Stanford University School of Medicine Stanford, CA
S. S. Butler1, H. J. No2, F. B. Guo3, G. Merchant4, N. J. Park1, S. Jackson1, D. E. Clark5, L. Vitzthum6, A. L. Chin1, K. C. Horst1, R. T. Hoppe1, B. W. Loo Jr7, M. Diehn6, and M. S. Binkley1; 1Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, 2University of Vermont, Burlington, VT, 3University of Rochester School of Medicine and Dentistry, Rochester, NY, 4Kansas College of Osteopathic Medicine, Wichita, KS, 5Department of Medicine, Division of Cardiovascular Medicine, Stanford University School of Medicine, Stanford, CA, 6Department of Radiation Oncology, Stanford University, Palo Alto, CA, 7Department of Radiation Oncology, Stanford University, Stanford, CA
Purpose/Objective(s): Atrial fibrillation (AF) has been associated with prior thoracic radiotherapy. However, AF risk after irradiation of specific cardiac substructures, namely the pulmonary veins (PVs), is still being characterized. We sought to investigate the relationship between PV irradiation and the development of clinically significant (grade=3) AF (AFG=3) and determine whether PV dose-volume cutoff values may predict for AFG=3 risk. Materials/
Methods: We conducted a retrospective study including serial patients who underwent definitive radiotherapy for localized cancers at a single academic institution from 2004-2022, with available dosimetric data. We calculated radiotherapy doses (mean, maximum dose [dmax], and absolute volume receiving =15 Gy [V15]) in 2 Gray (Gy) fraction equivalent dose to PVs and other cardiac substructures, including left atrium (LA), sinoatrial node (SAN), and left coronary arteries (LCAtot=left main+left anterior descending+left circumflex). AF incidence was calculated using Fine-Gray with univariable and multivariable analyses clustered by cancer type and adjusted for the competing risk of death to measure adjusted hazard ratios (aHR) and 95% confidence intervals (CI) for AFG=3 incidence. We used the validated “Mayo AF risk score” ([MAFRS] range, 0-12; greater risk with higher scores) to adjust for clinical factors known to predict for AF. Results: We identified 539 patients with median follow-up time of 58.8 months (range, 0-120). Cancer types included 43% [n=230] non-small cell lung cancer, 32% [n=174] breast cancer, 22% [n=119] Hodgkin lymphoma, and 3% [n=16]) esophageal cancer. Baseline factors included a median age of 58 years, 40% male, 52% never-smokers, 59% MAFRS 0-1, 42% PV dmax >20 Gy. 35 AFG=3 events occurred at a median of 22.9 months (range, 0.3-120). AFG=3 risk was greater among patients with higher PV dmax (per Gy; aHR 1.02, 95% CI 1.01-1.03, P<0.001), higher V15LCAtot (per cm3; aHR 1.14, 95% CI 1.06-1.23, P=0.001), higher LA volume (per cm3; aHR 1.01, 95% CI 1.00-1.02, P=0.02). Conversely, AFG=3 risk was lower among patients with MAFRS 0-1 compared to =2 (aHR 0.47, 95% CI 0.23-0.96, P=0.04) and greater LA dmax (per Gy; aHR 0.99, 95% CI 0.98-0.99, P<0.001). The hazard ratio for PV dmax remained comparably significant across baseline MAFRS groups (Pinteraction=0.14). Higher SAN dmax was not associated with AFG=3 risk (HR 1.01, 95% CI 0.99-1.02, P=0.36). Finally, in a spline analysis, patients with PV dmax >20 Gy (among several potential local maxima identified) had significantly higher risk of AF even when stratified by MAFRS. Conclusion: Pulmonary vein dmax, in addition to LCAtot dose, appear to be significant predictors of grade =3 AF — regardless of other underlying clinical risk factors. These findings provide new evidence to support the clinical relevance of these cardiac substructures with respect to radiation toxicity, and suggest that PV dose metrics may warrant further validation.
