par John R. Fischer
, Staff Reporter | September 13, 2019
The FDA has given GE Healthcare the go-ahead for its Critical Care Suite, a collection of AI algorithms embedded in its mobile Optima XR240amx system.
Deploying via the company’s Edison platform, the algorithms are designed to help reduce the average turnaround time for radiologists to assess a suspected pneumothorax, a difficult to interpret condition that causes lungs to collapse.
“The sooner a radiologist can be alerted of condition like PTX, the faster the communication to the attending physician can take place,” Katelyn Nye, general manager of global mobile radiography & artificial intelligence at GE Healthcare, told HCB News.
At RSNA 2019, Dunlee is announcing a new product development project to design CT replacement tubes for the next generation of GE CT scanners. The project promises to create an even broader portfolio of Dunlee CT replacement tubes. Read more>>>
PTX is a subtle condition that can be missed by ICU and ER clinicians, as well as radiologists, especially with trauma patients. But even if such a condition is found and prioritized as a "STAT" X-ray, the wait time for a radiologist to review it can be up to eight hours. This is partially due to the fact that 62 percent of exams are marked "STAT" or requested for urgent reading, despite some not being critical.
Critical Care Suite automatically analyzes images in search of a PTX while patients undergo their scan. If it identifies one, the system will send out an alert with the original chest X-ray directly to the radiologist for review through the provider’s PACS system, as well as an on-device notification to the technologist so they are aware of the prioritized case. Quality-focused AI algorithms simultaneously analyze and flag protocol and field of view errors, and auto rotate the images on-device.
Developers trained the PTX algorithm on thousands of unique patient images derived from six data sources across three countries and from 11 different X-ray detector manufacturers, including CR/DR equipment and AP/PA views from mobile and fixed X-ray devices. The strongest form of annotation was used to draw the contour of each PTX into a region of focus for the deep learning algorithm as a training basis.
Critical Care Suite generates AI results within seconds of the image acquisition and does not depend on connectivity or transfer speeds. No investments in additional IT infrastructure, security assessments or cybersecurity precautions are needed to route images offsite, and technologists can reject or reprocess cases right at the patient’s bedside and before the images are sent to the PACS system.
“Critical Care Suite is only intended to triage PTX cases at this time, although GEHC envisions to expand to other conditions in the future,” said Nye.
Additional partners in the development of Critical Care Suite include St. Luke’s University Health Network, Humber River Hospital, and CARING – Mahajan Imaging – India.