Document resource
Background To develop and validate a non-invasive predictive model for moderate-to-severe endoscopic activity in ulcerative colitis (UC) using intestinal ultrasound (IUS) and biomarkersMethods A retrospective study was conducted on patients with UC admitted to the Second Hospital of Hebei Medical University from August 2024 to February 2026. Patients were divided into remission/mild activity group and moderate/severe activity group according to the Mayo endoscopic score (MES). Clinical data, laboratory indicators, and IUS parameters (a total of 24 variables) were collected. Univariate logistic regression and LASSO regression were used to screen predictive factors. Subsequently, multivariate logistic regression analysis was performed to establish the predictive model, which was visualized as a nomogram. The discriminative ability of the model was assessed using the area under the receiver operating characteristic curve (AUC). Calibration curves and Bootstrap resampling (1000 iterations) were used to evaluate calibration and stability. The model was further validated using the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) score.Results A total of 211 patients with UC were enrolled, including 54 in the remission/mild group and 157 in the moderate/severe group. Univariate analysis revealed significant differences in 18 indicators between the two groups. Subsequent LASSO regression and multivariate logistic regression identified bowel wall thickness (BWT) (OR=2.047, 95%CI: 1.345-3.114), fecal calprotectin (FC) (OR=1.012, 95%CI: 1.006-1.019), white blood cell count (WBC) (OR=1.681, 95%CI: 1.241-2.278), and albumin (ALB) (OR=0.843, 95%CI: 0.735-0.965) as independent predictors, which were incorporated into the nomogram ( IDDF2026-ABS-0106 Figure 1. Nomogram for predicting moderate to severe endoscopic activity in patients with UC). The model demonstrated excellent discriminative ability for predicting moderate to severe endoscopic activity, with an AUC of 0.956 (95%CI: 0.928-0.983) (IDDF2026-ABS-0106 Figure 2. ROC curve of the prediction model). At the optimal cut-off value, the sensitivity and specificity were 87.3% and 96.3%, respectively. The calibration curve showed good agreement between predicted probabilities and actual observations (Brier score = 0.069) (IDDF2026-ABS-0106 Figure 3. Calibration curve of prediction model), and the Bootstrap internal validation yielded a mean absolute error of 0.139. When validated using the UCEIS score, the model maintained robust performance with an AUC of 0.921 (95%CI: 0.884-0.958) (IDDF2026-ABS-0106 Figure 4. ROC curve for evaluating the performance of the prediction model based on UCEIS).Conclusions A nomogram incorporating four non-invasive indicators (BWT, FC, WBC, ALB) accurately predicts moderate-to-severe endoscopic activity in UC, facilitating early risk stratification and treatment optimization.Abstract IDDF2026-ABS-0106 Figure 2Abstract IDDF2026-ABS-0106 Figure 3Abstract IDDF2026-ABS-0106 Figure 1Abstract IDDF2026-ABS-0106 Figure 4