Clinical Hypothyroidism After Proton Versus Photon Regional Nodal Irradiation: A Prospective Correlative Study Within the RADCOMP Randomized Trial.

PURPOSE: Thyroid dysfunction is a common complication of breast regional nodal irradiation (RNI). Proton therapy has been proposed as a means to spare normal tissue. We prospectively evaluated the incidence of hypothyroidism in a cohort of patients receiving RNI on a randomized trial comparing proton with photon modalities. A secondary objective was to evaluate the associations between thyroid dose-volume metrics and the risk of subsequent hypothyroidism.

METHODS AND MATERIALS: We conducted a single-institution prospective correlative study within the randomized phase 3 Radiotherapy Comparative Effectiveness (RadComp) trial. Thyroid function testing was obtained at baseline, annually for 3 years, and at 5 years. Clinical hypothyroidism was defined as elevated thyroid-stimulating hormone above institutional range with recommendation for or initiation of thyroid replacement therapy. The cumulative incidence of hypothyroidism after proton and photon RNI was compared using Fine-Gray regression. Thyroid dose-volume metrics were also explored as potential predictors of hypothyroidism, including previously validated metrics, thyroid Dmean > 21 Gy(RBE), and thyroid volume spared from 20 Gy(RBE) (VS20Gy(RBE)).

RESULTS

Ninety-two patients enrolled; 71 patients met criteria for analysis (proton 38, photon 33). Median follow-up was 66 months. The 3-year cumulative incidence of clinical hypothyroidism was 13%. By modality, incidence was 16% (proton) versus 9% (photon) (P = .14). Stratifying by technique, the incidence of hypothyroidism was 0% for 3-dimensional conformal radiation therapy (3DCRT) and 16% for both intensity modulated radiation therapy (IMRT) and pencil beam scanning (P = .26). Dmean > 21 Gy(RBE) (HR = 3.02; 95% CI, 1.02-8.98) and VS20Gy(RBE)<2.2 cc (HR = 8.80; 95% CI, 1.54-50.30) were associated with hypothyroidism.

CONCLUSIONS

After RNI, the 3-year incidence of clinical hypothyroidism was 13%, with no statistically significant difference between proton and photon modalities. Thyroid Dmean > 21 Gy(RBE) and VS20Gy(RBE) < 2.2 cc were associated with a higher risk of hypothyroidism. Inverse-planned RNI with superior coverage of the supraclavicular fossa may come at the expense of increased thyroid dose.