Precision Medicine Approaches in Oncology

Abstract

Precision oncology -- the practice of tailoring cancer treatment to the molecular profile of each patient's tumour rather than treating by organ site alone -- has moved from a theoretical ideal to a clinical reality over the past decade. Molecular tumour boards, next-generation sequencing panels, and basket trials have collectively demonstrated that matching patients to targeted therapies based on actionable genomic alterations can improve response rates, progression-free survival, and in some cases overall survival. Yet the fraction of cancer patients who actually benefit from precision approaches remains disappointingly small: across large institutional series, only 8-15% of patients harbour actionable mutations for which an approved or investigational targeted therapy exists. This study evaluated five precision oncology frameworks -- single-gene biomarker testing, multi-gene panel sequencing, transcriptomic subtyping, integrated multi-omic profiling, and AI-driven treatment matching -- across four solid tumour types (non-small cell lung cancer, colorectal cancer, breast cancer, and melanoma) in a retrospective cohort of 4,826 patients treated at five European comprehensive cancer centres between January 2016 and December 2021. Our Precision Oncology Utility Score (POUS) integrates actionability yield, treatment-response prediction, survival benefit discrimination, turnaround time, and cost-effectiveness. AI-driven treatment matching achieved the highest POUS (0.924) by integrating genomic, transcriptomic, and clinical features to predict treatment response across 142 regimens, while multi-gene panel sequencing (POUS 0.898) offered the best balance of yield and implementability for routine clinical use.