MR Elastography Characterization of Biomechanical Properties to Enhance Enterographic Fibrosis Diagnosis.

Abstract

Background Current diagnostic tools for fibrosis assessment in Crohn disease (CD) remain suboptimal. MR elastography could address this limitation by offering tissue biomechanical characterization. Purpose To evaluate multifrequency MR elastography for intestinal fibrosis diagnosis, using surgical histopathologic assessment and Piezo1 signaling as references, and to determine its incremental value over established MR enterography through machine learning. Materials and Methods This prospective study (December 2023 to December 2024) included participants with CD who underwent preoperative multifrequency MR elastography (30-60 Hz) and MR enterography. Elastographic parameters (shear-wave speed [SWS] and loss angle [φ]), advanced enterographic features (apparent diffusion coefficient and T1 mapping), and conventional enterographic features were measured. Histopathologic assessment included fibrosis scoring and collagen proportion analysis. Piezo1 level, which was evaluated using immunofluorescence, served as the stiffness reference standard. Three types of models (basic [conventional enterographic features], extended [plus advanced features], and complete [plus elastographic features]) were trained, validated, and tested using seven machine learning algorithms, with performance assessed using the area under the receiver operating characteristic curve (AUC) and DeLong test. Results This study included 56 participants (mean age, 31 years ± 10.8 [SD]; 39 men and 17 women; 146 intestinal specimens). SWS and φ were positively correlated with histologic fibrosis score (<i>r</i> = 0.64 and 0.41, respectively; both <i>P</i> < .001) and collagen proportion (<i>r</i> = 0.67 and 0.46; both <i>P</i> < .001), with AUCs of 0.87 (95% CI: 0.80, 0.94) for SWS and 0.73 (95% CI: 0.64, 0.83) for φ for differentiating moderate-to-severe from none-to-mild fibrosis. Piezo1 level positively correlated with histologic fibrosis score (<i>r</i> = 0.69; <i>P</i> < .001) and collagen proportion (<i>r</i> = 0.76; <i>P</i> < .001). Importantly, SWS also positively correlated with Piezo1 expression (<i>r</i> = 0.62; <i>P</i> < .001), supporting its role as a promising imaging biomarker of stiffness, for characterizing fibrosis. XGBoost consistently outperformed other algorithms within each feature set. Among the three optimal models (with XGBoost as the classifier), the complete model (AUC, 0.89) outperformed the basic model (AUC, 0.54; Bonferroni-corrected <i>P</i> = .006) and extended model (AUC, 0.62; Bonferroni-corrected <i>P</i> = .02) in an internal test set, confirming the added diagnostic value of MR elastography. Conclusion In participants with CD, MR elastography quantified tissue biomechanics, and incorporation of MR elastography improved the diagnostic performance of MR enterography for intestinal fibrosis stratification in CD. © RSNA, 2026 <i>Supplemental material is available for this article.</i>