Machine learning-enhanced microfluidic CircRNA detection platform for high-accuracy breast cancer diagnosis

The discovery and sensitive detection of breast cancer biomarkers are significant for cancer diagnosis. Here, to increase the accuracy of breast cancer diagnosis, we show that hsa_circ_044235 and hsa_circ_000250 are effective biomarkers for breast cancer diagnosis. Moreover, we present an integrated electrochemical microfluidic circRNA detection platform (ECMCDP) that combines gold platinum (AuPt) nanoparticle-modified screen-printed electrodes (SPEs), a catalytic hairpin assembly (CHA), an electrochemical microfluidic chip and a customized low-power electronic system for simultaneous detection of the two breast cancer-associated circRNAs. The limits of detection (LODs) are 0.12 fM and 0.01 fM, and the diagnostic accuracies are 92.50 % and 88.75 %, respectively, for clinical blood samples. The platform is validated using paired pre/postoperative blood and tissue samples. By combining five machine learning-based diagnostic models, the ensemble diagnosis model achieves a high accuracy of 93.75 %. This work is aimed at identifying novel breast cancer biomarkers and establishing an innovative circRNA detection platform to improve breast cancer diagnosis and support clinical prognosis assessment.