Ezra looks to change cancer screening with MRI scans and artificial intelligence
Cancer screenings are so important to the lives of so many people, but the effectiveness of those tests is alarmingly low. With lung cancer screenings, for example, 56 percent of tests have false positives, meaning that those patients are exposed to more testing and radiation as a result.
Ezra is a new company that is looking to put an end to that problem by changing the way we screen for cancer. Namely, it wants to use a full body MRI, along with artificial intelligence, to detect any abnormalities or potential problems, which it says is not only more effective, but also safer for the patient.
The company came out of stealth on Thursday, announcing the launch of its private beta, as well as a partnership with outpatient imaging company RadNet.
In addition, the company also announced that it raised a $4 million seed round led by Accomplice. Additional strategic investors include Founders Future, Credo Ventures, Seedcamp, Esther Dyson and a number of startup founders and angel investors.
Founded in 2017, Ezra is a direct-to-consumer cancer screening platform that uses a combination of MRI screenings and artificial intelligence to do more accurate cancer screenings.
Members pay Ezra $999 to get one screening a year, though they will eventually be able to pay more if they wish to do additional screenings. They can schedule an appointment with a RadNet imaging facility on the Ezra platform, and, when they arrive, they will be greeted by a member of the Ezra staff at the facility who will guide them through the process. The member gets their scan done and then the results will be available within their Ezra interface, and, with one tap of a button, they can schedule a call or a meeting to speak to the Ezra physician to be taken through the results.
"In a nutshell, the problem we’re trying to solve is that, today, there’s no easy way to get screening for cancer anywhere in the body. If you want to do a test once a year to tell you if you have cancer anywhere in the body, there’s no solution. Our mission is to solve that, to try to create a new way to screen for cancer anywhere in the body using MRI and AI," Emi Gal, CEO & co-founder of Ezra, told me in an interview.
Starting with prostate cancer
While the New York City-based company does eventually want to use its technology to detect all types of cancer, it is start small, initially only screening for prostate cancer "because the current way to screening for prostate cancer is pretty broken," said Gal.
Currently, prostate cancer is screened is first with a PSA blood test, followed by a prostate biopsy. On average, Gal said, 48 percent of the time, men will have cancer and it will be completely missed by this process. The accuracy for a prostate MRI scan, on the other hand, is 90 percent.
The reason that Ezra is going direct to consumer, rather than selling to hospitals or facilities, is that is the best way to help people detect cancer, Gal told me.
"We didn’t start Ezra to apply AI to medical imaging, we started Ezra because we wanted to help people detect cancer early, and so going direct to consumer and targeting consumers that are at risk, and showing them why Ezra can help them detect cancer, we think will get more people to find cancer early, rather than providing our technology to hospitals or imaging facilities where you still don’t solve the problem or people not knowing that they should get screened and not knowing that this is an available solution to them," he said.
The other reason to go direct-to consumer is that, right now insurance will not pay for an MRI scan for cancer treatment, as that is not the standard treatment.
"If you are a man at risk for prostate cancer, what insurance will today cover is a PSA blood test. They won’t cover a prostate MRI because it’s not standard of care. Our goal is to start by going direct to consumer and, in time, gather data to show that what we do has higher accuracy than the current standard of care, therefore it should be covered by insurance," said Gal.
"I expect that process is going to take a while so, in the meantime, we are targeting men who are worried about their risk of having prostate cancer and want to have a better way of doing it, or finding out whether they have it. And are willing to pay out pocket."
Using MRI to detect cancer is not only more accurate, he told me, but it's also safer, as it doesn't expose the patient to unnecessary radiation.
"The great thing about MRI is that is the best imaging modality for the body because it’s very high resolution and it doesn’t create any radiation on the body. Because of that, you can basically use MRI to image all the internal organs and tissue in the body and, therefore, you ca use MRI find abnormalities or lesions in any organ, anywhere in the body," he told me.
"The reason it’s not used today is because MRI scans are both to expensive acquire and expensive to analyze. So, we plan to use AI to bring costs down in order to make MRI-based cancer screen more affordable. And, ultimately, what we envision is you do a full body MRI scan that’s analyzed by AI, that tells you whether you have cancer anywhere in the body or not."
Of course, the long term mission is to beyond just prostate cancer, and to be able to scan the entire body for any type of abnormalities.
"Once we get to do that, think about it: you can get once scan a year and be comfortable and can be at peace that don’t have cancer and that this one scan can save your life, and I think that’s very important," said Gal.
Ezra will be rolled out the roughly 100 men over the next month in its private beta.
Using artificial intelligence
The other part of the Ezra platform is its artificial intelligence, though, as the company is currently in the process of applying for FDA clearance, the AI won’t actually be part of the platform until sometime next year.
Once the AI is cleared, the vision is to have the Ezra AI support radiologist in three use cases, the first of which is offering them clinical support when they analyze the scans for Ezra's members.
"What we will do, once our AI is cleared by the FDA, is provide our AI to RadNet's radiologists so that when they analyze your scan, they can automate a number of the things that they currently do manually," Gal explained.
"One of the things a radiologist needs to do is measure the size of the prostate, which takes a few minutes; we can do that automatically in a few seconds. If there are lesions, like abnormalities in the MRI, the radiologist needs to measure the size of the lesion, the location, which we also automate. Lastly, we also provide a heat map of what the AI sees; this way, if there are any abnormalities that the radiologist missed, we will point those out."
The second use case is to help the radiologist prioritize the cases that they sees in a day, so they know which ones are critical and require urgent action.
"If you’re a radiologist, you’re coming in in the morning and you’re looking at 50 patients that you need to analyze that day. What we’ll do is we’ll prioritize the cases based on the most abnormal ones, or the most important ones, based on what the AI has found, which then help the radiologist start in the morning with the most complicated cases, when they’re rested and potentially have a lower error rate," said Gal.
The last use case is radiologist QA, where every single patient will be analyzed by the radiologist and also by the AI, which means that the radiologist will always have a sounding board to consult in their analysis, which will potentially help them eliminate biases.
Ultimately, the AI will go toward making Ezra a better experience for its members, giving them more accurate results.
"Part of what we’re solving is not just building AI to make radiologist more productive and to lower the cost of prostate cancer screening, it’s also creating a much better customer experience," said Gal. "We’re making prostate cancer screening as easy as ordering an Uber which is how we think should be, and it’s definitely not the case right now."
The future of cancer screenings
The company will be using its funding, in part, to expand its team from five to 20 people by the third quarter of next year, specifically building out its engineering team. It will also be using the funding to continue getting more data in order to train the AI, either by acquiring it or through partnerships.
Finally, the company will use the capital to acquire members and run campaigns to "make sure that we identify men who are at risk of prostate cancer, maybe they’re not even aware of it, and get them to become Ezra members," said Gal.
That could mean targeting men on social media who are in the right age bracket, or who have Googled about prostate cancer. Or those who have liked a prostate cancer charity on Facebook.
"Our advertising and marketing will be mainly to educate on the importance of screenings and then on why what we do with Ezra is more accurate and potential of greater clinical significance than the current standard of care."
And that feeds into the company's main objective: to get people to detect their cancer as early as possible.
"Our belief is that, and this is supported by a lot of third party studies, if you detect cancer early, not only do people have a much higher chance of surviving the cancer, but it’s also much more cost effective to treat," he said.
The difference in cost for catching cancer early is stark: if found early, prostate cancer will cost the healthcare system $20,000. If detected late, the same cancer, once it’s expanded potentially to other regions in the pelvic area, the treatment can involve surgery plus radiotherapy plus chemotherapy, which can cost up to $1 million a year.
"What we think will happen once everyone’s detecting cancer early anywhere in the body is that the cost of treatment will go down, which means that the insurance premiums will stay flat or go down, which means that people will be spending less of their income on treatment," Gal said.
"Aside from the benefits of helping people survive longer, we think we’ll be able to help the healthcare system save money, which is important given that the healthcare system in the United States is 17 percent of GDP every year. We think that’s a very important mission to pursue and we think that early detection will help save costs in the medical ecosystem."