Abstract
Purpose :
To evaluate rescue of the transition zone (TZ) autofluorescence in six subjects undergoing unilateral subretinal AAV2.REP1 gene replacement therapy (NCT02077361).
Methods :
Bilateral serial 30-degree fundus autofluorescence (FAF) images (Spectralis BluePeak, Heidelberg Engineering) were acquired at baseline and at 6-month intervals post-intervention. Longitudinal images were aligned using the i2k Align Retina platform (Dual-Align LCC, Clifton Park, NY) to compensate for image distortion, changes in resolution, pixel density and focus. Semi-automated serial segmentation of stacked images was performed using custom code written in Matlab (MathWorks Inc.). Expert graders also manually segmented the images using the polygon selection tool in ImageJ (NIH) to determine the ground-truth. Change from baseline in FAF area was quantified at 6-month intervals and compared between treated and untreated eyes.
Results :
Baseline residual FAF areas ranged from 8.63 to 72.64 mm2 (median: 12.26 mm2), with high symmetry between the two eyes (r=0.83). The semi-automated algorithm showed on average 88.95% similarity (Dice similarity coefficient) with the manually segmented ground truth, with a test-retest repeatability of 3%. After 6 months, an nonsignificant increase in FAF area loss was noted in the treated eyes (8.1% vs. 4.5% in the untreated; p=0.30). FAF signal loss occurred exclusively at the TZ between healthy and degenerated retina. At the 12-month timepoint, change from baseline in FAF area was similar between the treated (11.86%) and untreated (9.86%) eyes.
Conclusions :
FAF allows accurate and reliable monitoring of the TZ activity in CHM. A semi-automated segmentation algorithm reduced significantly the measurement bias and error in FAF area calculation. In this limited cohort of CHM subjects, TZ appears not to be rescued by gene therapy. Longer follow-up is required to determine whether gene therapy shows any benefit in modifying the natural history of RPE loss in CHM.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.