Skip to main content

  • Assessing Photoreceptor Status in Pegcetacoplan-Treated Patients With AI

    By Jean Shaw
    Selected by Andrew P. Schachat, MD
    Add to My Bookmarks

    Journal Highlights

    Ophthalmology Retina, January 2023

    Download PDF

    Riedl et al. used spectral-domain OCT (SD-OCT) and artificial intelligence–based algorithms to assess the ability of intravitreal pegcetacoplan to inhibit photoreceptor (PR) loss and thinning in geographic atrophy (GA). They found that this strategy enabled them to report the condition of PRs with high precision, including both PR loss and thinning.

    For this post hoc analysis of 246 pa­tients (246 eyes), the researchers eval­uated 12-month data from the FILLY trial, which investigated the safety and efficacy of pegcetacoplan, which in­hibits complement component 3 (C3), for the treatment of GA. They limited this follow-up investigation to only those eyes imaged using the Spectralis OCT (161 eyes of 161 participants), using scans acquired at baseline and at months 2, 6, and 12.

    The difference in the change of the area of PR loss was compared among the trial’s cohorts (monthly treatment, bimonthly treatment, and sham in­jections), along with the change in PR thickness adjacent to the GA borders and the entire 20-degree scanning area. The main outcome measures were the area of PR loss, PR thickness, and ratio of PR loss to retinal pigment epitheli­um (RPE) loss.

    All told, 31,556 B-scans of 644 SD-OCT volumes were evaluated with a convolutional neural network that delineated the PR layer on each B-scan. Of the 162 participants, 52 had received monthly pegcetacoplan injections, 54 were given bimonthly injections, and 56 served as controls.

    The results of the analysis showed statistically significant less growth in the area of PR loss in the monthly treatment group than in controls at all time points. In addition, PR thin­ning was significantly reduced under monthly treatment than with sham injections within the GA junctional zone as well as throughout the 20-de­gree scanning area. Finally, the analysis found a trend toward greater inhibition of PR loss than RPE loss in those who were actively treated.

    In their discussion, the authors note that the results provide “objective proof of principle that complement inhibition can indeed preserve PRs as the major correlate of retinal function.” They also pointed out that automated quantification of PR loss and/or main­tenance, using OCT images, is a fast and accurate method of monitoring patient progression.

    The original article can be found here.