Lange said AI adoption has been slow so far. He feels that this will change rapidly in the next few years as more clinical evidence shows that AI can improve diagnostic accuracy and significantly speed up his workflow.
“This is happening slowly, but in the next few years there will be an explosion of AI for everyone. I don’t think you can, there will be misdiagnosis and the measurements won’t be too unreliable,” Lange explained.
How can AI improve the accuracy and workflow of echocardiography?
Lange said commercial AI algorithms include many methods that help automate time-consuming parts of echo inspection and post-processing. This is all the more important as echo labs across the country are now facing a growing shortage of sonographers. Lange says the time-saving tool will now be able to automatically identify and outline the anatomy of the heart, allowing sonographers and echocardiographers to draw lines on images. said he no longer had to spend time on Combining this with an AI that selects textbook sample images to perform automated measurements helps standardize quantifications such as ejection fraction (EF) and strain. AI can also take his full-volume 3D dataset and slice the views required for echo examinations to perform measurements.
He said this would standardize how measurements are performed and enable reproducibility, regardless of who performed the test or their level of experience. This also helps to produce more accurate and consistent measurements across consecutive patient exams, and provides a more accurate picture of disease progression rather than assuming that there was some variability in the measurements of the person who last scanned the patient. It helps in early and accurate identification.
“People would say, ‘There are no jobs right now,’ but I think people at Echo are just going to get different kinds of jobs,” Lange explained. “Instead of people doing these tedious measurements over and over again, we’re going to be able to do all sorts of different things with our brains.”
He said the first step for AI is perfecting anatomical identification and contouring of that anatomical structure. The technology has been FDA cleared and has been on the market for several years. Once we have that and it is reliable, the next step is to use the automatically contoured anatomy so that the AI can find the ideal frame and position within the anatomy. to enable automatic measurements, Lange said. He said the technology has been brought online commercially over the past few years. Now, vendors are taking the next step, explained Lange. AI can look at measurements, determine if they are normal or abnormal, and make suggestions for diagnosis.
He said all these steps were done in order. It is expected that many new echo systems entering the market in the near future will have these features.
“In the future, when you run Echo, all measurements will be displayed and you will be able to accept or change these measurements,” explained Lange.
Lange also said this should go beyond EF’s standard echocardiogram and include more sophisticated algorithms to evaluate patients for heart failure, amyloidosis, hypertrophic cardiomyopathy and other more complex conditions. Stated. A doctor could come to the same conclusions as an AI, but much more quickly if the AI took an initial reading and provided evidence as to why it thinks a patient has a particular disease state. can do. After that, it is up to the doctor to double it or not. Check AI and use its clinical insight to determine if the machine is correct or if there are other factors that need to be considered.
AI could help identify the type of amyloid a patient has
AI is also being developed to help classify different subtypes of disease, which is not always straightforward with today’s echocardiography. Disease is often detected by echo, but sometimes it is unclear to distinguish which subtype it is, or more advanced image processing is required.
One example is amyloidosis, where the subtype of the disease determines treatment.
“Amyloidosis is a very underdetected disease. You can say, ‘This is amyloidosis,’ and you can tell what type it is. We will be able to do that.” Amyloidosis,” Lange said.
