Integration of expression datasets to identify biomarkers for accurate Gleason scoring in Prostate Cancer

Prostate cancer is a significant global health issue with considerable mortality rates, emphasising the urgent need for advanced treatment options and improved diagnostic methods. Current diagnostic standards for prostate cancer, including PSA testing and digital rectal examination, often produce false positives, resulting in unnecessary biopsies for patients. This limitation highlights the critical requirement to incorporate more precise biomarkers to enhance diagnostic accuracy and reduce unnecessary procedures. This study aims to investigate biomarker candidates that can effectively determine prostate cancer aggressiveness. By integrating diverse prostate tissue expression datasets and employing machine-learning techniques, this approach seeks to refine diagnostics and provide insights into the molecular underpinnings of the disease, potentially transforming early detection and patient management strategies. Our proposed biomarkers achieve a minimum precision of 0.80, addressing the false positives limitations associated with classical prostate cancer biomarkers. Moreover, the ROC-AUC profiles of most of the candidates proposed in this study align with those exhibited by other innovative biomarkers recently proposed (ROC-AUC [≥] 0.70). We believe these biomarkers are promising candidates for further in vivo and in vitro investigation.