P560 Application of a clinical decision support tool for predicting outcomes with vedolizumab therapy in patients with inflammatory bowel disease: A KASID multicenter study

Abstract

Abstract Background A clinical decision support tool (CDST) has been developed and validated for predicting outcomes of vedolizumab (VDZ) therapy for patients with Crohn’s disease (CD) and ulcerative colitis (UC), respectively. We aimed to validate each CDST for predicting outcomes and need for drug optimisation in patients with inflammatory bowel disease (IBD). Methods We retrospectively analysed 166 patients with IBD (71 with CD, and 95 with UC) treated with VDZ from January 2017 through November 2021 at 6 tertiary referral centres in Korea. We assessed clinical remission (CREM) and response (CRES), corticosteroid-free clinical remission (CfCREM) and response (CfCRES), biochemical response (BioRES), endoscopic healing (EH), and need for drug optimisation by CDST-defined response groups (low vs. intermediate to high probability group with a cut-off of 13 points in CD, Table 1A, and 26 points in UC, Table 2A, respectively). CDST was evaluated by area under the receiver operating characteristics curve (AUC) and test performance. Results Patients were classified as low (27 with CD, and 28 with UC) and intermediate to high probability group (44 with CD, and 67 with UC). Among CD patients, there showed no significant difference in the response rate of any outcomes between CDST-defined groups. Among UC patients, the rates of CREM (41.0 vs. 18.5%, p=0.040) and EH (56.5 vs. 33.3%, p=0.045) at week 14 were higher among intermediate to high probability group. The rate of drug optimisation during maintenance therapy was also higher in low probability group (54.2 vs. 30.2%, p=0.038). At week 26, CDST for CD discriminated CfCREM with an AUC of 0.656 (sensitivity 87.5%, specificity 43.8%) (Table 1B). Meanwhile, CDST for UC identified CREM with an AUC of 0.606 (sensitivity 83.3%, specificity 37.9%) at week 14 (Table 2B). BioRES of FC was identified with an AUC of 0.649 (sensitivity 84.6%, specificity 45.2%) at week 26. BioRES of CRP was identified with an AUC of 0.660 (sensitivity 68.8%, specificity 63.2%) at week 26, and with an AUC of 0.642 (sensitivity 70.0%, specificity 58.3%) at week 54. Drug optimisation during maintenance therapy was identified with an AUC of 0.620 (sensitivity 54.2%, specificity 69.8%). Conclusion Both CDST for CD and UC can be used to help guide VDZ therapy for Korean patients to predict BioRES with a fair discriminant function and moderate sensitivity.