Graph Neural Networks for Medical Imaging Analysis and Biological Data: Integrating Topology, Geometry, Radiomics, and Generative AI.
Authors
Affiliations (11)
Affiliations (11)
- Department of Radiology, Mayo Clinic, Rochester, MN 55905, USA.
- College of Engineering and Information Technology, University of Dubai, Academic City, Dubai 14143, United Arab Emirates.
- Staticlysm LLC, Miami, FL 33157, USA.
- Department of Mathematics, Statistics and Computer Science, Faculty of Science, The University of Bertoua, Bertoua 00237, Cameroon.
- School of Computer Science and Applied Mathematics, The University of the Witwatersrand, Johannesburg 2050, South Africa.
- African Institute for Mathematical Sciences, Research and Innovation Centre, Kigali P.O. Box 6428, Rwanda.
- Department of Mathematics, The University of Texas, Austin, TX 78712, USA.
- UPMC Genome Center, Pittsburgh, PA 15213, USA.
- Cooper Medical School, Rowan University, Camden, NJ 08103, USA.
- Department of Surgery, Mayo Clinic, Rochester, MN 55905, USA.
- Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104, USA.
Abstract
Graph neural networks (GNNs) are increasingly used for medical imaging analysis and biological data modeling, where the integration of radiomics, topology, geometry, and generative artificial intelligence (AI) may improve representation learning from medical images and related biomedical data. Across the reviewed literature, GNNs show particular value for modeling spatial relationships, multimodal interactions, graph-structured biological networks, and non-Euclidean imaging features that are difficult to capture using conventional convolutional architectures alone. Topology- and geometry-aware approaches further expand this capability by encoding multi-scale structure, higher-order relationships, curvature, geodesic organization, and equivariant spatial priors. Hybrid graph-transformer models and generative graph methods represent emerging directions for modeling long-range dependencies, augmenting scarce datasets, supporting synthetic pretraining, and improving representation learning in low-label or heterogeneous biomedical settings. However, clinical translation remains limited by variability in graph construction, limited external validation, computational cost, scalability constraints, interpretability challenges, and uncertainty regarding the biological realism of synthetic data. Overall, this review highlights that GNN-based medical imaging analysis is most likely to advance when graph construction is biologically justified, model performance is evaluated across diverse clinical cohorts, and technical gains are paired with transparent validation, interpretability, and implementation strategies.