The first person to read your next X-ray could actually be a machine, thanks to a private equity-backed company.
Imaging Advantage LLC, a cloud-based imaging company backed by Goldman Sachs Group Inc., among others, is hoping to introduce artificial intelligence in the reading of X-rays.
The company, which also counts private equity firms Brightwood Capital Advisors and CRG as investors, teamed up with faculty members from the Massachusetts Institute of Technology and Harvard Medical School’s teaching hospital, Massachusetts General Hospital, to develop a machine learning initiative named Singularity Healthcare. The technology will take a first look at digital X-rays, highlighting potential areas of injury and disease, before sending the images to one of the 500 board-certified radiologists connected to Imaging Advantage’s cloud-based network.
The machine is named after the concept of singularity, in which artificial intelligence would become capable of self-improvement to the point at which it surpasses human control or understanding.
“The whole idea of deep learning is that the machine is becoming better looking at analogies,” said Naseer Hashim, founder and chief executive of Imaging Advantage. “If it detects a potential problem area in one patient, it will also have the knowledge, hopefully, of many more similar indications.”
He added that Singularity Healthcare will continuously learn from Imaging Advantage’s expanding database of 7 billion images.
“The machine will read what radiologists determine in other cases and come up with a potential response on what that is,” he said.
There has been a push to introduce technology in health care. Although electronic medical records have become commonplace for the past couple of years, health care lags behind other industries in the deployment of analytics, often takes on an assisting role in facilitating physicians in back-office functions and hasn’t played that big of a part in treatment.
The Singularity project—to be launched later this quarter—is still far from allowing machines to take over the diagnosis-making process, but it is the first step to let technology aid physicians and clinicians in the more commoditized part of their job, allowing for a faster turnaround of test results and potentially help eliminate errors or misses on X-ray reading.
Despite its name, however, Mr. Hashim said the technology isn’t set up to take over the jobs of radiologists.
“I don’t believe the market will be ready to have non-board certified radiologists reading exams from a patient perspective,” Mr. Hashim said. “I don’t think insurance companies will support it. You are looking into the distant future, even if the technology is perfect, before radiologists come out of the equation.”
Write to Amy Or at email@example.com; follow her on Twitter: @theamyor