Cardiac radiation dose has been identified as an important predictor of cardiac toxicity and overall survival (OS) for locally advanced non-small cell lung cancer (NSCLC). However, radiation-induced cardiac toxicity among early-stage NSCLC patients treated with stereotactic body radiation therapy (SBRT), which delivers markedly different dosimetry compared to stage III disease, is less well-characterized. The primary objective was to assess associations between cardiac radiation dosimetry and decreased OS in early-stage NSCLC patients receiving SBRT, and the secondary objective was to characterize cardiac adverse events following lung irradiation.
From February 2009 to September 2014, 153 early-stage NSCLC patients were treated with SBRT at a single institution following a regimen of 54 Gy in 3 fractions for peripheral tumors and 50 Gy in 4 fractions for central tumors. Patients were followed with clinical examination and CT imaging every 3 months for the first 2 years, and then every 6 months, to assess treatment response and toxicity. Patients with synchronous or metachronous tumors and thus received multiple courses of radiation and patients whose treatment plans were unavailable were excluded from analysis. Maximum dose, mean dose, V10, V25, and V50 to 15 cardiac substructures and the whole heart were measured and analyzed for association with OS using the Kaplan-Meier method. Because dosimetric variables were co-linear, artificial neural network (ANN) analysis was then performed to modulate out confounding behaviors to predict for OS as a binary outcome.
A total of 112 patients were analyzed, with median follow-up of 5.1 years for surviving patients. Upon univariate analysis, right ventricle V10 most negatively predicted for survival (r = -0.20, p = 0.04), where patients who received right ventricle V10 < 4% had significantly longer OS than patients who received right ventricle V10 > 4% (5.3 years vs. 2.4 years). Upon ANN analysis, 74 input features including cardiac dosimetry parameters predicted for survival with a test accuracy of 64.7%; repeat ANN analysis using dosimetry to a neutral structure which should not predict for OS (spinal cord) confirmed predictive power of cardiac dosimetry. Nineteen subjects (17.0%) had one or more ≥ grade 1 cardiac adverse events following lung irradiation at a median of 25 months. In total, there was 1 pericardial event, 11 arrhythmic events, 10 ischemic events, and 1 unclassified event (endocarditis).
Cardiac dosimetry was associated with decreased OS in early-stage NSCLC patients treated with SBRT, and there was a significant incidence of cardiac adverse events following lung irradiation. These data support the importance of minimizing radiation dose to cardiac substructures; additionally, cardiac follow-up in these patients should be strongly considered.
Author Disclosure: S. Chan: None. N. Shaverdian: None. D. Ruan: Independent Contractor; Omega BioSystems. Research Grant; Varian Medical Systems. M. Cao: None. G. Raghavan: None. P. Lee: Honoraria; Viewray; Committee Co-Chair.
© 2018 Published by Elsevier Inc.