Toxicity and Dosimetric Analysis of Reirradiation of Head and Neck Cancers with High Dose Rate Brachytherapy

Abstract

Reirradiation (reRT) of recurrent or second primary head and neck cancers (HNC) after prior curative-intent external beam radiotherapy (EBRT) has historically achieved local control (LC) rates of 40-50% and been associated with high grade toxicity rates estimated at 25-50%. This study evaluated the hypothesis that patients with small reRT target volumes could be selected for high dose rate brachytherapy (HDR-BT) reRT and achieve similar LC without excess toxicity. Included were all patients with HNC squamous cell carcinoma treated with HDR-BT after having previously received curative-intent EBRT for primary HNC from 2000-2021. Patients were selected by a multidisciplinary tumor board to be appropriate candidates for HDR-BT reRT without EBRT generally for definitive or adjuvant treatment of small primary tumors without neck nodal metastases. Univariate analysis was performed using the logistic regression model. Survival outcomes were estimated with the Kaplan Meier method. Twenty-three patients were evaluated. Median follow up time was 19mo. Median age at time of HDR-BT was 64 years. Thirteen patients (57%) were treated for recurrent HNC, of which 7 were in the oral cavity (OC) and 6 were the oropharynx (OPX). Ten patients (43%) were treated for a second primary HNC, of which 5 were in the OC and 5 were in the OPX. Median time from completion of EBRT to HDR-BT was 41 (IQR 14-73) mo. Within their reRT course, 11 patients (48%) were treated with HDR-BT after resection, 9 patients (39%) received concurrent hyperthermia, and 7 patients (30%) received chemotherapy. HDR-BT regimens included 600cGy x5 (N = 11), 600cGy x6 (N = 6), 450cGy x8 (N = 1), 1500cGy x1 (N = 1),1000cGy x1 (N = 1), 500cGy x8 (N = 1), and 700cGy x5 (N = 1). One patient who was treated with two implants received 450cGy x 3 followed by 475cGy x5. A median of 5 brachytherapy catheters were used. Actuarial 2-year LC and overall survival rate was 68% and 62%, respectively. Of the 17 HDR-BT reRT plans available for review, median (IQR) target volume was 15.8 (10.6-34.9) cc. Median (IQR) target V100% was 90.6 (89.4-90.0)%, V150% was 50.5 (49.7-54.4)%, and V200% was 25.4 (23.8-29.0)%. Median (IQR) target D90% was 30.1 (29.8-35.5) Gy, and median D1cc was 116.4 (100.5-171.4) Gy. The mandible dose [median (IQR)] was D2cc:15.1(9.48-18.9) Gy; D1cc:16.9(11.1-21.3) Gy; and D1%:18.8(13.4-22.7) Gy. Nine of the 23 patients (39%) experienced ≥G3 toxicity including fistula, soft tissue necrosis, osteoradionecrosis, ulcer, hemorrhage, and dysphagia requiring a chronic feeding tube. Target D90% was associated with ≥G3 toxicity (p = 0.045). For D90% greater than the median of 30Gy, 45% ≥G3 toxicity was observed. This study suggests that HDR-BT for reRT of small recurrent or second primary HNC can provide similar LC without excess high-grade toxicities as compared to historical outcomes with EBRT reRT. Delivery of equivalent doses higher than 30Gy in 5 fractions should be approached with caution.