Quantitative Analysis of Radiation Therapy-Induced Cardiac and Aortic Sequelae in Patients With Lung Cancer via Magnetic Resonance Imaging: A Pilot Study

Abstract

PurposeTo quantify early radiotherapy (RT)-induced cardiac and aortic changes in patients with lung cancer using cardiac MRI. Methods and MaterialsNine patients with lung cancer treated with RT completed MR scans at baseline (before RT), 3 months, and 6 months following RT. Cine, T1/T2, late gadolinium enhancement (LGE), and 4D-flow MRIs were acquired to assess biological and mechanical cardiovascular changes globally (i.e., over the entire left ventricle (LV) or aorta) and regionally (according to an American Heart Association model). ResultsRegional metrics demonstrated multiple significant changes and dose-dependent responses. Notably, LGE showed changes at 3 and 6 months over septal and high-dose regions (P<0.0458). Longitudinal strain changes were notable at septal and high-dose regions at 3 months and septal regions at 6 months (P<0.0469). Elevated T1/T2 signals (P<0.0391) and changes in radial/circumferential strain at the septum (P<0.0391) were observed at 3 months. Both T1/T2 signal and LGE were correlated with dose at 6 months (T1 signal also at 3 months), with significantly greater changes in regions receiving >50 Gy (P<0.0331). LV dose was not correlated with LV strain changes (P>0.1), but ascending aortic dose was correlated with strain changes at segments 1 and 2 of the LV (P<0.0362). Global metrics identified only two significant responses: increase in LGE volume at 6 months and a reduction in ascending aortic circumferential strain at 3 months (P<0.0356). ConclusionEarly MR-based changes following RT occurred primarily in high-dose regions and the LV septal wall. While several early signals resolved by 6 months, LGE and longitudinal strain changes persisted for at least 6 months. Dose-dependent responses/correlations were observed for T1/T2/LGE changes at 6 months, with the greatest effect in regions exposed to >50 Gy. Further investigations with larger cohorts and longer follow-up are warranted to confirm regional dose dependence and the association between aortic dose and LV strain observed in this pilot study.