P150 MICROBIOME RISK SCORE (MRS) IN THE DIAGNOSIS AND CLASSIFICATION OF INFLAMMATORY BOWEL DISEASE

Abstract

Inflammatory bowel disease (IBD) is a common, complex and chronic condition of the GI tract. Crohn’s disease (CD) and ulcerative colitis (UC) are the two main forms of IBD. Dysregulated immune response due to alterations in microbial composition in genetically susceptible individuals has gained momentum as the etiology of IBD. Studies have shown that pre-treatment stool microbial dysbiosis is present in IBD. A fecal microbial dysbiosis index using only few taxa has been proposed as a screening tool to diagnose and monitor IBD but the potential of this index to classify or monitor disease is limited. Therefore, comprehensive, validated methods are needed to classify disease states, outcomes, and therapeutic responses. In this study, we analyzed fecal microbiota of 31 newly diagnosed, treatment-naïve pediatric IBD patients (23 CD, 8 UC) and 20 controls using 16S rRNA data. Linear regression analysis was performed to identify individual bacterial groups that are associated with IBD using age, gender and race as covariates. Absolute z-scores obtained from the normalized read counts of IBD-associated taxa, at various taxonomic levels, were summed up to compute microbiome risk score (MRS). We used random forest classifier to demonstrate the discriminatory potential of MRS. We observed that MRS generated from all the IBD-associated taxa could classify IBD patients from controls (AUC=0.77). MRS computed at various taxonomic levels, phyla, class, order and family, demonstrated similar discriminatory potential in classifying cases from controls. Furthermore, MRS showed strong correlation with disease-relevant clinical (PCDAI, r=0.71 and PUCAI, r=0.58), serological (C-reactive protein levels, r=0.84) and biological (fecal calprotectin, r=0.69) markers. Collectively, these data indicate the potential of using fecal microbiome based method, MRS, in screening and classifying IBD. Future studies are needed for the utility of MRS in disease monitoring.