Robot-assisted versus conventional minimally invasive oesophagectomy for oesophageal squamous cell carcinoma (RAMIE): a multicentre, open-label, randomised, phase 3, non-inferiority trial
Mené sur 362 patients atteints d'un carcinome épidermoïde résécable de l'oesophage (durée médiane de suivi : 71,5 mois), cet essai randomisé multicentrique de phase III évalue la non-infériorité, du point de vue de la survie globale, d'une oesophagectomie mini-invasive assistée par robot par rapport à une oesophagectomie thoracoscopique
Background: Robot-assisted minimally invasive oesophagectomy is used increasingly worldwide. However, no large-scale, multicentre, randomised controlled trial has compared long-term survival as the primary endpoint between robot-assisted oesophagectomy and conventional thoracoscopic oesophagectomy. We aimed to confirm the non-inferiority in overall survival of robot-assisted oesophagectomy over thoracoscopic oesophagectomy in patients with resectable oesophageal squamous cell carcinoma.
Methods: This multicentre, open-label, randomised, controlled, phase 3, non-inferiority trial (RAMIE) was conducted at six hospitals in China. Patients aged 18–75 years, with biopsy-proven squamous cell carcinoma, Eastern Cooperative Oncology Group scores of 0–2, and with tumour and nodal classifications of cT1–4a, N0–2, M0, or M1 (supraclavicular lymph nodes metastasis) were eligible. Patients were randomly assigned (1:1) using a computer-generated randomisation list and stratified by neoadjuvant therapy to robot-assisted oesophagectomy or thoracoscopic oesophagectomy, both with at least two-field lymphadenectomy. The primary endpoint was overall survival, analysed in the intention-to-treat population. The non-inferiority margin was 9% for 5-year overall survival (the upper limit of 95% CI of the hazard ratio [HR] was 1·33). Harms were assessed in the per-protocol population, defined as all eligible participants undergoing resection. This trial was registered with ClinicalTrials.gov, number NCT03094351 and is completed.
Findings: Between Aug 2, 2017, and Dec 23, 2019, 362 patients were randomly assigned (183 to robot-assisted oesophagectomy and 179 to thoracoscopic oesophagectomy. 309 (85%) of 362 patients were men and 53 (15%) were women. Two patients in each group did not undergo resection and were excluded from the per-protocol population. Median follow-up of the planned final analysis was 71·5 months (IQR 63·9–81·8). At 5 years, overall survival was 69·4% (95% CI 62·1–75·6) with robot-assisted oesophagectomy versus 56·2% (48·5–63·2) with thoracoscopic oesophagectomy (HR 0·71, 95% CI 0·51–0·97), confirming non-inferiority (one-sided pnon-inferiority=0·0001; exploratory analysis of p=0·032 for superiority). Intraoperative conversions to open surgery were similar between study groups (seven [4%] of 181 patients in the robot-assisted group vs six [3%] of 177 in the thoracoscopic group). Postoperative grade 3 or higher complications were comparable (22 [12%] of 181 patients in the robot-assisted group vs 18 [10%] of 177 in the thoracoscopic group). One treatment-related death occurred in each group.
Interpretation: In patients with resectable oesophageal squamous cell carcinoma, robot-assisted oesophagectomy was non-inferior, and seemed superior, to thoracoscopic oesophagectomy in terms of 5-year overall survival.
The Lancet Gastroenterology & Hepatology , résumé, 2026