What question did this study set out to answer?

The aim was to develop and validate a predictive model for early biologic treatment failure in perianal fistulizing Crohn's disease using 3D-TRUS parameters.

April 25, 2026

Development and validation of a predictive model for early failure of biological therapy in perianal fistulizing Crohn’s disease based on three-dimensional transrectal ultrasound

Key Points

The aim was to develop and validate a predictive model for early biologic treatment failure in perianal fistulizing Crohn's disease using 3D-TRUS parameters.
Retrospective double-center study with 102 patients undergoing biologic therapy and assessed via 3D-TRUS and MRI.
Identified predictive factors using LASSO regression and multivariate logistic regression, constructing a nomogram.
Model performance was validated internally with Bootstrap resampling and externally.
Three independent predictors of early treatment failure were identified: main fistula length ≥ 2 cm, inflammatory mass size > 2 cm, and 1-mm increase in internal orifice-anal margin distance (P < 0.05).
The nomogram had an AUC of 0.955 (95% CI: 0.915-0.995) in internal validation.
External validation yielded an AUC of 0.875 (0.700-1.000), confirming the model's clinical applicability.

Abstract

Abstract Importance Despite the clinical use of three-dimensional transrectal ultrasound (3D-TRUS) in pfCD, no validated model exists to predict early biologic treatment failure. Objective Our objective was to develop and validate such a predictive model using 3D-TRUS-derived parameters. Design This study was a double-center retrospective study. Setting Patients with pfCD who underwent biologic therapy at two medical centers and completed both 3D-TRUS and magnetic resonance imaging (MRI) assessments prior to treatment initiation and at approximately 12 weeks post-treatment (range: 6-18 weeks) were included in the study. Participants A total of 102 patients were included in the analysis, with 80 patients from Hospital A formed the training/internal validation cohort; 22 from Hospital B comprised the external validation cohort. Exposure Biologic therapy for pfCD. Main Outcome(s) and Measure(s) The primary outcome was early treatment failure, defined as stable or aggravated disease (SD/AD) based on MRI criteria. Patients were divided into a failure group and a non-failure group. Predictive factors were identified using LASSO regression followed by multivariate logistic regression, and a nomogram was constructed based on the final model. Model performance was assessed through internal validation via the Bootstrap method (1000 resamples) and external validation. Results Multivariate analysis identified three independent predictors of early treatment failure (all P .05): main fistula length ≥ 2 cm, inflammatory mass size 2 cm, and internal orifice-anal margin distance (per 1-mm increase). The nomogram showed an AUC of 0.955 (95% CI: 0.915-0.995) in training cohort. Hosmer-Lemeshow confirmed good fit, and decision curve analysis (DCA) demonstrated clinical utility. Internal validation maintained an AUC of 0.955 (0.911-0.987). External validation yielded an AUC of 0.875 (0.700-1.000), with DCA supporting clinical applicability. Conclusions and Relevance 3D-TRUS features effectively predict early biologic failure in pfCD. The nomogram provides a tool for risk assessment to guide clinical decisions regarding treatment escalation or switch, supporting more individualized patient management.

Mark Helpful

Bookmark

Relay