AI streamlines patient selection for geographic atrophy trials, outperforming traditional methods
Artificial intelligence (AI) has shown significant potential in improving the recruitment process for clinical trials, particularly for geographic atrophy (GA), according to a study that found that an AI system outperformed traditional methods in identifying eligible patients with greater precision.
This cross-sectional study analyzed a retrospective dataset from the INSIGHT Health Data Research Hub, which included 306,651 patients (602,826 eyes) with suspected retinal disease who had undergone OCT imaging.
A deep learning model was trained to analyze OCT scans and identify patients who might be eligible for GA clinical trials. This AI-driven approach was then compared to traditional keyword-based searches of electronic health records (EHRs). Its predictions were compared to expert assessments using fundus autofluorescence (FAF) images to validate the AI’s effectiveness.
The AI system outperformed the traditional EHR search, identifying more eligible patients with higher precision. The AI method shortlisted 1,139 patients with a positive predictive value of 63%, compared to 693 patients with a 40% positive predictive value from the EHR search. When combined, the AI-EHR approach identified 604 eligible patients with an 86% positive predictive value. The AI also demonstrated a strong correlation (0.77) between the GA area segmented on FAF images and AI-segmented OCT scans.
Reference
Williamson DJ, Struyven RR, Antaki F, Chia MA, Wagner SK, Jhingan M, Wu Z, Guymer R, Skene SS, Tammuz N, Thomson B, Chopra R, Keane PA. Artificial Intelligence to Facilitate Clinical Trial Recruitment in Age-Related Macular Degeneration. Ophthalmol Sci. 2024;4(6):100566. doi: 10.1016/j.xops.2024.100566. PMID: 39139546; PMCID: PMC11321286.
This content is independent editorial sponsored by Astellas. Astellas had no input in the development of this content.
Related Content
