Dosimetric Predictors of Dysphonia after Intensity-modulated Radiotherapy for Oropharyngeal Carcinoma

Aims: To investigate dosimetric predictors of voice changes after whole-field intensity-modulated radiotherapy (IMRT). Materials and methods: Patients treated with whole-field IMRT for oropharyngeal/unknown primary tumours were selected for the present retrospective study having grossly uninvolved larynx at the time of radiotherapy and at least one follow-up visit. Voice changes were prospectively scored at each follow-up examination according to the Common Terminology Criteria for Adverse Events (CTCAE) v3.0 scale and self-reported by two items (HN4 and HN10) of the Functional Assessment of Cancer Therapy-Head and Neck Scale (FACT-HN) questionnaire. Predictors of toxicity were investigated at logistic regression, including various patient and tumour characteristics, as well as individual dosimetric data. Results: With a median follow-up of 18 months (range 3-46 months), peak CTCAE dysphonia was graded as 2 in 13 patients (10.5%), whereas 45 patients (36.3%) reported peak grade 0-1 voice changes according to FACT-HN4. Communication (FACT-HN10) was barely affected. At multivariate analysis, the mean laryngeal dose was an independent predictor of both grade 2 CTCAE dysphonia (odds ratio=1.10, 95% confidence interval 1.01-1.20, P=0.025) and grade 0-1 FACT-HN4 voice changes (odds ratio=1.11, 95% confidence interval 1.04-1.18, P=0.001). Further stratification optimised by a receiver operating characteristic (ROC) analysis showed that, to minimise the risk of grade 0-1 FACT-HN4 voice changes, the mean dose to the larynx has to be kept ≤49.4Gy. Conclusion: Voice changes after whole-field IMRT are common, but mild, and are strictly correlated to the dose received by the uninvolved larynx; in order to minimise the risk of side-effects, the mean dose to the larynx should be kept ≤50Gy.