Cardiovascular Disease Classification through Retinal Fundus Image Analysis and Clinical Feature Integration

Cardiovascular illnesses remain a primary cause of morbidity and mortality in the world, highlighting the need for early, reliable, and non-invasive risk assessment tools. The retinal fundus imaging could be a useful tool for cardiovascular screening as it reflects the microvascular and haemodynamic changes associated with cardiovascular diseases. However, the extent to which retinal image-derived features can be used to predict cardiovascular risk is unknown. We present a machine learning-based approach to estimate cardiovascular risk based on retinal fundus images in this article. 230 fundus pictures were analysed, 115 cardiac and 115 noncardiac individuals. Following image pre-processing, retinal features were extracted and used for training using machine learning classifiers. The independent test set and standard classification measures were used to assess model performance. The Random Forest model has achieved the highest accuracy of 84.78% among all the tested models. The results indicate that ensemble learning can be particularly effective for detecting non-linear retinal patterns associated with cardiovascular risk. Results in this study demonstrated that the retinal fundus image contains valuable information for cardiovascular risk assessment and can be a good starting point for future multi-modal assessment frameworks based on retinal imaging, combined with clinical biomarkers. This method serves as an example of the potential of retinal imaging as an early cardiovascular risk assessment tool that is non-invasive and scalable.