Automated Visual Field Analysis for Glaucoma

Transcription

1 Island of Vision Automated Visual Field Analysis for Glaucoma Normal Visual Field Parameters 60 superior 60 nasal 75 inferior 100 temporal Macula the central 13 Fovea the central 3 Visual field is limited by the size of the retina and margins of the orbit 1

2 Pearls on Static Visual Fields Most visual fields test 0-51 decibels decibels is outside human vision 1 diopter of refractive blur in undilated patient A little more than 1 decibel of depression of the hill of vision With Goldmann III stimulus Leave cylindrical errors of less than 2 diopters uncorrected Adjusted with spherical equivalent Above 2 diopters correct the astigmatism with trial lens Background of a visual field illuminated (31.5 apostilbs) Minimum brightness for photopic or daylight Cones are isolated, test photopic system More on contrast, less on absolute brightness Changes in pupil size, crystalline lens color and transparency have less effect on result 24-2 vs 30-2 vs 10-2 Static Visual Field 30-2 tests 76 locations 24-2 tests 54 locations Tests 30 degrees nasal Little diagnostic information lost in 24-2 Time is saved; Fewer trial lens and lid artifacts 24-2 has become the VF for glaucoma Only a small percentage of glaucomatous defects occur in the peripheral visual field alone Only down side, 30-2 can sometimes find progression earlier due to more test points 10-2: Measures 10 degrees temporally and nasally and tests 68 points. Used for macula, retinal and neuroophthalmic conditions and advanced glaucoma A thousand points of light SAP and SITA SAP- Standard Automated Perimetry Determines the threshold (how dim of light) can be seen at various points Various algorithms have been developed to determine this threshold using few to numerous individual points in a single visual field test SITA-Swedish Interactive Thresholding Algorithm Optimizes the determination of perimetry thresholds Continuously estimating what the expected threshold is based on the patient's age and neighboring thresholds Reduce the time necessary to acquire a visual field by up to 50%. Decreases patient fatigue and increases reliability SITA mode is now widely used in many computerized automated perimeters SITA- can be applied to: SAP- Standard Automated Perimetry SWAP-Short Wavelength Automated Perimetry (SWAP) SITA Standard versus SITA Fast SITA strategies are twice as fast as previous strategies SITA fast takes 67% the time of SITA standard Sita fast has larger retest variability Primary difference is between the two strategies is the amount of certainty that is required before testing is stopped SITA standard More precise More tolerate of mistakes Easier test as stimuli are brighter SITA -Faster now available 2

3 SITA Faster tests in 2 minutes or less without compromise to test results Two minute test for near normal patients ~50% faster than SITA Standard; ~30% faster than SITA Fast Clinically equivalent to SITA Fast and Standard Same SITA algorithm and normative data as Standard and Fast Removes unnecessary dead time during the test No Blind Spot or False Negatives - Uses Gaze Monitoring and False Positives for test quality monitoring Mixed SITA GPA Reports Allows mixing all SITA test strategies for GPA reports Helps immediately adopt SITA Faster Clinical equivalence of tests allows intermixing The patient needs a near correction. The perimeter will tell you what to use based upon the patient s age and distance correction When in doubt, look at the last field and use that trial lens. It must be adjusted as close to the eye as possible. 3

4 Interpreting the single field analysis Reliabiltiy parameters Raw data Grey scale Total deviation and probabiltiy display Pattern deviation and probability display Glaucoma hemifield test Global indices Gaze Tracker 4

5 Global Indices MD 54 spots on 24-2 All 54 spots reduced by 1 DB (54DB) MD 1DB MD and PSD 54 spots on spots reduced by 2 DB (54 DB) MD 1 DB 54 spots on spots reduced by 4 DB (54DB) MD 1 DB PSD Low PSD (Generalized loss) 1.00 DB Moderate PSD (More localized loss) 3.00 DB High PSD (Localized loss) 5.00 DB Thoughts on Mean Deviation (MD) What is the Mean Deviation on a visual field of a blind eye? A Word About the Grayscale- It s Useful 5

6 The Four Rs of Visual Fields Reliable Reliability parameters vs. gaze tracker vs. comments New thoughts dismiss Reliable and Unreliable Highly useful information to No useful information Very subjective determination Recognizable As glaucomatous defects rather than something else Relatable Correlate field to anatomy to support or refute defects Reproducible Reproducibility is perhaps the greatest indicator of reliability Reproducibility increases the likelihood that insignificant defects are indeed significant. Probability Plot Butterfly/Cloverleaf The patient is zoning out 6

7 Gaze Tracking Recognizable Relative scotomas Fluctuating scotomas Absolute scotomas Paracentral scotomas Nasal steps Arcuate scotomas (Bjerrum s scotomas) Altitudinal defects General depression and diffuse loss is actually very rare in glaucoma and more indicative of cataract, miosis, or other media/refractive issues. 7

8 Patternless Fields 8

9 Shallow fluctuating scotma What Does This Tell Us? Cataract Post-op 9

10 Recognizable 68 YOWF Treated for glaucoma at various facilities Old records obtained Pressure excellent Disc pallor OU Old longstanding familial optic atrophy Consider neuro-ophthalmology consult Continue current medications So, where is the lesion? So, where is the lesion now? Apparently, nowhere Don t forget to pull the omas out of glaucoma Relatable 10

11 Visual Fields Don t automatically call all depressions, isolated missed points, especially if there is a structural correlate GHT: Within Normal Limits Reproducible Don t automatically call all depressions, isolated missed points, especially if they are repeatable GHT: Within Normal Limits on all fields Avoid making a snapshot judgment based upon 1 field 20/50 11

12 20/20 OD, OS 10-2 SS OS 10-2 SS OD Visual Field Progression Angle recession glaucoma: IOP 47 mm Pt non-compliant and declines surgery 12

13 Is This Field Progressing? Guided Progression Analysis: GPA Designed to help identify clinically significant progression of visual field loss in patients with glaucoma Highlights changes from selected baseline examinations that are larger than typical clinical variability in patients with similar degrees of glaucoma. Identifies consistent and repeated patterns of loss Can be used on full threshold, SITA Standard, and SITA Fast strategies Event analysis and trend analysis 13

14 Visual Field Index (VFI) Central points weighted more heavily than those on periphery Reduces cataract contribution to the measurement of VF loss A B HFA GPA VFI Summary - Interpretation at a Glance VFI = 90% VFI = 81% B has more damaged central points and lower VFI than A. The VFI Regression Plot VFI plotted against age Extrapolates rate of change up to 5 years 14

15 HFA GPA VFI Summary - Interpretation at a Glance Slow progression may not necessarily be vision threatening. Loss to date Projected future loss 100% Event analysis (GPA alert) indicates Likely Progression. Slope is nearly flat and the confidence interval is narrow. Patient is 75 years old. The VFI Bar historical and projected VFI loss If current therapy is continued, what is this patient s risk of visual impairment? Updating baseline after significant treatment change Updating baseline after significant treatment change Severely depressed visual field on latest exam First 4 exams showed fast glaucoma progression, followed by change in treatment. Is the patient now stable, or is further treatment escalation appropriate? BEFORE, with default baseline selections AFTER adjusting baselines Excluding non-representative exams Life expectancy can be an important consideration. Event analysis (GPA Alert) indicates Likely Progression. Progression is 3.0 ± 0.9% /yr Age at most recent exam is 65 BEFORE, with poor exam included AFTER excluding the poor exam A progression rate that might be acceptable at age 85 may not be acceptable at age

16 % Loss 6/21/2018 Imaging Technology in Glaucoma Diagnosis and Management Structural/ Functional Relationship in Glaucoma as the Disease Progresses Visual Field changes occur late in the disease The Optic disc often changes before visual fields The RNFL usually changes before both the visual fields and optic disc VF Time Early Moderate Severe 122 Dominant Imaging Technologies Optical Coherence Tomography Time Domain vs Spectral Domain RNFL, Disc topography, retinal imaging, ganglion cell complex, macular thickness Scanning Laser Polarimetry GDx VCC; GDx ECC RNFL only Scanning Laser Topography Disc topography (minimal retinal applications) Optical Coherence Tomography (OCT): Interferometry summary Light beam is simultaneously sent to the eye and a reference mirror. The light penetrates through the ocular tissue layers and is reflected back. The returning light is compared to the reference light and allows a computer reconstruction of the underlying tissue. 16

17 Optical Coherence Tomography Optical Coherence Tomography (OCT) Non-invasive diagnostic imaging technique first described in 1991 by Huang et al. Provides high-resolution crosssectional and topographic images of retinal and optic nerve tissues. Analogous to ultrasound StratusOCT (Carl Zeiss Meditec, Dublin, CA) is the the 3 rd generation machine, which received FDA approval in May OCT Image of the Retina vs. Histology Carl Zeiss Meditec Cirrus Topcon Mark II Optovue RTVue-100 Nidek RS-3000 Optipol\Cannon Copernicus Heidelberg Spectralis OPKO/OTI Spectral OCT/SLO Bioptigen 129 Identification of Retinal Layers Faster Computer and Operating System Smaller Footprint and Improved OCT Image Quality NFL ILM GCL IPL OPL IS/OS RPE/CC Choroid Stratus OCT NFL: Nerve Fiber Layer IS/OS: Junction of inner and outer ILM: Inner Limiting Membrane photoreceptor segments GCL: Ganglion Cell Layer RPE: Retinal Pigment Epithelium CC: Choriocapillaris IPL: Inner Plexiform Layer OPL: Outer Plexiform Cross-sectional image of live tissue; a virtual biopsy

18 Posterior Segment Applications Vitreous/Vitreoretinal Interface Neurosensory retina, RPE Choriocapillaris Optic Nerve/RNFL/ GCC 133 OCT Database Information Spectralis: 201 patients All Caucasian Age New database more representative of US population Cirrus: 284 eyes Age Ethnic Groups: Causasian, Asian, African-American, Hispanic RTVue: 600 eyes Disc Size African-American, Chinese, Japanese, Caucasian, Hispanic, Indian Glaucoma RNFL Thickness Analysis OPTIC DISC CUBE SCAN The 6mm x 6mm cube is captured with 200 A-scans per B-scan, 200 B-scans. RNFL OU Analysis RNFL THICKNESS MAP shows the patterns and thickness of the nerve fiber layer within the full 6mm x 6mm area RNFL thickness and comparison to normative data is shown in circle, quadrants and clock hour display Old Printout* CALCULATION CIRCLE AutoCenter function automatically centers the 1.73mm radius peripapillary calculation circle around the disc for precise placement and repeatable registration. The placement of the circle is not operator dependent. Accuracy, registration and reproducibility are assured. RNFL DEVIATION MAP, overlaid on the OCT fundus image, illustrates precisely where RNFL thickness deviates from the normal range. Data points that are not within normal limits are indicated in red and yellow. RNFL thickness along the calculation circle is displayed in graphic format and compared to age-matched normative data 18

19 Optic Nerve Head and RNFL OU Printout Combined report using the Optic Disc 200x200 cube scan Contemporary printout* Optic Nerve Head Calculations The disc edge is determined by the termination of Bruch s membrane. This is validated in the literature. The rim width around the circumference of the optic disc is then determined by measuring the amount of neuro-retinal tissue in the optic nerve. This differs from other methods that determine the cup margin based on its intersection with a plane at a fixed distance above the disc. In this method, the disc and rim area measurements correspond to the anatomy in the same plane as the optic disc. BURGOYNE & CHAUHAN ARE NOW ADVOCATING OPTIC NERVE ANALYSES BASED UPON ANATOMICAL LANDMARKS THAT PREVIOUSLY COULD NOT BE IDENTIFIED CLINICALLY. SOME ACADEMICS ARE NOW ARGUING THAT FDOCT CAN DOCUMENT OPTIC NERVE & RETINAL ANATOMY BETTER THAN STEREO COLOR PHOTOS. Burgoyne, Chauhan, & colleagues reminded us all that 3-D is better than 2-D The clinically perceived disc margin is most likely not the innermost edge of Bruch s membrane detected by SD-OCT. (Ophthalmology APR2012) BMO-MRW (minimum rim width based upon BM opening) quantifies the neuroretinal rim from a true anatomical outer border... (IOVS APR2012) Bal Chauhan is at Halifax and is an optometrist. IOVS APR2012 Green = Stereo Photo Red = OCT Analysis Elements Analysis Elements RNFL thickness map with cup and disc mask OCT en face fundus image shows boundaries of the cup and disc, and RNFL calculation circle integrated with the RNFL thickness deviation map RNFL Peripapillary Thickness profile, OU Matched to normative data Nero-retinal Rim Thickness profile, OU Optic Nerve Head calculations are presented in a combined report with RNFL thickness data. Key parameters are displayed in table format RNFL Quadrant and Clock Hour average thickness - Matched to normative data 19

20 RTVue-100 Fourier-Domain Optical Coherence Tomography (OCT)

21 Spectralis OCT

22 Ganglion Cell Analysis Measures thickness for the sum of the ganglion cell layer and inner plexiform layer (GCL + IPL layers) using data from the Macular 200 x 200 or 512 x 128 cube scan patterns. RNFL distribution in the macula depends on individual anatomy, while the GCL+IPL appears regular and elliptical for most normals. Thus, deviations from normal are more easily appreciated in the thickness map by the practitioner, and arcuate defects seen in the deviation map may be less likely to be due to anatomical variations. Ganglion Cell Analysis The analysis contains: Data for both eyes (OU) Thickness Map - shows thickness measurements of the GCL + IPL in the 6mm by 6mm cube and contains an elliptical annulus centered about the fovea. Deviation Maps - shows a comparison of GCL + IPL thickness to normative data. Thickness table - shows average and minimum thickness within the elliptical annulus. Sector maps - divides the elliptical annulus of the Thickness Map into 6 regions: 3 equally sized sectors in the superior region and 3 equally sized sectors in the inferior region. Values are compared to normative data. Horizontal and Vertical B-scans. Ganglion Cell Complex (GCC) GCC at the Macula Has over 50% of all retinal ganglion cells - Nerve Fiber Layer = Ganglion Cell axons - Ganglion Cell Layer = Ganglion Cell bodies - Inner Plexiform Layer = Ganglion Cell dendrite GCC (all three layers) becomes thinner as Ganglion Cells die Ideal region to detect early loss or mild changes over time 22

23 Higher Resolution (Fourier) allows isolation of the GCC layer GCC versus NFL - RNFL distribution depends on individual anatomy - GCC complex appears regular and elliptical for most normals Thickness Map RTVue-100 Raw Data Significance Map RTVue-100 Thicker Maps statistical significance of change from normal W/in normal range p>5-95% Thinner Borderline results p<5% Outside normal limits p<1% Thicker GCC peri-foveally in normal eyes No GCC at fovea RTVue-100 Thickness Map Deviation Map Significance Map RTVue-100 Raw Data Percent loss from normal Maps statistical significance of change from normal

24 RTVue-100 Ganglion Cell Analysis Cirrus Cirrus Spectralis Spectralis Spectralis software for the Heidelberg Spectralis that looks specifically at the macular regions. 3 (OCT instruments from Zeiss and Haag-Streit have also added macular analysis software, although these focus specifically on the ganglion cell layer.) 173 This new software analyzes the central 20 degrees of the retina, with that area divided into an 8mm x 8mm grid of three-degree sectors. Each sector has a calculated thickness, and most importantly, each sector is compared to its comparable sector in the opposite hemifield of the same eye, as well as the same hemifield of the fellow eye

25 Spectralis Interpretation is a 3 step process 1. Understand what the printout says 2. Apply experience and value judgement 3. Correlate to the clinical findings Imaging Pearls You cannot make a diagnosis of glaucoma based solely upon imaging results. The use and overemphasis of imaging technology to the exclusion of additional clinical findings and assessment of risk will put patients in peril. Exactly how much confidence should an OCT give you as to whether or not a patient has glaucoma? Depends how much confidence you had before you imaged the patient Issues in Imaging What to look for when interpreting OCT scans Normative Database Signal Quality Blink/Saccades Segmentation Errors Media Opacities Axial Length Quality score Illumination Focus clarity Image centered Any signs of eye movement Segmentation accuracy B Scan Centration Missing data Media issues Maculopathy for GCC scans

26 RTVue-100 Eye movement 181 Cirrus Cirrus Accidentally find CSC when looking for glaucoma 184 RTVue

27 Cirrus Spectralis Spectralis Spectralis Floor effect

28 61 YOF referred for suspicious discs Don t make clinical decisions based upon bad data YOM Referred for glaucoma consult IOP 34 mm Hg OD; 26 mm Hg OS CCT: 492 OD, 483 OS 197 Fields Normal OD, OS RNFL defect 28

29 Normal disc and RNFL OS Sparse/ abnormal RNFL OD Normal RNFL OS Abnormal GCC OD (matches RNFL) and normal GCC OS; Dx: Pre-perimetric glaucoma OD and OHTN OS Its nice when things agree- RNFL via photo, RNFL via OCT, and GCC Help! The Diagnostic Imaging Doesn t Agree with my Diagnosis! Low risk OHTN Local OD wants imaging for baseline GCC may allow us to diagnose glaucoma even earlier than RNFL OCT RNFL normal but markedly abnormal GCC OS Same patient, same day, same quality, GCC now normal 29

30 63 YOF: Healthy Discs and Normal IOP

31 POAG CRAO Red Disease A New Clinical Non-Entity OCT Example of Red Disease A supratentorial, non-glaucomatous masquerade non-disease Afflicts the educated patient (especially with Internet access) with good health care plans and/or wealth Debilitating to the patient and painful for the visual care provider to treat 44 YOM No family hx 20/20 OD, OS Perrl (-RAPD) SLE/ gonio: normal CCT: 531, 540 TA: OD: 14 OS: Journal of Irreproducible Results and Senseless Studies 8/10 8/10 10/10 9/10 31

32 Green Disease An Insidious Clinical Entity A glaucomatous process masquerading as non-disease Afflicts inexperienced, poorly-educated, and lazy doctors who simply want a machine to make all clinical decisions for them Debilitating to the patient and painful for the visual care provider, but a boon for malpractice attorneys Journal of Irreproducible Results and Senseless Studies Help! The Diagnostic Imaging Doesn t Agree with My Diagnosis! 56 YOM- Glaucoma suspect since 2012 Is this person really a glaucoma suspect? A example of Green Disease 32

33 Green Disease Green Disease Green Disease Red + Green = Yellow Disease? 33

34 panomap Version 8.0 panomap Version Updated Guided Progression Analysis (GPA ) Optic Nerve Head information now included Updated Guided Progression Analysis (GPA ) Page 1 Page 2 Average Cup-to-Disc Ratio plotted on graph with rate of change information. RNFL/ONH Summary includes item Average Cup-to-Disc Progression. Printout includes an optional second page with table of values, including Rim Area, Disc Area, Average & Vertical Cup-to-Disc Ratio and Cup Volume. Each cell of the table can be color coded if change is detected. Miscellaneous updates to the report design Carl Zeiss Meditec, Inc Cirrus 6.0 Speaker Slide Set CIR.3992 Rev B 01/

35 OCT to Verify Glaucoma the Optic Nerve Head? New Ophthalmologist, September 2007, p Optic Nerve Exam to Verify OCT? OCT Imaging Take Home Points Serial overlays/imaging to determine baseline (intra-session) noise Good signal strength Good segmentation without errors Optic nerve head exam for disc hemorrhage, pallor, myopic, and tilted nerve heads Determine structure-function correlation Follow all ancillary tests visual fields and optic nerve head photos for progression 35

36 Imaging Take Home Points No one single parameter is more important than the others. Never base a clinical decision based upon only one piece of data. This is not a Silicon Valley Rumplestilskin. You cannot put in straw and get out gold! Interpretation is a three-step process Cautions About Imaging No current technology is better than the human eye and common sense No device is better than a skilled and experienced clinician. Beware of Red Disease Treat Real Disease and not Red Disease Likewise beware of Green Disease Know the limitations of the technology: normative database, reproducibility, resolution, quality of imaging Technologies come and go OCT and any other technology is only as good as the clinician using it. 36