The practical potential of generating synthetic data for cardiovascular disease research.

Abstract

Cardiovascular diseases (CVDs) remain a leading global health issue, contributing to significant mortality and reduced quality of life. While advancements in big data and precision medicine have improved diagnostic and prognostic tools, challenges persist due to high economic costs, data pri- vacy regulations and dataset imbalances, particularly in under- represented groups. Synthetic data, generated through techniques like generative adversarial networks (GANs) and differential privacy (DP), offer a promising solution. These methods allow the creation of large, diverse and anonymized datasets that mimic real-world data while ensuring patient privacy. Synthetic data can address gender and class imbalances, enhance model training and improve imaging quality in CVD research. However, limitations remain regarding data quality, trust in synthetic outputs and practical implementation. Collaborative efforts among clinicians, researchers and policymakers are essential to realise the full potential of synthetic data in overcoming current barriers to CVD research. This work highlights both the opportunities and challenges of using synthetic data, emphasizing its role as a future tool to advance cardiovascular research