A comparison of voice quality following radiotherapy or transoral laser microsurgery of T1a laryngeal carcinomas


Kinshuck, AJ ORCID: 0000-0003-4756-106X (2016) A comparison of voice quality following radiotherapy or transoral laser microsurgery of T1a laryngeal carcinomas. Doctor of Medicine thesis, University of Liverpool.

[img] Text
200635344_Oct2016.pdf - Unspecified
Download (4MB)

Abstract

Introduction: Patients with laryngeal carcinoma often present early due to the change in their voice. The treatment for T1aN0M0 carcinoma varies throughout the world, but whether radiotherapy (RT) or endolaryngeal laser excision is performed both result in excellent local control of the tumour and five year survival rates. There are advantages and disadvantages of either treatment but there are no appropriately powered randomised controlled trials comparing them. Over recent decades external beam RT has become the more popular choice and this is partly due to a perception of poor voice outcomes from surgical excision. However with the development of technology allowing surgical precision, transoral laser microsurgery (TLM) has resulted in low morbidity and good voice outcomes. Objective: This research has three main objectives: a. To describe acoustic parameters of ‘normal’ voice; b. To compare voice outcomes in patients treated with TLM with those treated with radiotherapy for T1a SCC of the glottis; c. To investigate longitudinal changes in voice quality in patients undergoing TLM for T1a SCC of the glottis. Methods: The research was divided into three main parts. The first part was to analyse the acoustic parameters of ‘normal’ voice. To describe the parameters of ‘normal’ voice, adults with no history of voice disorders who scored zero on the voice questionnaire (Voice Handicap Index - 10) were included. The second part comprised a comparative cohort study of 40 patients with T1aN0M0 laryngeal carcinoma, treated with either TLM (20 patients) or RT (20 patients) to compare voice outcomes at least one year following treatment. The third part involved a prospective cohort study of 30 patients with T1aN0M0 laryngeal carcinomas who were treated with TLM, comparing voice qualities before and after treatment. All patients were recruited from those attending the regional Head and Neck centre in Aintree University Hospital. The same methodology was adopted for voice recordings for all three parts of the study. Participants were asked to read a phonetically balanced passage and produce a prolonged vowel sound. In a sound proof room the voice recording included simultaneous audio and electrolaryngograph readings. The voice recordings were scored according to the GRBAS voice scale by an experienced rater. Acoustic analysis was performed form the electrolaryngograph recording using the SpeechStudioTM software. Several objective acoustic parameters were calculated from both sustained vowels and connected speech. These include: fundamental frequency (Fx), jitter, shimmer, harmonics to noise ratio (HNR) and normalized noise energy (NNE). In the comparative study of TLM versus RT and the prospective TLM study, patients were asked to complete voice-specific and quality of life questionnaires. The voice-specific questionnaires were the Voice Symptom Scale (VoiSS) and the Voice Handicap Index-10 (VHI-10). The quality of life questionnaire adopted was the University of Washington Quality of Life (UWQoL) version 4. Results: In the acoustic analysis of sustained vowels in normal speech, females have a statistically significantly higher Fx than males (adjusted p=<0.05). There is no other statistically significant difference across the domains for sustained vowels in normal speech. In the analysis of connected speech, Fx is again higher in females (p<0.001). There is no statistically significant difference in amplitude (Ax) or contact quotient (Qx). In the comparison of voice post TLM and RT, there is no statistical difference in voice-specific questionnaires between the groups. The UW-QoL4 found a statistically significantly higher QoL score in the TLM compared with the RT group for appearance (p=0.003), recreation (p=0.048), chewing (p=0.015) and saliva (p=0.016), however these are not statistically significant when adjusted for age. Overall for QoL, the RT group have a statistically significantly lower median score compared to TLM in physical function (p=0.004) and this remains statistically significant when adjusted for age (p=0.036). There is no statistically significant difference for social function (p=0.441). There is no statistically significant difference in perceptual rating (GRBAS score) between RT and TLM groups (total mean 5.49 vs. 5.12, p=0.254). Most domains as part of the acoustic analysis of sustained vowels show no statistically significant difference between RT and TLM. The mean Fx analysis on connective speech is statistically significantly higher in the TLM group (161.2Hz vs. 131.1Hz, adjusted p=0.001). Coherence of frequency is statistically significantly higher in the TLM group (48.6% vs. 36.0%, adjusted p=0.027) and pitch irregularity is statistically significantly higher in the RT group (26.7% vs. 14.9%, adjusted p=0.013). There is no statistically significant difference in mean amplitude between the two groups. Coherence of amplitude is statistically significantly higher in the TLM group (adjusted p=0.006) and amplitude irregularity is statistically significantly higher in the RT group, (12.4% vs. 6.3%, adjusted p=0.005). There is no statistically significant difference in mean contact quotient (p=0.368), coherence (p=0.236) or irregularity (p=0.125) when comparing TLM and RT. In the comparison of voice pre and post TLM, there is no statistical difference in voice-specific questionnaires between the groups. There is no statistically significant difference in the UW-QOLv4 domain scores or composite scores in patients pre- and post- TLM. There was no statistically significant difference in mean score for ‘G’,’R’,’B’ and ‘S’ indicators as part of perceptual rating between pre and post TLM patients, although asthenia was statistically significantly lower post-TLM (0.97 vs. 0.94, adjusted p=0.015). There is no statistically significant difference in any of the domains in the acoustic analysis of sustained vowels pre and post TLM. In the acoustic analysis of connected speech, the mean DFx is statistically significantly higher in the post TLM group (adjusted p=0.001). There is no statistically significant difference in the coherence of frequency or pitch irregularity when comparing pre and post TLM. There is no statistically significant difference in the mean DAx (p=0.121), coherence (p=0.472) or irregularity of amplitude (p=0.184) when comparing pre and post TLM. There is no statistically significant difference in the mean DQx (adjusted p=0.904), coherence (adjusted p=0.293) or irregularity of the contact quotient (adjusted p=0.400) when comparing pre and post TLM. Conclusion: The treatment of T1a laryngeal carcinoma with either TLM or RT has been shown to have comparably good local control. There are advantages and disadvantages of both treatments, however TLM is often preferred by patient and clinician as it is a day case procedure, can provide histological clearance and leaves the option to use RT in the future. However voice outcomes of the procedures have been debated with various reports in the literature. There are challenges when comparing the two treatment modalities due to a number of tumour, patient and surgical factors. It is not surprising that the voice is affected by whatever treatment is performed to treat the glottic carcinoma. This study shows that voice quality is good, however it is measured, for after both TLM and RT.

Item Type: Thesis (Doctor of Medicine)
Divisions: Faculty of Health and Life Sciences
Depositing User: Symplectic Admin
Date Deposited: 22 Dec 2016 08:15
Last Modified: 19 Jan 2023 07:25
DOI: 10.17638/03004620
Supervisors:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3004620
Dimensions
Altmetric
CORE (COnnecting REpositories)