Factors impacting cardiac dose and overall survival in post-operative non-small cell lung cancer patients.

Abstract

The LungART trial randomized NSCLC patients with N2 disease to adjuvant chemotherapy ± radiotherapy. They found that patients who received radiotherapy had reduced overall survival and claimed that heart dose was the etiological factor, though without offering proof. Consequently, we evaluated cardiac dose and overall survival (OS) in a large cohort of patients treated with adjuvant radiotherapy for N2 disease. 155 patients comprised the final cohort, all treated for N2 disease. Cox proportional hazards testing for OS was performed across a wide range of clinical variables. Additionally, we evaluated factors that impacted cardiac dose. We performed multivariable analysis (MVA) using elastic net survival regression. We tuned the model's hyperparameters with Bayesian methods and conducted a variable importance analysis on the resultant model. Lastly, we analyzed relevant features after correcting for multiple comparisons. The internal target volume (ITV)/lung volume ratio was overwhelmingly the most important variable, although chemoradiotherapy sequencing played a small role. Cardiac dose significantly increased with increasing tumor volume and involvement of "peri-cardial" lymph node stations (4R/L and/or stations 7-12). After incorporating these factors, no cardiac dosimetric variables remained significant. Cardiac dosimetric variables did not impact upon OS in MVA. OS was overwhelmingly determined by the ITV/lung ratio. Furthermore, increasing cardiac dose was directly impacted by increasing tumor volume, number of "pericardial" lymph node stations involved, and/or treatment of hilar lymph nodes, all of which are eponymous for increasing tumor stage.