An Integrated Microfluidic Microwave Array Sensor with Machine Learning for Enrichment and Detection of Mixed Biological Solution

In this work, an integrated microfluidic microwave array sensor is proposed for the enrichment and detection of mixed biological solution. In individuals with urinary tract infections or intestinal health issues, the levels of white blood cells (WBCs) and Escherichia coli (E. coli) in urine or intestinal extracts can be significantly elevated compared to normal. The proposed integrated chip, characterized by its low cost, simplicity of operation, fast response, and high accuracy, is designed to detect a mixed solution of WBCs and E. coli. The results demonstrate that microfluidics could effectively enrich WBCs with an efficiency of 88.3%. For WBC detection, the resonance frequency of the sensing chip decreases with increasing concentration, while for E. coli detection, the capacitance value of the sensing chip increases with elevated concentration. Furthermore, the measurement data are processed using machine learning. Specifically, the WBC measurement data are subjected to a further linear fitting. In addition, the prediction model for E. coli concentration, employing four different algorithms, achieves a maximum accuracy of 95.24%. Consequently, the proposed integrated chip can be employed for the clinical diagnosis of WBCs and E. coli, providing a novel approach for medical and biological research involving cells and bacteria.