2813 - Dosimetric Predictors of Dysphonia in Long-Term Survivors of Locally Advanced Head and Neck Cancer Treated with Radiation Therapy

A. Hubler¹, C. Cooper¹, K. Heinzman¹, C. Cardenas², R. A. Cardan², E. Hapner¹, and A. M. McDonald²; ¹University of Alabama at Birmingham, Birmingham, AL, ²Department of Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL

Purpose/Objective(s): Voice changes are common late effects in those receiving radiation (RT) to the upper aerodigestive tract. Dosimetric predictors for many late RT effects are described, but there is a lack of data evaluating patient-reported dysphonia. The purpose of this study was to evaluate dosimetric measures predictive of dysphonia as measured by a validated patient-reported inventory. Materials/

Methods: Long-term survivors (>2 years) of locally advanced (T3+ or N+) head and neck cancer (LAHNC) treated with =60 Gy at a single academic institution were enrolled. Dysphonia was benchmarked with the validated 10-question voice handicap index (VHI-10), with significant dysphonia represented by scores =11. Organs at risk (OAR) evaluated were oral cavity, pharynx, glottis, and larynx. Dosimetric variables examined were the mean, max, and V40Gy for each OAR. Univariable and multivariable logistic regressions were used to identify variables associated with VHI-10 = 11.
Results: Of the 229 patients enrolled, 139 (60.7%) completed the VHI-10 questionnaire and had complete dosimetric information available at the time of this analysis. Median age was 57 years (range 19-82 years) at an average of 7.4 years from completion of RT. Surgery and chemotherapy were incorporated into 55% and 65% of patients’ treatment, respectively. The three most common disease sites were oropharynx (54%), larynx (13%), and salivary gland (8%). Patient-reported dysphonia (VHI-10 =11) was reported by 32% of patients. For the entire cohort, univariable logistic regression noted glottis V40Gy to be the only factor predictive of dysphonia (OR 1.01; 95% CI 1.00 – 1.02; p = 0.03). In patients undergoing surgery, univariable model noted glottis mean (OR 1.04; 95% CI 1.01 – 1.08; p = 0.02) and V40Gy (1.01; 95% CI 1.00 – 1.03; p = 0.03), larynx mean (OR 1.04; 95% CI 1.01 – 1.09; p = 0.03) and V40Gy (OR 1.02; 95% CI 1.00 – 1.03; p = 0.04), and pharynx max (OR 1.15; 95% CI 1.05 – 1.29; p = 0.01) to predict dysphonia. Multivariable model only noted pharynx max (OR 1.14; 95% CI 1.03 – 1.31; p = 0.03) to be predictive. In patients not undergoing surgery, no variable predicted dysphonia.
Conclusion: In this cohort of long-term survivors of LAHNC treated with RT, dysphonia was common and associated with the proportion of glottis receiving at least 40 Gy. Pre-specified dosimetric variables for the oral cavity, larynx, and pharynx were not associated with patient-reported dysphonia. This study adds to the limited understanding of late voice toxicity by using patient reported outcomes in a large population of LAHNC survivors involving various head and neck subsites.