AI-industrie.png

Smart Manufacturing & Materials-iStock-1135159658-BEWERKT.jpg

<p><strong>A robot that turns a patient over in bed? It’s still impossible today, but that robot will definitely become a reality, says Bram van Ginneken, professor of functional image analysis (Radboudumc). Van Ginneken is a pioneer in the application of artificial intelligence (AI) in healthcare. He was one of the pioneers of computer-based tuberculosis screening—the very first AI-based discovery recommended by the World Health Organization. He explains how AI can help make healthcare more efficient.</strong></p><p><img src="https://cms.aihub-oost.nl/uploads/PD_Ri3_Vm_X_Mx_T3kjd_CD_85y_Rq_A3pm_J8_W2_H3w_J_Li5k_GM_1603cb9772.jpeg" alt="PDRi3VmXMxT3kjdCD85yRqA3pmJ8W2H3wJLi5kGM.jpeg" srcset="https://cms.aihub-oost.nl/uploads/thumbnail_PD_Ri3_Vm_X_Mx_T3kjd_CD_85y_Rq_A3pm_J8_W2_H3w_J_Li5k_GM_1603cb9772.jpeg 234w,https://cms.aihub-oost.nl/uploads/small_PD_Ri3_Vm_X_Mx_T3kjd_CD_85y_Rq_A3pm_J8_W2_H3w_J_Li5k_GM_1603cb9772.jpeg 500w," sizes="100vw" width="509" height="339"></p><p>Van Ginneken doesn’t have to think long about the importance of AI for the medical world. “AI is the most important technology of this century. We’re going to build machines that are more intelligent than humans. That changes everything—including in the medical world: a great deal of work currently done by doctors or other healthcare staff will eventually be taken over by computers or robots. Take, for example, turning patients over in the intensive care unit (ICU). That requires two people and is a complex procedure, because ICU patients are connected to all kinds of equipment. But in a few decades, this will actually be a reality.”</p><p>That doesn’t mean, however, that doctors will become redundant. “No one in healthcare needs to fear for their job. The nature of the work will change. For example, there will be more time available to consult with patients,” Van Ginneken reassures us.</p><h2>Faster and Cheaper Tuberculosis Screening</h2><p>Van Ginneken has been working for years on the development of AI for medical applications. He was instrumental in developing software that detects tuberculosis on X-rays. Last year, the World Health Organization recommended using this software for tuberculosis screening. This was the first time the WHO had recommended that medical images be interpreted by computers rather than by humans. The software is now being used in fifty countries, primarily in Africa. Van Ginneken: “It saves on training and personnel costs. Moreover, the computer works much faster than a human. After the image is taken, a green or red light comes on immediately. Green means everything is fine; red indicates a suspected case of tuberculosis, which requires an additional test. Thanks to AI, this screening has become faster and more cost-effective: you can now provide care to more people for the same amount of money.”</p><h2>AI for More Efficient Healthcare</h2><p>According to Van Ginneken, AI isn’t so much about providing better healthcare as it is about making healthcare more efficient. “By 2060, it’s expected that 1 in 3 people in the Netherlands will be working in healthcare. We can’t afford that. I see automation as the only way to reduce the enormous need for healthcare workers.”<br>Over the next ten years, Radboudumc will conduct research into reliable artificial intelligence. Partly to improve care—such as predicting which patients are at high risk of another heart attack—and partly to make care more efficient. For example, work is underway on AI applications for lung cancer screening, for which the EU has issued a recommendation. “Lung cancer is the deadliest form of cancer. If you detect it early, you can drastically reduce the mortality rate,” says Van Ginneken. The Radboudumc project is investigating how CT scans can be automatically evaluated with the help of AI.</p><p>Another<br>lab is investigating how AI can make the evaluation of MRI images for prostate cancer screening more efficient, as well as the scanning process itself. “An MRI is very expensive. Partly because a person can easily spend half an hour in the scanner. We believe we can shorten that time with AI. By incorporating prior knowledge about the body into the program, you can obtain a 3D reconstruction using less measured data. You then focus only on the abnormalities in the body. A patient would only need to lie in the scanner for ten minutes to get a scan of the same quality. That means, for example, that you wouldn’t need to purchase as many expensive MRI scanners for large-scale screening.”</p><h2>Medical Imaging</h2><p>MRI scans, CT scans, and X-rays are just a few examples of medical imaging—a collective term for techniques used to visualize the body, in which AI plays a key role. “Medical imaging is booming,” says Van Ginneken. “When I started at Radboudumc in 2010, we had a team of ten people. Now we have seventy. During that time, we also founded two companies, Thirona and ScreenPoint, each of which employs approximately 50 people.” And in the past year, two more new companies were founded: one for AI applications in ultrasound assessment, Ardim, and one for pathology images, Aiosyn.”</p><h2>Valuable AI Cluster</h2><p>Oost NL is a key player in putting AI on the map in the region, says Van Ginneken. “If we want to set something up or expand it in the region, the development agency supports that. We’ve now created an entire cluster of AI-related activity in the eastern Netherlands. And once you have a cluster like that—and a hospital that’s open to using data—it attracts new companies. In the past, those were pharmaceutical companies; nowadays, software companies are on the rise. That makes sense, because software is taking over more and more tasks.” In addition, a cluster like this ensures that a great deal of specialized knowledge is available. “There are plenty of people in the region with expertise in legal matters and patents, which is extremely valuable for the practical application of AI in the medical field.”</p><h2>Companies Are Indispensable</h2><p>To further develop AI and make it truly applicable, collaboration is therefore of great importance—with manufacturers, research institutions, and companies. “We regularly collaborate with the University of Twente—for example, on research involving the MRI scanner—but also with companies. The software we develop must be certified before we’re allowed to use it in clinical settings. In theory, we could do that ourselves, but our core mission isn’t to bring products to market. That’s where companies come in. For example, that tuberculosis screening in Africa is only possible thanks to Delft Imaging, a Dutch company that sells X-ray equipment in Africa along with that software. Without them, it would never have happened there.”</p><h2>Cross-Sector Collaboration</h2><p>Collaboration is not only essential, it’s also becoming increasingly easier, says Van Ginneken. “In medical image analysis, we use the same technology that industry uses for quality inspections, mechanical engineering, or the process industry. Thanks to the broad applicability of AI, collaboration with universities or other partners in the region now makes more sense than it did ten years ago. We can therefore work very effectively together in the region and ensure that the results of our research are truly applicable in society.”</p>

Bram van Ginneken, Radboudumc: "The robot that turns ICU patients over in bed? It's coming."

These entrepreneurs have paved the way for you

Meet our frontrunners

PDRi3VmXMxT3kjdCD85yRqA3pmJ8W2H3wJLi5kGM.jpeg

<p>A robot that turns a patient over in bed? It’s still impossible today, but that robot will definitely become a reality, says Bram van Ginneken, professor of functional image analysis (Radboudumc). Van Ginneken is a pioneer in the application of artificial intelligence (AI) in healthcare. He was one of the pioneers of computer-based tuberculosis screening—the very first AI-based discovery recommended by the World Health Organization. He explains how AI can help make healthcare more efficient.</p>