22. Advancing personalized prostate cancer care: Utilizing miRNA profiling and machine learning for metastasis prediction

Abstract

In the pursuit of advancing personalized medicine for prostate cancer treatment, the identification of critical biomarkers is crucial for tailoring therapies and improving patient outcomes. Building upon our prior research, where we conducted high-throughput small RNA sequencing on 38 post-operative prostate cancer patients matched by Gleason scores, this study aims to refine our understanding and enhance the accuracy of microRNA-based predictions through sophisticated computational biology techniques.

Through meticulous computational approaches and rigorous statistical analysis, we have identified microRNAs exhibiting significant expression differences between metastatic and non-metastatic cases post-surgery. This has led to the identification of six high-confidence microRNAs: miR-6761, miR-93-5p, miR-92a-3p, miR-149-5p, miR-429, and miR-671-5p, marking a significant advancement in post-operative care.

Expanding our dataset with an additional 100 supporting microRNAs, we are pioneering the training of a neural network machine learning algorithm. This innovative approach aims to accurately predict the risk of metastasis after surgery, providing a ground-breaking tool for personalized patient monitoring and treatment decision-making.

By integrating these biomarkers into a neural network model, we anticipate establishing a new standard in post-operative care for prostate cancer patients, ultimately guiding more effective monitoring strategies and improving quality of life. This study not only emphasizes the importance of microRNA profiling in prostate cancer prognosis clinical scenario but also showcases the potential of machine learning in revolutionizing cancer care.