Resource Sparing Curative Radiotherapy for Locally Advanced Squamous Cell Cancer of the Head and Neck: The HYPNO Trial

Worldwide, head and neck squamous cell carcinomas (HNSCC) constitute approximately 7% of all incident cancers, and 75% of these are seen in low- and middle-income countries. Life-style factors, in particular tobacco and alcohol consumption, are major etiological factors although an increasing number of HNSCC cases are associated with viral infections (Epstein Barr Virus and Human Papilloma Virus). Radiation therapy (RT) alone or combined with cytotoxic or molecular targeted agents is a mainstay as definitive treatment for previously untreated locally advanced disease. This therapy offers organ and functional preservation in many cases with an approximate 30 to 50% of cases obtaining longterm loco-regional tumor control. Accelerated RT, that is increasing the weekly delivered radiation dose above the conventional 10 Gy per week, has been shown in a large number of randomized controlled trials to be associated with an improved efficacy to toxicity ratio relative to standard fractionation provided a careful balance between total dose, dose per fraction and overall treatment time is chosen. The Danish Head and Neck Cancer Group trial DAHANCA 6/7 tested a schedule of accelerated normofractionated radiation therapy for HNSCC delivering 6 fractions of 2.0 Gy per week to a total dose of 66-68 Gy. This schedule gave a 10% improvement in loco-regional tumor control compared with the same dose delivered with 5 fractions per week. The DAHANCA 6/7 schedule was tested in the IAEA ACC trial, comparing again 6 vs. 5 fractions per week but allowing the total dose to range from 66 Gy to 70 Gy. The outcome of the IAEA ACC trial also showed a significant 13% improvement in loco-regional control in the accelerated arm. Thus, accelerating RT by delivering 6 fractions per week may reasonably be considered a standard of care for definitive RT alone. Clinical evidence points in the direction of a 4 week accelerated, hypofractionated schedule as being a radiobiologically interesting alternative to accelerated schedules using more fractions for treatment of HNSCC. A series of exploratory calculations suggested that an attractive schedule could be 55 Gy in 20 fractions (2.75 Gy per fraction) delivered over 4 weeks and this schedule was chosen for the test arm of HYPNO. There is historical experience with this schedule in the North of UK and in many countries in the British Commonwealth. The aim of this study is to test whether this resource-sparing 4-week course of accelerated hypofractionated radiotherapy is non-inferior to radiotherapy delivering 33 fractions over 5.5 weeks.

The study design is a stratified, balanced, and randomized study (phase III) recruiting patients with Stage I to IV squamous cell carcinoma of the pharynx, larynx and oral cavity with the exception of paranasal sinus, nasopharyngeal and stage I to II glottic carcinomas. Patients will be assigned to the test arm or the control arm by central, randomization, stratified according to the following characteristics:

• Institution
• Performance status, WHO 0-1 vs. 2
• Tumor sub-site within the head and neck: pharynx, larynx, oral cavity
• Chemotherapy (yes/no) Other important prognostic parameters are: 2D vs. 3DCRT vs. IMRT; Tumor stage: T1-2 vs. T3-4; Nodal stage: N0-1 vs. N2-3. These will be randomly distributed among the treatment arms as a result of the randomization.

Patients are reviewed at least once a week during treatment. Time and severity of early radiation reactions in mucosa and skin are noted. These data are registered on the Treatment Response Form. Patients are seen about 2 months after the end of treatment to record persistent early toxicities and early tumor response, both at the primary and nodal site. Afterwards, the patients are seen every 3 months for 2 years and 6 monthly for 3 years bringing the total trial follow-up to just over 5 years. CRFs are designed to record all the information for an individual study subject required by the study protocol. The purpose of the CRF is threefold: i) to ensure data collection in accordance with the study protocol; ii) to fulfill the regulatory requirements for data collection; iii) to facilitate the effective, comprehensive data processing and analysis, results reporting, and to promote structured collection of safety and efficacy data. The completed CRF's are submitted by email to the Data Center: hypno@humonc.wisc.edu where the data will be electronically transferred to the central trial database. HYPNO uses a rather novel design as a non-inferiority trial with dual primary endpoints. The primary efficacy endpoint is 3 year loco-regional control (LRC) - even though 5 year disease status will also be collected - and the primary toxicity endpoint is 3 year grade 2 or higher (moderate or severe) side effects (TOX). All times to events will be measured from the date of randomization. The test arm (HYPNO -T) will be declared non-inferior to the control arm if both i) LRC in the test arm is non-inferior to that of the control arm AND ii) Late toxicity in the test arm is non-inferior (i.e. not worse) than in the control arm. This design has been developed in collaboration with the HYPNO trial statistician, Prof. Richard Chappell, Department of Biostatistics and Medical Informatics, University of Wisconsin - Madison.

Statistical formulation of the non-inferiority hypotheses with dual primary endpoints:

HYPOTHESIS I:

H0: LRC in the TEST arm less than LRC in the CONTROL arm (Test arm inferior to control arm with respect to LRC). HA: LRC in the TEST arm greater than or equal to LRC in the CONTROL arm (Test arm non-inferior to control arm with respect to LRC).

HYPOTHESIS II:

H0: TOX in the TEST arm greater than TOX in the CONTROL arm (Experimental treatment inferior to control treatment with respect to late toxicity).

HA: TOX in the TEST arm less than or equal to TOX in the CONTROL arm (Experimental treatment non-inferior to control treatment with respect to late toxicity).

The required sample size in this kind of design depends heavily on the choice of the inferiority margin, with decreasing values leading to a steep increase in the sample size. It also depends to some extend on the baseline 3 year LRC and TOX rates. These were selected based on the published results from the IAEA-ACC trial. Thus, the target sample size is 836 patients. The primary analysis will be performed for all randomized patients using the intention-to-treat principle.

A QA committee will be formed and they will be charged with all aspects of quality assurance and credentialing of centers for participation in HYPNO. Each participating center should appoint a contact person for QA issues (physician or physicist).

Data exchange dummy run Before the first patient can be randomized in a center, a 'dry run' of the data exchange is conducted. To this end, a HNSCC patient from the center will be arbitrarily chosen by the local investigator and a set of CRFs filled in for this case and submitted electronically to the Data Center.

Machine level QA An associated GCP and QA program securing compliance with the protocol will be established between the IAEA, the Data Center and involved departments. Participation in the IAEA TLD audit is mandatory for centers participating in the HYPNO trial. A credentialing QA program will be developed for departments using IMRT.

Patient level QA Radiotherapy plans and dose-volume histogram (DVH) data for the first two patients on trial from a center will be submitted for central review, if possible in electronic form. During the period when a center is accruing patients for the trial, the Data Center will randomly select a subset of patients (5-10%, a smaller proportion reviewed for highly accruing centers) for a 'crossed' review where 3-4 centers will review cases. It is expected that the total number of cases reviewed for HYPNO will not exceed 50.

An Independent Data Monitoring Committee (IDMC) will be appointed. Recommendations to stop patient accrual early for safety or efficacy reasons can only be made by the IDMC. Participating centers will not be informed about comparative effectiveness or toxicity in the trial arms as long as the trial is open for accrual.

An associated translational research program is planned. Formally these studies will run as separate research protocol and separate informed consent from patients will have to be obtained according to the relevant national regulations.