Truace
Subscribe

Clinical significance of intraoperative dissection time and machine learning-based prediction in laparoscopic distal gastrectomy for gastric cancer

Background Establishing a clinically meaningful cutoff value for intraoperative dissection time and developing predictive models for this metric among patients with distal gastric cancer. Methods This retrospective multicenter study included 849 patients with distal gastric cancer who underwent laparoscopic distal gastrectomy between November 2020 and May 2025 at three tertiary centers. Using postoperative complications as the status variable, determine the optimal cutoff value for intraoperative dissection time…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1934· cite & verify
Clinical significance of intraoperative dissection time and machine learning-based prediction in laparoscopic distal gastrectomy for gastric cancer
Diagnosis-and-treatment-of-perforated-or-bleeding-peptic-ulcers-2013-WSES-position-paper-1749-7922-9-45-S4 by Di Saverio S, Bassi M, Smerieri N, Masetti M, Ferrara F, Fabbri C, Ansaloni L, Ghersi S, Serenari M, Coccolini F, Naidoo N, Sartelli M, Tugnoli G, Catena F, Cennamo V, Jovine E. CC BY 4.0 · https://creativecommons.org/licenses/by/4.0

In brief

Researchers analyzed 849 patients who had laparoscopic distal gastrectomy for distal gastric cancer at three tertiary centers between November 2020 and May 2025. They found a dissection time of 133 minutes best separated patients by complication risk, then trained six machine learning models to predict who would fall into the long-dissection group.

The work suggests surgical teams could flag patients likely to face prolonged dissection before surgery and plan accordingly. However, the models showed only moderate discrimination, the data were retrospective, and long dissection time was not an independent predictor of survival, so its clinical usefulness remains to be shown.

Main points

  1. The optimal dissection time cutoff was 133 minutes, derived using postoperative complications as the status variable.
  2. Patients in the long dissection group had more complications and worse unadjusted overall and recurrence-free survival.
  3. Long dissection time was not independently associated with prognosis in multivariable analysis.
  4. Logistic regression had the highest test-cohort AUC (0.770), while random forest had the highest validation AUC (0.759).

The gain

Machine learning models identified patients likely to need long dissection time (over 133 minutes), a threshold linked to more postoperative complications, which may help preoperative risk evaluation.

The rundown

The study was retrospective and multicenter, covering patients treated at three tertiary centers from November 2020 to May 2025. Patients were grouped into short and long intraoperative dissection time based on the 133-minute cutoff.

The long dissection group differed in baseline characteristics: higher body mass index, larger tumors, more advanced clinical stage, and a higher share of patients given neoadjuvant chemotherapy. These differences help explain why survival gaps appeared in unadjusted but not adjusted analyses.

Six machine learning models were developed and validated. Logistic regression led in the test cohort and random forest in the validation cohort, and the authors regarded random forest as the most stable.

Sources

  1. Peer-reviewedSurgical Endoscopy2026-10-16

The debate