Abstract
BACKGROUND: Conventional minimally invasive esophagectomy (C-MIE) is the mainstay for locally advanced esophageal cancer. However, the relationship among facility size, risk-adjusted mortality and morbidity in C-MIE remains unclear. This study aims to clarify whether C-MIE should be consolidated into high-volume centers in Japan. METHODS: Risk models for perioperative mortality and morbidity were created using the Japanese National Clinical Database (NCD) data. NCD data registered between January 2016 and December 2020, including 14 152 C-MIE records. The developed risk models were used to estimate the ratio of expected to observed events (perioperative deaths or complications) (O/E ratio) for each facility. RESULTS: Regarding the risk model performances, the C-indices of the perioperative mortality risk prediction models were 0.793. The O/E ratio and 95% confidence interval (CI) for perioperative mortality were facility size < 10 MIEs/year, O/E ratio: 1.368 and 95% CI: 1.140-1.597; facility size 10-29 MIEs/year, O/E ratio: 0.886 and 95% CI: 0.644-1.127; and facility size ≥ 30 MIEs/year, O/E ratio: 0. 61 and 95% CI: 0.342-0.892. Conversely, there were no significant differences in morbidity rate by facility size. CONCLUSIONS: The risk of perioperative mortality from C-MIE was lower in hospitals with larger facilities than those with smaller facilities; therefore, consolidating patients for C-MIE in high-volume hospitals is necessary.