ChestLink, an autonomous AI imaging suite by medical imaging company Oxipit, was created to exist alongside medical professionals “as an integral part of the clinical workflow.” ChestLink can scan X-rays for abnormalities and send patients reports when their X-rays appear normal. When the tool encounters a potential health concern, it sends the X-ray to a radiologist for manual review. If there’s any “doubt” from the AI that a patient’s X-ray may feature abnormalities, it sends that X-ray to a radiologist as a precautionary measure to ensure hidden health issues don’t slip through the cracks.
Oxipit received CE mark certification for the tool last week, classifying ChestLink as a class IIb health device (on par with a ventilator or any intensive care monitoring equipment) found to comply with all relevant EU legislation. The company plans on certifying ChestLink through the Food and Drug Administration (FDA) as well for eventual use in the United States.
Before it was certified, ChestLink spent more than a year reviewing 500,000 real X-rays across multiple pilot locations. “The sensitivity metric of 99 percent has translated to zero clinically relevant errors at our deployment institutions during the application piloting stage,” Oxipit CEO Gediminas Peksys said in a release.
ChestLink’s first clinical deployment is slated for early 2023. For safety and certainty’s sake, Oxipit will start out by assigning ChestLink retrospective analyses—AKA X-rays that radiologists have already reviewed. Once the tool passes that real-world test, ChestLink will be used for preliminary analyses under supervision by Oxipit and medical institution staff. The final stage of deployment will move ChestLink into autonomous prospective reporting as designed. In the rare case that a radiologist may need to manually review a patient’s X-ray after ChestLink gave them a clear bill of health, staff will be able to quickly trace the steps of application decisions through a real-time analytics page.
Oxipit states ChestLink will be particularly convenient in primary care offices, where around 80 percent of all X-rays turn up normal. The time saved is intended to be spent on direct patient care. Given a widespread concern among radiologists that their profession could become fully automated, technology like Oxipit’s has been contested for quite some time. Oxipit and some experts say it’s more likely that AI will continue to assist and improve healthcare instead of replacing people entirely—despite mathematical proof that AI tends to outperform human radiologists. But as Oxipit itself says, “it’s not a question of mathematical performance” when AI capabilities and human expertise can be combined.
Now Read:
- What Is Artificial Intelligence?
- Scientists Have Finished Sequencing the Human Genome. Again.
- North America’s First Skull Surgery Was At Least 3,000 Years Ago
from ExtremeTechExtremeTech https://ift.tt/9bVqGFI
ليست هناك تعليقات:
إرسال تعليق