Abstract
BACKGROUND: A novel bacterial gene marker, Lachnoclostridium (m3), has potential for non-invasive diagnosis of colorectal cancer (CRC) and adenoma. AIM: This study aims to determine the diagnostic performance of m3 in a multi-center cohort and to examine the effect of diet and drugs on fecal m3 in subjects with colorectal neoplasia. METHODS: In a prospective cohort study, we included 2563 subjects (330 CRC, 1119 adenomas, and 1114 controls) from two independent cohorts from Hong Kong and five cities in the Chinese mainland. Stool samples were collected before colonoscopies and bowel preparation. Dietary information was collected using a qualitative dietary questionnaire. Fecal m3 was measured using quantitative PCR (qPCR). Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic performance of fecal m3. The impact of diet and drugs on fecal m3 was investigated using in vitro experiments and multivariate analysis by logistic regression. RESULTS: Fecal m3 could significantly discriminate CRC and advanced adenoma (AA) from normal controls, with areas under the receiver operating characteristic curve (AUROC) of 0.701 and 0.604 (both p < 0.001), respectively. Three types of dietary components (meat, vegetables, and oil) and 15 drugs did not affect fecal m3 levels in vitro. In multivariate analysis, after adjusting for clinical factors, diet and drugs did not have a significant impact on fecal m3 levels in subjects with colorectal neoplasia. CONCLUSION: Fecal m3 levels and diagnostic performance were not influenced by diet and drugs. It can be considered a promising and reliable non-invasive diagnostic biomarker for colorectal neoplasia.