Knowledge Graph–based Approaches for Personalized Cancer Treatment in Medical Oncology: A Review and Conceptual Framework

Personalized management of cancer patients has become increasingly important in medical oncology, as significant genetic, molecular, and clinical heterogeneity exists among cancer patients. Conventional cancer treatments typically rely on standardized protocols, which often fail to account for individual patient variability, leading to heterogeneous therapeutic responses. In the last few years, knowledge graphs (KGs) have started to be used as a powerful platform for integration of diverse biomedical data sources and precision oncology applications. In this review, we examine how recent knowledge graph–based methods for personalized treatment of cancer focusing on their structural configuration, data analysis, and clinical relevance. Although prior works show that there has been progress in building diseasespecific and multi-omics KGs, most proposed systems are still confined to static representations, fragmented data integration practices and lack of real-world clinical validation. Extra considerations regarding interoperability, interpretability of AI-based models and a clinician backed usability limit their translational potential. Upon critical review of existing literature, this study delineates the research gaps and proceeds to describe a consensual abstract model for an end-to-end operationalized knowledge graph pipeline talking about clinical, genomics, and research data. The proposed framework emphasizes interpretable learning and incremental knowledge updating to support clinically meaningful and personalized treatment recommendations.