Accuracy of the Discriminatory Ability of Combined Fecal Microbiota Panel in the Early Detection of Patients with Colorectal Cancer

Abstract

Background: Colorectal cancer (CRC) screening and detecting it at an early stage is an effective way to decrease mortality from CRC. Colonoscopy, considered the gold standard (GS) for diagnosing the disease in many countries, has several limitations. Therefore, the main focus of this literature is to investigate the ability of combining candidate gut microbiota for early diagnosis of CRC, both in the presence and absence of GS test outcomes.

Methods: We analyzed the data derived from a case-control study, including 83 screening colonoscopies conducted on subjects aged 18-92 years in Tehran, Iran. The candidate gut microbiota including, ETBF, Enterococcus faecalis, and Porphyromonas gingivalis were quantified in samples using absolute qRT PCR. The Bayesian latent class model (LCM) was employed to combine the values from the multiple bacterial markers in order to optimize the discriminatory ability compared with a single marker.

Results: Based on Bayesian logistic regression, we discovered that family history of CRC, physical activity, cigarette smoking, and food diet were all significantly associated with an increased risk of CRC. When comparing ETBF and E. faecalis to P. gingivalis, we have observed that P. gingivalis exhibited greater predictive power in detecting high-risk individuals with CRC. As such, the sensitivity, specificity, and the area under the receiver-operating characteristics curve of combining ETBF, E. faecalis, and P. gingivalis were 98%, 96%, and 0.97, respectively.

Conclusions: This study suggests that the combined use of the three markers markedly improves classification performance compared to pairwise combinations, as well as individual markers, both with and without GS test outcomes. Noticeably, the triple composition of the fecal markers may serve as a reliable non-invasive indicator for the early prediction of CRC.

Keywords: Colorectal cancer; Fecal microbiota; Latent class analysis; Prediction.