Paper Details

Name: Shashikumar R
Country: India

Year: 2025
Volume: Volume-12, Issue-1 (January-June)
Page Number: 142-159
DOI: https://doi.org/10.5281/zenodo.22876017

ABSTRACT
Healthcare is inherently relational: patients interact with healthcare providers and systems, diseases co-occur, drugs interact, genes regulate one another, proteins form functional networks, and clinical events unfold through interconnected temporal pathways. Graph theory offers a rigorous mathematical language for representing these relationships through nodes, edges, paths, weights, connectivity and network structure. Over time, this foundation has evolved into network analytics, knowledge graphs, graph representation learning, graph neural networks (GNNs), graph transformers, explainable graph artificial intelligence (AI), and graph–large language model (LLM) systems. This paper examines that evolution and synthesizes emerging applications in healthcare and biomedical research. It discusses patient similarity networks, disease-propagation networks, drug–drug interaction networks, protein–protein interaction networks, gene–disease networks, healthcare knowledge graphs, clinical risk prediction, precision medicine, and biomedical discovery. Particular attention is given to temporal, heterogeneous, multilayer, hypergraph, self-supervised, explainable, and multimodal graph learning. The paper then considers the emerging convergence of graph intelligence with LLMs and agentic AI, proposing a transition from static graph analysis toward systems capable of relational reasoning, evidence retrieval, hypothesis generation, and decision support under human oversight. Critical challenges concerning privacy, bias, explainability, data quality, interoperability, external validation, clinical utility, and governance are examined. A Graph Intelligence Continuum is proposed to connect classical graph theory with contemporary graph-based intelligence. The paper concludes with a research agenda focused on trustworthy, dynamic, multimodal, explainable, and human-centred graph intelligence for healthcare and biomedical research.

Keywords: graph theory; graph intelligence; healthcare analytics; biomedical networks; graph neural networks; knowledge graphs; graph transformers; explainable AI; large language models; agentic AI