Skeletal muscle loss and substantial left ventricular mass changes on follow-up imaging were strongly associated with poorer survival after curative radiotherapy.
Automated LAD contours reproduced the association between coronary dose and survival in 460 stage III NSCLC patients, although performance varied by model.
The model calibrated well only when non-cancer death was defined consistently; an overall-survival version validated in both external cohorts.
Planning-CT calcification and age predicted atherosclerotic events, while anthracyclines and respiratory toxicity predicted other cardiac complications.
Automatically predicted coronary habitats captured most arteries and enabled substantial dose reductions in three retrospectively replanned thoracic cases.
Symptomatic grade ≥2 cardiac toxicity occurred in 8% after central or ultracentral lung SBRT, typically developing more than one year after treatment.
A maximum left anterior descending coronary artery dose of at least 12 Gy was independently associated with higher long-term cardiac risk.
A 40 Gy equivalent dose to at least 18% of the left anterior descending artery was associated with more than twofold risk of cardiac events.
The framework achieved 85.5% external-validation accuracy and reduced estimated review time from 822 hours to 2.3–4.8 hours.