Why is AI so good at Radiology?
Hello, fellow health movers and shakers.
First, let me tell you the story of Doctor Marjorie, a radiologist.
Each day, she scans hundreds of images, spending her time on a patient for about 5-15 minutes, much more if the image shows something distinctively abnormal; sometimes, her working diagnosis impacts the direction of care—like whether this is a malignancy or not.
It is her experience that gives value to that determination.
And usually, she has to call the physician or the nurse taking care of the patient to get the context of the patient’s history.
Over the past decade, the FDA has reviewed and authorized a growing number of devices with AI/ML across many different fields of medicine.
And guess what?…
Through the end of July 2023, 79% of devices authorized in 2023 are in Radiology, 9% in Cardiovascular, 5% in Neurology, 4% in Gastroenterology and Urology, and 2% in Anesthesiology— that’s just to give an idea of how AI and machine learning is applied currently at the FDA authorization level.
In addition to having the largest number of submissions, Radiology has experienced the steadiest increase of AI/Machine Learning-enabled device submissions of any specialty.
When did this all start?
The application of AI in radiology dates back to 1992, when it was first utilized for detecting microcalcifications in mammography.
Today, these are some pf the applications where AI is beneficial:
1. Enhancing cardiac imaging.
2. Classifying brain tumors.
3. Spotting vertebral fractures.
4. Detecting Alzheimer's disease.
5. Diagnosing ALS.
6. Assisting with radiology reporting & data-related tasks.
7. Detecting breast cancer.
8. Dose optimization in radiation therapy and imaging.
Look how we have gone so far…
In 2020, the American College of Radiology. conducted a survey to try to understand how its members were using AI in their clinical practices.
They had 1,427 responses and found that AI was being used by a third, 33.5%, of radiologists.
A third and growing. Where are you, my dear hospital institution?
In a study published in the Radiology Journal, the authors reported a 99.1 sensitivity rate for AI on abnormal radiographs, compared to a 72.3 percent sensitivity for radiologist reports.
AI also yielded a 6.3 percent higher sensitivity than reporting radiologists for critical abnormal X-rays (99.8 percent vs. 93.5 percent).
But of course, this was simple Chest X-ray stuff. It’s more complex with MRI and CT scans.
But, we are close to the “a Computer defeats the Grandmaster in Chess” moment.
So does this mean we are heading towards a medical imaging world, where AI would be working round the clock all 365 days a year,
to report the most complex radiological investigations accurately, at a breathtaking pace, in cranky basements,
without getting fatigued, distracted, or bored by the monotonous nature of the work
and, also without demanding leaves or pay hikes?
This would seem like the stuff dreams are made of to the corporate sector investing in healthcare
and also for the general population, who would get their reports within minutes of the test.
But not so fast.
Radiologist humans should not be replaced in the process.
The best practices involve something like this:
1. AI checks out the boring stuff, like regular screening images,
2. spits out a recommendation or highlights a finding the clinician should correlate with.
3. The human radiologist assesses the significance and context of the finding. and lastly
4. The Human specialist approves the findings and recommendations.
In a more in-depth study, the N.I.H., in trying to study the accuracy of AI with the radiologists as a team,
revealed an average sensitivity of 84% and specificity of 61.5% of the AI system performance.
With the radiologist alone, it was 73%.
Is AI better at radiology? According to the Radiological Society of North America, in 2023.
“AI systems seem very good at finding disease, but they aren't as good as radiologists at identifying the absence of disease, especially when the chest X-rays are complex,”
“Too many false-positive diagnoses would result in unnecessary imaging, radiation exposure, and increased costs.”
The role of radiologists is important, to balance the ability to find and exclude disease,
avoiding both significant overlooked diseases and overdiagnosis.
Put differently, What is the role of artificial intelligence in radiological image interpretation?
Indeed, errors made by a radiologist can lead to delays or missed diagnoses,
which can cause unfavorable patient outcomes.
Thus, the application of AI in radiology ensures faster, more reliable, and cheaper image interpretation,
in conjunction with utilizing electronic health system records and digital imaging databases.
This increased accuracy benefits patients by reducing misdiagnoses
and allows radiologists to focus on more complex cases,
ultimately improving patient care and outcomes.
AI is also streamlining radiologists' workflows by automating time-consuming tasks.
Will AI ever take over radiography?
No. not at his time. It is highly likely that in the future, the work of human radiologists will be necessary to solve challenging problems and to oversee diagnostic procedures.
AI will absolutely become part of their routine in diagnosing basic cases and helping to assist with repetitive jobs.
AI, particularly its subset machine learning, is radically improving radiology, strengthening image analysis, and mitigating diagnostic errors.
The Future of AI in Radiology
The past and present advancements in artificial intelligence for healthcare are evidence enough to believe that the future is promising.
There is a possibility of unemployment due to artificial intelligence gaining traction. Although it is a valid concern, many argue that artificial intelligence is only meant to support radiologists, not replace them.
The only worry for practitioners is that radiologists who use AI will certainly replace those who don’t.
AI is here to stay; the sooner radiologists benefit from it, the better for everyone.
So, fellow nurses, take note.
This is Robert Domondon of Nurse Intelligence, showing the Brain of AI and the Heart of a Nurse.
Follow and rate this podcast. Check out all our links. Thanks for listening.
