LAY SUMMARY

A type of gullet cancer (oesophageal adenocarcinoma) has become the 5th commonest UK cause of cancer death. Unfortunately, by the time patients have symptoms, the cancer is often incurable. People with Barrett's oesophagus (change of gullet lining occurring in some with acid reflux) at risk of this cancer can have regular check-ups, involving examination through an endoscope (an instrument inserted by mouth, under mild sedation if required). A small proportion of people with Barrett's develop further changes (which might become cancer) in the gullet lining; if they do, it is important to remove the affected tissue before cancer develops, or when it is at an early stage.

There are several ways of removing this tissue but the investigators do not know which is best. The standard treatment is surgery, but there is a small risk of dying from the operation, and patients often suffer complications affecting them for a year or more afterwards. Two endoscopic treatments do not involve surgery. Both involve removing visible abnormalities by a technique called endoscopic resection, followed by cauterising the remaining Barrett's gullet lining by 1 of 2 techniques. One is recommended by the National Institute for Health and Clinical Excellence, but it is expensive and less widely available than the second. No-one has compared these treatments with each other, nor with surgery, in randomised trials (the most reliable way of deciding which is best). Patient groups say they would prefer to avoid surgery if the alternative works, and have encouraged us to do trials.

This feasibility study is a vital step towards two trials: (a) a trial to compare the two non-surgical techniques and (b) a trial comparing surgery with endoscopic treatment. It will help us find out whether it will be possible to enroll and retain enough patients by using several centres, and to identify/resolve any other potential barriers to recruitment and retention, including exploring viewpoints of patients and surgeons.

The study will comprise 2 parts:

Firstly, randomising up to 100 suitable patients over a 1 year period, identified at the upper gastrointestinal cancer specialist multidisciplinary team meeting in 6 expert English centres. Patients will have either high grade dysplasia (HGD) or early cancer in Barrett's oesophagus (BE) and will be randomised to two curative endoscopic non-surgical therapies (endoscopic resection [ER] and argon plasma photocoagulation [APC] versus ER and radiofrequency ablation [RFA]). All techniques are used in current clinical practice, but have never been directly compared.

Secondly, 2 qualitative studies in which we will examine clinicians' and surgeon's attitudes towards a trial of oesophageal surgery compared to endoscopic treatment (ER and ablation) by questionnaire. Qualitative interviews with a purposive sample of patients will explore their views of randomisation, recruitment and participation to help identify and pre-empt problems in subsequent planned trials of endoscopic treatments compared to each other and to surgical removal of all or part of the gullet.

Participants will be randomised to the ER plus RFA or ER plus APC group at enrollment. The randomisation schedule will be managed by the University of Leicester Clinical Trials Unit, using a computer generated random assignment.

Randomisation will be stratified for length of Barrett's epithelium (< 5; 5-10; > 10 cm).

DETAILS OF SAMPLE SIZE

Currently available estimates of eradication rates have poor precision, being derived from small case series. Consequently the sample size has been chosen to allow estimation of the quantities of interest whilst not exposing too many patients to trial procedures.

10- -15 new cases per million population per year are treated. Each participating centre is based on an upper gastrointestinal cancer treatment centre serving 1.5- - 2 million people.

Since each centre would expect to see 15- - 20 new patients a year, we aim recruit up to 100 patients over 1 year from 6 UK centres.

ENDOSCOPIC INTERVENTIONS

ER, RFA and APC are all techniques used in current practice, with which all investigators are familiar. We will ensure that all investigators are using the same techniques and that histological assessment is scrutinised by an an external expert pathology panel (see quality control below).

Best practice is to perform ER of all visible lesions. We will therefore aim for complete resection of all visible lesions initially and ablate residual flat Barrett's mucosa at subsequent treatment sessions (at 2 monthly intervals up to 8 months after the initial treatment by ER).

The treatment phase of the trial terminates at 8 months. At T= 12 months, diagnostic high resolution endoscopy is performed with targeted biopsies of any abnormal areas as well as 4 quadrant biopsies at 2 cm intervals of the area containing or previously containing the BE. The biopsies are in order to identify 'buried' Barrett's glands under the new squamous epithelium. The clinical endpoint is at 1 year when recurrent or persistent HGD or cancer will be assessed. Any residual BE (not containing HGD or cancer) will be assessed and recorded.

QUALITATIVE STUDIES

A purposive sub-sample of patients (sampling strategy designed to include a wide range of views and experiences, including patients from each centre) will be interviewed by telephone after being invited to take part in the feasibility trial, using a topic guide (developed collaboratively with lay representatives, with an emphasis on encouraging patients to describe their own perspectives freely - a preliminary topic guide has been developed and is included with this application), exploring views on being invited to participate in research in which they are randomised to different treatment options. The interview would also invite views on conducting a trial of endotherapy versus surgery. Interviews will be audio-recorded with separate consent and transcribed verbatim by transcribers working to professional standards of confidentiality.

Analysis of Qualitative Data

Analysis of the interview transcripts will be based on the constant comparative approach and managed by NVivo software. The findings will thus include a set of issues that are important from a patient perspective, that can help/hinder recruitment and retention, to inform subsequent trials.

Surgeon/clinician Questionnaires

A purposive sample of surgeons and clinicians at expert centres will receive questionnaires to explore views to endotherapy and surgery for the treatment of Barrett's oesophagus with HGD or early cancer and on the proposed trials (including 'free text' options inviting views on randomization to a trial comparing surgery to endoscopic treatment). Questionnaires will be distributed to up to 100 oesophagogastric surgeons and clinicians (lead MDT gastroenterologists or expert endoscopists) from the participating centres and other centres in the UK as advised by our surgical lead. The questionnaires will be administered by e mail. A second e mail will follow after a month in the case of non-response, but those who chose not to respond thereafter will not be further contacted.

Quality of life assessments and healthcare cost assessment

Quality of Life will be assessed at baseline and at the 6 and 12 months visits, using the EORTC Quality of Life Questionnaire version 3.0 (EORTC QLQ-C30) and module QLQ-OES18 to assess specific relevant aspects of quality of life related to oesophageal disease. General quality of life will also be assessed using EQ-5D. Additional data on healthcare utilisation will be collected at baseline, 6 and 12 months using a questionnaire we have developed which patients will be asked to complete at these scheduled visits for endoscopy.

The perspective of the economic analysis will be that of the NHS and personal social service. Costs to be obtained will include the costs of endoscopic treatment and continued endoscopic surveillance in both groups, as well as any additional intervention after the initial 8- month therapeutic phase of the trial. The costs of any salvage treatment at any stage will be included. We will test the validity and feasibility of administering the economic evaluation questionnaire (patient healthcare utilisation questionnaire and EQ-5D), examining the response rate achieved and levels of missing data. EORTC QLQ- C30 will also be tested to directly estimate QALYs using EORTC- 8D. Both EQ-5D and EORTC will be used to obtain the appropriate information for cost per QALY analysis in future economic evaluation. The economic analysis will inform sample size calculation and other necessary information to calculate the cost savings for future economic evaluation.

INCLUSION CRITERIA:

• Histology: high grade dysplasia (HGD) or early cancer with a maximum depth of invasion on endoscopic resection (ER) of T1m3
• Endoscopic ultrasound if any endoscopically visible abnormality: negative for T2 invasion or greater, and for suspicious lymph nodes.
• CT scan (thorax & top 1/3 of abdomen): negative for evidence of locally advanced or metastatic disease (done at the discretion of the multidisciplinary team, for invasive cancer only - T1m disease); PET-CT will not usually be required but may be carried out if indicated at the discretion of the multidisciplinary team.
• Suitability for trial agreed at local upper gastrointestinal cancer multidisciplinary team (MDT).
• Able to give informed consent
• Able (if applicable) to discontinue Clopidogrel for 7 days before & after endotherapy i.e. 14 days in total.
• Able (if applicable) to discontinue Warfarin with or without a bridging plan using low molecular weight heparin. The Warfarin can be restarted 1-7 days after endotherapy according to the local endoscopist's usual clinical practice.

EXCLUSION CRITERIA:

• Histology: depth of invasion beyond muscularis mucosae histologically (> T1m), poorly differentiated T1m cancers or lymphatic invasion or vascular invasion.
• Short tongues (<2 cm) of Barrett's epithelium that could be completely removed by Endoscopic Resection
• No localised endoscopically identifiable abnormality by high definition endoscopy (with or without magnification or chromo-endoscopic techniques)
• Prior oesophageal endoscopic therapy: e.g. Photodynamic Therapy, Endoscopic resection, prior ablation by other techniques such as argon ablation.
• Existing symptomatic stricture or one caused by the study diagnostic ER unless this can be dilated and the patient is then judged to be suitable for endoscopic treatment by the expert endoscopist.
• History of: radiation to mediastinum, oesophageal surgery (except fundoplication without complication), oesophageal varices or coagulopathy.