Voice Handicap Following Curative Therapy for Locally Advanced Head and Neck Cancer

Abstract

Voice changes are a common late effect experienced by head and neck cancer (HNC) survivors, but most prior studies have focused primarily on patients with larynx tumors. The purpose of this study was to characterize voice handicap in a broader population of long term HNC survivors. We also sought to determine the sensitivity of a single question screener for voice handicap in this population. This study enrolled patients with locally advanced HNC (T3+ or N+) who received ≥60 Gy at a single academic institution and survived at least 2 years from the completion of radiation therapy (RT). Voice handicap was assessed using the more comprehensive and validated 10-question Voice Handicap Index (VHI-10). Clinically significant voice handicap was defined as a total VHI-10 score of >11. Patient reported hoarseness was assessed by the single question "Have you had problems with hoarseness?" from the EORTC QLQ-HN35 form. Descriptive statistics and univariate binomial logistic regression were used to identify factors associated with significant voice handicap. Missing data were not imputed and denominators were adjusted for each analysis. All analyses were performed using RStudio (PBC, Boston, MA). A total of 199 patients were enrolled in this study and included in the analysis; 182 (91%) completed the VHI-10 and 189 (95%) completed the EORTC QLQ-HN35. The median time from completion of RT to QoL assessment was 5.6 years (range: 1.7 - 28.9 years) and the median age was 65 years (range: 25 - 88 years). The primary tumor was most commonly within the oropharynx (51%) followed by larynx (13%) and oral cavity (11%). Surgery was performed in 64% and 63% systemic therapy. Clinically significant voice handicap was present in 34% of patients and the median VHI-10 score was 8 (range: 0 - 40). Primary tumors of the larynx (OR 7.18; 95% CI 2.52-23.8) and oral cavity (OR 3.29; 95% CI 1.21-9.35; p = 0.02) were associated with a higher odds of significant voice handicap than oropharynx tumors. Other factors associated with voice handicap were African American race (OR = 2.78 versus white race, 95% CI 1.13-7.14), female sex (OR 2.34; 95% CI 1.21-4.55), and T3+ tumor stage (OR = 3.53, 95% CI 1.48-9.19). The sensitivity and specificity of the single item hoarseness question from the EORTC QLQ-HN35 for predicting significant voice handicap ranged from 64.6% and 69.7% when a "A little bit" was considered a positive response to 24.6% and 98.2% when "Very much" was considered a positive response. Significant voice handicap was present in more than one-third of long term HNC survivors in this cohort and primary tumors of the larynx and oral cavity were associated with a higher odds of voice impairment. A single question about hoarseness was not sensitive for detecting significant voice impairment, potentially because voice quality is impacted by a range of factors besides hoarseness. Results of this study highlight the need to improve our understanding of how voice handicap impacts the QoL of HNC survivors.