Modeling Dose Response for Late Dysphagia in Patients With Head and Neck Cancer in the Modern Era of Definitive Chemoradiation.

Abstract

To develop personalized multivariate dose-response models for late dysphagia in patients with head and neck cancer treated in the modern era of combined chemotherapy with intensity-modulated radiation therapy. The analysis included 424 patients (oropharyngeal cancer [n = 295] and nasopharyngeal, hypopharyngeal, or laryngeal cancer [n = 129]) who received definitive chemoradiation between January 2004 and April 2009. The superior, middle, and inferior pharyngeal constrictor muscles were contoured. We calculated generalized equivalent uniform dose (gEUD) for each and the total constrictor muscle volume, with the volume effect parameter a varying from log10 a = -1 to +1 in steps of 0.1. We used the National Cancer Institute Common Toxicity Criteria for Adverse Events (version 3.0) to grade late dysphagia and logistic regression to evaluate the correlation of gEUD( a) with grade 2 or higher (≥ G2) and grade 3 or higher (≥ G3) late dysphagia at each value of a. Median follow-up was 33.3 months (range, 6 to 69 months). There were 41 cases (10%) of ≥ G2 dysphagia and 22 cases (5%) of ≥ G3 dysphagia. Mean doses to the total constrictor ranged from 30.1 to 85.7 Gy (median, 61.2 Gy). The predicted rate of ≥ G2 dysphagia increased by approximately 3.4% per Gy at the mean dose, for which the probability of ≥ G2 dysphagia is 50%. The threshold mean total constrictor doses that limited rates of ≥ G2 and ≥ G3 dysphagia to < 5% were < 58 Gy and < 61 Gy, respectively. Other significant factors in the multivariate predictive model included disease site, mean dose to total constrictor muscle, and patient age. Incidences of both ≥ G2 and ≥ G3 dysphagia were dependent on the mean radiation dose to the total constrictor muscle volume, disease site, and patient age. Limiting the total volume of constrictor muscle to < 58 Gy could keep the predicted rate of ≥ G2 dysphagia to < 5%.