The addition of a gut microbiome analysis to fecal occult blood testing improved the accuracy of standard colorectal cancer screening by more than 50-fold.
The expanded analysis not only differentiated healthy subjects from those with lesions, it also separated those with adenomas from those with invasive cancers, Dr. Joseph Zackular and colleagues reported in the Aug. 7 issue of Cancer Research and Prevention (Canc. Prev. Res. 2014 Aug. 7 [doi:10.1158/1940-6207.CAPR-14-0129]).
"We found that failure to detect at least one of the [strains measured] served as a signal of the presence of adenoma. The probability of having an adenoma rose more than 50-fold with this added information about microbiome," wrote Dr. Zackular of the University of Michigan, Ann Arbor, and his coauthors.
The investigators used gene sequencing to identify bacterial strains in the fresh stool of 90 subjects: 30 healthy controls, 30 with colorectal adenomas, and 30 with invasive colorectal cancers.
They established a microbiome signature for each group, and striking differences quickly emerged.
Compared with healthy subjects, those with adenomas had a greater abundance of strains of Ruminococcaceae, Clostridium, Pseudomonas, and Porphyromonadaceae.
They also had a lower relative abundance of strains affiliated with Bacteroides, Lachnospiraceae, and Clostridiales; and greater abundances of strains associated with Fusobacterium, Porphyromonas, Lachnospiraceae, and Enterobacteriaceae.
The microbiome composition also differentiated those with adenoma and carcinoma. Compared with those with adenomas, subjects with carcinomas had higher abundances of strains associated with Fusobacterium, Bacteroides, Phascolarctobacterium, and Porphyromonas. Those with adenomas had higher abundances of strains affiliated with Blautia, Ruminococcus, Clostridium, and Lachnospiraceae.
The research team then selected specific strains to include in a logistic regression model to differentiate healthy subjects from those with disease. This they compared with a model consisting of the clinical characteristics that best differentiated the groups.
For adenoma, a model that included the five strains most common in healthy subjects, plus the risk factors of age and race, improved the sensitivity by 4.5-fold above that of age and race alone.
For carcinoma, a model that included age, race, body mass index, and six bacterial strains improved the probability of detecting carcinoma by 5.4-fold over the clinical risk factor model alone.
The model that best differentiated adenomas from carcinomas included body mass index and four bacterial strains. It also significantly improved the sensitivity of the clinical characteristic–only model.
Dr. Zackular and his associates also compared the predictive value of their microbiome-driven risk models with that of the standard fecal occult blood test.
When comparing the clinical risk model alone with the fecal occult blood test alone in the adenoma and carcinoma groups, the odds ratio for carcinoma was 3.76; that test carried a positive predictive value of 61%.
The microbiome analysis alone carried a positive predictive value of 95%. When that was added to a combination model of the fecal test and BMI, the positive predictive value rose to 97%.
Finally, the authors validated the combination model against the age-specific rates of colorectal cancer extracted from the Surveillance, Epidemiology, and End Results (SEER) cancer database.
"Because likely candidates for colorectal cancer screening would target identification of early-stage disease (adenoma), we designed a preliminary screening test based on the five [strains] which were enriched in healthy subjects," Dr. Zackular and his associates explained.
Those who had any detectable level of the five strains were more likely to have a healthy colon and these tests were read as negative. They also calculated the sensitivity and specificity of the test in each 4-year age group ranging from 35 to 85 years old – the sensitivity for any lesion with the combination test was 23% and specificity 100%.
"When we applied the [microbiome] test to [the total] age group, the probability of adenoma was 10.67% ... (1 in 9 chance of having an adenoma). For people 50 years of age, the results suggest a 1 in 26 chance of having an adenoma with a positive test, and for adults 80 years of age, a positive test yielded a 1 in 5 chance of having an adenoma."
The fecal occult blood test alone, however, was much less accurate, with a 6.5% probability of adenoma in those aged 65 years – a 1 in 15 chance of having an adenoma – compared with the 10.7% probability and 1 in 9 chances in the same patient using the microbiome test.
"This [test] has considerable importance in secondary prevention because screening for early-stage colorectal cancer hinges on the ability to detect early pathologic changes," they noted. "The probability of having an adenoma rose more than 50-fold with this added information about microbiome."