Purpose

To study the hypothyroidism risk after adjuvant radiation therapy (RT) and the association of different RT targets with hypothyroidism risk.



Methods

We studied 4073 women treated with adjuvant RT for breast cancer from 2007 to 2016. The primary endpoint was hypothyroidism development after RT. Patients were divided and analyzed into 3 groups: whole breast (WB)–alone (n = 2468), regional node irradiation (RNI)–Lv.4 (n = 215; cranial border at the subclavian artery, according to the European Society for Radiotherapy and Oncology consensus guideline), and RNI-supraclavicular lymph node (SCL) (n = 1390; cranial border at the cricoid cartilage). In general, RNI-Lv.4 was used in the patients with high-risk pN0 and pN1 breast cancer. In auxiliary analysis, the mean thyroid dose was estimated in each group (total n = 600, 200 from each group). All the doses were converted to the equivalent dose in 2 Gy fractions (EQD2) with α/β ratios of 3.



Results

The median follow-up duration was 84 months (WB-alone, 84 months; RNI-Lv.4, 44 months; RNI-SCL, 91 months). The 3-year hypothyroidism incidence rate differed significantly between the RNI-SCL and WB-alone groups (2.2% vs 0.8%; Bonferroni corrected P [Pc] < .001) but not between the RNI-Lv.4 and WB-alone groups (0.9% vs 0.8%; Pc > .05). The Cox model revealed an adjusted hazard ratio of 2.25 (95% CI, 1.49-3.38) for RNI-SCL vs WB-alone, 1.69 (95% CI, 1.12-2.56) for adjuvant systemic therapies, and 2.07 (95% CI, 1.07-3.99) for age <60 years. In the subgroup analysis, the hypothyroidism risk became more prominent in patients aged <60 years. The mean exposure doses to the thyroid were 0.23 versus 1.93 versus 7.89 Gy (EQD2) for the WB-alone versus RNI-Lv.4 versus RNI-SCL groups (P < .001). No statistically different locoregional recurrence rates were seen between groups (5-year rate: <3%).



Conclusions

The risk of hypothyroidism increases after RNI-SCL for breast cancer but not after RNI-Lv 4. These data support routine contouring of the thyroid in the RNI setting, and future studies are required to develop optimal dose-volume constraints.