64.02 Learning curve of minimally invasive Ivor-Lewis esophagectomy

F. Van Workum1, G. H. Berkelman3, A. E. Slaman4, M. Stenstra1, M. I. Van Berge Henegouwen4, S. S. Gisbertz4, F. J. Van Den Wildenberg5, F. Polat5, M. Nilsson2, T. Irino2, G. A. Nieuwenhuijzen3, M. D. Luyer3, C. Rosman1  1Radboudumc,Surgery,Nijmegen, GELDERLAND, Netherlands 2Karolinska Institutet,Surgery,Stockholm, -, Sweden 3Catharina Hospital,Surgery,Eindhoven, BRABANT, Netherlands 4AMC,Surgery,Amsterdam, NOORD HOLLAND, Netherlands 5Canisius-Wilhelmina Ziekenhuis,Surgery,Nijmegen, GELDERLAND, Netherlands

Introduction: Totally minimally invasive Ivor-Lewis esophagectomy (TMIE-IL) has a learning curve but the length of the learning curve and the extent of learning curve associated morbidity for surgeons experienced in TMIE-McKeown is unknown.

Methods: This study was performed in 4 high volume European esophageal cancer centers from December 2010 until April 2016. Surgeons experienced in TMIE-McKeown changed operative technique to TMIE-IL. All consecutive patients with esophageal carcinoma undergoing TMIE-IL with curative intent were included. Baseline, surgical and outcome parameters were analyzed in quintiles and were plotted in order to explore the learning curve. Textbook outcome (the percentage of patients in which the process from surgery until discharge was <21 days and uneventful in terms of complications, interventions, mortality and oncological aspects) was also analyzed. CUSUM analysis was performed in order to determine after how many cases proficiency was reached. An area under the curve analysis was performed to calculate the learning associated anastomotic leakage and costs.

Results: Four hundred and sixty eight patients were included. In one hospital, ASA classification was significantly higher in quintile 2 and 3 (p=0.01) and in one hospital, more distal esophageal tumors were operated in quintile 4 and 5 (p=0.01). In the pooled curve analysis, anastomotic leakage decreased from 26% at introduction of MIE-IL to 8% at the plateau phase which occurred after 121 cases. Textbook outcome increased from 39% to 60% and the plateau phase occurred after 128 cases. Learning curve associated anastomotic leakage occurred in 42 patients and this excess morbidity was associated with more than € 2 million in healthcare costs.

Conclusion: TMIE-IL has a significant learning curve. Learning curve associated morbidity and costs are substantial, even for surgeons experienced in TMIE-McKeown. The length of the learning curve was more than 100 operations.