Modified BEST-J Score Model Predicts Bleeding after Endoscopic Submucosal Dissection with Fewer Factors

Area of Science:

• Gastroenterology
• Oncology
• Surgical Innovation

Background:

• Endoscopic submucosal dissection (ESD) is a minimally invasive procedure for early gastrointestinal cancers.
• Post-procedural bleeding is a significant complication following gastric ESD.
• Existing prediction models, like the BEST-J score, can be complex to apply in clinical practice.

Purpose of the Study:

• To develop and validate a simplified prediction model for post-ESD bleeding.
• To incorporate the prognostic nutritional index (PNI) into a novel scoring system.
• To compare the accuracy of the new model against the established BEST-J score.

Main Methods:

• A retrospective cohort study involving 449 patients who underwent gastric ESD.
• Patients were divided into a model cohort (n=299) for model construction and a validation cohort (n=150).
• Independent risk factors for post-ESD bleeding were identified, including P2Y12RA use, tumor size > 30 mm, lower gastric lesion location, and PNI ≤ 47.9.

Main Results:

• The modified BEST-J score, incorporating the identified risk factors, demonstrated a sensitivity of 73.9% and specificity of 78.1%.
• In the validation cohort, the modified BEST-J score showed comparable accuracy to the original BEST-J score (ROC-AUC 0.77 vs. 0.75, p=0.81).
• Post-ESD bleeding occurred in 8.4% of patients in the model cohort.

Conclusions:

• The simplified, modified BEST-J score effectively predicts post-ESD bleeding.
• This new model offers a more straightforward clinical tool with equivalent predictive performance to the BEST-J score.
• The prognostic nutritional index is a valuable component in assessing bleeding risk after gastric ESD.