Impact of Tumour Location and Incidental Radiation Dose to the Internal Mammary Lymph Nodes on Breast Cancer-Specific Mortality
Abstract
Purpose
To investigate the effect of tumour location and incidental radiation dose delivered to the internal mammary lymph nodes (IMLNs) in breast radiotherapy (RT) on breast cancer-specific mortality (BCSM).
Methods
Patients curatively treated with breast-conserving surgery and RT between 2005 and 2014 were identified in a population-based database. Inclusion criteria were pathologically node-positive, CT-based planning, and a contoured seroma. Exclusion criteria were neoadjuvant systemic therapy and therapeutic IMLN irradiation (Dmean>40Gy3.5). Internal mammary vessels in the first 3 intercostal spaces were contoured with Limbus Contour (version 1.8.0-B3) and edited as needed. Mean IMLN dose (Dmean) and minimum distance from seroma to IMLN and lung contours were extracted. The seroma-to-lung distance was used as a surrogate for tumour distance from the deep mammary plexus that drains to the IMLNs. Absolute doses were converted to equivalent dose in 2Gy fractions using α/β=3.5. Subdistribution hazard ratios were estimated using multivariable Fine-Gray competing-risk regression to evaluate the effect of baseline patient, tumour, and treatment variables on BCSM. Interaction terms of the mean IMLN dose quartiles and distances were included to determine if the association between distance and BCSM differed by the dose received.
Results
The 475 study patients had a median follow-up time of 13.7 (range: 0.8–20.3) years. The median of Dmean was 9.4 (first quartile: 3.1, third quartile: 24.3) Gy3.5. The median seroma-to-IMLN distance was 71.5 (range: 7.4–158.9) mm and the median seroma-to-lung distance was 10.9 (range: 1.3–96.7) mm. In patients with incidental doses within the first quartile, the risk of BCSM was lower for longer seroma-to-IMLN (sHR=0.96, p<0.001) and seroma-to-lung (sHR=0.64, p<0.001) distances.
Conclusion
For patients with lower incidental IMLN doses, shorter seroma-to-IMLN and seroma-to-lung distances were associated with increased risk of death from breast cancer. This effect was absent as the incidental IMLN dose increased.