The Ritz Herald
St. Luke’s Lehighton Campus Radiologist Dr. Eric Lee, M.D. (L) talks with a patient about a chest x-ray. © St. Luke's University Health Network

St. Luke’s Lehighton Pioneers Artificial Intelligence in X-Rays


74,000 Americans suffer collapsed lungs each year

Published on September 24, 2020

St. Luke’s Lehighton Campus is the first hospital in Pennsylvania using new artificial intelligence technology (AI) to help radiology specialists find dangerous lung conditions quickly and accurately before they become critical.

The Carbon County hospital last month launched GE Healthcare’s Critical Care Suite AI X-ray software, which enhances clinicians’ ability to find a collapsed lung (pneumothorax) within seconds of image acquisition.

The algorithm-based software was uploaded to the hospital’s GE X-ray machine. This software enables the machine to identify an X-ray that may contain a pneumothorax and alert the imaging team to prioritize that study, so a physician will quickly see the image and diagnose the abnormality before it becomes a danger to a patient. A collapsed lung occurs when air enters the space between a lung and the chest well. This condition is life-threatening, so it must be treated quickly.

Over the past two years, St. Luke’s University Health Network radiologists worked with an international team fielded by GE Healthcare to pioneer this innovative AI technology, a first of its kind in the industry. The FDA approved the GE Health Critical Care Suite product in September 2019.

St. Luke’s vice chair of radiology Robert Fournier, MD, said the novel software “will help radiologists look at the thousands of images we see in the reading room every day.”

Chest X-rays are the most common diagnostic imaging test performed in hospitals, frequently done on critical care units or in emergency rooms. Many are ordered “STAT,” or urgent, some for reasons other than true clinical necessity. As a result, scores often must wait in a long queue to be reviewed by a radiologist in the order they’re received regardless of their suspected or eventual findings, or patient symptoms. Delayed diagnosis of a large or hard-to-spot pneumothorax could be life-threatening. At St. Luke’s University Health Network, 130,000 chest X-rays are performed each year.

According to global healthcare knowledge provider BMJ, some 74,000 Americans suffer collapsed lungs each year, resulting from trauma, cigarette smoking, drug abuse and certain medical conditions.

In helping develop the AI technology with GE, radiologist Karl Yaeger, MD, and the team at SLUHN reviewed hundreds of X-rays without and with pneumothoraxes of all sizes and degrees of severity. These were then processed with the deep-learning software algorithm to determine its accuracy.

He says he’s enthused to be a part of this pioneering effort ushering in this “new wave” of healthcare technology. He’s optimistic at the potential of the Optima XR240amx featuring Critical Care Suite to serve as an “untiring and undistractable partner.”

Staff Writer