Abstract
Purpose :
Progression detection in advanced glaucoma is often challenging. Retinal Nerve Fibre Layer Optical Texture Analysis (ROTA), a novel algorithm, uncovers optical texture and trajectories of axonal fibre bundles with high sensitivity and specificity. In this 5-year longitudinal study, we investigated the performance of ROTA in identifying progression and characteristics of axonal fiber bundle losses in advanced glaucoma.
Methods :
95 eyes of 84 patients with advanced glaucoma (vertical cup-to-disc ratio ≥0.85 and visual field mean deviation ≤-12 dB) were followed for at least 5 years. ROTA and retinal nerve fibre layer (RNFL)/ganglion cell-inner plexiform layer (GCIPL) thickness measurements were obtained from the Triton optical coherence tomography (OCT) (Topcon, Japan) and Cirrus HD-OCT (Carl Zeiss Meditec, USA), respectively. Progressive loss in axonal fibre bundle optical texture in ROTA was detected and confirmed by three observers. Progressive RNFL/GCIPL thinning was determined by Guided Progression Analysis (GPA).
Results :
Of 95 eyes, progression was identified in 43 eyes (45.3%) by ROTA and 32 eyes (33.7%) by GPA of RNFL thickness and/or GCIPL thickness. Progressive RNFL loss on ROTA was found most often over the superior papillofoveal bundle (10 eyes) (23.3%) followed by the inferior arcuate bundle (9 eyes) (20.9%), the inferior papillomacular bundle (8 eyes) (18.6%), the inferior papillofoveal bundle (7 eyes) (16.3%), the superior arcuate bundle (6 eyes) (14.0%), and the superior papillomacular bundle (6 eyes) (14.0%). For eyes with concomitant progressive RNFL loss on ROTA and progressive RNFL/GCIPL thinning during the study period (n=26 eyes), 21 had progressive RNFL changes detected first by ROTA and 2 had progression identified on the same date; ROTA was on average 10.3 months (95% confidence interval 3.0-17.6 months) ahead of GPA of RNFL/GCIPL thickness in detecting glaucoma progression.
Conclusions :
ROTA could identify progressive axonal fibre bundle loss in advanced glaucoma that was not detectable in conventional progression analysis of RNFL/GCIPL thickness.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.