Biosensors based on single or multiple biomarkers for diagnosis of prostate cancer

Abstract

Prostate cancer is the second deadliest cancer among men and poses a threat to the health of elderly men. Current methods of diagnosing prostate cancer including digital rectal tests or determining the increase in prostate-specific antigen level in serum are still not effective and hence can lead to overtreatment. New prostate cancer biomarkers in blood, urine, or tissues are reported and the methods for their accurate detection are being pursued. Herein, we present a comprehensive review of the recent literature reporting the biosensors for prostate cancer detection. The focus of the review was to evaluate and compare the design and performance of biosensors based on single and/or multiple biomarkers. The continual emergence of new biomarkers promotes the specificity of biosensors. And the joint detection of multiple biomarkers promotes the accuracy of biosensors. However, it is necessary to correctly screen the biomarker types and combinations because having more biomarkers does not necessarily guarantee improved biosensing performance. Furthermore, this review especially highlights the potential of artificial intelligence and machine learning tools and methodologies in prostate cancer biosensing because of their ability to recognize weak and complex signals, which will effectively improve the specificity, sensitivity, and accuracy of biosensors. The combination of machine learning and multiple biomarkers biosensors is a trend in the development of prostate cancer diagnosis. However, most of the current work still focuses on the classification of non-cancer and cancer. The use of linear regression and other tools for quantification to distinguish different stages of cancer is urgently needed for development.