Evolution of Generative Adversarial Networks (GANs) in Medicine: A Systematic Review of Architectures, Applications, and Implementation Challenges

Wan Aezwani Wan Abu Bakar; Nur Laila Najwa Josdi; Mustafa Man; Evizal Abdul Kadir; Bishwajeet Kumar Pandey

doi:10.47852/bonviewAIA52026216

Authors

Wan Aezwani Wan Abu Bakar Faculty of Informatics and Computing, Universiti Sultan Zainal Abidin, Malaysia
Nur Laila Najwa Josdi Faculty of Informatics and Computing, Universiti Sultan Zainal Abidin, Malaysia
Mustafa Man Faculty of Computer Science and Mathematics, Universiti Malaysia Terengganu, Malaysia
Evizal Abdul Kadir Department of Informatics Engineering, Universitas Islam Riau, Indonesia
Bishwajeet Kumar Pandey Department of Computer Application, G.L. Bajaj Institute of Technology and Management, India

DOI:

https://doi.org/10.47852/bonviewAIA52026216

Keywords:

data fidelity, electronic health records (EHRs), generative adversarial networks (GANs), genomics, medical imaging, synthetic data

Abstract

Generative Adversarial Networks (GANs) have gained increasing attention in healthcare as a promising approach to addressing data scarcity, offering synthetic alternatives that support research while mitigating privacy risks. This review examines the landscape of GAN-based synthetic data generation in healthcare, with applications spanning medical imaging, electronic health records, genomics, and multimodal datasets. A systematic search guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework identified 81 peer-reviewed studies published between 2014 and 2025, ensuring comprehensive coverage of methodological and translational developments. The review maps the diversity of GAN architectures employed, synthesizes evidence on evaluation strategies, and outlines ethical, privacy, and regulatory considerations that influence adoption. Results indicate that GANs often achieve strong fidelity and downstream utility, with emerging fairness-aware models addressing demographic bias. However, inconsistent validation practices, limited clinical integration, and unresolved ethical and governance challenges continue to hinder translation into real-world settings. Overall, the review consolidates methodological trends, barriers, and future directions, highlighting the potential of GANs to serve as viable tools to overcome data scarcity in healthcare research and practice.

Received: 22 May 2025 | Revised: 27 August 2025 | Accepted: 12 November 2025

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Author Contribution Statement

Wan Aezwani Wan Abu Bakar: Conceptualization, Methodology, Formal analysis, Resources, Visualization, Supervision, Funding acquisition. Nur Laila Najwa Josdi: Software, Data curation, Writing – original draft, Writing – review & editing, Funding acquisition. Mustafa Man: Validation, Investigation, Supervision, Project administration. Evizal Abdul Kadir: Resources. Bishwajeet Kumar Pandey: Project administration.