Imaging-Based Detection and Prediction of Radiation-Induced Cardiac Toxicity: A Narrative Review

Abstract Radiation therapy (RT) can lead to late-onset cardiac toxicity, especially in patients receiving treatment for breast, thoracic, or hematologic malignancies. Early identification of subclinical cardiac injury is essential to guide surveillance and intervention strategies. This narrative review synthesizes imaging-based evidence published between 2014 and 2025 on the detection and prediction of RT-induced cardiac toxicity. We screened 1535 studies and included 78 studies that employed echocardiography, cardiac magnetic resonance imaging (CMR), positron emission tomography (PET) and single-photon emission computed tomography (SPECT), cardiac computed tomography (CT), or multimodal approaches. Echocardiographic strain imaging, particularly global longitudinal strain (GLS), consistently identified early subclinical dysfunction. CMR offered spatial and tissue-level characterization of myocardial fibrosis and inflammation, while PET/SPECT revealed perfusion and metabolic changes in irradiated myocardium. CT-based methods such as coronary artery calcium scoring and coronary CT angiography detected subclinical atherosclerosis with dose correlations. Several predictive models integrating imaging, dose, and clinical data demonstrated strong performance in identifying high-risk patients. Despite encouraging results from such models, variability in timing and endpoint selection limits clinical adoption. Standardized imaging protocols and prospective validation of predictive models are urgently needed. Imaging biomarkers may help tailor follow-up strategies to individual risk and guide the design of future cardio-oncology trials and radiation treatment planning strategies.