Emi Gal, founder and CEO of Ezra, on the VatorNews podcast

Steven Loeb · May 17, 2024 · Short URL: https://vator.tv/n/5894

Ezra offers a full-body MRI that can detect cancer and over 500 other conditions in 13 organs

On the VatorNews podcast, Steven Loeb and Bambi Francisco talk to Emi Gal, founder and CEO of Ezra, a company pioneering the use of full-body MRI to detect cancer and over 500 other conditions in up to 13 organs in the body.

Highlights from the call:

  • "I became very interested in cancer for very personal reasons: I'm personally at high risk for cancer and sadly, I've had cancer in my family, my mother passed away from cancer because she found cancer late. And so, with my computer science hat on, I was like, ‘well, could we create a way to help people find cancer early so that we can make cancer non-lethal, and so that we can give people a better chance at survival?’ And that's how Ezra was born.”
  • “Full-body MRIs have actually existed for quite a while, I want to say like a decade or longer. However, the problem with full-body MRIs before Ezra was that you would have to pay $20,000 and spend two to three hours in an MRI machine to get one. And so, the innovation that we came up with was to use AI technology and software to decrease the cost of a full-body MRI so that anyone can afford it.”

Ezra has three AI tools: Ezra Flash, which cleans up the image; Ezra Assist, which helps the radiologist find issues; and Ezra Reporter, which helps patients understand the findings.

  • “Imagine you have a photo camera and this photo camera is taking a photo, and it's an older type camera. The image resulting from the photo camera will be a little bit grainy, there will be this noise structure to it. If there's an object far in the distance you might not be able to see the object clearly. What we do is we take that image, and we just clean it, we remove the noise, we increase the resolution, we increase the sharpness, so that we can see that the object in the distance is your dog. And so, what this machine learning model does is it enables us to acquire much faster scans on an MRI machine, and then we've cleared a model with the FDA that enables us to just wipe all of that stuff away so that the radiologists can see the scan more clearly."
  • “When radiologists look at the images, they are looking for issues, for tumors, for abnormalities, for hypo or hyper intense areas in a particular slice of a particular MRI sequence. If they find a particular area that they think is interesting, they can consult with the AI to see what the AI thinks about that particular region. So, they can drag a bounding box around a particular area and say, 'do you see anything here too?' and if the AI also sees a lesion there, it will say, 'there's a lesion and here's the lesion border in this particular slice and in the next slice,' and so on. That enables us to assist the radiologist, that's why it's called Ezra Assist, in finding these things, making them potentially more accurate and faster because we can automatically measure these lesions, we can automatically measure the organ, we can automatically provide annotation and measurements and so on.
  • “AI number three, does something very simple: imagine I asked you to read Mandarin; you may as well say, 'read a radiology report.' For most people, it's like reading Mandarin. Our AI is able to take Mandarin and turn it into plain English. And so, that's what we do at every step of the way, non-technically; there's a lot of technical detail on how we do that, but in terms of the actual applications, we're cleaning the images, we’re assisting the radiologist, and we are explaining to the consumer what the findings mean.”
  • “There's always a human in the loop. The reason I refer to the AI as co-pilot, as an assistant to the radiologist, is because the report is generated by the radiologist, it just happens that they're supported by the AI. And so, at the end of the day, it's the radiologist who's signing on the radiology report to send it to consumers.”
  • “Ezra Reporter not only translates radiology reports into plain English, but it also assigns a score to every single finding and then we use that score as a guide to help people determine what they should do about our findings. And in doing so we help minimize the unnecessary follow-ups and downstream procedures because if it's a one or two, we'll just say, 'this is a benign bone island in your C6 of the spine, nothing to be concerned about,' or if it's a three, we'll say, ‘this is something that looks like a liver hemangioma, let's just monitor it over time to make sure that it's stable in size.’ Or if it's a four, we'll say, ‘this is very likely prostate cancer, you need to follow up with a diagnostic scan of the prostate, potentially a biopsy and then go from there.’”
  • “We've now scanned many thousands of people, we found cancer for 6% of our members, so quite a significant number. And it is cancers that would have otherwise gone undetected until they would have become symptomatic, which would have been too late. Every week, we have a medical review with our medical team and we talk about different cases of cancer that we found and from just the past couple of weeks, we have found kidney cancer, we have found brain tumors that ended up being malignant, we have found one breast cancer, even though technically we don't cover breast, we recommend that women do a mammogram. We have found uterine polyps, we have found fatty liver disease, and this is just from the past couple of weeks. In all of the organs we cover, we have helped people find cancer in those organs over the years.”
  • “Sensitivity rate gives you the false negative rate and the specificity gives you the false positive rate. MRI has very high sensitivity for certain organs, and that's why we chose those organs. MRI, everybody agrees, has almost no false negative rate: we're talking 97 or 98% sensitivity, which means like a 2% false negative rate. As a screening exam, it's almost insignificant. What's great about MRI is it also has relatively high specificity and for the organs we've chosen, the average specificity on MRI is about 90%, which means a 10% false positive rate. In screening, you're going to expect some false positive rate: mammogram, which is the most adopted, most well established screening procedure for breast, the specificity is around 88% or 89%, so it has an 11% false positive rate. So, if we're okay with screening exams that have around 11%, false positive rate, which clearly we are with mammograms, we should be okay with other screening exams that have an equivalent false positive rate such as an MRI, and that's the approach we've taken.”
  • “My goal is within the next three to five years everyone's getting a scan every year. Obviously, that means geographical expansion, so we need to partner with all of these health systems. It also means costs, because the more affordable the Ezra scan the more people can buy one and then the more likely it is that payers will ultimately reimburse it. And so, we're pretty confident internally, based on some AI breakthroughs that we've had, that we can get to a 10 to 15 minute full-body scan, full-body MRI within about two years, and a 10 to 15 minute full-body MRI, we can probably price that around $600. At that price point, a lot of people will be able to afford it and payers will be very likely to reimburse it.”

Support VatorNews by Donating

Read more from our "Podcast" series

More episodes

Related Companies, Investors, and Entrepreneurs



Joined Vator on

Ezra’s mission is to enable early cancer detection for everyone by using Artificial Intelligence and the most advanced medical imaging technology available.


Emi Gal

Joined Vator on

I’m a tech entrepreneur and software engineer based in New York, currently working on a new startup in the healthcare space called Ezra (in stealth mode).