Purchase this article with an account.
Jonathan Denniss, Allison M McKendrick, Andrew Turpin; High resolution suprathreshold perimetry for mapping residual visual field in advanced glaucoma. Invest. Ophthalmol. Vis. Sci. 2020;61(7):3887.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Conventional perimetric sensitivity measurements in areas of low sensitivity are unreliable and time-consuming. In advanced glaucoma, an alternative approach measuring spatial extent of residual vision may be advantageous. Here we test the hypothesis that a high-resolution suprathreshold approach can more accurately and efficiently map the spatial extent of residual visual field than existing procedures in advanced glaucoma.
Empirical data from 97 patients with Mean Deviation worse than -10dB were used as input to computer simulations comparing two high resolution (1.5° grid, central 27°) suprathreshold procedures to interpolated Full Threshold (FT) 24-2. Each procedure was simulated 200 times per patient. Spatial Binary Search (SpaBS) presented at locations bisecting seen/unseen points in any direction until all neighbours’ seen status matched or tested points were adjacent. SupraThreshold Adaptive Mapping Procedure (STAMP) presented stimuli in spatial regions with maximum entropy and modified the status of all local points after each presentation, stopping after 200 presentations. Both new procedures presented stimuli only at 20dB. Accuracy was defined as the percentage of locations whose seen status matched that of the interpolated input data thresholded as seen/unseen at 20dB. Repeatability was defined as abs(2 x proportion seen over repeat tests -1), such that 1 represented perfect repeatability and 0 represented chance.
Full Threshold used median 310 (interquartile range [IQR] 275-335) presentations vs. 284 (IQR 259-309) for SpaBS and 200 for STAMP (both p<0.001). Mean accuracy was slightly better with STAMP (median 93%, IQR 89-96%) compared to FT (median 91%, IQR 87-94%, p=0.03), but was worse with SpaBS (median 83%, IQR 80-86%, p<0.001). Mean repeatability was similar between FT (median 0.89, IQR 0.82-0.93) and STAMP (median 0.89, IQR 0.81-0.93, p=0.79), but was reduced for SpaBS (median 0.68, IQR 0.63-0.73, p<0.001).
The present STAMP procedure maps the spatial extent of advanced visual field defects slightly more accurately and within substantially fewer presentations compared to conventional tests with post-hoc spatial interpolation. Additional advantages in patient comfort with suprathreshold testing may also be conferred. Further work is necessary to refine STAMP, particularly in the trade-off between speed and accuracy.
This is a 2020 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only