Malnutrition Risk Prediction Model for Crohn’s Disease Achieves High Internal Discrimination

A novel prediction model combining logistic regression and machine learning with inflammatory markers and disease characteristics identified malnutrition in Crohn’s disease with AUC 0.967 on internal validation (training 0.987). Retrospective single-database study; external vali…

• Developed on 800 patients from IBDCD, validated with bootstrap resampling and an independent hold-out subset of 280 patients, all using ESPEN 2015 criteria.
• Decision curve analysis showed clinically meaningful net benefit across relevant threshold probabilities (exact figures not reported in source).
• Limitation: single-center, retrospective design with no external multicenter validation; generalizability unproven.

RadPigeon summaries are original and for information only. They are not clinical advice.