P0364MR Elastography Characterizing Biomechanical Properties to Enhance Enterographic Fibrosis Diagnosis

Background Current diagnostic tools for fibrosis assessment in Crohn’s disease (CD) remain suboptimal.[1] MR elastography may advance toward addressing this limitation by offering tissue biomechanical characterization. [2] To evaluate MR elastography for intestinal fibrosis diagnosis using surgical histopathology and Piezo1 signaling as reference, and determine its incremental value over established MR enterography through machine-learning modelling. Methods This prospective study analyzed 146 intestinal specimens from 56 CD participants undergoing preoperative MR elastography and enterography. Elastographic (shear-wave speed [SWS], loss angle [φ]), advanced (DWI, T1 mapping) and conventional enterographic parameters were measured. Histopathological assessment included fibrosis scoring and collagen proportion analysis. Piezo1 served as stiffness reference standard which was evaluated by immunofluorescence. Three types of models (basic: conventional; extended: +advanced; complete: +elastographic) were developed using seven machine-learning algorithms, with performance assessed via AUC (DeLong’s test). Results SWS and φ significantly correlated with histological fibrosis score (r=0.642/0.408) and collagen proportion (r=0.666/0.461) (all P<0.01), with AUCs of 0.870 (95% CI: 0.803-0.937) for SWS and 0.734 (0.640-0.828) for φ in differentiating moderate-to-severe from none-to-mild fibrosis (Figure 1). Piezo1 level correlated with histological fibrosis score (r=0.687, P<0.001) and collagen proportion (r=0.760, P<0.001). Importantly, SWS also showed significant correlation with Piezo1 expression (r=0.615, P<0.05), supporting its role as a promising imaging biomarker of stiffness-characterizing fibrosis. XGBoost consistently outperformed other algorithms across all models. Complete modelXGBoost outperformed basic modelXGBoost (AUC, 0.892 vs. 0.540; Bonferroni-corrected P=0.006) and extended modelXGBoost (AUC, 0.892 vs. 0.619; Bonferroni-corrected P=0.024) in independent test set, confirming the added diagnostic value of MR elastography. SHAP analysis of complete modelXGBoost highlighted SWS as the dominant predictor, exhibiting its strongest relationship with fibrosis probability (mean |SHAP|=0.715), substantially exceeding other parameters (range: 0.003-0.146). A web-based calculator incorporating the selected features was developed for external validation of MR elastography’s incremental diagnostic value for CD fibrosis.(Figure 2) Conclusion MR elastography serves as a promising tool for quantifying intestinal fibrosis in CD, providing real-time biomechanical profiling that synergistically enhances MR enterography’s precision in fibrosis severity stratification. References: 1. Rieder F, Fiocchi C, Rogler G. Mechanisms, Management, and Treatment of Fibrosis in Patients With Inflammatory Bowel Diseases. Gastroenterology. 2017;152(2):340-350.e6. 2. Chirra P, Sleiman J, Gandhi NS, et al. Radiomics to Detect Inflammation and Fibrosis on Magnetic Resonance Enterography in Stricturing Crohn’s Disease. Journal of Crohn’s and Colitis. 2024;18(10):1660–1671. Conflict of interest: Dr. Wang, Yangdi: No conflict of interest Qiapeng, Huang: All authors declare no relevant relationships. Fang, Zhuangnian: No conflict of interest Ruonan, Zhang: No conflict of interest Xiaodi, Shen: No conflict of interest Li, Xuehua: No conflict of interest