Integrating inflammatory and coagulation biomarkers for surgical risk stratification and treatment benefit assessment in Crohn's disease - PubMed
3 hours ago
- #Biomarkers
- #Machine Learning
- #Crohn's Disease
- Integrating inflammatory and coagulation biomarkers with disease-related features improves surgical risk stratification in Crohn's disease.
- A gradient boosting machine (GBM) model outperformed traditional Cox regression, achieving a C-index of 0.816 for predicting surgery-free survival.
- Inflammatory and coagulation biomarkers contributed about 70% to the model's explainability, with fibrinogen mediating a significant portion of inflammation-associated surgical risk.
- Biologic therapy showed significant benefit in intermediate- and high-risk patient groups, but not in low-risk patients, based on inverse probability weighting analysis.
- An online platform was developed to support individualized risk stratification and treatment assessment for clinicians.