Interpreting SITA Plots - OPT505 Lecture 11 PDF

OPT505 Lecture 11: Interpreting SITA plots Ellie Livings Intended Learning Outcomes Understand what is meant by SITA and ZATA Understand how to set-up px for automated perimetry Understand how various factors can affect results Which testing programme to select How to identify reliable...

OPT505 Lecture 11: Interpreting SITA plots Ellie Livings Intended Learning Outcomes Understand what is meant by SITA and ZATA Understand how to set-up px for automated perimetry Understand how various factors can affect results Which testing programme to select How to identify reliable results How to interpret results Work through case scenario  NA-UP-JR Sir, we need to do a fields test……  NA-UP-JR Recap Frequency of seeing curve, threshold v supra-threshold  NA-UP-JR Why use a threshold test? Supra-threshold are normally screening tests Threshold provide more detail and are more sensitive to more subtle loss Good to monitor condition, and detect early loss Finding the threshold can take a long time-recap, staircase method  NA-UP-JR Humphrey Field Analyzer (HFA) SITA is effectively the gold standard for field evaluation in glaucoma, and is widely used in research and trials Humphrey Field Analyzer (HFA) NA-UP-JR  NA-UP-JR What is SITA? SITA 24-2 Swedish Interactive Threshold Algorithm Name of the Design of algorithm pattern: 24 2 tests 24 second test degrees pattern: straddles A way to speed up the test by starting closer to the threshold: temporally H&V meridians Uses a Bayesian approach combined with staircase algorithm Does not use catch trials for false positives Uses adaptive inter stimulus interval  NA-UP-JR What is SITA? Continuously estimates next response based on px age and adjacent thresholds. Quicker (up to 50% wrt other threshold methods) Less px fatigue Increased reliability Do not need to perform ful threshold: this is just as accurate, but much quicker  NA-UP-JR Henson 9000, ZATA NA-UP-JR  NA-UP-JR SITA v ZATA SITA standard SITA fast ZATA (Zippy adaptive threshold algorithm) Type of test Threshold threshold threshold Test pattern 24-2 or 30-2 24-2 or 30-2 24-2, 30-2 (can extend) Stimuli White on white White on white White on white algorithm 4-2dB step, uses estimation 4dB step, uses estimation Uses prior px data when of sensitivity, requires more of sensitivity, accepts less available, focusses on accuracy at individual accuracy at individual accuracy in areas of loss. locations locations Normal points/v poor points not tested as deeply. Pros/cons Slower, but more accurate Less adaptive, quicker but Shallow defect poss less for individual thresholds less reliable* than SITA obvious wrt SITA, very fast standard Additional points in central Not many central points 10 degrees tested Machine HFA HFA Henson  NA-UP-JR 1. Heijl A, Patella VM, Chong LX, et al. A new SITA perimetric threshold testing algorithm: Construction and a multicenter clinical study. Am J Ophthalmol 2019;198:154-165.  NA-UP-JR HFA SITA  NA-UP-JR Set-Up Key points: ✓ Trial lens: Working distance is about 30cm so THE NEAR RX !! Including cyls if over about 1.50DC, may be plano, or even Patients get fatigued minus. Some perimeters calculate for you. easily, especially the ✓ Lens position: Full aperture, close to px (so field not longer the test goes on obstructed: Rim defect) The usefulness of ✓ Px position: Check they are comfy, chin rest, table height, perimetry declines if the can they see the fixation light? px’s ✓ Eye patch: is the elastic in their eye, are they properly reliability/concentration occluded? is poor ✓ Lighting: Make sure the lights are off as you set them up Think about ✓ Button: put it in their hand, ‘beep’ it so they see how. environment: Noisy, ✓ Px instructions: clear and simple, frequent reminders plus crowded, other px’s encouragement. (Remember, they worry about ‘missing Correct set-up makes a points’) HUGE difference ‘’Don’t worry if you miss a light, or you’re not sure. The most important thing is to look at the main light all the time, ok? ‘’ Tip: Read the manual ……. NA-UP-JR  NA-UP-JR Quick, Suprathreshold test: 40 points over central 30° Useful for screening where no defect C40 suspected (baseline normal) If defect is detected, may need to investigate further Also for gross defects e.g. hemianopia If px can’t manage longer test Longer (approx. 2-5 mins) Threshold test (SITA algorithm) More data-tests 54 points over central 24° (out to 30° nasally) points 6° apart Sita 24-2 Grid of points straddles midlines: nasal step/hemianopia Statistical analysis Used for glaucoma, monitoring, neurological defects  NA-UP-JR What field defect are you trying to catch?  NA-UP-JR Where do we test? Point pattern How do we test? Testing strategy  NA-UP-JR Fields analysis: Before jumping into the field plot, consider all clinical evidence: History Family history: Glaucoma? Onset: sudden, slow? Which eye? Monocular or binocular? Has the px spontaneously reported the loss? Pain? Medical issue: Stroke/trauma/diabetes? Clinical findings ONH appearance: Swollen? Cupped? Pale? Tilted? Myopic? Media opacities: cataract, corneal pathology Normal pupils? Retinal appearance: macula pathology? Haemorrhages? Refractive findings Amblyopia VA: Sudden drop? Light perception? Difference between eyes? Colour vision  NA-UP-JR How not to catch rubbish!  NA-UP-JR Interpreting the fields: is it good data? Sometimes px are rubbish at fields. They don’t concentrate, they don’t listen to instructions, they fidget… Don’t keep the test going ( esp longer ones like SITA 24-2) if they have multiple fixation losses. Our local glaucoma shared care scheme POCGS requires 20% fixation loss is unreliable – Blind spot method Blind spot is located at the start of the test. The machine will periodically present stimuli to the blind spot. Fixation loss is assumed if px responds. – Gaze tracking method Monitors position of eye and no. of blinks not good if small pupils, dry eye, ptosis  NA-UP-JR Reliability: Catch trials Can you rely on the ‘clicks’? False positive False negative Periodically the perimeter produces a sound but light stimulus is Periodically a stimulus is presented where the threshold has not presented. already been determined (you saw it before, but not now) If a patient responds to a non-existent stimulus, a false positive is This stimulus is brighter than the measured threshold recorded If the patient fails to respond to this stimulus, a false negative is A patient is called ‘trigger happy’ when too many false positives, recorded often because they are anxious about the test and trying to ‘do well’ > 33% is unreliable > 15% is unreliable and you may need to restart the test and Often indicates px is getting tired repeat instructions. May occur in px with significant field loss Reassure px it doesn’t matter if they don’t see all the lights, just clink if they think they’ve seen one  NA-UP-JR Deviation plots Total deviation: Patten deviation Is how much you varied from the database of age-matched ‘normals’ Once we account for factors such as media opacites/small pupils, this is what is left over. The top plot is the difference, in dB, between the numeric scale values and the age-corrected normal values (how deep) Plot is adjusted for any changes in the height of the hill of vision. The lower plot is the probability of the deviations occurring by The most useful bit to look at, as it gives the true ‘shape’ of any chance. The darker boxes are more likely a ‘real’ defect defects. (not the depth) It often looks ‘less bad’ than the total deviation. Important to compare these two plots to differentiate focal loss from diffuse loss P values shown on legend: significance of the field defect  NA-UP-JR Global indices – diffuse loss Mean deviation (MD) How different is this field from normal? MD is the difference between the average measured threshold values and the population normal threshold values Values which are greater than ± 2.00 dB are usually abnormal P value: significance of diffuse defect MD becomes more negative in the presence of cataract and large areas of focal loss  NA-UP-JR Global indices – focal loss Pattern Standard Deviation (PSD) How irregular/damaged is this field? PSD is an index which is sensitive to non-uniformity (localized or focal defects) in the visual field Values which are greater than + 2.00 dB are usually abnormal P value: significance of focal defect PSD becomes more positive in the presence of focal loss  NA-UP-JR Global indices - Glaucoma Hemifield Test (GHT) Could this be glaucoma? GHT compares groups of corresponding points above and below the horizontal meridian to assess for significant difference which may be consistent with glaucoma. It can be: ‘Within normal limits’ (all areas similar) ‘Borderline’ when some differences between areas are present but is not significant: p>0.01 but p 33%)? Is there a lens rim artifact or uncorrected ptosis? If you didn’t do the fields, ask the technician how they did. If the fields appear reliable, continue and analyze data. Look at the sensitivity map to determine whether the field is within normal limits. If the fields are within normal limits, there is no further analysis. If one or both of the eyes exhibit abnormal fields, continue.  Interpreting the data Identify the shape of the visual field defect. Refer to the visual field pathway chart to determine the likely region of damage to the visual pathway. Compare these visual fields with each of the patient's previous visual field tests to identify progression of visual field loss. Do not take a shortcut by comparing these fields to only the most recent visual field, as this may be misleading. Generally six or more visual field tests are necessary to evaluate disease progression. Consider the findings in the context of the physical exam findings and the results of other tests and imaging. If there is uncertainty, consult with colleagues. https://eyerounds.org/tutorials/VF-testing/ Px found to have occludable angles on OCT: Referred for ……. Interpreting the fields: Reliability Interpreting the fields: Global Indices OS OD Deviation plots Total deviation: Patten deviation Is how much you varied from the database of age-matched ‘normals’ Once we account for factors such as media opacites/small pupils, this is what is left over. The top plot is the difference, in dB, between the numeric scale values and the age-corrected normal values (how deep) Plot is adjusted for any changes in the height of the hill of vision. The lower plot is the probability of the deviations occurring by The most useful bit to look at, as it gives the true ‘shape’ of any chance. The darker boxes are more likely a ‘real’ defect defects. (not the depth) It often looks ‘less bad’ than the total deviation. Important to compare these two plots to differentiate focal loss from diffuse loss P values shown on legend: significance of the field defect Global indices – diffuse loss Mean deviation (MD) How different is this field from normal? MD is the difference between the average measured threshold values and the population normal threshold values Values which are greater than ± 2.00 dB are usually abnormal P value: significance of diffuse defect MD becomes more negative in the presence of cataract and large areas of focal loss  Global indices – focal loss Pattern Standard Deviation (PSD) How irregular/damaged is this field? PSD is an index which is sensitive to non-uniformity (localized or focal defects) in the visual field Values which are greater than + 2.00 dB are usually abnormal P value: significance of focal defect PSD becomes more positive in the presence of focal loss  Global indices - Glaucoma Hemifield Test (GHT) Could this be glaucoma? GHT compares groups of corresponding points above and below the horizontal meridian to assess for significant difference which may be consistent with glaucoma. It can be: ‘Within normal limits’ (all areas similar) ‘Borderline’ when some differences between areas are present but is not significant: p>0.01 but p