Modeling pegcetacoplan treatment effect for atrophic age-related macular degeneration with AI-based progression prediction.

Abstract

Background: To illustrate the treatment effect of Pegcetacoplan for atrophy secondary to age-related macular degeneration (AMD), on an individualized topographic progression prediction basis, using a deep learning model.

Methods: Patients (N = 99) with atrophy secondary to AMD with longitudinal optical coherence tomography (OCT) data were retrospectively analyzed. We used a previously published deep-learning-based atrophy progression prediction algorithm to predict the 2-year atrophy progression, including the topographic likelihood of future retinal pigment epithelial and outer retinal atrophy (RORA), according to the baseline OCT input. The algorithm output was a step-less individualized topographic modeling of the RORA growth, allowing for illustrating the progression line corresponding to an 80% growth compared to the natural course of 100% growth.

Results: The treatment effect of Pegcetacoplan was illustrated as the line when 80% of the growth is reached in this continuous model. Besides the well-known variability of atrophy growth rate, our results showed unequal growth according to the fundus location. It became evident that this difference is of potential functional interest for patient outcomes.

Conclusions: This model based on an 80% growth of RORA after two years illustrates the variable effect of treatment with Pegcetacoplan according to the individual situation, supporting personalized medical care.