Radiation-Induced Hypothyroidism after Radical Intensity-Modulated Radiation Therapy for Oropharyngeal Carcinoma

Abstract

This work aims to apply an existing normal tissue complication probability (NTCP) model for radiation-induced hypothyroidism (RHT) on a large cohort of oropharyngeal carcinoma (OPC) patients, to identify the clinical and dosimetric parameters for a more robust multivariate NTCP model. The relationship between dose, volume, and thyroid function is explored. Head and neck cancer (HNC) patients treated with radical RT from 2005 to 2013 were reviewed. We identified 1180 HNC patients treated with RT without thyroidectomy. OPC patients treated with retrievable IMRT plan/dose DICOMs and available baseline and follow-up thyroid function tests were included. Mean dose (Dmean) to the thyroid gland (TG) and its volume were calculated. Biochemical HT was defined as a serum TSH level > the normal value for at least two subsequent labs. Clinical HT was defined as grade ≥ 2 HT per CTC-AE, v4 grading system. Univariable and multivariable analyses were carried out for predictors of clinical HT. Dmean and volume of TG for those with HT were compared to others, using Wilcoxon rank-sum test. Other dosimetric parameters including; the percentage of thyroid volume exceeding 10, 20, 30, 40 and 50 Gy (V10, V20, V30, V40 and V50) were considered. Receiver Operating Characteristic (ROC) curves and area under the curve (AUC) for the fitted model vs the Boomsa et al model were calculated. Four hundred seventy-one OPC patients were included. The median age was 57 years and 396 were males. Most tumors (52%) originated from base of tongue, and (93%) were node positive. 61% had received concurrent and 36% induction chemotherapy, respectively. IMRT-split field was utilized in 94%, and median RT dose was 69.96 (range 60-72) Gy. 295 patients (63%) developed clinical HT after RT. Univariate analysis revealed that positive nodal disease, higher Dmean to TG and smaller TG volume were associated with clinical HT (p=0.002, <0.0001 and 0.004, respectively). On multivariate analysis Dmean (Odds Ratio 1.06 (1.04-1.09)) and TG volume (OR 0.879(0.827-0.932)) remained statistically significant, (p<0.0001). Dmean was significantly higher in patients with clinical HT vs. those without (50 vs.45 Gy, p<0.0001). Patients with HT had smaller TG volume compared to those without (11.7 compared to 12.96 cc, p<0.0001). AUC of 0.67 (0.61-0.71) for fitted model vs. 0.67 (0.62-0.72) for the applied Boomsa et al model to current cohort was identified. Volume and Dmean of the TG are important predictors of clinical HT and shall be integrated in NTCP models for RHT. Dmean should be considered during the IMRT plan optimization without compromising target coverage in OPC patients. More effort is needed to develop more robust NTCP model for RHT that accounts for more potential clinical and dosimetric predictors of RHT.