Publication Details
Abstract
The diagnosis of cancer is shifting towards paradigm with integration of multi-omics that includes genomics, transcriptomics, proteomics, metabolomics and epigenomics. Integrated platforms can deliver a comprehensive molecular map of the tumor, with important insights into early detection, prognosis stratification and the accuracy of therapeutic response prediction. But, despite the huge potential, multi-omics biomarkers are not yet in daily use in clinics due to a number of systemic problems. These include pre-analytical variability in sample preparation, lack of standardized analytical protocols, high-dimensional bioinformatics and considerable regulatory and economic hurdles. In this regard, artificial intelligence (AI) and machine learning (ML) have proven to be game-changers, providing the computational capacity needed to combine heterogeneous data and uncover nuanced biomarker patterns that underlie personalized oncology. Multi-omics tools, through their ability to provide longitudinal monitoring, allow to adapt the therapeutic approach to the individual's biological profile, maximizing the benefits of the treatment and minimizing of the ineffective treatments. To achieve this in a clinical lab setting, automated workflows and stringent quality control protocols must be created and implemented to meet regulatory compliance requirements. This review aims to discuss the present status of multi-omics in oncology, explore current gaps in translation and present a roadmap for the successful integration of AI-driven precision medicine into next-generation oncology.