Multimodal MRI-based Neurophenotype Reflecting Brain-gut Interactions to Predict Intestinal Disease Progression in Patients With Crohn's Disease

Brain-gut axis plays a crucial role in the pathogenesis of Crohn's disease (CD); however, CD neurophenotype and its impact on intestinal disease progression remain unclear. We aimed to develop a novel multimodal neuroimaging-based model to characterize the neurophenotype of CD patients and assess its ability for predicting disease progression, using multiomics data to interpret the model. This study enrolled CD patients who underwent baseline testing (including neuroimaging, psychological scales, MR enterography, and ileocolonoscopy) and faecal/blood samples collection. The neurophenotypes of patients were characterized using a neuroimaging-based model. The predictive ability of neurophenotype model for disease progression was evaluated using Cox regression analysis. Multiomics data (including faecal microbiome, faecal/blood metabolomics, intestinal permeability, blood-brain-barrier permeability, and blood neurotransmitter levels) were used to elucidate how neurophenotypes reflect brain-gut interactions.