Prediction of vedolizumab efficacy in ulcerative colitis: a nomogram incorporating pathological feature and serological marker.

Abstract

Vedolizumab (VDZ) is a humanized, gut-selective biologic used in the treatment of ulcerative colitis (UC). However, data on predictive factors for treatment response are limited. This study aims to develop a nomogram to predict VDZ treatment responsiveness in UC. We retrospectively collected clinical data from patients with moderate-to-severe active UC who received VDZ induction therapy at Chongqing General Hospital from December 2020 to March 2024. Full-slide images of colon biopsies from UC patients prior to VDZ treatment were analyzed to quantify mean mucosal eosinophil density (MMED). Based on clinical response 14-week post-treatment, patients were categorized into responsive and non-responsive groups. In total, 84 UC patients were analyzed, with 58 responding to VDZ treatment and 26 not responding. Significant differences were observed in pathological indices, with MMED showing a statistically significant difference between the groups (p < 0.001). Serum biomarkers, including C-reactive protein (CRP), also showed a significant difference (P = 0.015), as did the CRP/albumin (CRP/ALB) ratio (P = 0.018). Additionally, UCEIS scores differed significantly between the groups (P = 0.025). Independent risk factors identified through multivariate logistic regression analysis were used to establish a predictive model, presented as a nomogram. The area under the curve (AUC) for the combined MMED and CRP predictive model was 0.867 (95% CI: 0.781-0.953, p < 0.001), indicating high accuracy in predicting VDZ efficacy. These data are easily accessible even in primary healthcare settings, allowing our predictive model to support improved treatment decisions for patients.