Deep learning-based fundus image analysis for cardiovascular disease: a review

AI Generated Summary*

Drawing mainly on 19 studies from the past five years, this review covers deep learning applied to color fundus photographs for cardiovascular risk assessment. One model, trained on UK Biobank and EyePACS images, estimated age within 3.26 years, identified sex with an AUC of 0.97, and flagged major adverse cardiac events at 0.70 versus 0.72 for the SCORE calculator. Others predicted carotid atherosclerosis (0.713), coronary artery calcium presence (0.742), chronic kidney disease (above 0.80), type 2 diabetes (0.731 from images alone, 0.810 with risk factors) and anemia (0.88 to 0.93). Such tools might complement existing stratification, particularly where resources are scarce, though weak interpretability, dataset bias, regulatory uncertainty and few head-to-head comparisons with established calculators leave prospective trials necessary.

*This summary was generated by AI and is published unedited. Oku does not alter these summaries. It may contain errors or omissions and is provided for general informational purposes only. Oku does not guarantee its accuracy, completeness, or reliability. For authoritative information, please refer to the original, peer-reviewed article.