NTCP model for hypothyroidism after supraclavicular-directed radiation therapy for breast cancer.

BACKGROUND AND PURPOSE:Hypothyroidism (HT) is a well-known complication of radiation (RT) that includes supraclavicular (SCV) fields. We analyzed breast cancer patients who received SCV-directed RT to evaluate predictors of HT and developed the first normal tissue complication probability (NTCP) model for HT specific to breast cancer patients. MATERIALS AND METHODS:192 breast cancer patients received SCV-directed RT between 2007 and 2019 and met inclusion criteria. Individual dose-volume histograms were analyzed to determine thyroid volume within and outside specific isodose lines as well as minimum, mean, and maximum doses. Multivariable logistic regression was performed to assess potential clinical and treatment factors for the development of hypothyroidism. An NTCP model was created, and model validation was performed. RESULTS:Thirty-seven patients (19.3%) developed HT following SCV-directed RT at a median 25 months (range: 2-83 months). Multivariable analysis revealed longer length of follow-up (p = 0.015) and larger thyroid volume receiving less than 20 Gy (CV20Gy[cc]; p = 0.045) were significant prognostic factors (p = 0.039). IMRT was not associated with an increased risk of hypothyroidism (p = 0.28) despite lower CV20Gy[cc] (p = 0.0002). On NTCP modeling, CV20Gy[cc] ≥ 8.5 cc was associated with a risk of HT < 15%. For smaller thyroids, mean dose and thyroid volume were found to be predictive of HT risk. Model validation demonstrated comparable performances between our model and other published models (AUC 0.69-0.72). CONCLUSION:NTCP modeling within our patient cohort suggested that greater than 8.5 cc thyroid volume receiving less than 20 Gy may be a recommended dosimetric guideline to minimize HT risk in breast cancer patients receiving SCV-directed RT.