DNV has been assigned a central role in the EU's new research project, SYNTHIA, aimed at accelerating the use of artificial intelligence in the healthcare sector. Generative artificial intelligence will be used to create synthetic data that mimics real patient data. The project aims to demonstrate that synthetic data can be the solution to the lack of high-quality, real datasets.
The project will provide reliable tools and methods for synthetic data generation, which will be tested on cancer tumors, blood cancer, Alzheimer’s, and metabolic diseases. DNV is tasked with developing a framework for quality assurance of synthetic data to ensure safe and secure use. Artificial intelligence has been highlighted as a solution to challenges in the healthcare sector related to repetitive tasks. The algorithms require large amounts of high-quality data, while maintaining data privacy. The project will contribute to making synthetic data—created using generative AI to mimic real patient data—a viable solution to today’s privacy and data access challenges. The project is funded by the EU’s Innovative Health Initiative, a public-private partnership where this project has the aim to revolutionize AI-driven research and innovation, ensuring patients receive the best possible treatment while protecting personal information. “Trust in AI systems is crucial, and we are proud to take on the role of ensuring that the data foundation and the use of generative AI are fit for purpose and unbiased. Using synthetic data to build large datasets has significant potential to enhance research and development in healthcare by complementing already available data. By generating synthetic databases with AI, it is possible to maintain privacy while offering new tools,” says Serena Marshall, DNV’s project lead for SYNTHIA. “We will work to build trust in the technology and the generated data. This is essential to achieve the project’s goal, which is to use generative AI to accelerate medical discoveries and make personalized healthcare more accessible and efficient for everyone,” Marshall adds. Establishing a Workflow Platform The SYNTHIA project is developing validated tools and methods for synthetic data generation across various data types, including lab results, clinical notes, imaging, and genomics (the study of human DNA to understand genetic function and variation). A key component of the project is a platform that will serve as a central resource for research communities. The platform will offer synthetic data generation workflows tailored to specific needs, along with robust frameworks for assessing privacy, quality, and applicability. Each dataset will be clearly labeled for its suitability across different research areas. The SYNTHIA consortium includes 32 partners from 16 European countries, encompassing the medtech, pharmaceutical, academic, medical research, and healthcare sectors. Alongside DNV, participants include Novo Nordisk, GE Healthcare, Pfizer, Janssen, Gates Ventures, and a number of universities and research institutes. The project spans 60 months with a budget of 22.4 million euros. DNV’s role includes quality-assuring the use of synthetic data for various diseases, identifying trust gaps among stakeholders affected by the use of synthetic data, ensuring sustainable use through a lifecycle-based understanding of quality, privacy, and regulatory needs, and exploring how DNV, as an independent third party, can bridge trust gaps by leveraging regulatory knowledge and expertise in risk management, Medical Device Regulation, the EU AI Act, and harmonized standards.
https://www.dnv.com/news/dnv-to-ensure-quality-assurance-of-synthetic-data-use-in-healthcare/
2024/11/14 19:46:00