Exploring the learning curve in minimally invasive esophagectomy: a systematic review

Abstract

Summary Minimally invasive esophagectomy (MIE) has been shown to be superior to open esophagectomy with reduced morbidity, mortality, and comparable lymph node (LN) harvest. However, MIE is technically challenging. This study aims to perform a pooled analysis on the number of cases required to surmount the learning curve (LC), i.e. NLC in MIE. PubMed, Embase, Scopus, and the Cochrane Library were systematically searched for articles from inception to June 2022. Inclusion criteria were articles that reported LC in video-assisted MIE (VAMIE) and/or robot-assisted MIE (RAMIE). Poisson means (95% confidence interval [CI]) was used to determine NLC. Negative binomial regression was used for comparative analysis. There were 41 articles with 45 data sets (n = 7755 patients). The majority of tumors were located in the lower esophagus or gastroesophageal junction (66.7%, n = 3962/5939). The majority of data sets on VAMIE (n = 16/26, 61.5%) used arbitrary analysis, while the majority of data sets (n = 14/19, 73.7%) on RAMIE used cumulative sum control chart analysis. The most common outcomes reported were overall operating time (n = 30/45) and anastomotic leak (n = 28/45). Twenty-four data sets (53.3%) reported on LN harvest. The overall NLC was 34.6 (95% CI: 30.4–39.2), 68.5 (95% CI: 64.9–72.4), 27.5 (95% CI: 24.3–30.9), and 35.9 (95% CI: 32.1–40.2) for hybrid VAMIE, total VAMIE, hybrid RAMIE, and total RAMIE, respectively. NLC was significantly lower for total RAMIE compared to total VAMIE (incidence rate ratio: 0.52, P = 0.032). Studies reporting NLC in MIE are heterogeneous. Further studies should clearly define prior surgical experiences and assess long-term oncological outcomes using non-arbitrary analysis.