Voice Outcomes After Radiation for Early-Stage Laryngeal Cancer

Radiation treatment for laryngeal cancer has been shown to cause tissue changes to the vocal folds, which can result in degradation of voice. Our objective in this study was to investigate changes in perceptual, acoustic, and patient-reported outcomes over an extended period of follow-up after radiation.

Retrospective review.

All patients treated with radiation for early-stage laryngeal carcinoma (in situ, T1, or T2) by a single surgeon from 2011-2018 were reviewed. Demographics and treatment information were recorded. Only patients with at least two dates of follow-up with acoustic data (cepstral spectral index of dysphonia [CSID]) and patient-reported surveys (voice handicap index-10 [VHI-10]) were included. Voice samples were rated by two senior speech-language pathologists on the grade, roughness, breathiness, asthenia, and strain scale.

Of 115 patients with early-stage laryngeal cancer, 31 patients met inclusion criteria. The average follow-up from time of treatment was 9.6 years (range 3.0-20.3 years), and the average time in between the first and last voice recordings was 2.6 years (range 0.3-5.5 years). The lesions represented were carcinoma in situ (n = 4), T1 (n = 22), and T2 (n = 5). The VHI-10 scores worsened slightly (mean increase +0.27, median +1) from first to last measurements as did the CSID score (median increase +7.0, median +7.4), though neither reached statistical significance when correlated with time since radiation (P = 0.269 and P = 0.0850). Perceptual analysis as rated by two speech-language pathologists raters showed excellent inter-rater reliability (Cronbach's alpha = 0.84), with no significant change over time (mean +0.39, median, with P = 0.347). Grade, roughness, breathiness, asthenia, and strain, VHI-10, and CSID were all correlated (all pairwise comparisons P < 0.001).

Perceptual, acoustic, and patient-reported outcomes years after radiation for early-stage laryngeal cancer do not show voice degradation over time in this preliminary analysis. Further research with a larger cohort may elucidate voice changes in this population.