Imbalances in the microbiota (dysbiosis) have been associated with numerous diseases, including diabetes mellitus, obesity, neuropsychiatric or neurodegenerative disorders, and even cancer. Metabolites produced by the gut bacteria enter the bloodstream early. With this in mind, a new study has sought to establish the profile of serum metabolites linked to the gut microbiota. The goal? To find a serum metabolite signature associated with the gut microbiota in people with colorectal cancer (CRC) or adenomas. This non-invasive, accurate, and rapid detection method would allow the early diagnosis of these conditions.
Metabolomic alterations at all levels
The analysis of serum samples from a discovery cohort (31 healthy individuals, 12 patients with adenoma and 49 with CRC) identified 885 serum metabolites whose relative abundance differed between adenoma or CRC patients and healthy individuals.
We know gut microbiota alterations in patients with colorectal anomalies can reprogram the fecal metabolome; but can they reprogram the serum metabolome? To determine the potential of these markers to predict colorectal anomalies, the researchers performed an analysis of serum and fecal metagenomic metabolites of the gut microbiota in 11 healthy individuals and 33 abnormal colorectal patients. 322 metabolites were found to be associated with the gut microbiota, including species known to be associated with CRC onset and progression (Fusobacterium nucleatum, Parvimonas micra, etc.). An algorithm was then used to accurately identify 8 serum metabolites that distinguished the healthy individuals in the cohort from those with adenomas and CRC (area under the curve 0.96). These metabolites were selected as a predictive panel for colorectal disease: gut microbiome-associated serum metabolites (GMSM).
Towards a predictive model?
This model was tested on a modeling cohort (72 healthy individuals and 120 with colorectal disease) and an independent validation cohort (53 healthy individuals and 103 abnormal colorectal patients) and reliably distinguished adenoma and CRC patients from healthy individuals (area under the curve 0.98 and 0.92, respectively). Lastly, this model was compared to other commonly used detection methods, the carcinoembryonic antigen (CEA) test and fecal occult blood test (FOBT). The GMSM panel was superior to the CEA in discriminating patients from healthy individuals in the validation cohort (area under the curve of 0.92 vs 0.72), and also beat the FOBT in discriminating between the two groups (sensitivity 83.5% vs 65.2%).
