WCN26-5475 INTEGRATING FUNDUS IMAGING AND CLINICAL DATA USING MACHINE LEARNING TO PREDICT BIOCHEMICAL ABNORMALITIES IN CHRONIC KIDNEY DISEASE

Fundus imaging has emerged as a valuable noninvasive tool for assessing systemic diseases. Retinal photographs can capture microvascular changes that reflect underlying biochemical and metabolic abnormalities. This study aimed to establish a computer-aided diagnostic framework integrating fundus images with clinical data using machine learning to predict key biochemical markers in patients with chronic kidney disease (CKD), focusing on vascular calcification, inflammation, and cardiac remodeling.