Are Synthetic Data Useful for the Diagnosis and Treatment of Diseases?

Implementing and systematising processes for generating synthetic data that can faithfully reproduce real data—thereby safeguarding patient privacy while at the same time delivering high-quality results to accelerate scientific progress and improve public health— is the goal of the SYNTHIA project, in which the University of Bologna is one of the partners.

Developers and experts in healthcare, legal, and socio-economic fields will create new technologies and enhance existing ones in order to generate multimodal and longitudinal synthetic data of various types: laboratory results, clinical notes, genomic data, and imaging data. These data must be realistic and usable in healthcare settings, since it is their quality—comparable to that of real data—that makes them suitable for clinical decision-making.

By focusing on six specific conditions (lung and breast cancer, multiple myeloma, diffuse large B-cell lymphoma, Alzheimer’s disease, and type 2 diabetes), SYNTHIA will demonstrate the usefulness of synthetic data in advancing personalised medicine.

The project will also identify specific metrics and methodologies for the assessment and validation of these data, while the analysis of the legal implications of their use will ensure compliance with current regulations.

The project results—comprising widely reusable datasets as well as the methodologies used to generate and validate them—will be made available through a dedicated platform. Based on extended open-source frameworks, the platform will ensure interoperability with other infrastructures, promote the responsible use of synthetic data, and contribute to the sharing of information within the European Health Data Space.