Abstract
Purpose :
Intact EZ provides candidate outcome measures for RP clinical trials. This study compared EZ measurements acquired using different scan settings (SSs) to inform the design of NAC Attack a phase-3 multicenter RP trial.
Methods :
Spectralis OCT images were acquired from 11 RP patients enrolled at Wilmer Eye Institute. 2 foveal-centered volume scans (VSs) (30° x 25°/121B-scans & 20°x20°/ 97B-scans) were acquired on the same day on each eye, with ART 16 and HR mode. Duke Reading Center graded 2 EZ parameters after automated segmentation with manual correction of the intact EZ on each B-scan: the enface EZ area calculated from segmented B-scans; and the EZ width measured as the horizontal width of intact EZ on the centermost B-scan. Pearson correlation coefficients (PCC) between EZ-width and EZ-area were estimated. Agreement between SSs was assessed using intraclass correlation coefficients (ICC(1,1)) and Bland-Altman (B-A) plots. Mean differences (MD) between SSs were estimated using mixed effect models.
Results :
Demographics and best corrected acuity are summarized in Fig 1. On 20°x20° scans, mean EZ-width and EZ-area was 1419.3 µm (coefficient of variation [COV]=0.705) and 1.77 mm2 (COV=1.20), respectively. On 30° x 25° scans, mean EZ-width and EZ area was 1467.8 µm (COV=0.714) and 1.85 mm2 (COV=1.18), respectively. PCC between EZ-width and EZ-area was 0.94 in either SS. Fig 2 shows B-A plots. MD between SSs was -47.5 (95%CI -156.9, 61.8)µm for EZ-width and -0.08 (95%CI -0.21, 0.06)mm2 for EZ-area. The ICC was 0.92 for EZ-width and 0. 87 for EZ-area.
Conclusions :
With either SS, EZ-width and EZ-area were highly correlated; EZ-width had smaller variability than EZ-area. These data support foveal centered EZ width as an endpoint in RP clinical trials. Between SSs, both parameters had high ICC but measurements were generally smaller on 20°x20° scans than those from 30° x 25° scans. With the small sample size, the MDs between SSs were not statistically significantly different from 0, but the data suggest that measurements may not be comparable between the 20°x20° and 30°x25° SSs. Therefore, it may be preferable to use a single SS throughout a clinical trial. Selection between SSs should consider the tradeoff of longer acquisition time associated with the larger scan region vs. its ability to cover a larger region and hence observe earlier-stage disease with more preserved EZ.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.