Purchase this article with an account.
Eleanor Kim, Ashley Sun, Zhichao Wu, Carlos Gustavo de Moraes, Robert Ritch, Donald C Hood; A Comparison of Qualitative and Quantitative Assessment of Glaucoma Progression Using Optical Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2018;59(9):2089.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare a qualitative to a common quantitative assessment of glaucoma progression using spectral domain optical coherence tomography (SD-OCT)
There were 2 sets of eyes: a “variability group” (VG) with scans 1.1±0.3 months apart (95 eyes from 94 patients, age=52.3±19.4 yrs) from the MAPS  database; and a “longitudinal group” (LG) with scans 17.3±5.3 months apart (79 eyes from 47 patients, age=61.4±13.4 yrs). The VG consisted of 54 glaucomatous or glaucoma suspect eyes and 41 healthy eyes, while all LG eyes were glaucomatous or glaucoma suspects, as determined by glaucoma specialists. All eyes had 3.5 and 4.7mm circle scans using Heidelberg Spectralis SD-OCT ART (Automatic Real-time Tracking), which allows for alignment of optic nerve head structures between exam dates. For the quantitative method, we calculated the difference between the global cpRNFL thickness on the 1st and 2nd scans. For the qualitative method, an experienced OCT reader rated the probability of progression on a scale of 0 to 100. For either method, LG eyes showing change exceeding the 5% lower limit of normal test-retest variability for both circles were called “progression.”
Similar sensitivities and specificities were observed for both methods, 20.3% (16/79) and 98.9% (94/95) for qualitative and 24.1% (19/79) and 97.9% (93/95) for quantitative (p=0.303). However, both methods only agreed on “progression” for 10 eyes. Examining circle scans for the 6 eyes identified as “progression” only by the qualitative method suggested that in 5 of these eyes the quantitative method missed clear local damage (e.g., Fig. 1). A similar examination of the 9 eyes identified only by the quantitative method suggested that 6 of these eyes were falsely classified as “progression” due to segmentation errors (e.g., Fig. 2).
The quantitative (global cpRNFL thickness) and qualitative methods (visual inspection) agreed in only 40% (10) of all (25) eyes identified as “progression” by at least one method. A post-hoc examination of scans indicated that the quantitative method was prone to errors due to segmentation problems and an inability to identify local progression, suggesting that a qualitative method shows promise for improving the accuracy of detecting progression.1. De Moraes et al., ARVO, 2017.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.
This PDF is available to Subscribers Only