RESUMO
PURPOSE: A cohort study analysing phases and outcomes of the learning curve required to master minimally invasive total adventitial resection of the cardia. METHODS: Data from 198 consecutive oesophagectomies performed by a single surgeon was collected prospectively. Patients' stratification reflected chronologically and technically the four main phases of the learning curve: open surgery (open total adventitial resection of the cardia (TARC), n = 45), hybrid Ivor Lewis oesophagectomy (HILO, n = 50), laparoscopic-thoracoscopic assisted (LTA, n = 56) and totally minimally invasive TARC (TMI TARC, n = 47). Operating time, hospital stay, specimen lymph nodes and resection margins were analysed. Five-year survival was the main long-term outcome measured. RESULTS: Overall 5-year survival was 45%. Perioperative mortality was 1.5% (n = 3). Hospital stay was 22 ± 23 days. Specimen lymph node median was 20 (range: 15-26). Resection margins were negative (R = 0, American College of Pathologists) in 193 cases (97.4%). Five-year survival in the four phases was 37.8%, 44.9%, 42.9% and 55.3%, showing a positive trend towards the end of the learning curve (p = 0.024). Median specimen lymph nodes was 20 (range: 15-22) for open TARC, 18.5 (13-25) for HILO, 19.5 (15-25) for LTA and 23 (18-30) for TMI TARC (p = 0.006). TMI TARC, adenocarcinoma, R >0, T >2, N >0 and LyRa (ratio positive/total specimen nodes) were associated with survival on univariate analysis. T >2 and LyRa independently predicted worse survival on multivariate analysis. CUSUM analysis showed surgical proficiency gain since laparoscopy was introduced. CONCLUSION: Mastering minimally invasive TARC requires a long learning curve. TMI TARC is safe and oncologically appropriate and may benefit long-term survival: it should be validated by randomised trials as a standardised anatomical resection for tumours of the oesophagogastric junction.