Born of the discovery of the x-ray by Wilhelm Roentgen in 1895, the specialty of radiology is suddenly facing perhaps its greatest challenge with advent of artificial intelligence (AI) and machine learning.
Radiologists who have adapted to all manner of new diagnostic modalities over the generations now find themselves facing the prospect of machines which read and interpret imaging results quicker and more accurately than they can.
The Canadian Association of Radiologists (@CARadiologists) recently held its 80th annual meeting in Montreal and among the posters was one advertising that next year’s meeting would focus on AI. One wonders, given the speed with which AI and other technologies such as 3D printing are transforming radiology and medicine in general, whether next year is maybe too late to grapple with the issues raised.
The program committee of the CAR may have unconsciously acknowledged this as many sessions at this year’s meeting dealt with advanced technologies and at least two speakers dealt directly with the future role of the radiologist.
In addition, one of the highlights of the exhibit hall was IBM Watson Health (@IBMWatsonHealth). Last year, IBM created the Watson Health medical imaging collaborative, a global initiative with more than 26 leading health systems, academic centres, and imaging companies (they are currently looking Canadian participants) to bring cognitive imaging into daily practice. Earlier this year, IBM launched the Watson Imaging Review to reduce practice-pattern variation and reconcile differences between a patient’s administrative record and his or her clinical diagnosis.
One of the IBM staff was heard telling a CAR delegate, Watson “is not here to take your job away, it’s here to make your job easier.”
Presentations on AI and machine learning were matched by discussions of 3D printing, another technology currently transforming radiology and health care delivery.
Dr. Frank Rybicki, chief of radiology at the University of Ottawa and chief of medical imaging at The Ottawa Hospital, gave a comprehensive overview of how 3D printing is transforming many areas of medicine. The Ottawa Hospital recently opened the first 3D printing program based at a Canadian hospital.
Dr. Rybicki predicted that soon every hospital would have such a program as 3D printing moves from niche interventions to a leading role in reconstructive surgery, and cardiovascular and neurological interventions as well as supplying models to improve physician-patient communications and reducing peri- and post-operative complications.
“3D printing is the information delivery system of the current radiology generation,” he said, such as the young radiologists from Memorial University of Newfoundland who presented a paper showing the value of 3D printing of blood vessels to help educate the public and medical students.
It was a past-president of CAR, Dr. Ted Lyons from the University of Manitoba who bluntly outlined the future facing radiologists as a result of all these changes.
He noted that one of the fundamental roles of radiologists – reading and interpreting x-ray film – has almost already totally disappeared with the advent of PACS (picture archiving and communication systems). By 2035, he predicted, all x-rays, CT scans and MRIs will be read and interpreted by machines.
The way forward for radiologists is twofold according to Dr. Lyons; firstly by becoming more than a “faceless name” who interprets images in a darkened room and being more directly involved and engaged with other clinicians and patients at the bedside. The second fundamental need, he said, is for radiologists to become data scientists and lead the integration of AI into radiology practice.
Dr. Lyons presentation was complemented later in the meeting by a presentation from a SickKids Hospital, Toronto radiologist Dr. Erika Mann who reiterated how radiologists need to become more patient-centred if they wish to remain relevant.