KEY POINTS
- The systematic review included 10 observational studies of thoracic radiotherapy. Eight studies contributed nine cohorts to the quantitative incidence analysis.
- Across all eligible cohorts, 293 of 3,192 patients developed study-defined atrial fibrillation. Random-effects pooled incidence was 9.1% (95% CI 6.3–12.9%), but heterogeneity was extreme (I²=90%).
- Because postoperative atrial fibrillation after esophagectomy is an important confounder, the main analysis excluded two esophageal cohorts with substantial perioperative influence. Among the remaining 2,766 patients, 217 developed AF, corresponding to a pooled incidence of 7.6% (95% CI 5.6–10.3%).
- Heterogeneity remained substantial even after this restriction (I²=80.3%), and the prediction interval ranged from 2.9% to 18.8%. Alternative statistical models produced very similar pooled estimates of approximately 7.5–7.6%.
- Dosimetric analyses repeatedly implicated structures involved in atrial electrophysiology, particularly the pulmonary veins, atria and sinoatrial node. For example, one mixed thoracic cohort reported a 22% increase in grade ≥3 AF risk per 10 Gy increase in pulmonary-vein Dmax.
- Whole-heart metrics were less consistent. Some studies found associations with mean heart dose, heart V40 or maximum heart dose, whereas another found no significant association between mean heart dose and AF (p=0.094).
- The evidence is not sufficient to derive clinical planning constraints. AF definitions, cardiac contouring, dose metrics, baseline AF exclusion and rhythm surveillance varied substantially, and all included studies were observational.
CLINICAL TAKEAWAY
Thoracic radiotherapy-associated cardiac risk may not be adequately represented by mean heart dose alone. The pulmonary veins, atria and conduction system deserve prospective study and routine reporting may eventually prove useful, but there is currently no validated AF-specific dose constraint that should be imposed on clinical plans.