194 - Perceptual Voice Outcomes of Long-Term Head and Neck Cancer Survivors: A Multi-Listener Large Cohort Study

Purpose/Objective(s): Voice changes are common following radiotherapy (RT) for locally advanced head and neck cancer (HNC) and are strongly linked to patient quality of life (QoL). Although the importance of voice outcomes is widely appreciated, more work is needed to classify subtypes of dysphonia in long-term HNC survivors and better identify contributing factors.

Materials/Methods: This cross-sectional study enrolled participants who previously received >50 Gray of RT for locally advanced HNC and survived at least 2 years thereafter. Participants completed a standardized audio recording of the Rainbow Passage read aloud, sustained vowel phonation, a pitch range task and as well as QoL questionnaires including the Voice Handicap Index-10 (VHI-10). Nine voice specialized speech language pathologists used the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V). The overall visual analog scale of voice impairment (VAS; 0-100 scale) was averaged across all 9 listeners. A primary aberrant feature and presence of resonance changes were deemed present when =5 of 9 of reviewers were in agreement. Generalized linear models were used to assess factors associated with CAPE-V voice deviance scores and VHI-10 composite scores.

Results: Voice recordings and QoL assessments were completed by 198 HNC survivors yielding a total of 1,782 perceptual voice assessments at a median of 3.9 years (range: 2.0-28.9 years) from completion of RT. The oropharynx (51.5%) was the most common subsite followed by larynx (11.6%) and oral cavity (11.1%). Mean CAPE-V VAS scores indicated moderately impaired voice (VAS: 40-65) for 39 (19.7%) participants and severely impaired (VAS: 66-100) for 9 (4.5%) participants. The most common primary aberrant feature was roughness for 128 (64.6%) participants. Resonance change was identified for 36.4% of participants with oral cavity tumors, 34.8% with larynx tumors, and 16.3% of remaining participants. Larynx subsite (p=0.008) and presence of resonance change (p=0.038) were associated with higher CAPE-V VAS scores. Primary tumor stage T3+ (p=0.009), female sex (p=0.008), CAPE-V score (p<0.001), and presence of resonance change (p=0.039) were associated with higher VHI10 composite scores.

Conclusion: This study is among the largest trained listener studies of voice outcomes in long-term HNC cancer survivors to date. Perceptually assessed impaired voice quality was present in nearly one quarter of participants. Although CAPE-V scores were highly related to larynx subsite, the presence of resonance changes was also a significant contributor. In addition to CAPE-V scores and resonance changes, further variables such as sex and T-stage were associated with VHI-10 scores, supporting the hypothesis that perceived voice handicap is impacted by factors in addition to objective dysphonia. Further analysis of this and other data sets is needed to characterize dysphonia phenotypes and their contributing factors to inform future strategies to prevent and mitigate voice handicap.