1 MD change 25 0.4 dB/year 15 Structural progression 2001 2008 2001 2008 suggested at 6 to 8 Slope: 0.4dB / Yr (p<0.5%) Slope: 0.8dB / Yr (p<0.5%) o'clock position DD Diffuse defect LD Local defect 0 0 2001 2008 Large progression at 25 15 6 to 8 o’clock position 2001 2008 2001 2008 Slope: -0.2dB / Yr Slope: 0.8dB / Yr (p<0.5%) 2001 2008 • Grayscale series shows expansion of superior nasal defect to a superior arcuate defect from 2001 to 2008 and mild inferotemporal sensitivity loss due to BRVO • Signi icant (P < 1%) but slow MD worsening at 0.4 dB/year due to fast progression in affected superior clusters (Cluster MD change 1.1 to 2.1 dB/year) • Large (up to 30 dB) progression at 6 to 8 o’clock position in Polar Trend Analysis • Rim thinning and RNFL loss spreading from 7 o’clock position towards 6 and 8 o'clock position • Clear relationship between fundus and visual ield progression con irming glaucomatous progression 268 Chapter 13 | Clinical cases 12 EARLY TO MODERATE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 53-year-old male, no family history • High IOP identi ied during visit related to eye pain IOP/VA corr • 25 mmHg/1.2 - 0.75 (sph), - 1.0 (cyl) x 90° FUNDUS • 2002 Rim thinning at 1 to 2 o’clock position. Rim notching at 5 o’clock position. RNFL loss at same positions. Optic disc hemorrhage at 6 o’clock position. • 2008 Rim thinning from 1 to 6 o’clock position MD Mean defect sLV Loss variance 0 0 1.1 0.6 0.2 2.5 0.1 S 30 20 10 N T 0.1 [dB] I 0.0 0.2 15 0.0 -0.4 Floor effect Structural progression 25 15 (near absolute sensitivity loss) 2002 2008 2002 2008 suggested at Slope: 0.5dB / Yr (p<0.5%) Slope: 0.1dB / Yr 5 o'clock position DD Diffuse defect LD Local defect 0 0 2002 2008 25 15 2002 2008 2002 2008 Large progression at Slope: 0.2dB / Yr Slope: 0.5dB / Yr (p<0.5%) 5 o’clock position 2002 2008 • Grayscale series shows expansion of inferior arcuate defect to superior nasal side from 2002 to 2008 • Signi icant (P < 1%) but slow MD worsening at 0.5 dB/year due to fast progression in affected superior clusters (Cluster MD change up to 2.5 dB/year) • Large (~28 dB) progression at 5 o’clock position in Polar Trend Analysis • Rim thinning and RNFL loss spreading from 1 to 2 o'clock position towards 6 o'clock position • Clear relationship between fundus and visual ield progression con irming glaucomatous progression Glaucoma | Trend 269 13 EARLY TO MODERATE GLAUCOMA (NORMAL TENSION GLAUCOMA) PATIENT • 51-year-old male, no family history • Patient reported a blind spot in visual ield of left eye during reading and a visual ield defect temporally near ixation upon ixation of distant objects IOP/VA corr • 15 mmHg/1.0 – 6.0 (sph), - 1.25 (cyl) x 160° FUNDUS • 2001 Small disc. RNFL loss (including papillomacular nerve iber) from 2 to 5 o’clock position. Temporal alpha zone and beta zone peripapillary chorioretinal atrophy (PPA). • 2004 Challenging to identify changes because of small disc and severe myopia MD Mean defect sLV Loss variance 0 0 1.1 0.8 0.8 0.8 5.4 S 30 20 10 N T 3.3 [dB] I 0.5 0.4 15 0.6 Very fast progression 0.6 25 15 in central visual Large progression suggested 2001 2002 2003 2004 2001 2002 2003 2004 field clusters at inferior temporal optic disc Slope: 1.2dB / Yr (p<0.5%) Slope: 0.8dB / Yr (p<0.5%) DD Diffuse defect LD Local defect 0 0 2001 2004 25 15 2001 2002 2003 2004 2001 2002 2003 2004 Slope: 0.5dB / Yr Slope: 1.0dB / Yr (p<0.5%) 2001 2004 • Grayscale series shows expansion of superior paracentral defect towards ixation from 2001 to 2004 • Signi icant (P < 1%) and fast MD worsening at 1.2 dB/year due to very fast progression in affected central clusters (Cluster MD change 3.3 and 5.4 dB/year) • Challenging to asses structural changes, but large (up to 30 dB) progression at 5 o’clock position in Polar Trend Analysis corresponding with RNFL loss in fundus image suggests glaucomatous progression • Relationship between fundus and visual ield progression con irming glaucomatous progression 270 Chapter 13 | Clinical cases 14 EARLY TO MODERATE GLAUCOMA (PRIMARY OPEN-ANGLE GLAUCOMA) PATIENT • 74-year-old female • Patient showed advanced disc damage at presentation Suboptimal IOP control under topical medication, but patient refused surgery IOP/VA uncorr • 16-22 mmHg (28 mmHg pre-treatment)/1.0 OCT • 2008 Pathologically low peripapillary RNFLT in inferotemporal sectors • 2013 Statistically signi icant further RNFLT decrease both infero- and superotemporally FUNDUS • 2008 Advanced disc damage (C/D=0.95) MD Mean defect sLV Loss variance Fast superior 0 paracentral 0 progression 2.6 2.5 0.4 -0.5 1.7 S 30 20 10 N T 0.7 [dB] I -0.1 0.2 15 0.0 MD change -0.0 25 1.0 dB/year 15 2013 2013 Slope: 1.0dB / Yr (p<0.5%) Slope: 0.4dB / Yr (p<0.5%) DD Diffuse defect LD Local defect 0 0 2008 2013 No diffuse progression Significant local progression 25 15 2013 2013 Slope: 0.2dB / Yr Slope: 1.1dB / Yr (p<0.5%) 2008 2013 • Grayscale series shows progression of superior arcuate and both superior and inferior paracentral defects from 2008 to 2013 • Local progression apparent from signi icant (P < 1%) sLV increase and LD worsening due to very fast progression in superior arcuate and superior and inferior paracentral clusters (Cluster MD change up to 2.6dB/year) • Up to 30 dB progression at infero- and superotemporal test locations in Polar Trend Analysis spatially related to further RNFLT loss between 2008 and 2013 • Relationship between OCT and visual ield progression con irming glaucomatous progression Glaucoma | Trend 271 15 END-STAGE GLAUCOMA (EXFOLITATIVE GLAUCOMA) PATIENT • 79-year-old female • Patient presented with end-stage glaucoma, iltration surgery was performed with no further medication during follow up • Patient reported only minimal central visual ield worsening during follow-up IOP/VA corr • 08 - 14 mmHg (43mmHg pre-treatment)/1.0 + 1.0 (sph) OCT • 2008 Severe peripapillary RNFLT loss • 2013 No change in the average peripapillary RNFLT FUNDUS • 2008 C/D=0.99 MD Mean defect sLV Loss variance 0 0 -0.0 Floor effect 0.0 (MD > 20 dB, no progression) -0.0 -0.0 1.4 S 15 30 20 10 N T 2.5 [dB] I -0.9 -0.0 Fast superior 0.0 -0.0 and inferior 35 15 paracentral 2013 2013 progression Slope: 0,1dB / Yr Slope: -0,2dB / Yr (p<10%) DD Diffuse defect LD Local defect 0 0 2008 2013 Floor effect (DDc > 20 dB, no progression) 25 15 2013 2013 Slope: -0,0dB / Yr Slope: 0,0dB / Yr 2008 2013 • Grayscale series shows very dense visual ield loss with little remaining sensitivity in macula • MD appears stable, but cannot be interpreted for progression because of loor effect (exceeding perimeter’s measurement range) • Signi icant (P < 1%) superior and inferior paracentral progression (Cluster MD change 1.4 and 2.5 dB/year) • 12 to 25 dB progression at 8 to 10 o’clock position (papillomacular bundle) in Polar Trend Analysis not apparent in OCT results due to the loor effect of OCT in end-stage glaucoma • Polar and Cluster Trend Analysis indicate late-stage glaucomatous progression 272 Chapter 13 | Clinical cases 16 CEREBRAL INFARCTION (BILATERAL) PATIENT • 65-year-old male, no family history • Patient experienced occipital headache and optic agnosia of name, letters, etc. • Diagnosed with cerebral infarction in left temporal lobe • Previous central serous chorioretinopathy in left eye Demo John, 1933/01/01 (65yrs) Left eye (OS) / 1999/07/12 / 11:36:23 Seven-in-One Grayscale (CO) Values 26.3 24 0.4 -0.1 25.7 95%..100% 24 27 26 1 83%...94% 71%...82% 26 25 26 59%...70% 27 24 28 5 47%...58% 27 28 35%...46% 30 29 1 23 30 24 2 1 23%...34% 30 31 21 18 1 11%...22% 29 31 27 29 30 0%...10% 27 27 27 26 29 30 27 28 30 28 28 24 -1.0 7.9 28.4 19.0 29 26 Comparison Corrected comparisons Defect curve Rank + + 1 74 -5 + + + 23 + + + 23 + + 0 5% + + + + + + + 23 + + + 23 5 + + + + Defect (dB) 95% + + 30 + + 30 10 9 9 + 7 29 28 + 7 29 28 + + 9 11 26 + + 9 10 26 15 + + + + + + + + + + Sensitivity + + loss + + + + + + 20 at fixation + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Vertical drop characteristic for Probabilities Corrected probabilities quadrantanopia P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 409 / 2 30° Parameters: 4 / 1000 asb III 100 ms Duration: 15:54 MS [dB]: 18.3 Catch trials: 0/20 (0%) +, 4/21 (19%) - RF: 9.7 MD [< 2.0 dB]: 8.4 Refraction S/C/A: +0.75/-1.5/90 VA: 0.4 sLV [< 2.5 dB]: 12.4 Pupil [mm]: 6.4 IOP [mmHg]: 14 CsLV [dB]: 12.5 SF [dB]: 1.5 Comment: Classification: OCTOPUS® OCTOPUS 101 Neuro | Single field 273 IOP/VA corr • OD 19 mmHg/ 1.0 + 0.5 (sph), - 2.0 (cyl) x 100°; OS 20 mmHg/ 0.4 - 1.5 (cyl) x 90° FUNDUS • No abnormality Demo John, 1933/01/01 (65yrs) Right eye (OD) / 1999/07/12 / 11:00:42 Seven-in-One Grayscale (CO) Values 26.3 25 2.0 -0.1 24.1 95%..100% 26 23 24 83%...94% 71%...82% 24 1 25 27 59%...70% 28 28 25 3 47%...58% 27 27 27 28 35%...46% 28 30 1 29 29 31 20 23%...34% 28 29 29 26 21 18 11%...22% 30 24 13 30 27 0%...10% 29 28 25 24 29 28 27 28 28 25 28 25 -1.4 6.1 28.3 21.4 28 25 Comparison Corrected comparisons Defect curve Rank + + 1 74 + + + + + + -5 + 26 + 25 0 5% + + + + + + + 22 + + + 22 5 Defect (dB) + + + + + + + + + + 30 + + 29 10 95% + + + + 11 + + 11 + + + + 9 9 + + + + 9 9 15 + 5 13 + 5 13 + + + + + + + + + + + + 20 + + + + 25 + + + + + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 424 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 16:01 MS [dB]: 19.9 Catch trials: 0/21 (0%) +, 6/22 (27%) - RF: 13.9 MD [< 2.0 dB]: 6.8 Refraction S/C/A: +0.5/-2.0/90 VA: 1.0 sLV [< 2.5 dB]: 11.3 Pupil [mm]: 6.0 IOP [mmHg]: 14 CsLV [dB]: 11.3 SF [dB]: 2.0 Comment: Classification: OCTOPUS® OCTOPUS 101 • Superior homonymous quadrantanopia sparing ixation on right side of vertical meridian due to cerebral infarction in left temporal lobe • Signi icant sensitivity loss at ixation in left eye due to previous central serous chorioretinopathy with decrease in visual acuity (0.4) 274 Chapter 13 | Clinical cases 17 LEBER HEREDITARY OPTIC NEUROPATHY (BILATERAL) PATIENT • 31-year-old male, no family history • Patient reported decreased visual acuity in right eye • Patient diagnosed with central serous chorioretinopathy and retinal hemorrhage • After referral, patient diagnosed with optic neuropathy based on MRI indings • Patient diagnosed with Leber hereditary optic neuropathy based on maternal mitochondrial DNA test Demo John, 1973/01/01 (31yrs) Left eye (OS) / 2004/10/21 / 12:51:41 Seven-in-One Grayscale (CO) Values 16 17 8.5 18.7 19.3 8.8 95%..100% 21 15 14 7 20 13 83%...94% 71%...82% 18 22 21 22 10 12 9 14 59%...70% 22 21 24 26 22 13 2 1 17 16 47%...58% 22 24 23 19 3 1 1 15 35%...46% 23%...34% 22 22 22 19 4 11%...22% 22 27 25 28 11 2 0%...10% 21 23 21 20 20 5 12 10 19 24 24 22 18 15 7.8 20.5 21.4 8.0 20 21 23 24 Comparison Corrected comparisons Defect curve 9 7 + + Rank 120 105 90 75 60 1 74 6 11 13 19 6 12 + + 5 12 + 5 -5 135 45 10 6 7 6 19 16 19 12 + + + + 11 8 11 + 0 5% 150 30 6 7 5 + 8 17 28 28 10 9 + + + + + 9 20 20 + + 5 6 5 7 13 29 31 27 11 + + + 5 21 23 19 + Defect (dB) 10 95% 165 15 7 7 9 13 23 + + + 5 15 15 6 + 5 + 20 29 + + + + 13 21 180 10 30 40 50 60 70 80 90 0 20 8 6 9 10 10 24 16 17 + + + + + 16 9 9 25 10 5 5 7 10 11 + + + + + + Diffuse defect [dB]: -1.1 195 345 8 7 5 + + + + + 210 330 Probabilities Corrected probabilities 225 315 240 255 270 285 300 P>5 P<5 V4e P<2 P<1 I4e P < 0,5 I3e I3a Programs: 32 Standard White/White / Normal Questions / repetitions: 585 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 19:20 MS [dB]: 14.2 Catch trials: 0/29 (0%) +, 4/30 (13%) - RF: 6.7 MD [< 2.0 dB]: 14.0 Refraction S/C/A: // VA: sLV [< 2.5 dB]: 9.2 Pupil [mm]: 6.0 IOP [mmHg]: CsLV [dB]: 9.2 SF [dB]: 1.9 Comment: Classification: OCTOPUS® OCTOPUS 101 Neuro | Single field 275 IOP/VA corr • OD 10 mmHg/ 10 cm, inger counting; OS 10 mmHg/ 30 cm, hand motion FUNDUS • Pale optic discs in both eyes CENTRAL CFF • OD 32 Hz; OS 42 Hz Demo John, 1973/01/01 (31yrs) Right eye (OD) / 2004/10/21 / 13:17:49 Seven-in-One Grayscale (CO) Values 23 18 20 19 4.7 12.6 22.7 15.2 95%..100% 22 24 23 25 24 18 83%...94% 71%...82% 22 24 24 26 22 23 19 21 59%...70% 21 24 24 26 23 22 24 1 20 47%...58% 22 23 25 22 20 4 15 35%...46% 23%...34% 23 22 25 24 22 7 1 19 11%...22% 22 23 22 23 20 22 24 24 18 23 0%...10% 23 25 24 23 21 23 23 25 21 25 23 22 23 24 5.6 10.2 22.9 19.1 24 25 25 24 Comparison Corrected comparisons Defect curve + 7 5 6 + + + + Rank 1 74 120 105 90 75 60 + + + + + 9 + + + + + + -5 135 45 + + + + 6 5 9 7 + + + + + + + + 0 5% + + 5 + 7 8 5 27 8 + + + + + + + 22 + 5 150 30 Defect (dB) + 5 5 10 12 28 13 + + + + 7 22 8 10 95% 165 15 + 6 6 7 11 25 28 10 + + + + 6 20 23 5 15 + 5 7 8 11 9 6 6 11 5 + + + + 5 + + + 6 + 20 180 10 30 40 50 60 70 80 90 0 + + 5 7 9 7 6 + + + + + + + + + 25 5 + 6 7 6 5 + + + + + + Diffuse defect [dB]: -1.1 195 345 + + + + + + + + Probabilities Corrected probabilities 210 330 225 315 240 255 270 285 300 P>5 P<5 P<2 V4e P<1 P < 0,5 I4e I3e Programs: 32 Standard White/White / Normal Questions / repetitions: 572 / 0 30° I3c Parameters: 4 / 1000 asb III 100 ms Duration: 18:29 MS [dB]: 20.0 Catch trials: 0/29 (0%) +, 3/29 (10%) - RF: 5.1 MD [< 2.0 dB]: 8.2 I2e Refraction S/C/A: // VA: sLV [< 2.5 dB]: 7.6 Pupil [mm]: 6.2 IOP [mmHg]: CsLV [dB]: 7.5 SF [dB]: 1.8 Comment: Classification: OCTOPUS® OCTOPUS 101 • Dense sensitivity loss in center of both eyes • Additional inferior nasal visual ield loss from 20 to 50° • Asymmetrical visual ield defect, central and peripheral scotomas more severe in left eye 276 Chapter 13 | Clinical cases 18 BILATERAL OPTIC NEURITIS (MULTIPLE SCLEROSIS) PATIENT • 25-year-old female, no family history • Patient reported dif iculty in seeing for two weeks Demo Jane, 1975/01/01 (25yrs) Left eye (OS) / 2001/11/30 / 14:19:18 Seven-in-One Grayscale (CO) Values 22 13 16 19 6.7 4.1 21.4 23.8 95%..100% 16 21 23 22 26 24 83%...94% 71%...82% 18 24 22 25 25 24 25 23 59%...70% 26 22 23 23 23 28 25 25 23 23 47%...58% 20 20 24 26 29 27 25 25 22 35%...46% 23%...34% 24 26 12 9 28 27 28 24 22 11%...22% 22 20 14 14 13 26 26 25 24 22 0%...10% 15 18 24 18 27 26 24 24 21 18 20 25 26 25 11.8 4.1 17.8 24.8 20 18 24 22 Comparison Corrected comparisons Defect curve + 12 9 6 + 7 + + Rank 120 105 90 75 60 1 74 11 6 + 5 + + 6 + + + + + -5 135 45 10 + 7 + + + + + 5 + + + + + + + 0 5% 150 30 + 6 6 7 8 + 6 + 5 + + + + + + + + + + + 5 8 9 7 7 + 5 5 + 5 + + + + + + + + + Defect (dB) 10 95% 165 15 5 + 20 24 5 5 + 5 6 + + 15 18 + + + + + 15 7 9 17 17 19 5 5 5 5 5 + + 11 12 14 + + + + + 180 10 30 40 50 60 70 80 90 0 20 14 12 7 12 + + 5 + 9 7 + 7 + + + + 25 8 11 10 + + + + 6 5 + + + Diffuse defect [dB]: -1.1 195 345 9 10 + 5 + 5 + + 210 330 Probabilities Corrected probabilities 225 315 240 255 270 285 300 P>5 P<5 V4e P<2 P<1 I4e P < 0,5 I3e I3b Programs: 32 Standard White/White / Normal Questions / repetitions: 627 / 2 30° Parameters: 4 / 1000 asb III 100 ms Duration: 20:51 I2e Catch trials: 0/31 (0%) +, 1/32 (3%) - RF: 1.5 MS [dB]: MD [< 2.0 dB]: 22.0 6.6 Refraction S/C/A: +0.25/-0.5/80 VA: 1.0 sLV [< 2.5 dB]: 4.4 I2b Pupil [mm]: 6.3 IOP [mmHg]: 11 CsLV [dB]: 3.9 SF [dB]: 2.6 Comment: I1e Classification: OCTOPUS® OCTOPUS 101 Neuro | Single field 277 IOP/VA corr • OD 13 mmHg/ 0.7 + 0.25 (sph), - 1.0 (cyl) x 85° OS 11mmHg/ 1.0 + 0.25 (sph), - 0.5 (cyl) x 80° FUNDUS • No abnormality CFF • OD 34 Hz; OS 44 Hz MRI • Demyelinated plaque at optic chiasm Demo Jane, 1975/01/01 (25yrs) Right eye (OD) / 2001/11/30 / 14:45:35 Seven-in-One Grayscale (CO) Values 26 16 24 25 2.1 2.0 25.8 26.2 95%..100% 22 27 24 23 26 27 83%...94% 71%...82% 26 27 25 26 23 26 25 27 59%...70% 26 26 27 28 29 29 30 29 26 26 47%...58% 25 27 28 30 29 28 27 27 28 35%...46% 23%...34% 27 28 27 30 19 26 26 29 29 11%...22% 26 27 27 27 15 18 20 15 27 28 0%...10% 24 27 25 26 7 15 23 23 25 29 26 19 20 20 3.6 8.3 25.3 21.3 26 24 21 23 Comparison Corrected comparisons Defect curve + 9 + + + 7 + + Rank 1 74 120 105 90 75 60 + + + + + + + + + + + + -5 135 45 + + + + 6 + + + + + + + + + + + 0 5% 150 30 + + + + + + + + + + + + + + + + + + + + 5 Defect (dB) + + + + + 5 + + + + + + + + + + + + 10 95% 165 15 + + + + 14 7 6 + + + + + + 12 + + + + 15 + + + + 16 14 11 16 + + + + + + 14 12 9 13 + + 20 180 10 30 40 50 60 70 80 90 0 + + + 5 23 16 7 6 + + + + 21 13 5 + 25 + + + 11 9 9 + + + 9 7 7 Diffuse defect [dB]: -1.1 195 345 + + 7 6 + + 5 + Probabilities Corrected probabilities 210 330 225 315 240 255 270 Intact 285 peripheral 300 visual field P>5 P<5 P<2 V4e P<1 P < 0,5 I4e I3e Programs: 32 Standard White/White / Normal Questions / repetitions: 568 / 4 30° I3b Parameters: 4 / 1000 asb III 100 ms Duration: 19:38 Catch trials: 0/28 (0%) +, 2/29 (7%) - RF: 3.5 MS [dB]: MD [< 2.0 dB]: 24.7 3.9 I2e Refraction S/C/A: +0.25/-1.0/85 VA: 0.7 sLV [< 2.5 dB]: 4.7 Pupil [mm]: IOP [mmHg]: 13 CsLV [dB]: 4.5 I2b SF [dB]: 1.9 Comment: Classification: I1e I1a OCTOPUS® OCTOPUS 101 • Sensitivity loss on lower temporal side of vertical meridian in both eyes (i.e., mild bitemporal hemianopia) • MRI shows demyelinated plaque , thus bitemporal hemianopia is attributed to multiple sclerosis at optic chiasm 278 Chapter 13 | Clinical cases 19 TUBERCULUM SELLAE MENINGIOMA (BILATERAL) PATIENT • 64-year-old male, no family history • Patient reported dif iculty in reading books and newspaper Demo John, 1941/01/01 (64yrs) Left eye (OS) / 2005/10/12 / 12:52:19 Seven-in-One Grayscale (CO) Values 18 18 24.7 1.6 0.9 23.7 95%..100% 1 15 24 23 19 83%...94% 71%...82% 2 22 24 24 21 59%...70% 27 25 25 23 22 47%...58% 30 30 27 28 25 35%...46% 23%...34% 30 27 29 27 26 11%...22% 28 27 27 25 25 0%...10% 1 27 25 24 25 9 26 24 21 26.5 0.9 0.6 25.4 2 22 23 Comparison Corrected comparisons Defect curve 5 + + + Rank 120 105 90 75 60 1 74 24 10 + + + 21 7 + + + -5 135 45 24 5 + + + 22 + + + + 0 5% 150 V4e 30 + + + + + + + + + + 5 Defect (dB) + + + + + + + + + + 10 95% I4e 165 15 + + + + + + + + + + 15 + + + + + + + + + + I3e 180 10 30 40 50 60 70 80 90 0 27 + + + + 25 + + + + 20 25 I2e 195 345 18 + + + 16 + + + Diffuse defect [dB]: -1.1 24 + + 22 + + I1e210 330 Probabilities Corrected probabilities 225 315 240 255 270 285 300 Absolute defect P>5 stopping at P<5 vertical midline P<2 P<1 P < 0,5 Programs: 32 Standard White/White / Normal Questions / repetitions: 415 / 0 30° Parameters: 4 / 1000 asb III 100 ms Duration: 16:34 MS [dB]: 13.0 Catch trials: 0/21 (0%) +, 7/21 (33%) - RF: 16.6 MD [< 2.0 dB]: 13.1 Refraction S/C/A: +1.0/-1.0/100 VA: 1.0 sLV [< 2.5 dB]: 12.4 Pupil [mm]: 5.7 IOP [mmHg]: 15 CsLV [dB]: 12.4 SF [dB]: 1.4 Comment: Classification: OCTOPUS® OCTOPUS 101 Neuro | Single field 279 IOP/VA corr • OD 12 mmHg/0.15 – 2.0 (sph) OS 13 mmHg/1.2 – 1.5 (sph), - 1.0 (cyl) x 100° FUNDUS • Pale optic disc with slight cupping • Slight bending of blood vessels CFF • OD 25 Hz; OS 40 Hz MRI • Meningioma in tuberculum sellae Demo John, 1941/01/01 (64yrs) Right eye (OD) / 2005/10/12 / 13:14:55 Seven-in-One Grayscale (CO) Values 25.3 25.6 0.1 0.0 95%..100% 83%...94% 71%...82% 59%...70% 47%...58% 2 35%...46% 23%...34% 16 14 23 26 24 11%...22% 20 19 18 24 22 0%...10% 18 15 18 19 8 13 12 9.8 27.1 16.6 0.0 4 7 Comparison Corrected comparisons Rank 1 74 120 105 90 75 60 -5 135 45 0 5% 150 30 5 Defect (dB) 28 9 95% 10 165 15 9 12 6 + 7 + + + + + 15 + 7 9 5 7 + + + + + 20 180 10 30 40 50 60 70 80 90 0 7 11 9 9 + + + + 25 16 13 15 + + + Diffuse defect [dB]: -1.1 195 345 20 19 + + Probabilities Corrected probabilities 210 330 225 315 240 255 270 285 300 P>5 P<5 P<2 V4e P<1 P < 0,5 I4e I3e Programs: 32 Standard White/White / Normal Questions / repetitions: 363 / 0 30° I2e Parameters: 4 / 1000 asb III 100 ms Duration: 15:20 MS [dB]: 4.3 Catch trials: 0/18 (0%) +, 8/19 (42%) - RF: 21.6 MD [< 2.0 dB]: 21.8 I1e Refraction S/C/A: +0.5// VA: 0.15 sLV [< 2.5 dB]: 7.8 Pupil [mm]: 6.0 IOP [mmHg]: 13 CsLV [dB]: 7.7 SF [dB]: 2.1 Comment: Classification: OCTOPUS® OCTOPUS 101 • Complete sensitivity loss (heterononymous hemianopia) temporally of vertical meridian • Additional absolute defect in superior nasal quadrant of right eye 280 Chapter 13 | Clinical cases 20 AGE-RELATED MACULAR DEGENERATION PATIENT • 64-year-old male, no family history • Patient reported decreased visual acuity in left eye IOP/VA corr • 13 mmHg/ 0.2 – 1.75 (sph), - 0.75 (cyl) x 80° FUNDUS • Exudative age-related macular degeneration in macula area Demo John, 1942/01/01 (64yrs) Left eye (OS) / 2006/09/28 / 14:38:48 Seven-in-One Greyscale (CO) Values 15.8 14.6 13.8 15.2 95%..100% 29 28 83%...94% 28 28 25 29 71%...82% 30 23 17 27 59%...70% 23 17 2 10 11 30 47%...58% 25 11 8 2 9 7 4 25 25 1012 4 9 1114 16 35%...46% 21 6 7 9 9 6 5 4 7 18 20 1 4 11 6 7 18 23%...34% 24 11 9 13 12 11 3 26 26 6 16 13 9 1811 31 11%...22% 22 25 13 11 27 30 0%...10% 27 24 25 29 27 27 28 28 12.7 12.0 27 30 17.4 18.3 Comparison Corrected comparisons Defect curve Rank 1 74 + + + + -5 + + + + + + + + 0 + 6 12 + + + 8 + 5% 6 12 2820 19 + + 8 2316 15 + 5 Defect (dB) + 192329222327 + + 151824181922 + 5 201927232017 14 + 161523181513 9 95% 10252523 23252727 5 20201819212222 10 25 1311302821262413 21 8 7 2624172220 9 6 202219192128 5 + 161815151624 + 15 + 251519221320 + + 20101418 9 15 + 7 5 1820 + + + + 1316 + + 20 + 6 5 + + + + + + + + + + + + + 25 Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 M-pattern Programs: M Standard White/White / Normal Questions / repetitions: 759 / 0 12° Parameters: 4 / 1000 asb III 100 ms Duration: 23:03 MS [dB]: 16.4 Catch trials: 0/38 (0%) +, 2/38 (5%) - RF: 2.6 MD [< 2.0 dB]: 13.6 Refraction S/C/A: -1.0// VA: 0.3 sLV [< 2.5 dB]: 10.1 Pupil [mm]: 6.1 IOP [mmHg]: 19 CsLV [dB]: 10.2 SF [dB]: 3.1 Comment: Classification: OCTOPUS® OCTOPUS 101 • M-pattern (10°) used for a high resolution of the macula • Dense visual ield loss within central 5° of macula, no visual ield loss from 6° to 10° Retina | Single field 281 21 BRANCH CENTRAL RETINAL ARTERY OCCLUSION PATIENT • 51-year-old female, no family history • Patient reported sudden loss of vision in superior visual ield of left eye IOP/VA corr • 14 mmHg/ 1.0 – 4.0 (sph), + 0.75 (cyl) x 80° FUNDUS • Ischemia-induced retinal edema in area of blood vessels caused by occlusion of the downward branch of the central retinal artery Demo Jane, 1956/01/01 (51yrs) Left eye (OS) / 2007/05/16 / 11:37:22 Seven-in-One Greyscale (CO) Values 11.1 20.2 19.3 10.3 95%..100% 11 6 83%...94% 19 23 22 18 71%...82% 12 2 20 23 59%...70% 22 1 20 23 47%...58% 20 1 2 4 13 25 7 15 9 8 32 2529 29 35%...46% 8 2725 16 29 3230 28 33 26 3329 33 33 2529 31 23%...34% 27 3131 33 31 3232 30 31 3333 29 30 3232 31 11%...22% 26 32 33 32 31 31 0%...10% 32 31 32 30 30 30 29 28 0.5 0 .8 29 29 30.4 30.2 Comparison Corrected comparisons Defect curve Rank 1 74 120 105 90 75 60 18 22 16 21 -5 135 45 10 6 7 11 9 + 6 9 0 18 28 10 6 16 27 9 5 5% 150 30 7 30 11 7 6 29 9 5 5 Defect (dB) 10 3029 2718 5 9 2928 2617 + 24 162324 + 7 + + 22 152223 + 5 + + 95% 24 5 8 17++ + + + 22 + 6 15++ + + + 10 165 15 5 + + ++ 8 + + + + + ++ 7 + + + + + ++ + + + + + + ++ + + + + + + ++ + + + + + + ++ + + + 15 + + ++ + + + + ++ + + 20 180 10 30 40 50 60 70 80 90 0 + + + + + + + + + + + + + + + + 25 + + + + Diffuse defect [dB]: -1.1 195 345 Probabilities Corrected probabilities 210 330 225 315 240 255 270 285 300 P>5 P<5 P<2 V4e P<1 P < 0,5 I4e I3e I2e Programs: M Standard White/White / Normal Questions / repetitions: 703 / 0 12° Parameters: Catch trials: 4 / 1000 asb III 100 ms 0/35 (0%) +, 3/36 (8%) - Duration: RF: 21:19 4.2 MS [dB]: 23.1 I1e MD [< 2.0 dB]: 7.7 Refraction S/C/A: -2.5/-0.75/80 VA: 1.0 sLV [< 2.5 dB]: 10.3 Pupil [mm]: IOP [mmHg]: CsLV [dB]: 10.1 SF [dB]: 4.0 Comment: Classification: OCTOPUS® OCTOPUS 101 • M-pattern (10°) used for a high resolution of the macula • Dense to absolute visual ield loss in superior visual ield corresponding with ischemic region of downward retinal artery • Fixation is spared, corrected visual acuity of 1.0 is maintained • Kinetic perimetry shows absolute defect outside 10° nasally 282 Chapter 13 | Clinical cases 22 MACULAR HOLE PATIENT • 67-year-old female, no family history • Patient reported distorted vision in right eye IOP/VA corr • 12 mmHg/ 0.2 – 1.5 (sph), - 2.5 (cyl) x 80° FUNDUS • Macular hole with luid cuff in surrounding region Demo Jane, 1939/01/01 (67yrs) Right eye (OD) / 2006/07/31 / 13:23:40 Seven-in-One Greyscale (CO) Values 25.6 25.1 4.0 4.3 95%..100% 22 25 83%...94% 22 28 27 27 71%...82% 27 27 27 26 59%...70% 26 29 25 25 27 27 47%...58% 29 2825 26 23 2527 30 29 2526 2420 2425 29 35%...46% 30 2724 17 20 252226 16 282423 21 18 1824 28 23%...34% 29 2627 2423 2526 28 28 2526 25 23 2727 28 11%...22% 28 30 28 25 30 26 0%...10% 28 30 29 28 28 28 27 26 3.6 4.6 28 25 26.4 25.3 Comparison Corrected comparisons Defect curve Rank 1 74 5 + + + -5 6 + + + + + + + 0 + + + + + + + + 5% + + 55 + + + + ++ + + Defect (dB) 5 + + 5 57 5 + + + + + +5 + + + + 6 5 711 7 5 + + + + 58 5 + + 95% + + 7 1512 7 9 5 + + 5 13 9 + 6 + 10 + 7 16 9 111413 7 + + 5 6 814 1211 5 + + 5 5 79 6 5 + + + + 56 + + + + 5 5 68 + + + + + + +6 + + + 15 + + +5 + + + + ++ + + 20 + + + + + + + + + + + + + + + + 25 + + + + Diffuse defect [dB]: -1.1 Probabilities Corrected probabilities Signficant foveal defect at P < 0.5% P>5 P<5 P<2 P<1 P < 0,5 M-pattern Programs: M Standard White/White / Normal Questions / repetitions: 611 / 0 12° Parameters: 4 / 1000 asb III 100 ms Duration: 18:13 MS [dB]: 25.5 Catch trials: 1/30 (3%) +, 2/31 (6%) - RF: 4.9 MD [< 2.0 dB]: 4.3 Refraction S/C/A: +1.0/-2.5/80 VA: 0.2 sLV [< 2.5 dB]: 3.8 Pupil [mm]: IOP [mmHg]: 14 CsLV [dB]: 3.6 SF [dB]: 1.9 Comment: Classification: OCTOPUS® OCTOPUS 101 • M-pattern (10°) used for a high resolution of the macula • Signi icant visual ield loss in the central fovea leading to decreased visual acuity (0.2) due to macular hole Retina | Single field 283 23 BRANCH CENTRAL RETINAL VEIN OCCLUSION PATIENT • 76-year-old male, no family history • Patient reported decreased visual acuity in left eye, blurred and double vision IOP/VA corr • 10 mmHg/ 0.2 + 3.75 (sph), - 2.0 (cyl) x 170° FUNDUS • Retinal hemorrhage and soft exudate along RNFL in lower retinal arcade Demo John, 1928/01/01 (76yrs) Left eye (OS) / 2004/03/24 / 15:44:51 Seven-in-One Greyscale (CO) Values 16.7 16 16 16.7 8.7 8.7 95%..100% 18 18 12 18 15 17 83%...94% 71%...82% 14 14 23 21 12 16 59%...70% 21 14 14 15 18 47%...58% 20 11 19 20 18 15 35%...46% 24 11 16 25 17 20 22 24 24 23%...34% 22 24 24 24 10 18 11%...22% 20 24 23 23 23 0%...10% 23 22 24 22 21 22 20 22 22 20 24 22 5.1 4.8 21.6 21.4 22 22 Comparison Corrected comparisons Defect curve Rank 7 6 + + 1 74 5 6 12 6 8 6 + + 6 + + + -5 11 12 6 7 0 5% + 5 13 8 + + 8 + 5 Defect (dB) + 13 14 12 10 + 7 8 7 + 5 17 10 7 7 8 + 12 + + + + + 18 14 + + 13 9 + 10 95% 14 9 8 8 6 5 + + + + + 5 5 + 16 6 + + + + 11 + 15 5 + 5 + + + 5 5 + + + + 7 5 + + + + + + + + 20 Diffuse & 25 local defect 5 + 5 6 + + + + + + + + Diffuse defect [dB]: -1.1 + + + + Probabilities Corrected probabilities P>5 P<5 P<2 P<1 P < 0,5 Programs: G Standard White/White / Normal Questions / repetitions: 474 / 1 30° Parameters: 4 / 1000 asb III 100 ms Duration: 16:31 MS [dB]: 19.1 Catch trials: 1/24 (4%) +, 0/24 (0%) - RF: 2.0 MD [< 2.0 dB]: 6.9 Refraction S/C/A: +7.0/-2.0/170 VA: 0.2 sLV [< 2.5 dB]: 4.1 Pupil [mm]: 7.2 IOP [mmHg]: CsLV [dB]: 4.0 Very reliable test SF [dB]: 2.1 Comment: Classification: OCTOPUS® OCTOPUS 101 • Sensitivity loss in superior visual ield corresponding with inferior retinal hemorrhage • Diffuse visual ield defect associated with poor visual acuity (0.2) • Signi icant local visual ield loss in superior paracentral area due to macular edema 284 Index 285 INDEX 07 pattern 70, 72-73 constricted visual ield 60-61, 65, 148, 271 10-2 pattern 72, 244 Corrected Cluster Analysis 116, 118, 143-144, 1-level test (1 LT) 91-92 152-155 24-2 pattern 64, 244 Corrected Cluster Trend Analysis (CCTA) 166-167, 2-level test (2 LT) 94-95 183-186 32/30-2 pattern 64, 244 Corrected Comparison 115-117, 143-144, 149-151 Corrected Probabilities 116-117, 143-148 A Corrected square root of Loss Variance (CsLV) 119, ability testing 74-78 121 abnormal visual ield 11, 19, 145-148 critical fusion frequency (CFF) 198-199 absolute defect 101-102, 261 cupola perimeter 214-215, 236, 243 age-related macular degeneration (AMD) 70-71, 280 altitudinal defect 60-61 D apostilb (asb) 14-15 D pattern 73 arcuate defect 60-61, 259, 265-268, 270 data import 239-242 artifactual visual ields, see untrustworthy visual ϔield dB, see decibel asb, see apostilb DD (diffuse defect) 115-118, 121-122, 179, 181-182 automated kinetic perimetry 230-231 DD Trend Analysis 166-167, 179, 181-182 decibel (dB) 14-15 B Defect Curve 109-110, 141-143 background 47 defect, see sensitivity loss background luminance 47, 200, 236 deviation from normal, see sensitivity loss baseline tests 250-251 diabetic retinopathy 70-71, 73 beeping sound 39 diffuse defect 60-61, 100, 115-118, 140-144, binocular visual ield 8-9 178-182 blepharoptosis, see ptosis disability testing 78 blind spot 69, 90-91, 205 driving license test 74-76 blue-on-yellow perimetry, see SWAP drug-induced maculopathy 70-71 dynamic range 195, 201-202 C dynamic strategy 83, 85-86 caecocentral defect 68 candela per m2 (cd/m2) 14-15 E cataract 140-141, 181, 260 Esterman test 74-75 catch trials 123-124 examination parameters cd/m2, see candela per m2 ixed 47-48 central defect 68, 70-71, 280, 282 patient-speci ic 48-57 chiasmal defect, see heteronymous defect exfolitative glaucoma 271 Cluster Analysis 110-113, 152-155 eye patch 31 cluster defect 146-147, 152-155, 183 EyeSuite Progression Analysis 166-190 Cluster Trend Analysis (CTA) 166-167, 183-186 Comparison 103-106, 149-151 286 Index F Grayscale false negative answers 124-125, 138-139, 225, 248 of Comparison 105-106, 149-151 false positive answers 45, 123, 125, 138-139, 177, of Values 102, 106, 245 224-225, 247 G-Screening pattern 66, 92-93 fatigue effect 38 Guided Progression Analysis (GPA) 250-252 ixation 34-36 ixation loss 35, 39, 45, 248 H ixation targets 34 Heijl-Krakau blind spot monitoring 248 ixed examination parameters 47-48 hemianopia 67-68, 206, 231, 278 licker perimetry 52-53, 198-199 Hertz (Hz) 198-199 loor effect 171, 175, 184, 195, 201-201,268, 271 heteronymous defect 67-68, 278-279 luctuation 20-22, 136-140, 172-173, 243 HFA, see Humphrey Field Analyzer frequency-of-seeing curve 21-22 hill of vision 10-11, 13, 16, 18-19, 207-210 function-speci ic perimetry 52-53, 193-196 homonymous defect 67-68, 272-273 Humphrey Field Analyzer 235-252 G hydroxychloroquine, see drug-induced maculopathy G pattern 39, 62-63, 65 gaze tracker 248 I glaucoma intensity, see luminance advanced stage 65, 148, 265, 271 isopter 207-210, 219-223 altitudinal defect 60-61 arcuate defect 60-61, 259, 265-268, 270 K constricted visual ield 60-61, 65, 148, 271 kinetic perimetry 48-51, 205-232 diffuse defect 60-61 early stage 153-154, 158, 257-262, 266-270 L exfoliative 271 LD (local defect) 122, 180-182 moderate stage 263-264, 266-270 LD Trend Analysis 166-167, 180-182 nasal step 60-61, 143-144, 258, 260, 266-267 learning effect 37-38, 243 normal tension 257-258, 261, 263, 267, 269 lens rim artifact 43, 45 paracentral defect 60-61, 258-259, 261, 269-270 lens, see trial lens partial arcuate defect 60-61, 262-263 lid artifact, see also ptosis 43-44, 77, 177 primary open-angle 259, 262, 264-266, 268, 270 linear regression analysis 172 progression 181-182, 185-186, 190, 266-271 local defect 100, 115-118, 140-143, 178-182 temporal wedge defect 60-61 local sensitivity loss, see local defect test patterns 62-66 low vision 53, 65, 86-87, 148, 265, 201-202, 212, typical defects 60-61 271 Glaucoma Hemi ield Test (GHT) 247-249 low vision strategy 83, 86-87 global indices 118-122 luminance Global Trend Analysis (GTA) 166-167 background 47, 200, 236 Goldmann general 14-15, 47-48 perimeter 214-215 maximum stimulus intensity 47-48, 236-237 stimulus intensities 217-218 stimulus intensity 14-15, 47-48, 217-218 stimulus size 52, 216-217 GPA, see Guided Progression Analysis M G-Periphery pattern 65-66 M pattern 65, 71-72, 280, 282 macular hole 282 Index 287 manual kinetic perimetry 230 patient instructions 29-30, 199, 201 maximum stimulus luminance 47-48, 236-237 patient monitoring 35-36 MD Trend Analysis 166-167, 174-176, 178-179 patient positioning 31-34 Mean Defect (MD) 119, 159-160, 168, 174-176, patient-speci ic examination parameters 48-57 178-179 Pattern Deviation 245-246 Mean Deviation (MD) 247 Pattern Standard Deviation (PSD) 247 Mean Sensitivity (MS) 119 pattern, see test pattern monocular visual ield 8-9 perimeter cupola perimeter 214-215, 236, 243 N screen-based perimeter 236, 243 N pattern 69 set-up 28 nasal step 60-61, 143-144, 258, 260, 266-267 perimeter set-up nerve iber bundle defect 68 eye patch 31 neurological disease patient 31-34 caecocentral defect 68 pupil 33 central defect 68 trial lens 33-34 disk edema 67 perimetrist, see visual ϔield examiner hemianopia 67-68, 206, 231, 278 Perimetry 7 heteronymous defect 67-68, 278-279 point-wise event analysis 250-251 homonymous defect 67-68, 272-273 point-wise trend analysis 187-188 idiopathic intercranial hypertension (IIH) 67 Polar Analysis 113-115, 155-158 nerve iber bundle defect 60-61, 68 Polar Trend Analysis (PTA) 166-167, 187-190 optic nerve head drusen 67 Post chiasmal defect, see homonymous defect optic neuritis 67, 276-277 primary open-angle glaucoma 259, 262, 264-266, optic neuropathy 67, 274-275 268, 270 quadrantanopia 67-68, 227-229, 272-273, 279 Probabilities 107-108, 145-148 stroke 67, 272-273 progression 165-190 test patterns 67-70 selection of visual ields 176-178 tumor 67, 231, 278-279 ptosis 43-44, 76-77, 177, 230 typical defects 67-68 pulsar perimetry 52-53, 196-197 normal strategy 83-84 pupil 33, 44 normal tension glaucoma 257-258, 261, 263, 267, 269 normal visual ield 8-10, 18-22, 145-148 Q normative data base 18-20, 237-238 quadrantanopia 67-68, 227-229, 272-273, 279 normative values 18-20, 237-238 qualitative strategy 56-57, 81-82, 90-95 quantitative strategy 56-57, 81-90 O optic neuritis 67, 276-277 R optic neuropathy 67, 274-275 rate of progression 170, 174-176 reaction time compensation 226-227 P refractive error 28, 40-42, 135-136 paracentral defect 60-61, 258-259, 261, 269-270 Reliability Factor (RF) 124-125 partial arcuate defect 60-61, 262-263 reliability indices 123-125 patient data retinal disease date of birth 28, 40, 135-136 age-related macular degeneration (AMD) 70-71, refraction 28, 40-42, 135-136 280 288 Index branch central retinal artery occlusion 281 SAP (standard automated perimetry) 51-53, branch central retinal vein occlusion 283 193-196 diabetic retinopathy 70-71, 73 SWAP (Short-Wavelength Automated Perimetry) drug-induced maculopathy 70-71 52-53, 200-201 macular hole 282 strategy 56-57, 81-82, 96 retinitis pigmentosa (RP) 70-71, 211 1-level test (1 LT) 91-92 test patterns 70-73 2-level test (2 LT) 94-95 typical defects 70-71 dynamic 83, 85-86 retinal nerve iber layer (RNFL) 60, 62, 155-156, 187 low vision 83, 86-87 retinitis pigmentosa (RP) 70-71, 211 normal 83-84 RNFL, see retinal nerve ϔiber layer qualitative 56-57, 81-82, 90-95 quantitative 56-57, 81-90 S screening 92-93 SAP (Standard Automated Perimetry) 51-53, 193-196 Tendency-Oriented Perimetry (TOP) 83, 87-90 scotoma 207-209, 221-223 stroke 67, 272-273 screen-based perimeter 236, 243 structure-function relationship 155-156, 187, screening 62, 92-93, 244 258-271 semi-automated kinetic perimetry 231-232 subjectivity 20-21, 25, 243 sensitivity loss 18-19, 100, 103-106 suprathreshold test, see qualitative strategy sensitivity threshold 12-17, 100-102, 236-237 SWAP 52-53, 200-201 sensitivity to light 9, 11-14 sensitivity, see sensitivity threshold T setting up perimeter 28 technician, see visual ϔield examiner set-up errors 28, 40, 135-136 temporal wedge defect 60-61 Short-term Fluctuation (SF) 124-125, 140, 224-225 Tendency-Oriented Perimetry (TOP) 83, 87-90 Short-Wavelength Automated Perimetry, see SWAP test duration 140 52-53, 200-201 test parameters, see examination parameters signi icance of change 171-174 test pattern 54-55, 59 SITA fast 244 07 70, 72-73 SITA standard 244 10-2 72, 244 slope 170, 174-176 24-2 64, 244 sLV Trend Analysis 166-167, 180-182 32/30-2 64, 244 spot checking 207, 209 60-4 244 square root of Loss Variance (sLV) 119-120, 160-162 BG (blindness) 78 src 197 BT (Blepharoptosis) 76-77 Standard Automated Perimetry, see SAP D 73 static perimetry 12-14, 48-51, 205-206 Esterman 74-75 stimulus exposure time 48 G (glaucoma) 39, 62-63, 65 stimulus luminance 14-15, 47-48, 217-218 G-Periphery 65-66 stimulus speed 218 G-Screening 66, 92-93 stimulus type 51-53 M (macula) 65, 71-72, 280, 282 licker 52-53, 198-199 N 69 function-speci ic 52-53, 193-196 three-zone strategy, see 2-level test Goldmann sizes 52, 217 threshold test, see quantitative strategy low vision 53, 201-202 Threshold Values 245 pulsar 52-53, 196-197 threshold, see sensitivity threshold Index 289 Total Deviation 245-246 central 18, 60-64, 67 trend analysis 168-174 eccentricity 18-19 in luence of luctuation 172-173 interpretation 127-162 in luence of number of tests 173-174 monocular 8-9 linear regression analysis 172 normal 8-10, 18-22, 145-148 ordinary least square it 172 peripheral 18, 65, 67, 69, 72-73, 206, 211, 213 point-wise 187-188 progression 165-190 rate of progression 170, 174-176 spatial extent 8-9 selection of visual ields 176-178 spatial resolution 17-18, 54 signi icance of change 171-174 visual ield examiner 27 slope 170, 174-176 Visual Field Index (VFI) 247, 250 trend line 168-170 visual function 22 t-test 170-171 visual impairment 78 trial lens 29, 33-34, 41-42 trial lens calculator 41-42 W trigger-happy 40, 138, 142-143 white-on-white perimetry, see SAP t-test 170-171 widespread defect, see diffuse defect tumor 67, 231, 278-279 two-zone strategy, see 1-level test U unreliable visual ields, see untrustworthy visual ϔield untrustworthy visual ield 136-140 false negative answers 124-125, 138-139, 225, 248 false positive answers 45, 123, 125, 138-139, 177, 224-225, 247 fatigue 38 ixation loss 35, 39, 45, 248 incorrect patient age 40-41 lack of attention 40 learning 37-38, 243 lens rim artifact 43, 45 lid artifact 43-44, 77, 177 pupil size 44 refractive error 40-42 set-up errors 28, 40, 135-136 trigger-happy 40, 138, 142-143 V Values 101-102 vector 207-209, 219-223 visual ield abnormal 11, 19, 145-148 age-dependency 18-19 binocular 8-9, 74 MEMBERS OF HAAG-STREIT GROUP HAAG-STREIT Holding AG CLEMENT CLARKE Ltd. www.haag-streit-holding.com www.clement-clarke.com HAAG-STREIT AG, Diagnostics ComLab AG www.haag-streit.com www.comlab.ch HAAG-STREIT AG, Verkauf Schweiz HS DOMS GmbH www.haag-streit.ch www.hs-doms.com HAAG-STREIT Deutschland GmbH IPRO GmbH www.haag-streit.de www.ipro.de HAAG-STREIT Far East John Weiss Ltd. www.haag-streit-fareast.com www.johnweiss.com HAAG-STREIT Medtech AG Möller-Wedel GmbH & Co KG www.haag-streit-medtech.com www.haag-streit-surgical.com HAAG-STREIT Surgical GmbH Möller-Wedel Optical GmbH www.haag-streit-surgical.com www.moeller-wedel-optical.com HAAG-STREIT AG HAAG-STREIT UK OptoMedical Technologies GmbH Gartenstadtstrasse 10 www.haag-streit-uk.com www.haag-streit-surgical.com 3098 Koeniz Switzerland HAAG-STREIT USA Reliance Medical Inc. Phone +41 31 978 01 11 www.haag-streit-usa.com www.haag-streit-usa.com Fax +41 31 978 02 82

Asetronics AG SPECTROS AG www.haag-streit.com www.asetronics.ch www.spectros.ch