P624 Development and Validation of a Clinical Decision Support Tool to Predict Outcomes of ustekinumab Treatment in Patients with Crohn’s Disease

Abstract

Abstract Background Ustekinumab (UST) has been demonstrated for the treatment of Crohn's disease (CD) with a high efficacy and safety profile. There is limited data concerning disease progression after UST treatment. This study aims to establish a clinical decision support tool (CDST) to identify Chinese patients with CD who are less likely to experience disease progression under UST treatment. Methods A multicentre, retrospective observational study was conducted among Chinese patients with CD initiating UST between 1 May 2020 and 20 October 2022. Disease progression (including CD-related surgery, hospitalization and complications) was evaluated at week 24 (±4w). Logistic regression analysis was performed to identify factors associated with disease progression. The CDST, established based on these factors, was both internally and externally validated. The performance of this tool was tested using receiver operating characteristic (ROC) area under the curve (AUC) analyses. Results A total of 519 patients were enrolled in this study. In the derivation analysis, four factors including prior biologics usage, prior immunosuppressant usage, baseline faecal calprotectin, and baseline haemoglobin were identified to correlate with progression. A good predictive ability of this model was confirmed by AUC as high as 0.76 for the internal validation cohort and 0.72 for the external validation cohort, respectively. The CDST stratified patients into low, intermediate, or high risk of progression with cut-off values of 6 and 16. Conclusion We developed and validated a scoring system to identify CD patients less likely to experience progression at week 24 with UST therapy. This CDST will facilitate therapy decision-making and precision medicine.