NEW YORK (GenomeWeb) – Folding gut microbiome analysis in with traditional risk factor evaluation improves colorectal cancer screening and makes it better at distinguishing healthy people and those with precancerous or cancerous lesions, researchers from the University of Michigan reported in Cancer Prevention Research yesterday.
UMich's Patrick Schloss and his colleagues characterized the gut microbiomes of 90 people — 30 healthy people, 30 people with adenomas, and 30 with carcinomas — and noted shifts in the gut microbial populations that corresponded with cancer development.
"We found that the composition of the gut microbiome allowed us to identify who in our study had precancerous adenomatous polyps and who had invasive colorectal cancer," Schloss said in a statement. "If our results are confirmed in larger groups of people, adding gut microbiome analysis to other fecal tests may provide an improved, noninvasive way to screen for colorectal cancer."
While colonoscopies have contributed to the decline of colorectal cancer cases in the US, it is invasive and many patients opt to forgo the procedure. Because of that, the researchers said there is a need for cost-effective, noninvasive screening methods to prioritize patients for colonoscopies. The fecal occult blood test, they noted, is a noninvasive screening tool, but it has a low sensitivity.
Schloss and his colleagues collected stool samples from 90 people and sequenced the V4 region of the 16S rRNA genes present in the samples using the Illumina MiSeq platform. They assigned the reads they generated into operational taxonomic units (OTUs) based on sequence similarity, and calculated the relative abundance of each OTU.
Using logistic regression models, the researchers compared the microbiome composition of the different groups.
From this, they found that the relative abundance of different OTUs varied between healthy people and people with adenoma as well as between healthy people and those with carcinomas.
As compared to the healthy group, people with adenomas had higher relative abundances of OTUs associated with the Ruminococcaceae, some Clostridium, Pseudomonas, and the Porphyromonadaceae, but had lower relative abundances of OTUs linked with the Bacteroides, Lachnospiraceae, Clostridiales, and other Clostridium groups.
"This has considerable importance in secondary prevention because screening for early-stage colorectal cancer hinges on the ability to detect early pathologic changes," Schloss and his colleagues noted.
Similarly, people with carcinomas had higher relative abundances of OTUs associated with Fusobacterium, Porphyromonas, Lachnospiraceae, and Enterobacteriaceae, and lower relative abundances of OTUs affiliated with the Bacteroides, other Lachnospiraceae, and Clostridiales, as compared to the healthy controls.
There appears, the researchers said, to be an increase of pathogenic populations associated with colon tumorigenesis and a concurrent decrease in potentially protective bacteria.
Additionally, the relative abundance of OTUs could distinguish healthy people from those with any kind of colonic lesion and between people with adenomas and those with carcinomas.
For each, Schloss and his colleagues found that the model that could best tell the various groups apart also included demographic data such as age and race, and sometimes BMI and gender. For instance, adding information regarding the abundance of six OTUs to a model including age, race, and gender improved its ability to distinguish healthy people from those with any type of colonic lesion; without that information, the AUC for the model was 0.754, but after that OTU data was added, it increased to 0.936.
Further, incorporating guaiac fecal occult blood test (gFOBT) results into the microbiome-based models also improved their ability to tell the various patient groups apart.
Alone, gFOBT has a high specificity — nearly 100 percent in the researchers' example — but it was less able to distinguish healthy people from those with colonic lesions, as compared to the microbiome data alone, an AUC of 0.617 versus 0.952.
But the model combining BMI, gFOBT, and microbiome data had an even better discriminatory ability, the researchers reported, with an AUC of 0.969.
To test whether the state of the gut microbiome could be used as a screening tool, Schloss and his colleagues drew on age-specific colorectal cancer incidence data from the Surveillance, Epidemiology, and End Results (SEER) Program at the National Cancer Institute.
They designed a preliminary screening tool of five OTUs — including Clostridiales, Clostridium, Lachnospiraceae, and two Bacteroides OTUs — that were more common in healthy people than in those with adenomas. People with detectable levels of these OTUs, the researchers said, were more likely to have a healthy colon and received a negative score on the screening test.
Using a Bayesian model, the researchers calculated the positive likelihood ratio for this screening test and applied it to population probabilities of SEER for different age groups.
Based on this, they reported that people who are 65 years old had a pretest probability of 0.17 percent for having an adenoma, but after the microbiome-based test was applied, that probability rose to 10.67 percent, or a one in nine chance of having an adenoma, the researchers reported.
By comparison, a positive gFOBT test gave a one-in-15 chance that a 65 year old with a positive result had an adenoma.
This, the researchers said, indicates that though both the microbiome and gFOBT tests have high specificity, the sensitivity of the microbiome-based test is better.
"Our data show that gut microbiome analysis has the potential to be a new tool to noninvasively screen for colorectal cancer," Schloss said. "We don't think that this would ever replace other colorectal cancer screening approaches, rather we see it as complementary."