The body in focus: multi-million grant for medical visualizations AI will make patients' personal health situations more understandable

Thanks to a 3.5 million euro grant from AiNed, doctors will soon be able to easily visualize a patient’s personal medical condition. The Academic Alliance Radboud university medical center – Maastricht UMC+ together with companies will develop AI to visualize things like DNA, brain signals, or hearing. This will help patients better understand their own health.

According to researchers, communication between healthcare providers and patients regarding medical data will drastically change. Instead of a doctor explaining black-and-white scans step by step, AI will in the future generate clear illustrations of a patient's anatomy and any abnormalities. Trained computer models, similar to ChatGPT, will explain and even visualize medical data in an understandable way, enabling simple interpretation of complex sources like DNA analyses, brain signals, and hearing tests.

A consortium of 27 partners, led by ENT specialist Marc van Hoof from Radboudumc, has received €3.5 million to further develop this type of AI. The subsidy comes from AiNed, the Dutch growth fund dedicated to the development of AI with both economic and societal importance. Van Hoof: 'I expect our AI will lead to new insights, better diagnostics, and greater patient understanding of their health.'

Large Scale

Privacy legislation currently poses a barrier to AI development, as large amounts of data are necessary but patient-identifiable data cannot be freely used. The new consortium, called PROSPER InnovationLabs, has devised a solution. ‘We’ll build large synthetic datasets derived from real data, but fully anonymized and therefore non-traceable’, explains Joost Stultiens, a researcher at Maastricht UMC+.

This approach enables data exchange on a large scale between Dutch hospitals, while still complying with privacy laws. With data that no longer reveals patient identity, developers can quickly start their AI projects without requiring years of approvals and contractual negotiations.

This scale is only possible due to the partnerships built among Dutch hospitals. ‘We aim to collaborate in enabling real innovations that improve healthcare’, says Professor of ENT Ronald Pennings from Radboudumc.

Broad Application

In this system, patients can manage their data online themselves via a dynamic opt-in or opt-out consent procedure, depending on the nature and sensitivity of the data. This combination of anonymous health data and dynamic patient consent may be a breakthrough in acquiring the data needed for AI training in healthcare. Thus, no identifiable data needs to be shared with commercial entities, and patients control who can use their data and for what purposes.

Van Hoof and his team will initially implement this approach in ENT but expect rapid expansion into other medical fields. ‘We’ll demonstrate this by integrating various medical domains, from neurosurgery to at-home breast cancer diagnostics. This not only saves time and energy but also reduces the need to collect vast amounts of patient data for each application, strengthening both healthcare and the commercial standing of the Netherlands in AI health technology.’

Partners

The PROSPER InnovationLab consortium includes the following partners: Radboudumc, Maastricht UMC+, UMCG, TNO, Superconnectors, Enatom, SURF, WSK Medical, IDS International BV, Quantitas Solutions, Earhelp, Zuyderland Medical Center, Catharina Hospital, Deventer Hospital, Elisabeth-TweeSteden Hospital, Isala Clinics, Rijnstate, Bernhoven Hospital, Elkerliek Hospital, Treant Healthcare, AI-hub East NL, AI-hub Brightlands, Brainport, ScreenPoint Medical, Dutch Head Neck Society (NWHHT), and Eyehelp.

• Want to know more about these subjects? Click on the buttons below for more news.

  News homepage news & mediaArtificial intelligence