Ezra is Revolutionizing Cancer Detection with Computer Vision and Artificial Intelligence
/
Ezra CEO and Co-founder Emi Gal (courtesy Emi Gal)
When Emi Gal was searching for the cofounder for his New York-based cancer detection startup Ezra, he pored over 2,000+ online profiles of individuals with expertise in medical imaging and deep learning. From there, he reached out to 300 individuals, conducted 90 interviews, and walked four finalists through a four-month project. The entire process took nine months.
All while still at his “day job.” For Emi, that day job was his European startup that had just been acquired by one of its largest American clients. Not your typical founder’s story, but as his meticulous cofounder search shows, Emi is not one to take a blind leap.
It’s this willingness to go methodically down the rabbit hole and to leave no stone unturned that’s defined Emi’s approach, from one venture to the next.
To date, Ezra has raised $4 million in its first round, led by Accomplice and with participation from Founders Future, Seedcamp, Credo Ventures, Esther Dyson, Taavet Hinrikus, Alex Ljung, Daniel Dines, and many others. Emi has since brought on a head of talent to help expand the team from two to 12 in four months.
In this article, Emi talks about how Ezra is changing the game for cancer detection through visual technology. Read on to learn about his scientific research-driven approach, the long-term possibilities for Ezra.
Photo courtesy Ezra
MISSION-DRIVEN: MAKING CANCER SCREENING MORE AFFORDABLE, ACCURATE AND NON-INVASIVE
The name Ezra means “help” in Hebrew.
It’s a spot-on moniker for a company that, although powered by artificial intelligence and visual technology, Emi says is more mission-driven than technology-driven.
Multiple components in cancer detection rely on the painstaking study and analysis of visual inputs, making visual technology ripe for leveraging.
In its current stage, Ezra is focused on detecting prostate cancer through full-body MRIs that are then analyzed by artificial intelligence. Ezra’s subscription-based model offers an MRI and access to medical staff and support at $999 for one year.
At LDV Capital, we invest in people building visual technology businesses. Every year, we produce a proprietary LDV Insights report, where we deep dive into a sector where visual technology will revolutionize businesses. Learn about the latest advances in visual technologies and the role of artificial intelligence in medical diagnostics in our LDV Insights on Healthcare.
Here’s a sneak peek:
The full-body MRI is a huge change compared to the most prevalent detection method for cancer that kills 1 in 41 men: getting a biopsy, which is painful, uncomfortable, and can have unpleasant side effects.
Magnetic resonance imaging, on the other hand, eliminates the discomfort and is more accurate than biopsies or blood tests. MRIs, however, are not without their costs: about $1,500 if individual books one himself, sans Ezra. And then there’s the time a radiologist needs to scan it.
Radiologist vs AI detection of cancer in MRI scans (Courtesy Ezra)
“We’re trying to make MRI-based cancer screening affordable,” says Emi. “The cost of getting screened with MRI has two parts: the scanning and the analysis. You need to get a scan, and a radiologist needs to analyze the scan and provide a report based on the scan. Those two things each have a cost associated with them. We’re using AI to help drive the costs down by assisting the radiologist, and by accelerating the MRI scanning time.”
The first thing a radiologist does is make a bunch of manual measurements of the organ in question — in Ezra’s case, the size of the prostate. If you have an enlarged prostate, you have a higher likelihood of having cancer. If there’s a lesion like a tumor in the prostate, radiologists need to measure the size and location of the tumor. They need to segment the tumor so they can make a 3D model so the urologist knows what to focus on. All of those measurements and annotations are currently done manually, which makes up about half of a radiologist’s workload.
“What we’re focusing on is using the Ezra AI to automate all of those trivial, manual tasks, so the radiologist can spend less time per scan doing manual things, and instead focus on analysis and reporting. That will potentially make them more accurate, as well.”
For the future, says Emi, the team is already considering how to use AI to accelerate the scanning process as well.
“The reason this is possible now and it wasn’t possible before is that deep learning has given us the ability to be as good or better than humans at these things, which means it’s now feasible to create these types of technologies and implement them into the clinical workflow.”
THE SEED OF A NEW IDEA: PLANT IT BEFORE YOU’RE READY
While it looks like Emi has seamlessly gone from one successful venture to the next, the reality is a lot more nuanced.
It was while he was still running Brainient before it was acquired, that he started plotting his next move. In 2015, Emi was introduced to Hospices of Hope in Romania, which builds and operates hospices that care for terminally ill cancer patients. During his visits with doctors and patients, the seed of Ezra was born.
Cancer struck a personal chord. As a child, Emi had developed hundreds of moles on his body, which put him at a very high risk of melanoma. He started getting screened and going to dermatologists regularly from the age of 10 years onwards to make sure they weren’t cancerous. While he hasn't yet had any maligned lesions, he’s experienced the discomfort of biopsies firsthand, and he’s always been very conscious about the importance of screening.
“I realized while speaking to doctors that one of the biggest problems is that people get detected late. I started looking into that and realized that [this is] because there’s no fast, affordable, accurate way to screen cancer anywhere in the body,” says Emi. From there, he began researching different ways in which you can screen for cancer. A computer scientist by education, he spent two years on what he calls, “learning and doing an accelerated undergrad in oncology, healthcare, genetics, and medical imaging.”
That accelerated educated is supplemented with an incredibly impressive team. It’s no surprise that Ezra co-founder Diego Cantor is equally curious and skilled, and brings an enormous technical repertoire to the table: an undergraduate education in Electronic Design and Automation Engineering, a master’s degree in Computer and Systems Engineering, a Ph.D. in Biomedical Engineering (application of machine learning to study epilepsy with MRI), and Post-doctoral work in the application of deep learning to solve medical imaging problems. The scientific team is rounded out with deep technical experts: Dr. Oguz Akin (professor of radiology at Weill Cornell Medicine and a radiologist at Memorial Sloan-Kettering Cancer Center), Dr. Terry Peters, director of Western University’s Biomedical Imaging Research Centre), Dr. Lawrence Tanenbaum (VP and Medical Director Eastern Division, Director of MRI, CT and Advanced Imaging at RadNet), and Dr. Paul Grewal (best-selling author of Genius Foods).
Ezra’s deeply technical team trained AI with data sets from the National Institute of Health marked up by radiology experts. On new data sets, the Ezra was 90% accurate at agreeing with the experts.
LEARNING (AND STARTUP SUCCESS) IS ALL ABOUT COURSE CORRECTING
As a lifelong learner who actively chronicles his year-long attempts to gain new skills and habits, Emi has picked up a thing or two about doing new things for the first time.
“Learning anything of material value is really, really hard,” says Emi, who’s done everything from training his memory with a world memory champion to hitting the gym every single day for one year.
This focus on the process and being comfortable with being uncomfortable came in handy when Emi, who studied computer science and applied mathematics in college, pivoted into cancer detection.
Emi cycled through twelve potential ideas before deciding on Ezra’s current technology. At every turn, he researched methodically and consulted with experts.
One of the promising ideas that Emi considered — shelved for now — is DNA-based liquid biopsies. “We’re at least a decade away from DNA-based liquid biopsies being feasible and affordable,” says Emi, who was searching to make an immediate impact.
Emi had just sold Brainient and was on his honeymoon when he came up with the winning idea in November 2016. “I had this idea: what if you could do a full-body MRI and then use AI to lower the cost of the full-body MRI, both in terms of scanning time as well as analysis in order to make a full-body MRI affordable for everybody? My wife loved the idea, and that’s always a good sign.”
In January 2017, Emi discovered a research paper that was comparing the current way to screen for prostate cancer — a Prostate-Specific Antigen (PSA) blood test followed by a biopsy — with MRIs as an alternative. “An MRI was by far more accurate and a much better patient experience, and so I was like, this is it. It can work. And that’s how Ezra was born.”
THE FUTURE
Emi has big plans for Ezra going forward, and this year’s LDV Vision Summit is one step in that direction. He hopes to meet people working in the vision tech space, particularly within healthcare.
Although Ezra has been live just since January 7th of this year — 20 people were scanned in the first three weeks — its early results are very promising.
“The first person we scanned had early-stage prostate lesions. That really makes you wake up in the morning and go at it,” says Emi.
Out of the first 20 scanned, three had early-stage prostate lesions they were unaware of. Two early users came in with elevated PSA levels, but the MRIs showed no lesions, obviating the need to do prostate biopsies.
The long-term potential for Ezra — going beyond prostate screenings — is also clear.
“We helped one man who thought he was dying from cancer find out that he had no cancer but that he was likely an undiagnosed diabetic,” says Emi. “We gave him the information for his urologist and physician to make that diagnosis. We checked with him a month later and he’s over the moon happy. We helped him get peace of mind that he doesn’t have cancer, as well as diagnose a disease he wasn’t aware he had.”
Even though these results have been powered by AI and MRIs, Emi is emphatic that Ezra is not an AI company. “We want to help people detect cancer early...and we will build any technology necessary. We think of ourselves as a healthcare company leveraging AI, not the other way around,” he says.
While Ezra’s current focus is prostate cancer, expanding to other cancers that affect women is on the horizon. After all, as Emi points out, “Women are the most preventative-focused type of individual, for themselves and their families.” To underscore that point, Emi says that many of the early adopters for prostate screenings have come at the encouragement of men’s female partners.
“The way we are approaching expansion is based on cancer incidence. We’re starting with the cancers that are most prevalent across society, with prostate being one of them, breast, lungs, stomach, and our ultimate goal is to be able to do one scan a year and find cancer in any of those organs.
“Our ultimate goal is to do a full-body MRI analyzed by AI and find any and all cancer. In a decade, I hope we will have gotten there.”
Check out our latest post about Ezra and their AI-powered cancer screening.
If you’re building a unique visual tech company, we would love for you to join us. Through our annual LDV Vision Summit and monthly community dinners, we bring together top technologists, researchers, startups, media/brand executives, creators, and investors with the purpose of exploring how visual technologies leveraging computer vision, machine learning, and artificial intelligence are revolutionizing how humans communicate and do business.
