KEY POINTS
- This single-institution analysis included 452 patients treated with curative-intent IMRT or VMAT for middle or lower oesophageal cancer between 2009 and 2024. Eligible schedules delivered at least 40 Gy in 16, 20, 23 or 25 fractions, encompassing definitive, chemoradiotherapy and neoadjuvant treatment approaches.
- Planning CTs and dose distributions were deformably mapped to a common anatomy and converted to EQD2 using an α/β of 3 Gy. Voxel-wise dose was then compared between patients alive and deceased at 12 months, with permutation testing to control for multiple comparisons.
- The strongest survival-associated region was located in the posterior central heart, partially overlapping the left atrium and the previously described Cardiac Avoidance Area developed from lung radiotherapy cohorts. The spatial signal exceeded the 99th percentile of the permutation-derived null distribution.
- The Cardiac Avoidance Area includes the right atrium, aortic valve root and proximal left and right coronary arteries. Using the cohort median maximum dose of 43.5 Gy as a threshold, median survival was 713 days below 43.5 Gy versus 550 days at or above 43.5 Gy (p<0.005).
- Maximum Cardiac Avoidance Area dose remained an independent predictor of survival (p<0.005) after adjustment for clinical and dosimetric variables. Mean lung dose, performance status and nodal stage were also associated with survival.
- The association persisted after excluding patients receiving neoadjuvant chemoradiotherapy followed by surgery. Definitive chemoradiotherapy and definitive radiotherapy-only subgroups showed similar effect estimates, although the smaller subgroup analyses did not individually reach statistical significance.
- The study had no cardiac-event or cause-of-death data, so it cannot demonstrate that the survival difference was caused by radiation-induced cardiac injury. Dose was planned rather than delivered, and deformable registration, treatment heterogeneity and correlations between cardiac dose, tumour location and target volume remain important potential confounders.
CLINICAL TAKEAWAY
The study strengthens the case that where dose lands within the heart may matter more than whole-heart dose alone in oesophageal radiotherapy. The 43.5 Gy value should currently be viewed as a hypothesis-generating planning signal, not a validated clinical constraint.