Postradiation hypothyroidism in head and neck cancers: A Department of Veterans Affairs single-institution case-control dosimetry study.

We performed a case-control study to characterize the dose-volume relationship and other variables leading to hypothyroidism after head and neck (H&N) cancer radiation therapy (RT) in a homogenous Veterans Affairs (VA) population. All records of patients receiving RT for various H&N cancers at a single VA medical center between 2007 and 2013 (n = 143) were screened for post-RT thyroid stimulating hormone (TSH) levels (n = 77). The thyroid gland was contoured on each slice of the planning computed tomography scan when available (hypothyroid: n = 18; euthyroid > 2 years: n = 16), and dose-volume histograms based on physical dose and biologically equivalent dose (BED) were compared systematically to find the significant dose-volume thresholds that distinguish the patients who developed clinical hypothyroidism. Dosimetric and clinical variables were considered in univariate and multivariate analysis. Preirradiation prevalence of hypothyroidism was 8 of 143 (5.6%). After RT, 36 of 77 (47%) screened patients had abnormally high TSH, of which 22 of 36 (61%) had clinical hypothyroidism after 1.29 ± 0.99 years. The median follow-up durations were 3.3 years and 4.7 years for euthyroid and hypothyroid patients, respectively. Compared with the euthyroid cohort (n = 41), these hypothyroid patients displayed no significant difference in age, gender, primary tumor site, thyroid volume, hypertension, diabetes, or use of chemotherapy, surgery, or intensity-modulated radiation therapy (IMRT). They were more likely to have had stage 3 or 4 cancer than euthyroid patients (86.5% vs 73.2%, p = 0.01). The odds ratios of hypothyroidism for stage 3 + 4 cancers and V50Gy < 75% were 5.0 and 0.2, respectively (p < 0.05). Equivalent BED threshold of V75Gy < 75% gave an odds ratio of 0.156 for developing hypothyroidism (p = 0.02). The prevalence of post-RT clinical hypothyroidism was relatively high for patients with H&N cancers and warrants routine surveillance, especially in those with higher stage malignancy. V50Gy < 75% may be a useful guideline to avoid hypothyroidism. We also show BED data which could be used for unconventionally fractionated schemes, and V75Gy < 75% may be a useful guideline.