Emerging medical AI trends in radiology

Every year in the beautiful cultural city Vienna, there is conference ECR, the European Congress of Radiology (this year 1-5 March). This is a chance for the movers and shakers in medical imaging to get together to exchange ideas, make pitches, and launch new products. In recent years, AI has become more and more prominent at ECR, to the point where it is now quite rare to find an exhibitor not talking about AI. 

I wanted to briefly sample some of the most interesting companies, to give an overview of where medical AI is right now. What are they working on? What stage are they at? How is it going? I put these interviews together in an episode of The Medical AI Podcast.

Overall, however, my feeling coming away from the conference was very different than last year. Last year my impression was that much medical AI technology is ready for commercial deployment, but that it is not finding its way into real-world clinical practice. That felt a little frustrating. Some of my main questions for people last year were what are the blockages? And what is holding things up? This year at the end of the conference my overall feeling was quite different, more a case of Wow this is really starting to happen. 

The reasons underlying the pace of real-world commercial deployment of medical AI, is a highly complex issue. Last year I was focussed on understanding something of the fundamental, structural issues involved. I still think there are fundamental, structural blockages. At the same time, I now think I also lacked patience because medical AI is now evolving at a breakneck speed. Also, regulators and governments seem to be increasingly alive to the potential benefits of medical AI and how to manage this revolution.

With the National Health Service in the UK being under unprecedented strain, it's easy to feel pessimistic about the future of healthcare. However, medical AI however offers a bright beam of hope.