Conversion during Minimally Invasive Left Pancreatectomy

Abstract

OBJECTIVES: To identify risk factors for conversion, develop a predictive Conversion Risk Score (CRS), and assess the association between conversion and severe postoperative complications. BACKGROUND: Conversion occurs in 15-30% of minimally invasive left pancreatectomies (MILP). Risk factors and potential negative impacts on postoperative outcomes are poorly described. METHODS: Retrospective, nationwide, multicenter study including all MILP (laparoscopy and robot) performed between 2010 and 2021. Risk factors for conversion were identified by multivariate mixed model, and a CRS was developed on a "training-set" and validated (calibration diagrams and ROC curves) on a "validation-set." The association between severe complications and conversion was assessed using a propensity score based on the main risk factors for severe complications: age, sex, BMI, ASA score, tumor malignancy, multi-organ resection, operative duration, blood loss, splenectomy. RESULTS: 2104 patients included from 55 centers. Conversion occurred in 15.6% of MILP. Its risk factors were male sex (OR=1.67;P=0.048), BMI≥25 kg/m2 (OR=2.15;P=0.004), history of laparotomy (OR=2.9;P<0.001), initial pancreatitis (OR=3.58;P=0.007), tumor size≥40 mm (OR=2.12;P=0.003), planned splenectomy (OR=2.63;P<0.001), unplanned splenectomy (OR=4.05;P=0.028), portal vein resection (OR=36.3;P=0.002), multi-organ resection (OR=12.97;P<0.001). A predictive CRS was created based only on preoperatively available variables (the first six), with scores ranging from 0 to 7, corresponding to a conversion risk of 2% to 100%. No association was observed with tumor malignancy, robotic approach, or pancreatectomy volume. Conversion was significantly associated with severe complications [OR=1.80(1.16-2.54)], independent of other risk factors for complications. CONCLUSIONS: Conversion during MILP can be predicted by CRS, aiding surgeons in decision-making, given its significant association with severe complications.