Gut microbiome-based supervised machine learning for clinical diagnosis of inflammatory bowel diseases

Abstract

Despite the availability of various diagnostic tests for inflammatory bowel diseases (IBD), misdiagnosis of IBD occurs frequently, and thus there is a clinical need to further improve the diagnosis of IBD. As gut dysbiosis is reported in IBD patients, we hypothesized that supervised machine learning (ML) could be used to analyze gut microbiome data for predictive diagnostics of IBD. To test our hypothesis, fecal 16S metagenomic data of 729 IBD and 700 non-IBD subjects from the American Gut Project were analyzed using five different ML algorithms. Fifty differential bacterial taxa were identified (LEfSe: LDA > 3) between the IBD and non-IBD groups, and ML classifications trained with these taxonomic features using random forest (RF) achieved a testing AUC of ~0.80. Next, we tested if operational taxonomic units (OTUs), instead of bacterial taxa, could be used as ML features for diagnostic classification of IBD. Top 500 high-variance OTUs were used for ML training and an improved testing AUC of ~0.82 (RF) was achieved. Lastly, we tested if supervised ML could be used for differentiating Crohn's disease (CD) and ulcerative colitis (UC). Using 331 CD and 141 UC samples, 117 differential bacterial taxa (LEfSe: LDA > 3) were identified, and the RF model trained with differential taxonomic features or high-variance OTU features achieved a testing AUC > 0.90. In summary, our study demonstrates the promising potential of artificial intelligence via supervised ML modeling for predictive diagnostics of IBD using gut microbiome data.

Machine Learning Prediction Model for Inflammatory Bowel Disease Based on Laboratory Markers. Working Model in a Discovery Cohort Study

“…Nevertheless, AI in IBD is usually employed for translational research and IBD diagnosis based on microbiome analysis [ 18 ]. For instance, machine learning model, specifically random forest, yielded high accuracy (area under the curve, (AUC) > 0.9) in prediction of IBD using 117 differential bacterial taxa [ 19 ], whereas Khorasani et al developed a model based on the expression of 32 genes expressed in colon samples [ 20 ].…”

Challenges and opportunities in the application of artificial intelligence in gastroenterology and hepatology

“…In the diagnosis of IBDs, a recent study developed five different AI/ML-based models using data from the gut microbiome to classify patients with IBD[ 51 ]. The Random Forest (RF) model demonstrated the highest performance with AUROCs of 0.80 when bacterial taxa were used and 0.82 when operational taxonomic features were used[ 51 ].…”

“…In the diagnosis of IBDs, a recent study developed five different AI/ML-based models using data from the gut microbiome to classify patients with IBD[ 51 ]. The Random Forest (RF) model demonstrated the highest performance with AUROCs of 0.80 when bacterial taxa were used and 0.82 when operational taxonomic features were used[ 51 ]. Finally, the models were tested in distinguishing between Chron’s disease (CD) and ulcerative colitis (UC), with the RF model demonstrating an AUROC > 0.9[ 51 ].…”

“…The Random Forest (RF) model demonstrated the highest performance with AUROCs of 0.80 when bacterial taxa were used and 0.82 when operational taxonomic features were used[ 51 ]. Finally, the models were tested in distinguishing between Chron’s disease (CD) and ulcerative colitis (UC), with the RF model demonstrating an AUROC > 0.9[ 51 ]. A multicentered, genome-wide association study used data from single-nucleotide polymorphisms to develop several AI/ML-based models to classify patients with CD and UC[ 52 ].…”

“…AI/ML models employing non-invasive parameters that provide reliably accurate diagnosis could substitute endoscopy or significantly minimize its use, thus ameliorating the impact of endoscopy-related complications, significantly decreasing the cost for diagnosis, and providing an alternative to an unpleasant intervention for the patient. Current efforts include the diagnosis of gastroesophageal reflux disease[ 34 ], gastric cancer[ 47 ], celiac disease[ 48 , 50 ], and IBDs[ 51 ].…”