Circulating non-coding RNA cluster predicted the tumorigenesis and development of colorectal carcinoma

Carcinoembryonic antigen (CEA) is the most significant plasma biomarker in colorectal cancer (CRC), which is mainly used to diagnose and monitor the recurrence of CRC. However, due to the low sensitivity of CEA, it is more recommended for postoperative surveillance rather than early diagnosis. It is necessary to find efficient biomarkers for CRC. In this study, the expression of plasma non-coding RNAs was confirmed in three independent cohorts with total 1201 participants. First, 12 non-coding RNAs were screened from 9 plasma samples by using microarray. The expression of selected non-coding RNAs was further validated by multiphase detection and risk score analysis. We found that miR-20b-5p, miR-329-3p, miR-374b-5p, miR-503-5p, XLOC_001120 and ENSG00000243766.2 were significantly elevated in CRC plasma, and the AUC in training and validation set was 0.996 and 0.954, respectively. Moreover, miR-20b-5p, miR-329-3p and miR-503-5p were found elevated in plasma from larger tumors (5 cm as the cutoff) in CRC patients, and the merged AUC in training and validation set was 0.896 and 0.881. In conclusion, a panel of 6 non-coding RNAs showed their important clinical value for the early diagnosis of CRC. Among, miR-20b-5p, miR-329-3p and miR-503-5p might be the potential markers for evaluating larger tumor size of CRC.