Abstract
The symmetry of macular atrophy growth-rates between eyes of patients with bilateral nonexudative AMD was assessed to determine if both eyes could be used in clinical trials to test therapies that might slow the growth of atrophy.
Retrospective review of prospectively collected SS-OCTA images.
Patients with bilateral non-exudative AMD with at least one large hypertransmission defect (hyperTD) in both eyes were included.
All large hyperTDs, defined by a greatest linear dimension (GLD) ≥ 250 μm, were identified by two independent graders and annotated using a semi-automated algorithm. Growth-rates were assessed using the area, square-root (sqrt) area, and the perimeter-adjusted growth-rate strategies. Baseline lesion sizes were stratified using a size breakpoint of 0.9 disc area (DA). Inter-eye comparisons of annual hyperTD growth-rates were performed. Power analyses were calculated to determine the sample sizes needed for different clinical trial designs.
Inter-eye concordance of growth-rates was evaluated using Lin’s concordance correlation coefficients, and inter-eye growth-rate agreement was evaluated using Bland-Altman analyses. Sample sizes for a two-arm clinical trial design were calculated to detect a 30% reduction in growth-rates in the treatment arm with 80% power and a two-sided α = 0.05.
Sixty-four subjects (128 eyes) were followed for a mean of 4.2 ± 1.9 years, providing 942 annualized growth-rate intervals. Prior to stratification by baseline size of 0.9 DA, inter-eye concordance was high (rc = 0.76) for the total area growth-rate, but lower for the sqrt and perimeter-adjusted growth-rates (rc = 0.51, rc = 0.60, respectively). Upon stratification, lesions < 0.9 DA showed low correlation (0.39-0.46) across the three growth-rates, but lesions ≥ 0.9 DA showed high correlation regardless of the transformation strategy (0.73-0.78).
High inter-eye concordance in hyperTD growth-rates was observed for lesions ≥ 0.9 DA using three different growth-rate strategies. Regardless of the growth-rate strategy used, recruitment of fellow-eyes in a paired-eye study for localized treatments or both eyes for systemic treatments resulted in a marked reduction in sample sizes with improved clinical trial efficiency.