FDA publishes advice on the use of AI and machine learning to close large vessels in the brain

The Food and Drug Administration (FDA) has advised the use of imaging software for patients with suspected large arteries, suggesting that some radiologists be unaware of the targeted use of computed tomography (CAD) devices. (LVO) in the brain.

Emphasizing the proper use of CADt software, the FDA notes that these devices are not intended to replace radiological diagnostic assessments. While CADt devices can help in showing and prioritizing brain imaging with findings that are suspicious for LVO, counselors note that LVO is a common cause of severe ischemic stroke, even if it is detected by CADt imaging software. not given, may be present.

If there is any possibility of overdosing on CADt software, Vivek Bansal, MD, said it could come from a group of medical care providers who can do the right thing for a patient in a time-limited environment. While interventionists, neurosurgeons, and neuropathologists all have a strong knowledge of brain vessels, said Dr. Bansal, head of the National Neuroradiology Department at Radiology Partners, there can be different levels. He added that while these specialists look closely at the images they take in the operational package, “they can’t look at the actual CT images to the same level.”

Regarding the image, Dr. Bansal said that one can look at the small branched vessels that sink into different parts of the images up and down, and you have to move up and down to actually remove them from the vessel through the vessel. follow up. Dr. Bansal noted that this work can be difficult and especially challenging to do on a smartphone in a well-lit room.

“The clock is ticking, and time is the brain. We try to race with the clock, because every minute we get diagnosed, brain cells can die (if the patient has a clot). The sooner we diagnose them and the patient goes to the catalog laboratory, the better the results for the patient. I think that’s the biggest challenge: trying to do something that is very accurate in a very short amount of time, ”explained Dr. Bansal.

The FDA board also noted that awareness of the design capabilities of various CADt devices is important, many of which have artificial intelligence (AI) or machine learning technology. containers. Dr. Bansal said this is an important difference with AI tools.

“Although some AI instruments are very good at seeing the M1 occlusion, which is the proximal part of the middle cerebral artery, new AI instruments are able to see M2 occlusions with the anterior cerebral artery (ACA) and the posterior cerebral artery (PCA). All of these things are important when it comes to patient care, ”said Dr. Bansal, who is affiliated with the East Houston Pathology Group in Texas.

Dr. Bansal said the key to this is to understand the role of devices with AI enabled and their value in pilot cases.

“At any time, I can have 40 statistical exams on my list. I’ll take them as soon as possible, but if the AI ​​instruments say, ‘Hey, look at this queue,’ be it a possible closure of a large vessel or a brain hemorrhage?” to begin with, this is very helpful, ”said Dr. Bansal. “… Where we are now, I think the only way we can look at AI is to look at it as a test tool.”

Leave a Comment