6.6 C
Niagara Falls
Monday, February 26, 2024
Dr. Brown: Is AI a useful partner for physicians and radiologists?
An AI image of a human and robot doctor. RICHARD HARLEY/ FIREFLY

In my time at University Hospital in London, Ont., from 1972 to 1992, radiologists were very much a part of the clinical scene.

The radiology suite was a busy place where neurologists, neurosurgeons and trainees often gathered and anyone with an interest and something to add, could join in the discussion of cases to the benefit of everyone. On a larger scale, the same happened at grand rounds.

The latter were well-planned events on a rotating basis among the three major neurological centres in the city. The cases were presented by the junior resident on the related service, after which discussion began initially with other residents taking the lead, followed by staff members.

If those presentations were not enough, there was usually an invited speaker from somewhere in the Western world who spoke for half an hour or so. The whole weekly round usually lasted three hours, including a chat and coffee break in the middle. 

Throughout rounds neuroradiologists played vital roles, because with the advent of CT and, eventually, MRI scans, they were well-placed to help neurosurgeons precisely localize lesions such as benign tumours located in tricky places. They could help the surgeon plan an approach that offered the best chance for completely removing the tumour, with the least risk of injuring blood vessels and the cranial nerves. 

For that reason and many others, neuroradiologists became essential members of the surgical team and equally helpful for stroke service, epilepsy units and multiple sclerosis clinics.

The introduction of AI to the whole field of imaging now threatens that close relationship between neuroradiologists and clinical neurologists or neurosurgeons because AI turns out to be very good at recognizing patterns in images. 

Examples include recognizing diabetic retinopathy or differentiating between skin lesions, especially malignant lesions such as melanomas and squamous cell cancers from benign lesions and doing so with a facility that approximates the skill of the best of retinal and dermatologic specialists. 

The impact of AI with plain X-rays and CT scans has been similar, with MRI not far behind, provided the data base on which AI is trained is of sufficient quality. That means  imaging equipment located in remote or underserviced areas could provide useful imaging services if there is technical support to carry out the studies and backup from a radiologist connected via the internet.

Radiologists, including neuroradiologists, are rightly worried that should AI-supported devices prove as capable as radiologists, there might not be much need for them, except as backup partners. The same could happen in clinical settings, where AI could partner with physicians to operate MS clinics, for instance.

There’s no doubt in my mind that properly trained, AI would be an asset in busy MS clinics given the fact that it would be privy to all the patient’s medical records, a library’s worth of information from excellent sources about MS and, if programmed to read MRIs, would be capable of detecting changes that might well elude neuroradiologists and MS physicians.

AI is also well-equipped to recognize longitudinal changes in MS reflected in the patient’s symptoms, standardized physical testing and perhaps images of the retina. 

But however competent AI might become in specialty and general clinics, what’s missing in current versions would be the human touch and empathy, although the latter could be programmed in AI devices in the future.

But if AI isn’t all we would like it to be emotionally, remember most of us have run into physicians on off days when they were grumpy, impatient, even rude. Even the best, aren’t at their best all the time.

The best can be out of sorts for any number of reasons and show it in ways we recognize. When that happens, sometimes it helps to turn the arrow of empathy arrow around from patient to physician or nurse.

It might not be acknowledged but you will have done the right thing and sometimes overworked staff will realize what happened and thank you – provided they’re not in too much of a tizzy that day. 

When David Elkins and I put on the Meditation and Mindfulness series last September I realized how “mindless” I had been at times in the past and my guess is that physicians and nurses are just as much in need of some version of mindfulness as the public they serve. But they might not realize it. That’s my confession. 

Dr. William Brown is a professor of neurology at McMaster University and co-founder of the InfoHealth series at the Niagara-on-the-Lake Public Library.


Subscribe to our mailing list