Fecal MicroRNA-Based Algorithm Increases Effectiveness of Fecal Immunochemical Test-Based Screening for Colorectal Cancer.

Background & Aims: An algorithm based on fecal levels of 2 microRNAs (miR-421 and miR-27a-3p), fecal hemoglobin concentration, and patient age and sex can identify patients with advanced colorectal neoplasia. We investigated whether this algorithm, called miRFec, could increase effectiveness and efficiency of fecal immunochemical test (FIT)-based colorectal cancer (CRC) screening programs.

Methods: We obtained data and fecal samples from 767 persons with a positive result from the FIT who then underwent colonoscopy examination while participating a population-based CRC screening program, from March 2011 through May 2017 in Barcelona, Spain. Fecal miRNAs were isolated from the buffer contained in the original FIT collection device and analyzed by quantitative reverse transcription PCR. Aims were to evaluate the usefulness of the miRFec algorithm in identifying persons at greatest risk for CRC who should be prioritized for colonoscopy examination and individuals at low risk for whom colonoscopy could be avoided.

Results: Of the 767 study subjects, 414 (54.0%) were found by colonoscopy to have advanced colorectal neoplasia (67 with CRC and 347 with advanced adenomas) and 353 (46.0%) were found to have either non-advanced adenomas (n = 136) or a normal examination (n = 217). MiRFec algorithm scores (1-4) were independently associated with the presence of advanced colorectal neoplasia (P < .001). The miRFec algorithm differentiated patients with CRC from those with non-advanced adenomas or normal colonoscopy with an area under the receiver operating characteristic curve of 90% (95% CI, 86-94). Subjects with miRFec scores in the 4th quartile (above 3.09, high-risk group) were 8-fold more likely to have advanced colorectal neoplasia than subjects with miRFec scores in the 1st quartile (below 2.14, low-risk group). Subjects in the low-risk group had a positive predictive value below 30% for detection of advanced colorectal neoplasia. When we used a 50% specificity cut-off value, the miRFec algorithm identified 97% of patients with CRC and would allow 264 subjects (34.4%) to avoid colonoscopy examination.

Conclusions: An algorithm based on fecal levels of 2 miRNAs and hemoglobin, patient age and sex (miRFec) differentiated patients with CRC from those with non-advanced adenomas or normal colonoscopy with an area under the receiver operating characteristic curve value of 90% and avoided 34% of colonoscopies. Inclusion of this algorithm in FIT-based CRC screening programs could increase their effectiveness and efficiency.