How Nucleai is using AI to change healthcare
This article is part of our “How AI is changing the world” event series, held in San Francisco, New York, and Tel Aviv from June to November 2019, featuring insights by leading scientists and entrepreneurs on how AI will change healthcare, communication, agriculture, travel, and other industries. Check out all 12 talks here.
* * * * *
Avi Veidman, who spent 20 years in Israeli military intelligence, where he served as head of AI and machine learning. Avi, a software engineer by training, recalls that he and his military team used AI in aerial image analysis before it was a big thing.
“We started with detecting different objects in satellite images,” he says. “First, we imitated human analysis of satellite and aerial images. The second stage was to discover things that human beings couldn’t.”
After retiring from the military in 2016, Avi and two colleagues teamed up to start Nucleai — with the goal of finding an area where they could use AI to impact human lives. They settled on pathology, a field of medicine that has used the same method of studying tissue through a microscope for 150 years.
“We said to ourselves, ‘You know what, we could do something better,'” he recalls. “Essentially, they’re looking at the slides, at the tissue. They’re looking for different angles that are benign or malignant, and today computers could do that quite well.”
Backed by $5 million in Series A funding, Nucleai has built solutions for colon, stomach, breast, and prostate biopsies — and plans to add skin biopsies soon. Its AI-based solution performs a quality analysis of a pathologist’s work to help locate areas that might have been missed. The solution also allows pathologists to share and view annotations–a kind of “Uber for pathologists”, Avi says.
Using supervised machine learning, Nucleai’s system identifies patterns in biopsies invisible to the human eye. Data is the most important aspect to the development of the company’s AI solution, which is why it has established partnerships with several Israeli hospitals. To date, Nucleai has collected millions of slides that are used to understand which treatments benefit which patients.
Nucleai’s technology may help improve efficiency as the number of active pathologists declines. By 2030, the number of active pathologists in the United States may drop 30 percent compared to 2010 levels, according to a study co-authored by Prof. Stanley J. Robboy, an advisor to Nucleai. As supply shrinks, machine learning solutions could help fill the void.