Cardiac Substructure Dose Quantification and Optimization in Left-Sided Lung Cancer Radiotherapy.

Abstract

As curative-intent radiation therapy provides excellent local control for lung tumors, patients are surviving longer to experience treatment-associated adverse cardiac events. Patients treated for left-lung tumors are particularly vulnerable due to the heart's anatomical proximity. While the whole heart is conventionally contoured as an organ-at-risk, emerging research highlights cardiac substructure doses better predict cardiotoxicity post-radiotherapy than mean heart dose. An institutional dataset of 50 contrast-enhanced CT scans was compiled from patients treated between 2015 and 2025 for left-sided locally advanced primary lung cancers with definitive photon radiotherapy to 60 Gy in 30 fractions. Twenty-one cardiac substructures were manually contoured on each CT per gold-standard atlases and novel protocols. Mean and maximum substructure doses were extracted from treatment dose-volume histograms, and associations between substructure dose, target volume, tumor location, and demographics were analyzed. Strategic treatment replanning of the 15 cases exceeding a mean whole heart dose of 9 Gy evaluated feasibility of substructure dose reduction. Mean substructure radiation doses were compared against the optimal institutional whole-heart mean dose constraint of 10 Gy. The four highest average mean dose results with 95% confidence intervals were: Aortic Arch 38.6 [32.7-43.8] Gy, Pulmonary Artery 33.3 [29.4-36.8] Gy, Pulmonary Veins 28.8 [24.9-32.4] Gy, and Left Circumflex Coronary Artery 27.3 [21.8-32.7] Gy. In the full cohort, average LAD V15 Gy was 52.7%; in the elevated heart dose cohort, it reached 83.8%. Following replanning, LAD V15 Gy was reduced to 13.7%. This is one of few quantifications of substructure dosimetry in left-sided lung radiotherapy. Eleven substructures received mean doses exceeding the institutional whole-heart threshold, underscoring the need to consider substructures treatment planning. Strategic replanning demonstrates clinical feasibility in reducing substructure doses, particularly the left-sided coronary arteries. The resulting contour and dose datasets provide a foundation for predictive modeling and substructure-sparing treatment strategies.