Multidimensional learning curve in laparoscopy-assisted gastrectomy for early gastric cancer.

Abstract

BACKGROUND: Laparoscopy-assisted gastrectomy (LAG) is a complex and time-consuming procedure, which is increasingly used for early gastric cancer (EGC). We provide a multidimensional analysis of the learning curve in LAG.

METHODS: Cumulative sum method was used to analyze outcomes of 109 patients undergoing LAG for EGC by one surgeon over a two year period; the influence of patient selection was evaluated. Target failure rate was set at 10%, with failure defined as open conversion, mortality, major morbidity, residual tumor, or inappropriate lymphadenectomy.

RESULTS: There were 19 failures--fourteen performance and five oncologic. The learning curve, which displayed a slight rising trend and three phases was achieved after 40 cases with selected patients; it was broken, however, by the introduction of advanced procedures and unselected patients.

CONCLUSIONS: Advanced procedures and broad indications in LAG should be delayed until a learning curve is completed under the target failure rate.