Artificial intelligence (AI) can make healthcare better and keep it affordable. Algorithms have the potential to find patterns in data that are not visible even to very experienced doctors. Moreover, AI can play a role in preventive care and optimizing care processes. But to develop appropriate AI for healthcare, there needs to be a lot of relevant and good data available. As a concrete result, this Proof of Concept (PoC) will develop a new AI prototype, using a large amount of medical images spread across five Dutch hospitals (Maastricht UMC+, Radboud umc Nijmegen, NKI Amsterdam, Isala Zwolle, UMC Groningen). This is done in a pioneering way, where the data does not leave the hospital and thus patient privacy is optimally protected. Philips, Medical Data Works and TNO are providing the technology for this.
The innovative AI application being developed and demonstrated will make it possible, based on a standard medical image (CT scan), to predict the susceptibility of a lung tumor to different treatment types and thus provide the best intervention choice per patient. This will quickly impact lung cancer patients through the involvement of industrial partners who can bring the AI to market.
In addition to this concrete result, the chain project together with stakeholders throughout the Netherlands will deliver the Health AI platform, which will launch the first market model for health data. In it, citizens, patients, governments, companies and knowledge institutions can use this data in a scalable, responsible and sustainable way, taking into account European values. It is crucial that this happens as soon as possible since the Netherlands and Europe are in danger of falling behind other world players such as USA, China, India, because AI innovators here find it much harder to obtain data. By improving the availability of health data now, more scientific and economic activity around 'Health AI' will be created. Resulting in better and more personalised care for Dutch patients thanks to the algorithms developed in this PoC.