AI Accurately Detects Key Findings in Chest X-Rays in 10 Seconds

^{Photo/Thumb ahve been modified. Courtesy of zapp2photo - stock.adobe.com.}

An artificial intelligence (AI) system accurately identified key findings in chest X-rays of patients in the emergency department suspected of having pneumonia in just 10 seconds, researchers from Intermountain Healthcare and Stanford University reported at the European Respiratory Society’s International Congress 2019.

Traditionally, it takes physicians 20 minutes or more to identify pneumonia from chest X-rays.

“In this initial study, we’ve demonstrated the algorithm’s potential by validating it on patients in the emergency departments at Intermountain Healthcare,” said Jeremy Irvin, a Ph.D. student at Stanford. “Our hope is that the algorithm can improve the quality of pneumonia care at Intermountain, from improving diagnostic accuracy to reducing time to diagnosis.”

Early diagnosis could lead to treatment starting earlier, which is vital for severely ill patients, the researchers noted.

The research team used CheXpert, an automated chest X-ray interpretation model that uses AI, to review images taken at emergency departments at Intermountain hospitals in Utah. The Stanford Machine Learning Group used 188,000 chest imaging studies to create CheXpert, which determines what is and is not pneumonia on an X-ray.

Researchers fine-tuned CheXpert to be used for patients in Utah by reading 6,973 images from Intermountain emergency departments.

Radiologists at Intermountain categorized chest images from 461 patients as being likely, likely-certain, unlikely-uncertain or unlikely to have pneumonia. The radiologists also identified the images they thought showed pneumonia in multiple parts of the lungs and if those patients had parapneumonic effusion.

In more than half of patients, the radiologists categorized them differently. The model performed similarly to the radiologists.

Traditionally, radiology reports at Intermountain are run through Cerner natural language processing, a support tool to get needed information from the report. The tool then feeds the information into ePNA, an electronic clinical decision support tool.

CheXpert outperformed the current system and created the report in less than 10 seconds, compared to 20 minutes to hours with the natural language processing tool.

Nearly 60% of errors made by ePNA were due to natural language processing, a study published in JAMA Internal Medicine found.

“Using the CheXpert system, we found the interpretation time was very swift and the accuracy of the report to be very high,” said principal investigator Nathan Dean, M.D., section chief of pulmonary and critical care medicine at Intermountain Medical Center in Salt Lake City.

CheXpert will be used live in select Intermountain Healthcare emergency departments this fall, the researchers reported.

“CheXpert is going to be faster and as accurate as radiologists viewing the studies,” Dean said. “It’s an exciting new way of thinking about diagnosing and treating patients to provide the very best care possible.”

Get the best insights inside digital health directly to your inbox.

Related

AI Could Boost Identification of Chron's Disease

Machine Learning Analyzes MRI Scans 186-Times Faster than Humans

Analysis: AI Is as Effective as Humans at Diagnosing Disease