How a data-driven “diagnostic cockpit” could help health systems.
Illustration has been cropped and resized. Courtesy of Brother UK.
Healthcare might be a step closer toward creating ironclad standards for the flood of data poised to stem from the use of artificial intelligence (AI) in medical imaging, thus boosting diagnostics. That’s because the federal National Institute of Standards and Technology (NIST) and the Academy for Radiology and Biomedical Imaging Research (ARBIR) hosted a workshop last week in which multidisciplinary experts strived to identify and begin crafting such guidelines.
Writing for an NIST blog, scientists Denis Bergeron, PhD, and Michael Garris, MS, said uniting computer science and medical imaging to develop industry standards will “calibrate disparate measurements” and spur interoperability. Consequently, the project could enable diagnostic teams of the future to prescribe tests that measure cholesterol levels, breast tissue density, and more, without a hitch.
“Physical measurement standards will ensure that the data these tests generate will mean the same thing across patients, over time, and when measured using devices from different manufacturers,” Bergeron and Garris wrote.
>> READ: The Dystopian Concerns of AI for Healthcare
Their vision is lofty, an extension of health tech’s persistent push toward interoperability at large. That campaign, of course, has proved challenging, but various arms of the federal government, from the US Department of Veterans Affairs to the Office of the National Health Information Technology Coordinator, have contributed to the cause. Congress, meanwhile, sent electronic medical record (EMR) adoption rates skyrocketing with the billions of dollars in incentives baked in to the Health Information Technology for Economic and Clinical Health Act of 2009.
But EMRs are not AI, meaning the NIST’s quest to pinpoint and produce standards for cutting-edge medical imaging is a journey all its own.
Radiologists, medical specialists, device manufacturers, researchers, data scientists, and government officials met last week in Gaithersburg, Maryland, to map the path forward. Their ultimate goal: to create a “diagnostic cockpit” for clinicians to access data crucial to diagnosing and treating patients with the help of AI.
A follow-up to a session last year, this workshop saw them explore standards governing new data formats, imaging rules, performance metrics, interfaces, and even how data might be visualized, according to the NIST.
What’s more, the experts agreed that coronary artery disease and breast cancer stand to produce troves of data, from a number of sources, which can then be merged and integrated to develop various treatment paths. The skinny: AI and medical imaging are looking at these conditions as early use cases, a sort of testing ground.
As many key opinion leaders have said in the past, this marriage of technology could go on to boost the AI itself, patient outcomes, precision medicine, and value-based care.
“Standardized data at scale will fuel machine learning and create new generations of analytic and diagnostic models,” Bergeron and Garris wrote. “With AI’s ability to perform millions of incredibly complex weighing and correlation-finding calculations in a short period of time, human diagnostic teams will be able to quickly identify patterns and associations that they would otherwise miss.”
But that won’t happen, at least at scale, until the players agree on the ground rules.
Get the best insights in healthcare analytics directly to your inbox.
What’s the Difference Between AI, Machine Learning, and Analytics?
AI Goes Live for Clinical Pathology
How AI Is Shaking Up Healthcare, Beyond Diagnostics