VoiceS: Voice Quality After Transoral CO2-Laser Surgery Versus Single Vocal Cord Irradiation for Larynx Cancer

Background and Rationale

Head and neck squamous cell carcinoma (HNSCC) is the 6th most common type of cancer worldwide. Of all head and neck cancers, approximately 30% originates from the larynx. In Europe, annually 52.000 patients are diagnosed with cancer of the larynx. About 50-60% of laryngeal squamous cell carcinomas arise from the glottic region and over 80% of these patients present in an early stage. The larynx has important roles in production of voice, coordination of deglutition, and respiration. Therefore, the treatment aim of laryngeal cancer is not only achievement of maximum disease control, but also maintenance of function. Transoral CO2-Laser Microsurgical Cordectomy (TLM) and radiotherapy are well-established standard treatment modalities for unilateral stage 0-I carcinoma in situ and squamous cell carcinoma of glottic larynx (for convenience, both will be mentioned as SCCGL throughout this protocol). Based on various clinical studies and meta-analyses, functional and oncological outcome (overall survival and local control) after both treatment modalities are comparable. However, no properly conducted randomized trials comparing these treatments exist. The only such trial with the endpoint of voice quality had to be prematurely closed due to low accrual.

The treatment option varies remarkably in different countries and among institutions. Especially in the case of unilateral SCCGL, other factors such as voice quality, cultural and socioeconomic factors, and patients' preference should be considered. Hoarseness is one of the main and early signs of SCCGL, whereas both radiotherapy and TLM influence voice quality by altering the vocal cord motion and anatomy. Another issue worth to consider is the treatment time. Surgery is performed in one day followed by a few days of hospital stay. In contrast, radiotherapy is applied in daily fractions. Although each fraction only takes few minutes, the patients have to visit the radiation oncology department over a course of 4 to 7 weeks, depending on the institutional preference of dose and fractionation. The post-treatment follow-up schedules are identical for both strategies.

The differences between both treatment modalities have been extensively reported. However, either due to no meaningful difference or due to the heterogeneity in the voice analysis techniques, several studies revealed no significant differences while better voice quality was reported after radiotherapy as compared with TLM in other studies.

The traditional radiotherapy involves the treatment of the whole larynx over a period of 4 to 7 weeks. Recently, a new radiotherapy technique was introduced by a team of researchers from Netherlands, where the treated target volume consists of involved vocal cord and therefore 8 to 10 fold smaller than the target volumes used for traditional whole larynx irradiation. The treatment is reduced to 16 fractions with higher dose per fraction which corresponds to 3 weeks and a day. The results of a prospective cohort (n=30) with single vocal cord irradiation (SVCI) were compared with the results of a historical prospective cohort (n=131) previously treated with whole larynx radiotherapy in the same institute. The median follow-up was 30 months. The voice handicap index (VHI) at all time points beginning from the 6th week after SVCI was significantly superior to the same time points with conventional radiotherapy. Moreover, a comparable local control with SVCI (100%) vs. conventional radiotherapy (92%) was reported at two years, p=0.2412.

In conclusion, the long-term voice quality and oncologic outcome (survival and local tumor control) are comparable with both TLM and traditional radiotherapy. Furthermore, the new SVCI technique offers a possibly superior long-term voice quality with an at least equal oncologic outcome compared to traditional radiotherapy. With this background, the investigators' main aim is to compare SVCI to TLM with the main focus of patient-reported voice quality.

Investigational Treatments and Indication

With the indication to treat early stage SCCGL in an organ preservation setting, both TLM and SVCI are going to be compared head-to-head, and therefore are considered as investigational treatments.

Clinical Evidence to Date

In well-recognized international cancer treatment guidelines, TLM and radiotherapy are considered as comparable modalities for the successful treatment of early stage SCCGL. Based on various clinical studies and meta-analyses, functional and oncological outcome (overall survival and local control) after both treatment modalities are comparable. However, no properly conducted randomized trials comparing these treatments exist. The only such trial with the endpoint of voice quality had to be prematurely closed due to low accrual.

The differences between both treatment modalities have been extensively reported. However, either due to no meaningful difference or due to the heterogeneity in the voice analysis techniques, several studies revealed no significant differences while better voice quality was reported after radiotherapy as compared with TLM in other studies. Nevertheless, most of the literature consists of reports about retrospective case series with the lack of level I evidence.

Justification of Choice of Study Population

In line with the eligibility criteria, the study population will include patients diagnosed with unilaterally situated stage 0-I SCCGL who are considered as eligible both for radiotherapy and surgery by a multidisciplinary tumor board. Both treatment modalities are well-established standards.

Study Objectives

Overall Objective The ultimate goal of the study is to compare the treatment outcome of the SVCI and TLM for unilateral stage 0-I SCCGL.

Primary Objective Comparison of patient-reported subjective voice quality after TLM and SVCI

Secondary Objectives

* Qualitative post-therapeutic comparison of the voice between study arms * Quantitative post-therapeutic comparison of the voice between study arms * Comparison of the oncological outcome between study arms. * Assessment of toxicity/morbidity among study arms with the listed toxicity items provided in Appendix A. CTCAE v.5.0 will be used for the classification of toxicities.

Safety Objectives - Assessment of toxicity/morbidity among study arms with the listed toxicity items provided in Appendix A. CTCAE v.5.0 will be used for the classification of toxicities.

Study Outcomes

Primary Outcome

\- Average of the VHI assessed at 6, 12, 18, and 24 months

Secondary Outcomes

* Perceptual impression of the voice via Roughness - Breathiness - Hoarseness (RBH) assessment at 6, 12, 18, and 24 months * Quantitative characteristics of voice by means of Jitter and Shimmer (JS), Glottal-to-Noise Excitation Ratio (GNE) and Singing Power Ratio (SPR), which will be assessed at 6, 12, 18, and 24 months * Loco-regional control of the disease at 2 and 5 years * Treatment toxicity at 2 and 5 years based on CTCAE v.5.0

Other Outcomes of Interest The patient, disease and procedural characteristics will be described.

Safety Outcomes Although the safety endpoints within the secondary endpoints will be published only at two time points (2 and 5 years), the toxicity will be systematically assessed during and after the treatment in a more frequent schedule: during the treatment, every 3 months until 24 months, and every 6 months between the 24th and 60th months. Assessment of toxicity/morbidity among study arms with the listed toxicity items provided in Appendix A. CTCAE v.5.0 will be used for the classification of toxicities.

Study Design

General study design and justification of design This is a prospective randomized multi-center open-label comparative phase III study with a superiority design (see Section 11 for the details of statistical considerations). Primary endpoint of this study is patient-reported subjective voice quality between 6 to 24 months after randomization.

The sample size calculation is based on the primary outcome, the VHI at 6 to 24 months (averaged). Based on the literature, the investigators consider the VHI to be comparable between TLM and traditional whole larynx radiotherapy. Therefore, the working hypothesis is that there is a difference between TLM and SVCI in regard to VHI. Al-Mamgani et al. reported standard deviations for VHI ranging from 2 to 10 score points12. A difference of 8 points between the two groups are regarded as clinically relevant and a conservative standard deviation of 8 points is assumed. Based on a two-sample means test, 34 patients (17 per group) are needed to detect a difference in VHI at a two-sided alpha-level of 0.05 with a power of 80%.

This sample size calculation is conservative. In the analysis, the average difference over four time points (6, 12, 18, and 24 months) will be modelled and additionally adjusted for the baseline VHI to yield more power. In each center, patients will be enrolled into the study by the local principal investigator. The local PI can be a head and neck surgeon or radiation oncologist. However, before patient accrual, the patient must be informed about the study at least by the attending surgeon and the radiation oncologist. It is highly recommended that the patient is evaluated and informed in a multidisciplinary tumor board setting before the accrual.

Interventions in the trial are considered as non-experimental standard treatments. No diagnostic tool or imaging modality that will be used during the trial is experimental. No interim analysis for futility or safety will be performed.

Methods of minimizing bias

Randomization A probabilistic minimization technique will be used for random treatment allocation between the two treatment arms using a 1:1 ratio. The computer chooses a treatment dynamically, based on the tumor stage (Stage 0 vs. I), and VHI at baseline (\<34 vs. ≥34) as two predefined stratification factors.

Allocation will be done via a dedicated website within the clinical trial management system also containing the electronic case report forms. Only system administrators who are otherwise not involved in the trial will have access to the algorithm and stored lists during the recruitment period. Investigators receive the allocation only after registration of a patient. The underlying randomization lists and details of the minimization algorithm will not be disclosed but kept securely at CTU Bern. All these measures will help to ensure concealment of allocation.

Blinding procedures Due to obvious differences between surgery and radiotherapy, it is not possible to have a blinded design in this study setting.

• Roughness - Breathiness - Hoarseness (RBH) [Time frame: 6, 12, 18, and 24 months]
• Jitter and shimmer (JS) [Time frame: 6, 12, 18, and 24 months]
• Glottal-to-Noise Excitation Ratio (GNE) [Time frame: 6, 12, 18, and 24 months]
• Singing Power Ratio (SPR) [Time frame: 6, 12, 18, and 24 months]
• Loco-regional control of the disease [Time frame: 2 and 5 years]
• Toxicity / Morbidity [Time frame: 2 and 5 years]
• Voice Handicap Index separately assessed at 6, 12, 18, and 24 months [Time frame: 6, 12, 18, and 24 months]