By: Hannah Lin (CC '23)
Could you talk a bit about your background?
My position here at Columbia is professor of biomedical engineering and radiology. I’m a PhD, not an MD, so I’m a researcher. In that capacity, I’m director of the Columbia Magnetic Resonance Research Center, which spans multiple footprints. The component facilities of that center are located on the main campus with the School of Engineering and Applied Sciences; on the Manhattanville campus in the Zuckerman Mind, Brain, Behavior Institute; on the medical school campus in Radiology; at New York State Psychiatric Institute; and up at the Nathan Kline Institute, also a New York State Psychiatric research institute, in Orangeburg, about 20 miles up the river. So we have MRI systems—we call them MR systems—at those 5 sites. Actually, I shouldn’t leave out one important one: we also have an animal cancer research system at the Cancer Center on the medical school campus. In this capacity, we try to share resources and talent. It’s a construct in coordination and sharing where we share resources for research using the MR tool.
I say MR (magnetic resonance) instead of MRI (magnetic resonance imaging) because magnetic resonance imaging is an important subset of our toolkit that everyone is familiar with, but we also have magnetic resonance spectroscopy. It’s a way we can look at biochemistry and metabolism of the living body non-invasively. We also have functional magnetic resonance imaging, which is a way we look at brain activation. We can do all 3 types of imaging with magnetic resonance.
A rapidly developing new frontier is just acquiring the biological signals with MR; not even bothering to make images or something that humans can look at and quickly interpret, but just collecting the raw data and searching the new data with the newest machine learning and artificial intelligence approaches to find biomarkers that correlate biochemistry, structure, or function with symptoms or behaviors.
Can you describe your COVID-19 research at the moment?
I want to be quick to point out that in the context of this center, this COVID research is a collaboration between physicians at the medical center, specifically Andrew Einstein, Nir Uriel, and Marco Di Tullio, and here on the main campus in the School of Engineering and Applied Sciences, Andrew Laine and myself. These are the principals, but we have a collaboration of about 30 people—physicians, staff, faculty—across Columbia. The title of our research is COVID-19 Longitudinal Multi-ethnic Bioimaging Assessment of Cardiovascular Sequelae. Now that’s a mouthful—obviously the physicians thought that one up [laughs]. They worked hard to come up with an acronym. If you use the right letters in the right sequence, that adds up to COLUMBIA CARDS registry.
A lot of people think of COVID by its first symptoms, like the acute respiratory distress syndrome (ARDS). So when people have a hard time breathing, they come into the hospital, possibly requiring a ventilator to handle the respiratory problem, which comes from edema in the lungs—swelling in the lungs because they fill up with fluid, water mostly—and the lungs stiffen somewhat. That combination makes it very hard for people to breathe. When the old and the weak in particular get tired of breathing, they have to be helped out by ventilators. We don’t have a cure for this, of course, so the goal is to help somebody to breathe long enough—to stay alive long enough—for their body to create the antibodies to fight the infection. So that’s what we hear about and see mostly in the news, and that’s what we’re first confronted with by most COVID patients in distress.
But it turns out that this virus is very idiosyncratic in that it attacks every individual in a different way. In some individual on some day in some way, it attacks basically every organ or tissue. Some of these effects last well after the initial respiratory distress. For example—and most of these are anecdotal observations so far—a lot of recovered patients will walk away with myocarditis, which attenuates your heart function, your ejection fraction. They might have damage to their kidneys, they might have scarring in their lungs, they might have microemboli in their brains or throughout their vascular system. In youth as well—this is what we’re seeing -looks like Kawasaki disease. People are careful not to call it that, but they just say it closely resembles that. In the Columbia-NY Presbyterian Hospital, which is shared with Cornell-Weill, we had such a large volume of patients. We’ve seen, in addition to respiratory distress, symptoms from all of these other ailments to the heart, lung, brain, vascular systems, and other organs as well. These need to be studied—we don’t know anything about this aftermath of COVID that many patients have to live with and continue to recover from, so we need to first understand them before we can come up with therapies to treat them.
Specifically, we are in the early stages of putting together a body of research and collecting preliminary data to start applying for grants with the title I mentioned. The first aim is to assess the effect on myocardial tissue structure of COVID-19 infection in a multi-ethnic spectrum of convalescent patients (patients that are recovered) and relate structural features to patient important outcomes. We will do this with MRI and echocardiogram (ultrasound). In fact, we first screen these with ultrasound because it’s a quicker, easier, and cheaper way to triage the people we want to look at. We’ll do these studies in both healthy normal people for control and in recovered patients. They might not have been patients in the formal sense of hospitalization, but they could be like me (I’ve had COVID too), who just tested antibody positive. So we’ll image and compare the differences between 2 cohorts: one healthy normal group and one COVID antibody positive group. We’ll use a number of different MRI, PET (positron emission tomography), and echo techniques to characterize myocardial inflammation: tissue edema, increased interstitial space fibrosis, microvasculature, and ischemia. Myocarditis is a general term for those, and it basically means a viral disease of the heart and resultant heart ailments.
The second aim is to assess the effect on myocardial function of COVID-19 infection in a multi-ethnic spectrum of convalescent patients and to relate functional features to the patient important outcomes. In other words, looking at a broad spectrum of people from different ethnic groups, walks of life, races (New York is great for these kinds of populations). And you see in the news all the time where certain ethnic or racial groups are disproportionately affected. Of course, we don’t know the reasons for that yet—it could be socioeconomic, it could be physiological differences—we just don’t know yet. We can’t treat them or correct them until we understand them. So that’s why we’re always using this multi-ethnic spectrum.
The third aim is to develop a publicly available image-driven repository of COVID survivors from the unique northern Manhattan population and develop tools for long-term infrastructure sharing and computing support as well as advanced innovative deep learning based analytic tools including metadata, search and curation, labeling of findings, annotation and classification of measured cardiac function, and outcome prediction. According to The New York Times and Google, we’re the first such biomedical imaging lab fully integrated on the Google Cloud Platform. The Cloud platform gives us essentially infinite storage space and on-demand computing. Medical imaging is 90% of the data amassed from a hospital or clinic, so it requires huge amounts of archival space. It outstrips any local servers. So we really have to look to the cloud, to the future, for archiving medical image data and also for looking at that data rapidly and broadly to correlate to symptoms or anything else. To interpret that data, we also need new kinds of cloud-based tools that you hear a lot about in the machine learning or artificial intelligence realm, and Google has some of those tools, while we’ve developed others. So data is the new gold in not just research anymore, but in the economy in general. Data’s the oil, the gold, the currency that so many people are beginning to trade and, frankly, makes a lot more sense than dollars and gold bars—it’s more useful. It’s knowledge. The more data we acquire from our huge, diverse patient population, the more information we have to ask big questions about certain symptoms, therapies, or differences between racial or ethnic groups, genders, or ages. We can ask all these questions if we have enough data to answer them, so we want to create a very large data repository of COVID survivors on a cloud platform.
So those are 3 aims of a grant application that we are now beginning to collect preliminary data for to study the aftermath of COVID.
Have you faced any setbacks yet, regarding the cloud, the technology you’re using, or the data you’re collecting?
Not really. Things have gone quite smoothly. The cloud integration part went really smoothly. We’re operating now on the Google Cloud Platform, and we’ve contracted a cloud integrator, a smaller company named Flywheel—one of the vice presidents is actually one of my former students. With Flywheel and Google, things have gone very smoothly as far as setting up a cloud-based platform, and with a good part of our MR Center, we’re already fully integrated, storing our data and operating on our data. So that part is not a big factor. What we’re waiting through now—and they’re necessary obstacles—is just getting started with new research in the middle of the COVID crisis, because at once, we’re trying to limit the number of investigators in any building or lab at any one time, we’re trying to do only essential studies to limit both the number of people occupying any space at any one time. And then for the people in those environments, trying to protect them, figuring out what kind of masks, gowns, and where to come up with those with resources that are quite limited, how to test people on the front end and which test to use, getting institutional review board approval for these studies, getting seed grants to pursue these studies with, buying enough imaging time to collect the preliminary data we need to submit grant applications. All of these are just standard elements of protocol in getting going on a body of research and that’s what we’re going through now. It’s just what you do to get from A to B.
What is the general timeline of your research? I know you said it’s longitudinal, so is it going to be going on for the foreseeable future?
Yes. We’re trying to focus on getting this first grant application out, focused on cardiac (heart) ailments because a lot of these patients aren’t dying from respiratory distress, they’re dying from heart attacks. So our first focus is just to look at the heart, but we have to answer the same kinds of questions and look at similar ailments and problems for the other organs in the body. We’re already discussing 2 additional grants in different directions from this one.
One would be on the engineering side, setting up a toolkit of more methods and technologies to use to better image the heart, brain, or other organs affected, so there are different ways to stimulate the signal and receive the signal response from an MR system. There are different kinds of technology—we have antenna-like devices that transmit a stimulus signal and receive a signal response from what we call coils. There are new coils that need to be developed to target some of these specific organs, looking for specific markers. Then there are specific protocols—radio frequency, excitation protocols that we use to stimulate signal responses—that we need to program. There are also other methods by which we acquire the data and interpret the data. Setting up the cloud platform, organizing it in the right files and folders, making it HIPAA compliant because all of the public health information we’re collecting has to be hyper-secure, and sharing and securing it the way we want to share and secure it. Just trying to organize this project in the cloud and figuring out how we’re going to manage the massive amounts of data, which we refer to as data curation. And then how we’re going to interpret the data, so specific approaches using machine learning techniques to look for biomarkers in the data that correspond to symptoms that we observe in the lab or clinic. That would be a more engineering intensive grant focused on tools and methods to apply to investigating the aftermath of COVID in recovered patients.
Another more clinical direction would be looking at other organs. Not just the disease pathologies themselves, but also how they affect our lives. If we’re looking at the brain for example, we’re seeing a lot of younger patients, some of them not even symptomatic, having essentially mini strokes. We look at their brains where there are what we call microemboli, which basically means mini strokes throughout their cerebrum that can affect those afflicted in very subtle ways, like behavioral changes. People are concerned—we can definitely see, can measure the mini strokes, and strokes in the wrong place in the wrong way at the wrong time to the wrong extent can affect behavior, memory, or other body functions. So all these things have to be investigated. We have the imaging tools—MR is a wonderful one-stop-shop imaging tool in that you can look at structure, like anatomy, to high degrees of resolution; you can look at biochemistry and metabolism; you can look at physiology and function. All of this can be done non-invasively, without interrupting the system in any way on an outpatient basis. We can study large populations of people easily this way, which we plan to do. We’ll start with the heart, and that will help us gain experience in looking at recovered COVID patients. From that experience, we’ll know better what tools we need to develop, both hardware and software, and that will lead to studying other organs. The next organ we will be studying will be the brain. In quick succession we’ll look at the lungs. Graham Barr, a very famous investigator here, has a huge grant for lung studies. We’ve got the tools to look carefully at the ravages that COVID is doing to our bodies. As I said, COVID affects everybody in so many different ways, so it’ll take a lot of work to get our hands and brains around this.
So as you focus on this new COVID-19 research, are you still keeping prior projects going?
Oh yeah. For MR research, we’ve had to handle this period in 2 ways. We have a lot of research going; across the Columbia campus we’ve got over 120 investigators using MRI or MR in some capacity, supported with over 120 million dollars in NIH funding. Most bodies of research had to be paused just to limit the number of people coming and going in all the buildings and labs on campus.
But there are a few studies deemed essential, such as longitudinal studies of disease progression. For example, one of them is Alzheimer’s and another is Huntington’s disease. Unfortunately, we can’t pause the progression of these diseases in patients; therefore, we can’t pause the experiments, imaging them to document their disease progressions. The Institutional Review Board for Columbia decides what research does get done and how it gets done. We’ve continued a small number of deemed essential studies through the COVID period carefully, regulating the number of people involved and how they’re escorted in and out of the facilities, masks, sanitizing all the surfaces before and after. So yes, we have continued a small amount of our “essential” research, but we've had to attenuate most of it, or just put it on pause, anything that we could.
You said you had COVID-19. Can you talk about that experience?
Sure. I was actually one of the earliest cases. I remember very specifically where, when, and how I got it. I was sitting on a plane out of LaGuardia, back in the cheap section, where everyone has to sit on everybody’s laps. I thought I was going to get lucky on a very crowded plane—I had a seat empty next to me, but at the last minute, this guy got on who had a very characteristic dry hack. And the whole time he was sitting there, he was coughing and had to keep getting up to go use the bathroom and having me hold his coffee cup. So I knew I was going to get sick from something then. But at the time, frankly, it was so early—February 14th—that I thought it was a cold or flu. You know, this was back when everybody was calling this the “Democratic hoax.” Clearly, it was already alive and well in New York before February 14th; I’m proof of that.
I was going to a weekend of collaboration at the University of Minnesota, and that went fine. I got home Monday night, and Tuesday morning I woke up with all the characteristic symptoms: headache, a very nonproductive dry cough, all the stomach ailments. It just felt like I’d been run over by a bus, my whole body ached. I was just thinking it was a nasty case of the flu the whole time. And I still should have stayed home, but very foolishly, I had all this super important work that needed to get done in my mind, even though nothing is as important as people’s health, so I’m faulting myself, necessarily. I continued to try to work through this thing, but I just felt awful.
Coughing was the worst part. Everybody’s affected in different ways; to me, coughing was the worst part because I couldn’t go to sleep at night, my chest and diaphragm ached everytime I coughed, it was just painful. So I stayed doped up on ibuprofen, aspirin, everything else. There was some rumor going around for a while to “not take ibuprofen when you have COVID, because everybody that dies with COVID is full of ibuprofen” like there was some correlation. I’m like, the correlation is that if you’re dying of COVID, you need pain relief [laughs].
Anyway, it was pretty bad, and I felt just rotten for about 10 days. My wife, also a professor at Columbia, caught it from me very quickly. Strangely, my daughter living with us, who’s a graduate student at Columbia, didn’t catch it from us and she’s still antibody negative. I don’t know how that happened, since she lives with us and we all live in a tiny Columbia apartment together. But 2 or 3 people in my environment at work ended up catching it, one way or another, from me or from someone else, I don’t know. But it lingered for about 6 weeks. Looking back, it was very textbook. It’s just that when I caught it, not much was known about the symptoms; we didn’t know the difference between this or flu. And I did get antibody tested as part of the healthcare providers. I got an early antibody test and it was positive, so a positive antibody test together with all these obvious symptoms pretty surely means I had it. Anyway, I asked all of my students if any of them had it. Of course, none of them had been tested, but none of them had any symptoms more than a 3 day sniffle or something.
Last question here: what is your perspective on the future—I know you talked a bit about this regarding your research, but also the broader impacts of COVID-19 on science in general, and society as well?
When this first started, I was traveling to China a lot up until that point. I had been working there a lot. On my Facebook page (another use of the cloud by the way, since I use it to store my massive quantities of travel photos), I had a blog going. Mid-January, just from what I gleaned from the press and the talk, I said there’s this new virus that has every making of a pandemic. I don’t know how I knew that in mid-January and our president didn’t know that until mid-March, but whatever [laughs].
We had plenty of warning. I said there’s this new virus coming out of Wuhan; people don’t know whether it’s coming from a dead fish or a live mammal (a bat) or a snake. At the time they were saying it was a type of snake called a krait or cobra. So they don’t know whether it’s coming from a reptile or a dead fish or a live bat. If scientists can’t recognize this thing, I know our body isn’t going to recognize it, and if our bodies can’t recognize it, this is pandemic material. Sure enough, here we are.
At the beginning of February, I noted in my blog the World Health Organization working together with scientists in China and scientists around the world, including the US. I was very impressed with the way the world, together with the World Health Organization, was working together across borders to get a handle on this virus very quickly. I saw real cooperation and I commented on that in a very hopeful way, because on so many issues, the world doesn’t work together and can’t cooperate. Countries like South Korea have become a model of cooperation and positive action. I know people fault China for getting a slow start and trying to control the PR on this thing just like our own country did, but they did respond strongly and correctly and were able to stop this thing in its tracks in a very largely populated part of the world.
So I did see real models of cooperation, following science and medicine and accordingly, and getting positive results from that. We do have recorded examples of what working together across borders in science and medicine can accomplish for all of us. Unfortunately, as you see, especially in our own country, this cooperation rapidly devolved and (I’m happy to be on public record for this) because of our lousy leadership, we see an example of the human tragedy that bad leadership, a lack of coordinated government response, and a lack of international collaboration in goodwill can cause. So this disease has produced examples of both—the best and the worst of world cooperation in science and medicine to confront a world problem. I think we’ll be able to look back at COVID collectively and learn very important lessons across the broad spectrum of what to do and what not to do. I hope we will learn.