DOP06 Development of simplified scoring system to predict two-year clinical remission of patients with Ulcerative Colitis after treatment with vedolizumab

Abstract

Abstract Background No prospective study had been conducted to demonstrate the efficacy of vedolizumab in patients with ulcerative colitis (UC). While several studies have identified predictive factors for achieving mucosal healing, a well-validated algorithm is lacking. We aimed to identify the predictive factors and develop a practical scoring system to assess the efficacy of vedolizumab in patients with UC. Methods We performed a logistic regression analysis from the data of withdrawal versus continuation of thiopurine in vedolizumab-treated patients with ulcerative colitis (VIEWS) study which is a prospective multi-center longitudinal cohort in UC patients in Australia who had clinical and endoscopic remission (Mayo score ≤ 1) at baseline and continued vedolizumab as maintenance treatment. Predictive factors of two-year steroid-free clinical remission were developed into a model that predicted the outcome of treatment. We then validated our scoring system in a separate retrospective cohort of patients with UC who received vedolizumab in routine practice during 2016-2021. Results The derivation cohort comprised 62 patients, with 48 patients (77.4%) maintaining clinical remission at 2 years. Patient characteristics are presented in table1. Independent predictive factors used in developing the model included combination of thiopurine continuation (OR 3.6, P-value 0.11), absence of exposure to anti-tumor necrosis factor (TNF) (OR 3.75, P-value 0.047), histologic remission at baseline (OR 10.79, P-value 0.002), and duration of UC < 5 years (OR 9.29, P-value 0.04). The final simplified score is as follows: 3x (thiopurine continuation) + 2x (absence of exposure to anti-TNF) + 3x (histologic remission at baseline) + 2x (duration of UC < 5 years) yielding the area under the receiver operating characteristic curve (AUROC) of 0.86 in the derivation cohort (figure 1). In the validation cohort, there were 43 UC patients who had been initiated and continued vedolizumab as maintenance treatment, with 30 patients (69.8%) remained in clinical remission at two years. The AUROC of validation cohort was 0.75 after applying simplified score. At a cut-off level of 4 points, the model can identify two-year clinical remission with 83% sensitivity and 54% specificity, respectively. Conclusion We developed the practical predictive scoring system to determine UC patients who are likely to remain in clinical remission after two years of treatment with vedolizumab. In high-risk UC patients, combination use of vedolizumab and thiopurine is strongly recommended.