Bethany M. Erb, Elaine Botros, Thomas F. Saunders, Anna-Maria Haas, Rick Voland, Rachel Linderman, Barbara A. Blodi, Amitha Domalpally, Karl G. Csaky

Abstract

Purpose: Change in geographic atrophy (GA) area as measured by fundus autofluorescence (FAF), a primary outcome in clinical trials, does not correlate with visual function. Preservation of photoreceptors, particularly in the foveal region, is critical for maintaining visual function. The integrity of the ellipsoid zone (EZ), a hyperreflective layer at the inner-outer photoreceptor junction, may reflect this preservation. This study introduces Ellipsoid Zone Integrity Index (EZII), a novel OCT biomarker that reflects the integrity of the EZ in the 1 mm and 3 mm perifoveal zones. This study assesses EZII correlation with best corrected visual acuity (BCVA) and low-luminance visual acuity (LLVA) to determine its utility as a clinical trial endpoint.

Methods: A manual edge-detection approach was used to delineate ellipsoid zone loss on Heidelberg Spectralis OCT images for 2 visits from 43 subjects enrolled in a GA clinical trial (GSK 341) with no treatment effect. The ETDRS grid was used to delineate 1 mm (area: 0.79 mm2) and 3 mm (area: 7.07 mm2) circles around the fovea. Intact EZ area in the 1 mm and 3 mm circle was measured as a percentage to define the EZII. GA area was similarly delineated and measured for the whole image. EZII correlation with BCVA and LLVA was analyzed and compared with that of GA area.

Results: Mean GA area was 9.40 mm2 at baseline and 11.02 mm2 at 12-month follow up. Mean GA growth rate was 1.62 (SD 1.32) mm2. Mean EZII at baseline was 42% in the 1 mm zone and 38% in the 3 mm zone. At 12-month follow up, mean EZII was 34% in the 1 mm zone and 30% in the 3 mm zone. In the 1 mm area, baseline EZII was significantly correlated with BCVA (r2= 0.29, p-value < 0.001), but not with LLVA (r2 = 0.02). In the 3 mm area, EZII was positively correlated with BCVA (r2=0.17, p-value <0.01) and LLVA (r2 = 0.17, p-value <0.01). Corresponding GA area on OCT was neither correlated with BCVA nor LLVA.

Conclusions: Positive correlation between EZII in the 1 mm zone and BCVA and EZII in the 3 mm zone with BCVA and LLVA suggests EZII may be an important endpoint in GA analysis. The study’s contrasting correlation between EZII and LLVA in the 1 mm and 3 mm zone is likely due to the perifoveal photoreceptor distribution pattern. Further research is needed to validate the ellipsoid zone integrity index as a structure-function biomarker.