Abstract
PURPOSE: To compare global trend-based and point-wise event-based analysis for detecting visual field progression in eyes with glaucoma. METHODS: The study included a cohort of 367 glaucoma eyes from 265 participants seen over a mean follow-up period of 10 years to develop a computer simulation model of "real-world" visual field results. Progression was evaluated with point-wise event-based analysis using the Guided Progression Analysis (GPA) and global trend-based analysis using the mean deviation (MD) and visual field index (VFI) measures. The specificities of the methods were matched based on a simulated dataset of stable glaucoma eyes, to allow an adequate comparison of their sensitivities for detecting progression. RESULTS: The 5-year cumulative false-positive rate for the GPA alert of "possible progression" and "likely progression" (significant change from baseline at two and three consecutive visits, respectively) were 34.0% and 7.0%, respectively. At matched specificities, 27.7% eyes were detected as having progressed after 5 years using the GPA "likely progression" criterion, while 24.6% and 23.8% were detected as having progressed using the global trend-based analysis with MD and VFI, respectively. There was a moderate level of agreement between the GPA and global trend-based analyses. CONCLUSIONS: Pointwise event-based and global trend-based methods had similar performances to detect glaucoma progression when rigorously matched for specificity. TRANSLATIONAL RELEVANCE: Although both point-wise event- and global trend-based analyses perform similarly, they could provide complementary information that could be exploited to improve the overall detection of progression in clinical practice and clinical trials.