Linking gut microbiome and metabolomics with MR enterography for predicting bowel damage in Crohn's disease

Cumulative bowel damage (BD) plays a crucial role in the progression and prognosis of Crohn's disease (CD). This study aims to integrate gut microbiome and metabolomic data with MRE features to uncover mechanistic insights and develop a model for improved BD prediction. This prospective study involved 309 CD patients, categorized into BD and non-BD groups based on LI. All participants underwent MRE, fecal 16S rRNA gene sequencing, and fecal/serum metabolomic analysis. Thirty healthy controls were also included for microbial and metabolic comparisons. The optimal prediction model was built using seven machine learning algorithms, paired with distinct combinations of multi-omics features, and evaluated using nested 5-fold cross-validation.