OM-sqPCR enables one-Step multiplex quantification of circulating small-RNA subclasses for colorectal cancer classification

Here we present OM-sqPCR, an AI-assisted one-step multiplex quantitative PCR platform that enables simultaneous detection of microRNAs, transfer RNA–derived sRNAs (tsRNAs), and ribosomal RNA–derived sRNAs (rsRNAs) in a single closed-tube reaction. The system integrates optimized enzyme–buffer chemistry for short RNA templates with machine-learning–guided primer–probe design, enabling synchronized amplification across multiple optical channels.

Applied to a multicentre clinical cohort (n = 651), OM-sqPCR identified a four-sRNA signature that discriminated colorectal cancer from controls with an area under the ROC curve (AUC) = 0.97, 85.9% sensitivity, and 93.1% specificity, outperforming serum carcinoembryonic antigen (CEA) and maintaining high accuracy in early-stage disease (stage I–II, sensitivity: 84.6%). The workflow also validated biomarker panels for renal and thyroid cancers with comparable accuracy (AUC > 0.98).