Precision Study of CLAiR AI Software

This prospective, multi-site precision study was designed to evaluate the analytical performance and reproducibility of the CLAiR AI software, a software as a medical device (SaMD) intended to identify individuals with elevated risk of atherosclerotic cardiovascular disease (ASCVD) using retinal fundus images. Participants were enrolled at U.S. clinical sites and underwent protocol-defined retinal imaging procedures using Topcon NW400 and NW500 non-mydriatic fundus cameras. Imaging was performed under controlled conditions, including predefined operator sequencing and repeated image acquisition, to assess reproducibility across operators and devices. The primary objective was to evaluate agreement of device outputs across repeated imaging sessions and camera models. Secondary objectives included assessment of intra- and inter-operator variability. The device under investigation was not FDA cleared or approved at the time of the study and was evaluated under institutional review board (IRB) oversight as a non-significant risk (NSR) device study. No therapeutic interventions were administered, and no clinical management decisions were made based on device outputs.