Development and validation of a novel clinical decision support tool for sustained remission in Crohn's disease: a multi-center real-world study.

BACKGROUND AND AIMS: Long-term outcomes for Crohn's disease (CD) patients treated with infliximab (IFX) remain suboptimal. We developed and validated a clinical decision support tool (CDST) to predict sustained remission in CD patients treated with infliximab (IFX).

METHODS

This multicenter observational study analyzed 746 CD patients across three cohorts. Sustained clinical remission (CREM) was defined as steroid-free Crohn's Disease Activity Index (CDAI) <150 at week 104. Using logistic regression, predictors were weighted by inverse variance. The CDST was internally validated (Cohort I, n = 113) and externally validated (Cohort II, n = 367).

RESULTS

Key predictors of CREM included: prior biologic exposure (-18 points), penetrating disease (B3 phenotype, -7 points), albumin (+0.5/gL), younger age (-0.3/year), and absence of antibody to infliximab formation (ATI, +17 points). The model demonstrated strong discrimination (AUC 0.791 [95% CI 0.708-0.875]) and calibration (Brier score: 0.191). External validation AUC was 0.611 (95% CI 0.546-0.675), indicating modest generalizability. Risk stratification via CDST categorized patients into low- (<6 points), intermediate- (6-25), and high-risk (>25) groups. A cutoff of 25 points predicted 2-year CREM with 60% (95%CI 53.2%-66.5%) sensitivity and 52% (95% CI 41.2%-1.8%) specificity.

CONCLUSIONS

We developed and validated a CDST to identify CD patients likely to achieve sustained remission on IFX therapy. By stratifying patients into distinct risk profiles, it guides personalized therapy initiation and monitoring.