Learning curve in laparoscopic liver surgery: A single-surgeon risk-adjusted CUSUM analysis for postoperative complications of the first 320 laparoscopic liver resections

Abstract

Background: The Learning curve and the amount of procedures required to reach the 'mastery phase' in laparoscopic liver resections(LLR) are unknown. This study aimed to evaluate the short-term outcomes and investigate the learning curve(LC) during the first 320LLR performed by a single surgeon. Materials & Methods: Retrospective review of a prospective database of all patients who underwent LLR between October 2011 and November 2018. The IWATE Difficulty Score(IDS) was used to calculate a logistic regression risk model do adjust CUSUM for postoperative complications. Results: During 7years, 320/480 liver resections (66.7%) were performed by a pure laparoscopic approach. During these years, percentage of LLR increased from 45.0% to 69.2%(p=0.33). Malignancy was present in 275/320LLL (85.9%). Laparoscopic repeat hepatectomy was performed in 45/320(14.1%) cases. Median blood loss was 100ml(IQR:50-250ml) and median operative time was 130min (IQR:90-180min). Conversion rate was 2.5% and blood transfusion rate was 2.8%. Major complication rate (Clavien-Dindo=III) was 3.1% with 0.94% 90-day mortality. R0 resection rate was 93.6%. The single-surgeon CUSUM LC for postoperative complications showed the presence of three phases: an initiation phase of approximately 50 cases,an intermediate phase and a mastery phase beginning at approximately 140procedures. Important differences were found in complication rates for patients with different IDS (3.4% in IDS=0 VS 39.1% in IDS=10). A risk-adjusted CUSUM (RA-CUSUM) analysis showed an initial learning phase of 20 procedures,a plateau phase and a mastery phase after 60 cases with very low complication rates after 100 procedures. Conclusion: Even after a dedicated fellowship, with excellent short-term outcomes in early experience, a learning curve is present in laparoscopic liver surgery. The mastery phase is reached at approximately 60 cases. CUSUM and RA-CUSUM have an important role in monitoring surgical outcomes. However, adjustment for patient mix is critical. Difficulty scores can be used to adjust the CUSUM chart for different patient characteristics that significantly afffect the risk. CUSUM and RA-CUSUM charts for postoperative complications