Radiomics for the Detection and Prediction of Cancer Therapy-Related Cardiotoxicity

Modern cancer therapies have improved survival but unmasked the growing challenge of cancer therapy-related cardiotoxicity, a leading cause of morbidity and mortality in cancer survivors. Current surveillance strategies, relying on serial echocardiography and cardiac biomarkers, are reactive, detecting cardiac injury only after clinical symptoms emerge. Identifying fundamental, imaging-based biomarkers enables timely recognition of subclinical cardiotoxicity. Radiomics, the high-throughput extraction of quantitative features from routine medical images, provides a noninvasive method to characterize tissue pathophysiology beyond the limits of human visual perception. This review explores the emerging evidence of radiomics for the real-time detection and prediction of cardiotoxicity. We identify distinct radiomic signatures across imaging modalities and examine the challenge of personalized risk stratification utilizing radiomics, which show significant promise. This review concludes that although this novel field is currently limited by small studies lacking external validation, radiomics is poised to enable a paradigm shift in the field of cardio-oncology.