Gut microbiome identifies risk for colorectal polyps

Abstract

Objective: To characterise the gut microbiome in subjects with and without polyps and evaluate the potential of the microbiome as a non-invasive biomarker to screen for risk of colorectal cancer (CRC).

Design: Presurgery rectal swab, home collected stool, and sigmoid biopsy samples were obtained from 231 subjects undergoing screening or surveillance colonoscopy. 16S rRNA analysis was performed on 552 samples (231 rectal swab, 183 stool, 138 biopsy) and operational taxonomic units (OTU) were identified using UPARSE. Non-parametric statistical methods were used to identify OTUs that were significantly different between subjects with and without polyps. These informative OTUs were then used to build classifiers to predict the presence of polyps using advanced machine learning models.

Results: We obtained clinical data on 218 subjects (87 females, 131 males) of which 193 were White, 21 African-American, and 4 Asian-American. Colonoscopy detected polyps in 56% of subjects. Modelling of the non-invasive home stool samples resulted in a classification accuracy >75% for Naïve Bayes and Neural Network models using informative OTUs. A naïve holdout analysis performed on home stool samples resulted in an average false negative rate of 11.5% for the Naïve Bayes and Neural Network models, which was reduced to 5% when the two models were combined.

Conclusion: Gut microbiome analysis combined with advanced machine learning represents a promising approach to screen patients for the presence of polyps, with the potential to optimise the use of colonoscopy, reduce morbidity and mortality associated with CRC, and reduce associated healthcare costs.

Keywords: biopsy; classification; colorectal cancer; machine learning; microbiome; polyp; risk assessment; sequencing; stool.

Conflict of interest statement

Competing interests: GK, TK, and PMG have founders’ stock and LK, RH, and JB have stock options in Metabiomics.

Figures

Figure 1

Receiver operating characteristic (ROC) curves for five classifiers using informative operational taxonomic units (OTU) for biopsy, home stool, and rectal swab data sets. The evaluated classifiers included Naïve Bayes (red), random forest (orange), K-nearest neighbour (yellow), logistic regression (green), and Neural Network (blue). The straight line represents the null model. (A) For the biopsy data set, the best performing classifier was Naïve Bayes with area under the curve (AUC) equal to 0.85. (B) For the home stool data set, the best performing classifiers were Naïve Bayes and random forest with AUC=0.83. (C) For the rectal swab data set, the best performing classifiers were random forest with AUC=0.81 and Naïve Bayes with AUC=0.80.

Figure 2

Naïve Bayes and Neural Network classification scores for the naïve predictions of the home stool samples.