How AI “will out-perform humans” in the fight against breast cancer and other diseases
Professor Nico Karssemeijer and Professor Sir Michael Brady formed ScreenPoint Medical seven years ago with the aim of reducing the death toll from breast cancer.
Now, ScreenPoint’s AI system Transpara®, is helping breast radiologists identify potential cancers faster and earlier in more than 20 countries worldwide.
Professor Karssemeijer, CEO of ScreenPoint Medical, says that this is “just the beginning” of the effective use of AI in healthcare.
“We may have reached the limit for image interpretation with human eyes alone. For example, AI algorithms for reading mammograms are now starting to outperform even the best radiologists in terms of accuracy.
“AI will enable more effective use of knowledge in a wide range of applications which will provide expert skills.
“However, availability of specialist radiologists is becoming a big problem in many places, due to a lack of training or problems with recruitment when experienced radiologists retire. This could become unmanageable, particularly at a time when the incidence of breast cancer is increasing worldwide. But AI is now providing a great way of maintaining high quality screening services, and it also allows radiologists to spend more time with their patients and perform diagnostic procedures.
“Not only that, AI can help clear screening backlogs caused by Covid delays by helping radiologists read mammograms faster and by measuring the presence or risk of cancer for every woman who has had a mammogram.
“AI has been around in a simpler form for a while but the big change has been with deep learning and convolutional neural networks. This has given us the ability to teach computers to understand image interpretation and opened the door to many new technologies but healthcare is by far the most important.
“In the future AI will enable more effective use of knowledge in a wider range of applications, providing expert skills and releasing highly specialised doctors to deal with other complex problems.
“Everyone will use AI sooner or later. It may not replace radiologists but it will allow for more quantitative analysis and understanding diseases that will result in better outcomes. Most importantly, it will help provide accessible healthcare for all.