A recent Popular Science article describes the obstacles the collaborators have faced since announcing the project in August 2018. Larry Zitnick, PhD, a research scientist in the Facebook AI Research division, or FAIR, explained that high-quality imaging data must not only be accurate, but “the radiologists have to like the image,” too.
To train the AI algorithm to meet those criteria, the team went through approximately 1,000 different model variations. Once properly trained, the algorithm’s imaging data was evaluated by NYU radiologists. The rigorous training process was crucial, since imaging data can directly influence decisions to perform surgery and other invasive procedures.
“It totally makes us nervous,” Dr. Zitnick told Popular Science. “It is important to get these things right, and that’s why we’re doing this in a very methodical way.”
The researchers are now preparing the study for academic review. The AI-MRI algorithm will eventually be made available to the public for further research purposes.
More articles on AI:
Microsoft forms AI research partnership for precision oncology
UT Health, Cardinal Health, Amazon Web Services partner on AI medical research
Penn Medicine algorithm flags patients most in need of advance care planning
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.