Learning curve for the abdominal procedure in total robot-assisted minimally invasive esophagectomy with simultaneous cervical approach: an initial experience

BACKGROUND: Robot-assisted minimally invasive esophagectomy (RAMIE) is increasingly being implemented; however, the feasibility and learning process of total RAMIE using robotic approaches for thoracic and abdominal phases remain to be fully evaluated, particularly with a dual-docking technique enabling simultaneous cervical and abdominal approaches. METHODS: We retrospectively analyzed 50 consecutive patients who underwent total RAMIE with retrosternal gastric conduit reconstruction and cervical anastomosis at a single institution between September 2024 and August 2025. Cumulative sum (CUSUM) analysis of abdominal operative time was used to assess the learning curve. Perioperative outcomes were compared between the phases. RESULTS: CUSUM analysis identified three phases: phases I (patients 1-6, learning phase), II (7-18, competence phase), and III (19-50, proficiency phase). Median abdominal operative time decreased from 139 min (124-154) in phase I to 104 min (80-137) and 87 min (66-146) in phases II and III, respectively (P < 0.001). Initially, this procedure was applied to cStage I patients without abdominal nodal involvement, and was gradually expanded to advanced disease. The number of harvested abdominal lymph nodes was 15 (6-27) in phases I and II and 17 (6-26) in phase III. Postoperative complications (Clavien-Dindo grade ≥ 2) occurred in 14% of patients, with no abdominal procedure-related morbidity observed. CONCLUSIONS: Procedural proficiency for the abdominal operative procedure was achieved in approximately 18 patients, after which operative efficiency improved without compromising oncologic thoroughness or patient safety. Total RAMIE with a two-team approach can be safely implemented in centers with experience.