There’s a movement in the radiology community to become more quantitative. When you think about radiology, it’s a picture of what’s going on in your body, and those pictures very as you go from one MRI scanner to another, and they can vary from day to day but still gives you a reasonably good picture of your body. If you really want to do something quantitative, like watch how much a tumor is growing or reducing in size when you’re taking treatment, it doesn’t do very well unless you implement some standards or ground value. Elizabeth Mirowski, CEO of QalibreMD, shares that their MRI medical imaging standards provide a ground truth baseline for measuring quantitative imaging biomarkers. They developed this technology jointly with the ISMRM, QIBA/RSNA, and NIST to deliver rigor and relevance to medical research and patient care.
Watch the episode:
Listen to the podcast:
Elizabeth Mirowski, CEO QalibreMD transforming MRI medical imaging standards helping reduce diagnoses variability.
We have Elizabeth Mirowski, Ph.D. We interview some of the best and brightest business owners and entrepreneurs in and around the state of Colorado. We talk about the ins and outs of running a business, of being an entrepreneur. You can think of this as your back of the napkin MBA by top business leaders and entrepreneurs in the state of Colorado. The folks that are doing it are not talking about it every day. We talk about what to do and as importantly what not to do about growing, running or starting a business. We’re incredibly fortunate Elizabeth has taken time out of her day to visit with us. She is running the entire operation at the moment. At QalibreMD, the strategy is to transform medical imaging from a simple diagnosis of a lesion to an exact diagnosis of what disease is present, at what stage and to what extent in the body it is occurring. Their superior imaging standards certify the accuracy and imaging biomarkers to reduce unnecessary repeat scans and variability in diagnoses from one health center to another. They focus on making their standards easily adoptable for widespread implementation. Elizabeth, thanks so much for taking time.
It’s my pleasure. Thank you for having me.
Tell me about QalibreMD.
QalibreMD is a spin-off of a parent company. We spun off of High Precision Devices because we were medically focused. We realized that it was a very different market from what the parent company was doing. We needed to ramp up our efforts to meet what we see as the need and demand for radiologists to have more exact results in their diagnoses. We are actually also in the process of raising funds to be able to move quickly with the market
What was the need that QalibreMD started to solve, or the problem they started to solve?
There’s a movement in the radiology community to become more quantitative. What you think about when you think about radiology, it’s basically a picture of what’s going on in your body. Those pictures vary as you go from one MRI scanner to another, and they can vary from day to day. It still gives you a reasonably good picture of your body, but if you really want to do something quantitative like, watch how much a tumor is growing or reducing in size when you’re taking treatment, it doesn’t do very well. Unless you implement some standards, some ground truth value that lets you know that, for example, if you think about a picture and you have colors, your shirt is blue, mine is orange. If we switched them and yours was orange and blue, it’s that whole dress issue that was in the media for a long time. Is it a blue dress or a gold dress? In that case they tweaked the numbers of what the colors look like and produced a totally different result. That can happen in MR. If you put in a blue, that says that blue is blue, then you can start relating everything off of that particular value or number or truth.
You see the way it does weighing on a scale, use zero for tare before you weigh. For most of us, and me in particular on the MRI, I don’t think I’ve ever been exposed to an MRI, didn’t need to be far as I know. You said MRI is an MRI, but the reality is they’re not. You could have an MRI on the East Coast and your MRI on the West Coast and their zero is not the same. How did QalibreMD get involved? Did they develop this technology themselves?
We developed it jointly with the larger radiology community, the International Society for Magnetic Resonance in Medicine and the Radiological Society of North America. Also, an important player in this is the National Institute of Standards and Technology. As you can see in their names, standards, things that are related back to NIST, give you a very good, well-known ground truth value. It’s the zero. It’s a true zero on that terror of that scale that you’re looking at.
In the MRI world, let’s say you’re a consumer that has a tumor of some description. Contrast if you would, the before use of your technology perhaps after adoption of your technology and what the consumer could expect on the outcome of information.
One of the things that you would see change is the ability for you to go to any MRI in the health system that you’re in, the one closest to your house or if you want to travel somewhere and get it done with slightly further away and send it to the radiologist. With our technology, those will come back the same. Therefore, you won’t have to do a repeat scan. You won’t get a misdiagnosis, you won’t get two diagnoses. If you go and get a second opinion, that second opinion is related to the fact that the image data comes back slightly differently. It will always be the same image data.
The contrast is if you went to an MRI A and they go, “Your tumor has not gotten bigger.” You could see that we could go to MRI B, and because of it’s not being zeroed or maybe slightly different than A, you could say it’s either grown or shrunk. Whether that’s true or not is in question.
It’s in question because you can make that MRI image look different, which changes the size simply by the parameters that you choose to acquire the image by. This would enable volume metric analysis. Seeing what the size of that tumor is more accurately. In today’s day and age, even I’ve had maybe four or five different pairs and have been in four or five different health systems in the past fifteen years. Every time I go to a new place, I would have to get all of the image data taken again because they would say, “That stuff was taken at a different facility, we can’t look at it, we can’t read it.” We’ll allow people to be able to do different payer systems and have that data be the same regardless of where they’re getting their image.
I think about the confusion between the care provider and the consumer. I don’t know what you have, and we have to start all over again. You can’t see progression at all. We have some of the devices here. One of them looks like something from Star Wars deal and the other two, they look like what they’re designed to do. Let’s talk a little bit about the three devices.
This one’s one together. We put it onto one plate. It’s specifically designed to fit into breast imaging coils. It’s for characterizing how well tumor volume is assessed in breast tissue. It’s specifically for a clinical trial that is looking to try to graduate develop new drugs to treat breast cancer, to reduce the size of the tumor to the point that when you extract it, you actually maintain as much of the breast as possible and can do a minimally invasive surgery at that point. In order to know whether or not this drug is effectively killing the tumor, you need to be able to have a well-known ground truth value that tells you this is what your machine is reading today. Therefore, the error that you have on your volume assessment of your tumor is less than two percent.
We were talking about the regimen for when you go in for mammography. You go to the physician, get a mammography ordered. You’d go in and get your mammogram. The fun starts if they find something. They try to assess, and then there’s the biopsy. In these particular devices, there’s more going on here than just the shape. What you were talking about is the materials that are inside here and what they replicated. Let’s dig into that a little bit.
What we’ve developed with this product is a fat tissue mimic and it mimics human fat tissue. One of the issues when you do breast imaging is that the fat can obscure what you see in the image data and your ability to measure tumor size. You need to be able to subtract that out. We’ve provided something in this particular device to be able to have people test how well their subtraction works in that image data. Then we also have other elements that we developed that mimic how much water diffuses in the body. When you think about a tumor starting to grow, when it first starts to grow, there’s a whole bunch of cells growing, and so it’s very dense. At that point, water diffuses slower and so you can measure that with the MRI. Then as it becomes, as it grows and becomes edematous, which is typically later stages of cancer. Edema is water, that’s too much water in the cells. You start to see the diffusion of water go much more quickly in the body. You can stage cancer using a measurement of how much water is diffusing in that cell.
When you get a diffusion measurement for the physician or the radiologist is you get a mathematical representation.
Yes, you get a mathematical representation of where your stage of tumor is. Sometimes what you’ll see in anatomy, too, is when that cancer tumor starts to die. I doesn’t mean that the whole volume gets smaller. You have that damaged dead tissue that’s still there. In that case, you can also identify that based off of water diffusion.
Then you can start to tell if you have an effect even if it hasn’t shrunk yet. If they’re out there going like, “I’m beginning to get in the weeds on the science.” In layman terms, what we were talking about before we started this show is the evolution between current mammograms and treatment and biopsy and all this other stuff to you could go in with a calibrated machine and eliminate or at least reduce great deal of that whole trail of treatment
The variability definitely reduces it all and gives a patient a better diagnosis.
For anybody that’s had cancer in her family, which I understand is one in six or one in four Americans, will either have it in their family or somebody they know. That’s effectively everybody. It’s been in my family. It’s been in your family. We focused a little bit on the mammogram side, but for the guys in the prostate space?
We have a device specifically tailored to evaluate prostate imaging techniques. When you think about right now, what the process is you have an elevated PSA, and one that’s suspicious. Then a prostate biopsy and that can be fourteen needles that randomly assess. They take samples of the prostate. You don’t really see the whole picture, you see wherever that needle went. There’s a lot of problems with that technique. It’s invasive. People oftentimes will get infections as a result of that. Our device is meant to use water diffusion, watch water diffusion in the prostate and you can then see the whole prostate and monitor it over time without having to do a needle biopsy. We’re in the stages of trying to prove that we can do that by still having the sampling of the tissue and then correlate it to the MRI image data. Once we can get that to be consistent enough where we can get a number from the MRI that’s consistent with prostate cancer of a specific stage, then we could reduce how many biopsies are needed.
I think about the old comments that there’s a lot of guys past 80 that will have prostate, you’ll die with it, not from it. There’s also the other version that’s the fast growing that will kill you for sure. There’s probably some level of resistance with the notion. Let’s say it’s fourteen needles for a biopsy and if your toes don’t start to point toward each other, there’s something wrong with you. You think about that as an impediment to get treatment. What you’re talking about is an enabler for more folks to feel more comfortable about getting the diagnosis and getting checked. These devices, you can see there’s different colors inside. That’s your fat mimic, which you’re patenting as well.
Yes, we have submitted a utility patent on that and we’re waiting.
That will be your standardization piece. Theoretically, if every MRI was to use this on day one, once they’re done, they’d all be in at least in tolerance. You could go from MRI A to MRI B and get similar readings across the country.
That’s the idea if also you have compliance, which is making sure that they use the correct protocols. MRIs can have a whole slew of different protocols that you can use.
For the folks that don’t know what the protocol and an MRI is, what’s the protocol?
For the simple way, it’s the knobs and where you turn the knobs on the machine to be able to get that image. You have maybe ten or fifteen different knobs that you can change on that. If one site has them set to one, two, three, four, and five, and then another site has it set to five, four, three, two, one, you’re going to get a different answer. Part of it is, along with our product, is giving the sites the best possible protocol for breast imaging, for example, and the best possible protocol that you can use for prostate imaging
We talked about random data. It’s just noise and we were talking about signal to noise. For a lot of these machines, if you could get the standardization like you’re talking about, then it cuts down on the noise and you could actually tell signal. Where are most of these devices found right now?
They’re at the top research hospitals in the US and throughout the world. If you were to type in and look at News Week and see who the top hospital and research institutions are, we would pretty much be at all of those.
At this juncture, the local hospital, unless a researcher is unlikely to be employing this.
Correct, and part of that is because we don’t have the exacting protocol that we would like them to use. We are in the process of developing them for specific cases like there’s neuro imaging, imaging of the brain and there are several different types of brain ailments that are out there that you can image. Which protocol is best for which disease is yet to be determined.
We went on the brain, and we have this large realm object. That thing is probably ten pounds, twenty pounds and it looks like a very colorful death star. There’s all these spheres inside this sphere. Talk a little bit about the construct of this thing and why it is like it is.
It’s shaped the way it is because it’s meant to fit into head coils. It has fourteen different spheres of different colors. You can see green, blue, and then the red and then the yellow. Each one of those contain a different solution that gives you a different number that you should read on the MRI. There’s 42 different numbers. When you get those all correctly done, and you want to look into the brain and go beyond imaging, just the anatomy and you want to start relating numbers to a tumor that you see in there, this is what you use.
This is the calibration? These are standards and they don’t change.
It tells you how much air from the machine there is so that when you then put the human in there, you can say that my machine isn’t the one that’s causing this number to be so high or so low.
Either you can tweak the machine to get the number proper or you learn to adjust the numbers that you get after the reading from this. Folks are going to MRI and I didn’t dig into that too very far. Maybe a quick description so folks can get comfortable with what the MRI is.That knowledge of knowing whether something is working quickly is very important and crucial. Click To Tweet
At the high level, it produces an image of your body. It shows you your soft tissue. It can show you some bone structure and basically tells you this is what your body looks like on the inside. The way you get that image is by using a high magnetic field. What you’re looking at is your body is filled with water and water chemically has two different atoms in it. It has a hydrogen atom and an oxygen atom. Specifically, what MRI does is look at the hydrogen atom in your body. That hydrogen atom has a signal that it sends out when you’re healthy and it has a different signal that it sends out when you’re not healthy. It tells you what the environment locally in your body looks like. You can see the image data. “There I see a tumor in my brain.” The question is, “Is that tumor growing or is it a dying? What’s going on with that?” You can’t tell right now, but the idea is that when you assign numbers, you can see whether there’s part of the tumors dying, whether it’s growing. If you’re undergoing treatment and you want to know how well your treatment is doing, you could use MRI to stage that treatment and basically help you get feedback faster. Which is really important when you think about cancer and how quickly cancer can grow. That knowledge of knowing whether something is working quickly is very important and crucial.
We talked about going from Ph.D. type to entrepreneur. Tell me a little bit about that thought process and journey.
I spent a lot of time working in the lab. I got my Ph.D. and then proceeded to work in the lab, a small business startup in Longmont. I spent two years at the National Institute of Standards and Technology during my post-doctoral work there. One of the things that I learned as I went to NIST is that I liked the fact that it’s under the Department of Commerce. What they’re doing is related to what industry is doing. My Ph.D. thesis was on a subject matter that I don’t think I will see realized as a product in my lifetime. I wanted something more, and so when I went to NIST, I felt good about the fact that I was doing more that was industry related. Then I started to figure out that I was interested in business and making products work and making things sustainable and jobs and understanding accounting, pretty much the whole gamut.
Did you have any of that in your course pursuit?
I didn’t have any of it, which is interesting. I had thought about, “Should I go back for an MBA?” Then I realized I didn’t want to do more schooling. After Ph.D., it’s a lot of schooling. I find myself thinking creatively outside the box. I gravitate towards the startup, the initiation of an idea than taking it to product form. That’s where I thrive emotionally. That’s where I find myself being happy.
Folks have impressions of Ph.D.’s and I don’t know if Einstein comes to mind or not. You said some of your hobbies, you like to bow hunt. You like to fish. You like to climb.
Anything outdoors-ey. I have a big nature side. A lot of people would notice that. All my friends would say, “If you’re not inside, you’re outside.”
You were talking about the creative process. Maybe when you were looking at this as a problem, take us through the thought process that led to the creation of some portion of this
Back in 2007, the larger radiology community got together and said, “We need to improve what we get out of this MRI. It has so much more potential.” What we’re seeing is variability in diagnosis. We’re seeing increased repeat scans that don’t need to be done, and it’s costing the healthcare system a lot. That was the inception point of it. Being here in Boulder, Colorado, and fairly close to the National Institute of Standards and Technology, we were lucky enough to be able to do the very first prototype that was jointly designed by the larger community, the International Society for Magnetic Resonance in Medicine and the Radiological Society of North America, and to commercialize that. We learned a lot in this prototyping. You have to make this product MRI compatible. There’s no metals, and we started off as a machine shop. I’m doing a lot of stuff with metals.
In this case, it was, “How do you get plastics that are also going to hold this chemistry that’s in there for a long enough time?” Building all of that was something that we contributed to it significantly and that’s kind of where that started. Then we continued to go down. For the breast standard, we ended up applying for a small business innovation research grant. We won that. We went through phase one and phase two and we now have a product that we’re selling to the market. We’re still looking to continue our grant work through the NIH, National Institute of Health, and try to really hone in on that application and what protocols people should use, what’s the best methodology for you to implement our standards and then apply that to the patient.
The colors are not just the colors. You have all the chemical compositions and then you talked about how do you preserve the composition for durability and the plastic doesn’t affect the composition. I asked you how was this built, and you were injection molding. There is no metal. It’s all gaskets and whatnot, then machining and plastic. For a location that’s using that particular device, what’s the durability of the device? How long does it last?
We’ve had a proven shelf life of five years on this because it’s a standard where we’re interested in not trying to make it last forever. Because then it increases as a manufactured product, the likelihood that it may have a problem. Then if you’re a standards company you can’t have that problem at all. What we’re actually doing is we’re developing a software platform to be able to rapidly analyze this and keep track of the daily data and make it as easy for the MRI tech to do the analysis each day. Just plug and play, put it in, run the protocols and then get a report out that says, “You’re okay to scan today.” We were looking at packaging that up as a service and the product.
Are you going to get the data back?
Yes, it’s a server that we have hosted on the cloud. The idea would be that we would get the data back. We would know about the data.
That would be good, not only from compliance and you’re coming to the end of the life of that particular device. What’s the variability across the universe on the machinery?
When you think about big data science and companies that are trying to mine image data, even though artificial intelligence in the computers can find some answer, you should at least make sure they’re finding the exact answer on a known physical standard that’s in that.
Otherwise, the quality of your data is suspect if it’s not standard. You’ve got this large machine and you’re a radiologist or a number of practices and you get your software upgrade package which makes anybody, or everybody should cringe, because you don’t know what you have. What are some of the experiences that you’ve seen from folks doing software upgrades?
We’ve had customers call us and tell us there’s something wrong with our product. We say, “No, there’s nothing wrong with the product, send us the image data.” There’s a halo of a rectangle in the image, and we trace it back to the software and how the software was seaming together the data from the different channels that it was looking at. It literally was the software, and when we told them to click one button off to turn something off, it went away.
As you think about the acceptance in feedback from the current users, what’s the biggest misconception that your current users had of this particular technology?
The biggest misconception is that the data coming from the MRI is always the same day to day and from site to site, and it’s not.
You think about practically speaking, like the tolerance on the car motor at 10,000 miles is not the same at 100,000 miles. Then you think, “Why would you expect a piece of machinery like an MRI to be consistent day in and day out? It seems strange to me.” You’re now in fundraise. You have a broad acceptance of your technology. What’s next?
When we get that funding, it will be hiring more people to really hit the ground running on developing the applications and getting one minimum viable product out there to the diagnostic market to eliminate variability in the diagnosis. Our goal is to stop that first problem. Then our second goal is that would reduce variability in just what the anatomy looks like for those images. The ultimate goal is to get those images to have numbers associated with a much like a blood panel where you have numbers that it says are normal, and then it compares what your blood number is. If we could then take image data and say, “Here’s the anatomy, and then here’s your tumor,” and we’re showing that 80% of it is dead and 20 % of it is alive and it’s dying. We have numbers that tell us that. That’s our end goal is to enable that technology.
In the hoping dreaming space, how long down the road do you think that will be?
It’s hard to predict, but definitely in my lifetime. There’s already some people using those numbers to stage prostate cancer. We’re trying to target those people and try to make sure that our numbers are as accurate as possible and get our product to that. It could be within the next few years as well.
For the folks out there, they’re going to go, “I want to make sure I’m going to go someplace that has this technology.” Their chances of finding this technology are greater at the big research hospitals than it is in your small local.
One of our goals is to try to see if there’s a payer out there willing to work together with us and have it accepted in the entire health system. That would be the best for us because then we could get to those small hospitals faster.
You talked about the process to acceptance by the healthcare system. How do you see that transpiring? What’s their thought process? What do they need to see in order for them to go, “Yeah, we need to do this?”
They always need to see that they are not wasting money on unnecessary diagnostics, on repeat scans, on that diagnostic variability. If we can show that we reduce their costs there, if there were any healthcare costs, then that’s the driving factor.
The things that drive that insight is gathering the data, being able to take and verify that you can say, “You don’t need to do that extra one because, you don’t need to do another one because. Your machine is out of sync. If you put it in sync, it’ll reduce the number of these that you do by this.” That would seem to be apparent to me. Is it that they don’t see here or it’s not enough data yet to demonstrate that?
It’s not enough data to demonstrate that. We’re looking at pilot study on a wide health system. One of our grants that we applied for through the National Institute of Health is specifically to look at 25 different sites and accumulate some data from that. We’ve created the product. We’ve shown that it has enabled researchers to do a lot of new and exciting things. For example, reducing scan times from 45 minutes down to five minutes. Our next step is we’ve enabled the research community, so we have to enable something in the diagnostic community. Once we’re done fundraising we’re going to be going towards that application.
You said this company was spun out. Take us to that thought process between this was being put together inside, and then there was a decision to spin it out and there’s a company in here. What was that decision like?
It was basically seeing that there is a larger market for this than any other product line that our parent company, High Precision Devices had. It was also related to the medical industry, which nothing else that we do is. It requires things like ISO 13485 certification, CE marking. There’s a lot more regulation associated in that realm. We realized that we wouldn’t want to put that onerous of a task onto the entire company because it’s not necessary and it’s not cost effective. That was primary driving factor for spinning it off and also fundraising. We can have people happy that they’re not owning High Precision Devices, but they’re actually owning QalibreMD.
What are you hearing from the potential investors? What’s their typical response or commentary?
They’re excited about the technology. Typically, the investors we’ve been speaking to want us to be further along than we are, which is always the case. It’s understandable. We’re limited to the accredited Angel investor. There aren’t as many avenues that way as there are.
You’re the CEO, CFO, and everything else.
I’m not the CFO. I do a lot of the sales and marketing. I do the research and development, writing of the grants, project management of the grants. That’s plenty. I luckily have help in the manufacturing and I have dedicated staff that put these together. I have access to engineers for refinements through HPD still, and they have been a tremendously great resource. We have a slightly different business model and in terms of that is we spin it off but we can still use the resources of HPD. I still know everybody in the purchasing department. I don’t do everything, everything. I do specific to this product line stuff.
I have this idea and then all of a sudden, your ideas come to fruition. You’re sitting here day two after the spin-out. What was that thought process like day one being the CEO of Qalibre?
It didn’t hit. I kept thinking, “This is what I have to do.” I don’t think about the titles as much as I think about the tasks that need to get done.
I don't think about the titles as much as I think about the tasks that need to get done. Click To TweetHow many of these types of units are out there on the planet right now?
We have about 150 units out there, which is 150 different sites throughout the world. We’re in Europe, South America, Australia, Japan, Korea, and China.
Somebody says, “I need to know more.” What’s the best way to find you?
You can contact us through our website, which is www.Qalibre-MD.com.
It’s a pleasure and you were very generous with your time and I appreciate it.
I appreciate your time and your patience.
Thanks very much.
About Elizabeth Mirowski
QalibreMD is a privately owned Boulder-based startup company that has been in business since 2017. We are a spinoff of High Precision Devices, who for the last 25 years have been fabricating exemplary manufacturing and engineering equipment.
Here at QalibreMD we are on a mission to transform medical imaging. We enable MRI to go beyond simply diagnosing a lesion to potentially making an exact diagnosis of what disease is present, at what state, and to what extent in the body. Our impeccable imaging standards offer a universal guideline to validate the accuracy of MRI scanners. This reduces unnecessary repeat scans and variability to diagnoses from one system to another and from the same system over a period of time. Our products have been used to develop and validate new fast MRI scans, reducing scan times to a mere 5 minutes.
We pride ourselves in our dedication to providing exceptionally high quality products focused on precision and ease of implementation.