FDA Approves Deep-Learning Tool for Radiology Triage

An artificial intelligence (AI)-driven radiology company has earned approval from the U.S. Food and Drug Administration for a deep-learning tool that can identify certain hemorrhages based on images of the brain.

Aidoc, an Israeli startup founded little more than two years ago, today announced the approval, which is for the first solution in a software suite. The technology examines CT scans for signs of acute intracranial hemorrhaging, helping radiologists with workflow triage, according to the company.

>> READ: Why Experts Are Developing AI, Medical Imaging Standards

“Radiologists are challenged with responding to large numbers of acute cases in a timely fashion,” Aidoc CEO Elad Walach said in a statement. “The key, we believe, is to be comprehensive. If you really want to impact the daily practice, you have to cover a significant portion of the radiologist’s workflow.”

This isn’t Aidoc’s first trip to market. Elsewhere in the world, the company’s solutions went live last December, making their way into more than 50 medical centers since then. The company expects its solutions to take part in more than 1 million analyses in the first year.

So, how does it work? When a patient undergoes imaging, the startup’s FDA-approved technology analyzes the document and then alerts radiologists of its findings. In turn, that information enables radiologists to prioritize the most important cases.

In a study, Aidoc found that the solution reduced report turnaround times and increased clinician confidence.

“Seeing the software in action emphasized the key aspects an AI solution needs to possess to have an impact on the radiologist day to day: seamless integration into the workflow and broad applicability,” said Barry D. Pressman, M.D., chair of imaging at Cedars-Sinai Medical Center, who worked with Aidoc’s tech during clinical trials. “With the evidence I’ve seen, in the not-so-distant future, it will almost unthinkable to practice radiology without the assistance of solutions like Aidoc.”

Radiology, of course, is the area of medicine that is most prime for the adoption of commercial, everyday AI solutions. For years, some radiologists have fought the technologies, which they viewed as a threat to their jobs. But a global shortage of radiologists and an across-the-board maturity of AI tools for medical imaging have changed some minds.

The market, meanwhile, is expected to blossom over the next few years.

Aidoc touted its FDA approval as validation of the company’s contributions to this industry and the clinic. As time goes on, however, its competition will likely increase.

Armed with $10.5 million in venture capital, the organization is “in the process” of securing FDA clearance for its other solutions, which also focus on the detection of acute pathologies.

Get the best insights in healthcare analytics directly to your inbox.

Related

How AI Is Shaking Up Healthcare, Beyond Diagnostics

AI Diabetes Tech Continues FDA’s Software Push

Gottlieb Takes to Twitter for AI, Next Generation Sequencing Announcements