Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
September 18, 2026Frontiers in PhysiologyOpen Access

Explainable random forest–SHAP framework for Montreal phenotype–stratified prediction of one-year complications in Crohn’s disease

View Full Paper
Ask AI
Bookmark
Share

Authors

KXKun XiaXi'an Jiaotong UniversityYSYing ShiFirst Affiliated Hospital of Jinan UniversityXHXiaohan HuangNanjing University of Chinese Medicine

Discussion

Loading...

Member takes

Implication

Multicenter cohort study reveals machine-learning prediction of one-year complications in Crohn's disease, suggesting interpretable algorithms aid phenotype-stratified risk assessment.

Key Points

  • To develop and validate an interpretable machine-learning framework that predicts one-year risks of bowel resection, perianal complications, and abdominal complications stratified by Montreal phenotypes in Crohn's disease.
  • Analyzed pooled clinical data from 370 patients across two centers using stratified 10-fold cross-validation.
  • Implemented LASSO feature selection, SMOTE-Tomek resampling, and evaluated seven machine-learning algorithms on discrimination, calibration, and decision-curve analysis.
  • Applied SHapley Additive exPlanations (SHAP) to interpret feature contributions and assess phenotype-stratified associations with complication risk.
  • Complications within one year post-discharge included bowel resection in 70 patients (18.9%), perianal complications in 142 (38.4%), and abdominal complications in 84 (22.7%).
  • Random Forest yielded the most consistent performance, achieving an AUROC of 0.86 (95% CI, 0.82–0.90) for bowel resection, 0.71 (95% CI, 0.64–0.78) for perianal complications, and 0.73 (95% CI, 0.69–0.78) for abdominal complications.
  • SHAP interpretation uncovered distinct outcome drivers and marked phenotypic variation, linking isolated small-bowel disease to elevated resection and abdominal risks, while isolated colonic disease carried the highest perianal complication burden.

Cite This Study

Xia et al. (2026) studied this question.

synapsesocial.com/papers/6aad0aaede0393d728b89012https://doi.org/10.3389/fphys.2026.1902465
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Development and validation of an explainable machine learning model for predicting postoperative pulmonary complications after lung cancer surgery: a machine learning study2025 · 31 citations
  2. 2Achievement of Endoscopic Remission After Induction Reduces Hospitalization Burden in Crohn’s Disease: Findings From a Pooled Post Hoc Analysis of Risankizumab and Upadacitinib Phase III Trials2024 · 4 citations
  3. 3Development and validation of a machine learning-based predictive model for clinical remission in Crohn’s disease patients receiving Adalimumab therapy2025 · 7 citations
  4. 4Patient-reported outcome measures poorly correlate with objective inflammatory bowel disease activity measures: a systematic review2025 · 4 citations
  5. 5ECCO Guidelines on Inflammatory Bowel Disease and Malignancies2022 · 319 citations