Artificial Intelligence is going to disrupt every industry, but its biggest impact is widely believed to be seen in the healthcare industry. While science and technology have advanced, healthcare is still not available in an equitable manner. AI is seen as the pathway to a stage where healthcare is available to every person on the planet.
In order to reach that stage, every major healthcare organisation and big nation are racing to create AI for health solutions. One of the countries where AI for health is being studied extensively in the Netherlands. We recently saw Erasmus MC set up two AI labs and now leading universities have come together to create an even bigger challenge called “Projects AI for Health“.
Projects AI for Health: What you need to know
AI for Health is an alliance of Eindhoven University of Technology (TU/e), Wageningen University & Research (WUR), Utrecht University (UU), and University Medical Centre Utrecht (UMCU). These four partners have the data, knowledge and facilities to develop solutions in the field of AI for health. The Projects AI for Health is an initiative aimed at facilitating and financially supporting cross-disciplinary collaborations between the four institutions.
The project offers grants to large applications related to AI for Health in the near future. The grant can also be used for business plan writing, conducting preliminary research and for projects aimed at societal added value. The applications must also address at least one category of the main themes of AI for Health and one cross-cutting theme or building block.
The main themes of AI for Health are:
- A1 Healthy and safe food supply chain
- A2 Precision/personalised health, nutrition & behaviour
- A3 Healthy living environment/smart cities
- A4 AI in clinical decision support
- A5 Medical imaging
Cross-cutting themes or building blocks of AI are:
- B1 Ethics & Legal AI, trustworthy AI, explainable AI, human-centred AI
- B2 AI methodology and algorithms
- B3 Data quality, data measurement error, data biases
- B4 AI infrastructure: data sharing, storage, harmonisation.
The AI for Health initiative awards a minimum of €15,000 and a maximum of €40,000 to the selected applicants. The total budget of the project for this year is set at around €120,000. The deadline for the application closed on 1st November and the project attracted a total of twenty high-quality proposals. The jury has announced the five winning proposals that cover AI research in the field of prevention, diagnosis, care, treatment, and AI building blocks. Here is everything you need to know about the winners.
AI-driven prediction model of complications after corneal transplant surgery
This project builds on the previous independent work done by the proposers to form a new collaboration in corneal biomarker development. The project proposed by Mitko Veta and Josien Pluim of TU/e, Gerko Vink of UU and Robert Wisse of UMCU will bring together multiple disciplines from three alliance partners.
The project will see UMCU’s complementary expertise in surgical ophthalmology join hands with AI for medical image analysis at TU/e and AI methodology and statistics at UU. The project has received a grant of €40,000 and could help detect complications such as bleeding, glaucoma and fluid leakage from the cornea or detached retina sooner using AI.
Towards the future of surgery: anatomical recognition in robot-assisted surgery
This project shows how the Dutch healthcare system is keen on leveraging its expertise in clinical practice. The project proposed by Jelle Ruurda of UMC Utrecht, Josien Pluim and Maureen van Eijnatten of TU/e, Fred van Eeuwijk of WUR, Frank van der Stappen and Natasha Alechina of UU envisions the future of surgery built on the expertise available right now.
With a grant of €40,000, the project will leverage clinical practice and research expertise in minimally invasive and upper-gastrointestinal surgery (UMCU), medical object recognition and AI in medical image analysis (TU/e), man-machine interfaces and autonomous intelligent systems (UU). It will also rely on AI, applied statistics and the realisation of digital twins with a focus on applications in life and environmental sciences developed by WUR.
AI for Healthy Family Life
This project aims to help patients change their lifestyle within the social dimension of a family and overcome the “intention-behaviour” gap. In other words, the project aims to help patients not fail at translating their intentions into action. For the project, the proposers have been awarded a grant of €40,000, where €20,000 is co-funded by the working group Preventive Health.
The project by Henk Wensink and Marielle Timmer of WUR, Pieter van Gorp of TU/e, Catharine Evers of UU and Marielle Jambroes of UMC Utrecht aims to start by leveraging Eindhoven University of Technology’s complementary expertise in developing a social health data-platform. It will also rely on capturing scientific validated data and knowledge on nutrition and health for personalised advice from WUR.
Further helping the project will be expertise areas of UU and UMC called psychology of human behaviour in self-regulatory abilities and assessing health of vulnerable groups and participatory action research respectively.
Establishing a Consortium
While the projects mentioned above aim to solve a clear issue or overcome a gap, this project aims to establish a European consortium. The goal of the consortium is to provide “effective, explainable, and transparent AI-aided tools in the hands of intelligent decision-makers in the healthcare domain.” The proposal also envisions writing a grant proposal for the Horizon Europe funding scheme.
The proposal by Rens van de Schoot (UU), Daniel Oberski (UMC Utrecht), Bedir Tekinerdogan (WUR), and Chris Knighting (TU/e) relies on their expertise in software engineering, artificial intelligence and their contribution to Open Science. The proposal has received a grant of €40,000 and could become a foundational block for many “AI for Health” science and research projects.
A talisman for epilepsy: Preparing an online transformer-based clinical research monitor
This project is proposed by Willem Otte and Eric van Diessen of UMC Utrecht, Marijn Schraagen of UU and Pieter van Gorp of TU/e. The project aims to explore the boundaries of a “potentially powerful AI-driven tool” using the new partnership between the alliance members.
The project aims to bring together UMC Utrecht’s complementary expertise in epilepsy research, explainable AI to facilitate the implementation of Natural Language Processing of TU/e and the recent transformer-based sequence processing and essential input in model fine-tuning developed by Utrecht University (UU). With a grant of €20,000, the project aims to bring the magic of AI to solving epilepsy.