RadboudUMC study: AI diagnoses prostate cancer better than current methodologies

AI is better at recognizing prostate cancer than radiologists and gives half as many false alarms, according to an international study coordinated by Radboud University Medical Center and published in The Lancet Oncology. With further development and proper safeguards, AI can ensure less workload for radiologists, more accurate diagnoses, and fewer unnecessary prostate biopsies, the researchers said.

AI expert Henkjan Huisman and radiologist Maarten de Rooij, project leaders of the PI-CAI study, brought together over 200 AI teams and 62 radiologists from 20 countries to develop AIs and evaluate over 10,000 MRI scans made available by centers in the Netherlands and Norway. The five AIs that proved best at analyzing the scans were combined into “a kind of super algorithm” and tested further.

The study, the first of its kind into AI in prostate cancer diagnosis, concluded that the AI detected almost 7 percent more significant prostate cancers than the group of radiologists, the researchers said. AI also had 50 percent fewer false alarms, pointing to a suspicious spot that later turned out not to be cancer. “This means that the number of biopsies can be halved when using AI,” the researchers said. “If these results are repeated in follow-up studies, this could help radiologists and patients a lot in the future.”

The developed AI still needs to be validated and is not available for patients yet.

AI expert Huisman pointed out that society still has little confidence in AI, largely because the easily accessible ones aren’t always great. He is working on a transparent test to fairly assess AI, as well as a quality management system similar to the one used in the aviation industry. “If aircraft come close to colliding, a committee will look at how they can improve the system so that it does not happen again in the future. I also want that for AI,” Huisman said.

“I want to develop a system that learns from every mistake so that AI is monitored and can continue to improve. This is how we build trust in AI for healthcare. Good, reliable AI can help make healthcare more efficient by relieving doctors of routine work.'