Short-term surgical outcomes of robotic versus laparoscopic radical gastrectomy for gastric cancer: a propensity score-matched retrospective cohort study

Abstract

Robotic gastrectomy (RG) is increasingly used for gastric cancer, but real-world comparative data from Chinese institutions during the early phase of robotic program implementation remain limited. We retrospectively identified patients who underwent radical gastrectomy between January 2015 and October 2025. The primary comparison included 54 patients who underwent RG and 162 who underwent Laparoscopic gastrectomy (LG). After exclusion of patients with missing propensity-score covariates, 53 RG and 143 LG patients were eligible for matching. One-to-two nearest-neighbor matching was performed using a caliper of 0.2 standard deviations of the logit of the propensity score, yielding 45 RG and 88 LG patients. The principal outcomes was the number of lymph nodes retrieved. After matching, all included covariates were balanced. The median number of lymph nodes retrieved did not differ significantly between RG and LG (30 [interquartile range, 25-42] vs 34 [27-47]; 95% Confidence Interval [CI]: -10.0–2.0, p = 0.181). RG was associated with a shorter postoperative length of stay (16 vs 18 days; 95% CI: -4.1–0.1, p = 0.016) and fewer positive lymph nodes (0 vs 2; 95% CI: -3.3–0.7, p = 0.007). Median operative time was 310 minutes for RG and 300 minutes for LG (95% CI: -20.5–40.5, p = 0.072). CUSUM analysis revealed a V-shaped recovery for LN retrieval, and confirmed operative efficiency convergence to the LG benchmark by the late phase (300 min, p = 0.053), together indicating a trend toward improved performance during the learning process. RG and LG achieved comparable lymph-node yields, while RG was associated with a shorter postoperative hospital stay. Exploratory sequential analyses suggested improvement in operative efficiency but did not establish a proficiency threshold. Complete outcome ascertainment and contemporaneous comparisons are needed. Clinical Trial: The study was registered retrospectively in the Chinese Clinical Trial Registry on August 27, 2026 (ChiCTR2600131011; http://www.chictr.org.cn).