Risk of Severe Laryngeal Toxicity Following Re-Irradiation With Head and Neck Stereotactic Body Radiation Therapy

Abstract

Survival rates are poor following recurrent head and neck cancers (rHNC) after prior radiation therapy. SBRT is an increasingly utilized technique for re-irradiation of rHNC. Prior studies have demonstrated late grade 3+ toxicity rates <20% overall, though this rate increased to 50% with larynx/hypopharynx recurrences. We sought to evaluate the association between laryngeal dose and the risk of severe toxicity in patients undergoing SBRT as re-irradiation for rHNC, as no validated dose-volume constraints exist to guide SBRT in this setting. We retrospectively reviewed patients with previously-irradiated rHNC treated from 2008-2013. Patients treated early in our experience with incomplete dosimetry were excluded. Patients were treated with linear accelerator-based SBRT to a median of 44 Gy in 5 fractions delivered on a twice-weekly basis with concurrent Cetuximab. SBRT plans were retrospectively reviewed, and the larynx was contoured. The maximum dose to 0.1cc (D0.1cc), 1cc (D1cc), 2cc (D2cc), and 5cc (D5cc) of the larynx and the mean larynx dose were recorded and analyzed for association with laryngeal toxicity, including dysphagia and airway compromise, using binary logistic regression. A total of 75 patients were identified, and 12 patients were excluded due to prior laryngectomy. Median follow-up was 8 months (range: 1-91) for all patients, and 37 months for surviving patients (range: 31-91]. Five patients (8%) received more than 1 course of SBRT, and cumulative larynx doses from fused summary plans were recorded. The median D0.1cc, D1cc, D2cc, D5cc, and mean larynx doses were 17.1 Gy [interquartile range (IQR) 0.6-45.4 Gy], 8.8 Gy [IQR 0.5-36.5 Gy], 6.3 Gy [IQR 0.5-30.0 Gy], 3.8 Gy [IQR 0.4-18.5 Gy], and 3.0 Gy [0.3-13.2 Gy], respectively. The overall rates of severe acute and late toxicity were 15.9% and 20.0%, respectively. The rate of severe laryngeal toxicity was 12.7%. There was no association between severe laryngeal toxicity and prior RT dose, time from RT to SBRT, SBRT dose, or planning target volume size. Laryngeal toxicity was associated with D0.1cc [Odds ratio (OR) 1.10, 95% CI 1.01-1.19, P=.031], D1cc (OR 1.11, 95% CI 1.02-1.20, P=.013) D2cc (OR 1.11, 95% CI 1.03-1.20, P=.006), D5cc (OR 1.10, 95% CI 1.04-1.17, P=.002), and mean larynx dose (OR 1.09, 95% CI 1.03-1.15, P=.004). On multivariable analysis (MVA), only D5cc was associated with laryngeal toxicity. The rate of severe laryngeal toxicity was 50.0% at a D5cc >30 Gy vs. 2.3% for <30 Gy. The results demonstrate a low risk of severe laryngeal toxicity with SBRT for re-irradiation of rHNC, which was associated with all dosimetric parameters on univariate analyses. Only D5cc remained significant on MVA, and the risk of toxicity was 50% with D5cc >30 Gy, which represents a potentially validated constraint to guide application of SBRT for future clinical trial design.