Intel lends its deep learning AI to DiA Imaging's heart ultrasound analysis tech

Artificial intelligence developer DiA Imaging Analysis has a fever, and the only prescription is more AI.

The Israel-based company has already landed a handful of FDA clearances for its cadre of AI technologies but is now planning to amp up the analytical abilities of its software by infusing it with Intel’s own deep learning programs.

In a collaboration announced this week, DiA said it would use Intel’s OpenVINO toolkit to speed up the processing time of its LVivo Seamless cardiac ultrasound analysis platform.

“Every minute counts in a busy echocardiography environment. Using an AI solution like LVivo Seamless that automatically runs ‘behind the scenes’ on hospital servers eliminates many manual and visual steps involved with both cardiac ultrasound view selection and measurement,” said Hila Goldman Aslan, DiA’s CEO and co-founder.

LVivo Seamless was cleared by the FDA in 2020 for its ability to automatically cherry-pick the most useful ultrasound views from the 40 to 60 captured in a standard echocardiogram exam.

From there, the software ropes in a slate of DiA’s other LVivo programs to analyze the selected images. That includes LVivo RV, which was recently proven to analyze right ventricle size and function on par with cardiac MRI scans, as well as applications to assess the heart’s ejection fraction and myocardial strain.

To speed up that entire process, DiA is looking to the Intel Distribution of OpenVINO toolkit—the Intel-supported version of the toolkit, which is also offered as an open-source system. OpenVINO gives software developers a plug-and-play framework to easily integrate deep learning technology into their own applications that run on Intel’s hardware.

So far, thanks to the team-up with Intel, DiA said it has already been able to speed up the processing time of its image selection and ultrasound analysis AI by about 43%, while still maintaining its usual levels of accuracy. Because DiA’s technology is installed on hospitals’ existing IT systems, the boost is applicable only to those running on Intel Core processors—which, fortunately, comprises many of the U.S.’s hospitals.

This isn’t DiA’s first partnership with a tech giant. In 2017, it linked up with GE Healthcare to develop a version of its ultrasound imaging analysis software that could be used at the point of care.

A year later, DiA unveiled a version of LVivo EF that could be used on GE’s Vscan Extend handheld ultrasound device. At the time, the companies said it was the first automated ejection fraction analysis software to be made available for portable ultrasound tech.