BackgroundLocalised oesophageal cancer can be curatively treated with surgery (oesophagectomy) but the procedure is complex with a risk of complications, negative effects on quality of life and a recovery period of 6–9 months. Minimal-access surgery may accelerate recovery.ObjectivesThe ROMIO (Randomised Oesophagectomy: Minimally Invasive or Open) study aimed to establish the feasibility of, and methodology for, a definitive trial comparing minimally invasive and open surgery for oesophagectomy. Objectives were to quantify the number of eligible patients in a pilot trial; develop surgical manuals as the basis for quality assurance; standardise pathological processing; establish a method to blind patients to their allocation in the first week post surgery; identify measures of postsurgical outcome of importance to patients and clinicians; and establish the main cost differences between the surgical approaches.DesignPilot parallel three-arm randomised controlled trial nested within feasibility work.SettingTwo UK NHS departments of upper gastrointestinal surgery.ParticipantsPatients aged ≥ 18 years with histopathological evidence of oesophageal or oesophagogastric junctional adenocarcinoma, squamous cell cancer or high-grade dysplasia, referred for oesophagectomy or oesophagectomy following neoadjuvant chemo(radio)therapy.InterventionsOesophagectomy, with patients randomised to open surgery, a hybrid open chest and minimally invasive abdomen or totally minimally invasive access.Main outcome measureThe primary outcome measure for the pilot trial was the number of patients recruited per month, with the main trial considered feasible if at least 2.5 patients per month were recruited.ResultsDuring 21 months of recruitment, 263 patients were assessed for eligibility; of these, 135 (51%) were found to be eligible and 104 (77%) agreed to participate, an average of five patients per month. In total, 41 patients were allocated to open surgery, 43 to the hybrid procedure and 20 to totally minimally invasive surgery. Recruitment is continuing, allowing a seamless transition into the definitive trial. Consequently, the database is unlocked at the time of writing and data presented here are for patients recruited by 31 August 2014. Random allocation achieved a good balance between the arms of the study, which, as a high proportion of patients underwent their allocated surgery (69/79, 87%), ensured a fair comparison between the interventions. Dressing patients with large bandages, covering all possible incisions, was successful in keeping patients blind while pain was assessed during the first week post surgery. Postsurgical length of stay and risk of adverse events were within the typical range for this group of patients, with one death occurring within 30 days among 76 patients. There were good completion rates for the assessment of pain at 6 days post surgery (88%) and of the patient-reported outcomes at 6 weeks post randomisation (74%).ConclusionsRapid recruitment to the pilot trial and the successful refinement of methodology indicated the feasibility of a definitive trial comparing different approaches to oesophagectomy. Although we have shown a full trial of open compared with minimally invasive oesophagectomy to be feasible, this is necessarily based on our findings from the two clinical centres that we could include in this small preliminary study.Trial registrationCurrent Controlled Trials ISRCTN59036820.FundingThis project was funded by the NIHR Health Technology Assessment programme and will be published in full inHealth Technology Assessment; Vol. 20, No. 48. See the NIHR Journals Library website for further project information.