Machine learning in cancer prognostic and diagnostic biomarkers: A promising approach for early cancer detection

Abstract

Cancer is a multifaceted disease that arises from both genetic and epigenetic alterations that disturb the regulation of cell growth and programmed cell death, also known as apoptosis. This imbalance results in unchecked cell proliferation, ultimately leading to tumor formation. Globally, cancer presents a significant public health issue, causing millions of deaths each year and placing immense pressure on healthcare systems around the world. Early cancer detection is crucial, as it dramatically improves the chances of successful treatment. Consequently, advancements in biomarker research and diagnostic technologies are vital for creating more effective detection strategies that reduce the impact of cancer on patients and healthcare systems. For biomarkers to be helpful in clinical practice, they must thoroughly evaluate their analytical and clinical validity and utility. Analytical validity focuses on the technical accuracy of biomarker assays, encompassing aspects such as sample handling, assay methods, and the reliability of the results. In recent years, machine learning (ML) and deep learning (DL) have become crucial tools for identifying biomarkers in healthcare engineering. The advancements in artificial intelligence (AI) have expanded these technologies' applications across various healthcare domains. This research highlights the latest innovations and methodologies, including advanced feature selection techniques and ML/DL algorithms. This review emphasizes the transformative impact of ML technologies in biomarker discovery. It also encompasses multiple facets of ML-based platforms, detailing their applications and effectiveness in biomarker discovery. Using sophisticated algorithms, these advanced computational methods allow researchers to analyze various data types, from genomic sequences to proteomic profiles. This approach facilitates the identification of novel biomarkers, enhances our understanding of cancer biology, and paves the way for improved patient care through personalized medicine. This review highlights exciting opportunities for future research, encouraging continuous innovation and collaboration across disciplines among data scientists, healthcare experts, and academic researchers.