Prediction of new-onset atrial fibrillation in patients with non-small cell lung cancer treated with curative-intent conventional radiotherapy
Fariba Tohidinezhad, Leonard Nürnberg, Femke Vaassen, Rachel Ma Ter Bekke, Hugo J.W.L. Aerts, Lizza Hendriks, André Dekker, Dirk De Ruysscher, Alberto Traverso
Abstract
Background Atrial fibrillation (AF) is an important side effect of thoracic Radiotherapy (RT), which may impair quality of life and survival. This study aimed to develop a prediction model for new-onset AF in patients with Non-Small Cell Lung Cancer (NSCLC) receiving RT alone or as a part of their multi-modal treatment. Patients and Methods Patients with stage I-IV NSCLC treated with curative-intent conventional photon RT were included. The baseline electrocardiogram (ECG) was compared with follow-up ECGs to identify the occurrence of new-onset AF. A wide range of potential clinical predictors and dose-volume measures on the whole heart and six automatically contoured cardiac substructures, including chambers and conduction nodes, were considered for statistical modeling. Internal validation with optimism-correction was performed. A nomogram was made. Results 374 patients (mean age 69 ± 10 years, 57 % male) were included. At baseline, 9.1 % of patients had AF, and 42 (11.2 %) patients developed new-onset AF. The following parameters were predictive: older age (OR=1.04, 95 % CI: 1.013–1.068), being overweight or obese (OR=1.791, 95 % CI: 1.139–2.816), alcohol use (OR=4.052, 95 % CI: 2.445–6.715), history of cardiac procedures (OR=2.329, 95 % CI: 1.287–4.215), tumor located in the upper lobe (OR=2.571, 95 % CI: 1.518–4.355), higher forced expiratory volume in 1 s (OR=0.989, 95 % CI: 0.979–0.999), higher creatinine (OR=1.008, 95 % CI: 1.002–1.014), concurrent chemotherapy (OR=3.266, 95 % CI: 1.757 to 6.07) and left atrium D max (OR=1.022, 95 % CI: 1.012–1.032). The model showed good discrimination (area under the curve = 0.80, 95 % CI: 0.76–0.84), calibration and positive net benefits. Conclusion This prediction model employs readily available predictors to identify patients at high risk of new-onset AF who could potentially benefit from active screening and timely management of post-RT AF.