Transcript
Announcer:
You’re listening to Clinician’s Roundtable on ReachMD. On this episode, Dr. Taro Minami will be talking about the role of AI in point-of-care ultrasound. Dr. Minami is the Chair of the Ultrasound and Chest Imaging Section at CHEST, the Director of the Intensive Care Unit at Kent Hospital, and a Professor of Medicine and Medical Science and Clinician Educator at The Warren Alpert Medical School of Brown University. He’s also a Clinical Professor at Kyoto University in Kyoto, Japan and a Visiting Professor at Tokyo Women’s Medical University in Tokyo, Japan. Let’s hear from Dr. Minami now.
Dr. Minami:
AI is still in progress, so you got to be really careful about whether what AI is telling us is true or not true. Having said that, AI is a great tool to guiding the novice learners. For example, there is image guidance. "Hey, you should move the probe this way or that way." There's a certain product. There are so many companies giving you so many products in different ways. But with image guidance and standardizing the quality by different providers, from the novice to the expert, AI might play an important role in the near future. It's already happening.
But again, you still have to be careful. AI could give you some answers to, say, what is the ejection fraction? What is the PTI? What are the IVC variations? It gives you some answers, but if you do not understand the true concept, and you simply believe the numbers, then you might be running the risk of making big mistakes, because I still see a lot of variations in the assumption of the ejection fractions or the PTI. You have to be really careful about the variability or the certainty of the numbers. Still, it's a mission in progress. Say that AI is telling you EF is 60 percent, and you simply believe that without double-checking by eyeball assessment, so to speak. Then you might be running a big risk. In general, we could underestimate or overestimate ejection fraction or underestimate IVC diameter or misinterpret it. For example, we show the image of the abdominal aorta and let the AI assess IVC variation. IVC variation might be so little, so you don't have to give fluid or volume resuscitation, when, in fact, what you're looking at is abdominal aorta, not IVC. AI might not be able to tell that's aorta versus IVC. So you have to be really careful. You always have to double and triple-check with your own eyes.
Announcer:
That was Dr. Taro Minami talking about how AI is transforming point-of-care ultrasound. To access this and other episodes in our series, visit Clinician’s Roundtable on ReachMD.com, where you can Be Part of the Knowledge. Thanks for listening!
