Scholarlink

Background <p>Delayed bleeding is a serious complication after endoscopic submucosal dissection (ESD) for early gastric cancer and precancerous lesions. Existing prediction models predominantly rely on pre-procedural factors, such as antithrombotic therapy (ATT), and overlook critical intraoperative variables. This study aimed to develop and validate a novel, procedure-centered risk model that integrates these intraoperative variables to improve predictive performance.</p> Methods <p>In this retrospective, single-center study, we enrolled 1523 patients undergoing gastric ESD (January 2018 to September 2025). The primary outcome was delayed post-ESD bleeding. We focused on variables available before or immediately after ESD. Independent risk factors were identified via multivariate logistic regression to construct the Bleeding after ESD Trend from China (BEST-C) score. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration, and decision curve analysis, with internal validation via bootstrapping. Its discriminative ability was directly compared with two established models (BEST-J and BEST-FH).</p> Results <p>Delayed bleeding occurred in 52 patients (3.4%). Multivariate analysis revealed that dual-ATT interruption, intraoperative bleeding, and the interaction between intraoperative bleeding and muscular layer injury were independent risk factors, whereas the use of hemostatic clips was identified as a protective factor. The BEST-C model demonstrated excellent discrimination (AUC = 0.868) and calibration, significantly outperforming the BEST-J and BEST-FH scores (AUC difference 0.321, <i>P</i> &lt; 0.001). The derived integer score effectively stratified patients into low- (bleeding rate 1.0%), intermediate- (11.6%), and high-risk (88.2%) groups.</p> Conclusion <p>The BEST-C score, which uniquely incorporates critical intraoperative variables, most notably a key interaction term, provides a robust tool, centered on intraoperative factors, for immediate post-ESD risk stratification of delayed bleeding after gastric ESD. It demonstrates superior predictive performance to existing models and shows promising clinical utility.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of delayed bleeding after endoscopic submucosal dissection in early gastric cancer and precancerous lesions by BEST-C score

  • 期刊论文

  • 发表日期：

    2026年2月23日

查看全文
  • Kaier Gu,
  • Tianer Gu,
  • Wei Xie,
  • Han Bu,
  • Tianhao Weng,
  • Juzeng Zheng,
  • Sisi Jin,
  • Yang Liu

摘要

Background

Delayed bleeding is a serious complication after endoscopic submucosal dissection (ESD) for early gastric cancer and precancerous lesions. Existing prediction models predominantly rely on pre-procedural factors, such as antithrombotic therapy (ATT), and overlook critical intraoperative variables. This study aimed to develop and validate a novel, procedure-centered risk model that integrates these intraoperative variables to improve predictive performance.

Methods

In this retrospective, single-center study, we enrolled 1523 patients undergoing gastric ESD (January 2018 to September 2025). The primary outcome was delayed post-ESD bleeding. We focused on variables available before or immediately after ESD. Independent risk factors were identified via multivariate logistic regression to construct the Bleeding after ESD Trend from China (BEST-C) score. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration, and decision curve analysis, with internal validation via bootstrapping. Its discriminative ability was directly compared with two established models (BEST-J and BEST-FH).

Results

Delayed bleeding occurred in 52 patients (3.4%). Multivariate analysis revealed that dual-ATT interruption, intraoperative bleeding, and the interaction between intraoperative bleeding and muscular layer injury were independent risk factors, whereas the use of hemostatic clips was identified as a protective factor. The BEST-C model demonstrated excellent discrimination (AUC = 0.868) and calibration, significantly outperforming the BEST-J and BEST-FH scores (AUC difference 0.321, P < 0.001). The derived integer score effectively stratified patients into low- (bleeding rate 1.0%), intermediate- (11.6%), and high-risk (88.2%) groups.

Conclusion

The BEST-C score, which uniquely incorporates critical intraoperative variables, most notably a key interaction term, provides a robust tool, centered on intraoperative factors, for immediate post-ESD risk stratification of delayed bleeding after gastric ESD. It demonstrates superior predictive performance to existing models and shows promising clinical utility.

Graphical abstract