Research: Hands-free AI Can Identify Coronary Blockages After Angiogram

Research: Hands-free AI Can Identify Coronary Blockages After Angiogram.

Share this
Two surgeons work in an operating room.

MHRI research has demonstrated that an algorithm can quickly measure fractional flow reserve values from an angiogram, efficiently helping identify lesions.


Patients who come to the cardiac catheterization lab with chest pain may face invasive fractional flow reserve (FFR) testing to search for blockages in the coronary arteries. FFR, an underutilized procedure, involves threading a thin wire into the blood vessels to measure the pressure inside along with a medication to simulate cardiovascular activity during exercise.While it is a very useful test, it is invasive and carries risks of inserting a wire into the heart vessels.

But a new study from MedStar Health Research Institute suggests that AutocathFFR, a novel AI software, could provide a safer, non-invasive and equally effective alternative—no wires or medications injected into the heart required.

Our researchers, in partnership with two international centers, found that the AI identified blockages in the coronary arteries based on angiogram images with 90% accuracy.

This innovation could significantly reduce the need for pressure wire testing, saving providers time and reducing patients’ discomfort. 

Accurate, reliable results from AI.

AI in healthcare is different from what you hear about in the media. Our algorithm is trained with thousands of angiograms so it knows how to identify lesions. But once it’s well trained, our AI stops learning. We put it into “hibernation” so it can’t continue to train itself. 

For this research, MHRI scientists partnered with two other research centers and MedHub Ltd., the software creator. MedHub’s algorithm calculates FFR values based on angiogram images—pictures taken from inside the coronary arteries—to identify blockages.

Our study compared AutocathFFR findings to FFR pressure wire results from nearly 300 patients. This retrospective study revealed that the algorithm provided FFR values that led to the automated identification of lesions with 90% accuracy. The algorithm requires an expert provider to interpret the results it efficiently generates.

We have submitted a report of our findings to the U.S. Food and Drug Administration and are eagerly awaiting its approval to use this new algorithm in clinical practice.

Benefits for patients and providers.

AI can’t replace a doctor, but it can make them more efficient. Using the new algorithm to measure FFR values offers ample benefits, including:

  • Reduced cost compared with traditional wire testing 
  • Time savings—pressure wire testing takes about 30 minutes, while the AI takes about 45 seconds 
  • Less discomfort and lower risk of complications 
  • Hands-free and noninvasive 
  • Easy for clinicians to adopt with limited training

At the forefront of impactful research.

MedStar Health Research Institute is making a real impact for patients. We’re developing our own AI algorithms to build efficiencies throughout healthcare, and we’re partnering up to bring these solutions to market.

Throughout our health system, AI is making us more efficient and supporting patient safety enhancements, bringing the future to our patients today.

Want more information about the MedStar Health Research Institute?

Discover how we’re innovating for tomorrow.

Explore With Us

Stay up to date and subscribe to our blog

Latest blogs