Assuta Medical Centers has launched an innovation center in collaboration with Nvidia. The center is intended to enable startups and medical researchers to carry out research and development of artificial intelligence-based solutions.
The center is run by Dr. Michal Gindi, director of imaging at Assuta Medical Centers. Its activity began about a year before the official launch. "We have already met 200-250 companies in the past year without issuing a call for proposals - they found us," says Gindi. "Today, we are working intensively with twenty companies, three of which have already received data using the Nvidia system , while four others will receive data within a few months."
Two of the companies can already be named: AIDoc and Rhino Health. The former has developed image processing technology that warns of problems affecting the prioritization of the patient in the wait for treatment. "In their original model," Gindi says, "they send the warning to the doctor. We helped them to realize that in fact it was better to send the warning to the imaging technician, who after all is with the patient. This is an insight that was discovered to apply particularly at Assuta, because it could be that at a large university hospital it would be more correct to send the warning to the doctor, but we are more similar to the smaller clinics with which the companies will eventually work.
"The need to empower technicians is very significant in hospitals today. The intention is that they should have information and feedback, and be able to make decisions themselves. Together with the company we examined this new use in the research that we planned, and we will also publish an article that follows on from it."
The other company, Rhino Health, deals in federated learning. Liron Freind Saadon, director of developer relations at Nvidia Israel, explains: "This is a new field in artificial intelligence. It can be likened to sending your algorithm to school instead of schooling it at home. Our aim is to expose the algorithm to as much data as possible of different kinds, but it's very hard to bring the data home, to within the company. So we basically send the algorithm to a place like Assuta, run it on the local system, and send it home, improved.
"This is a field that has developed a great deal during the Covid-19 pandemic. Hospitals did not want to send lung images of specific patients outside, so image interpretation companies that wanted to understand what the coronavirus looked like sent their models for training within the hospitals, and the result was found to be accurate.
"Rhino Health has developed that platform that facilitates such cooperation. The challenge is to arrange the data from the various sources in such a way that the algorithm can read it. This is very much the cutting edge of AI technology, and it’s a big advantage for Assuta to be part of this world."
"We're looking for startups that need the experience and knowledge of our doctors," says Gindi. "One of the most enjoyable things in the process is to find the doctor who really needs the product that the companies are developing and is really determined to advance it and to mentor the company in its interface with the hospital."
What is the center's relationship with the companies that use the technology?
Gindi: "In the case of AIDoc, for example, we are the customer for their product, and we also collaborate with them on research. The greater our contribution to the product, the more likely we are to have a holding in the company. Another possibility is simply to benefit from use of the product.
"Our agenda is not to make money on the technology, but to expose the doctors to the newest technology, and to promote personalized preventative medicine, in which our doctors are at the forefront."
Are you cooperating with a technology transfer company to commercialize these projects?
""Assuta has no classic technology transfer office. Our doctors aren't really 'ours', but independent. Nevertheless, if a venture suitable for commercialization emerges from this activity, we'll do it. We know how to do it taking into account all the different affiliations of the doctors."
In the future the center will be able to use Nvidia's new platform, called Nvidia Clara Holoscan, an AI computing platform for medical devices that is just becoming available.
Published by Globes, Israel business news - en.globes.co.il - on November 10, 2021.
© Copyright of Globes Publisher Itonut (1983) Ltd., 2021.