Improving and Expediting Care for Stroke Patients with Artificial Intelligence

Dr. Corinn Cross (Host): Advancing Health Personalizing Care. At Memorial Herman, this is our mission. This podcast shares the science and stories behind those efforts. Welcome. I'm Dr. Corinn Cross. Today's going to be a really interesting episode because we're going to discuss ai, artificial intelligence, and how it's being incorporated into medicine at Memorial Hermann. Memorial Hermann affiliated physicians with UTHealth Houston Neurosciences recently implemented Viz.ai, an artificial intelligence driven software platform with the purpose of improving and expediting care for stroke patients. Today I'm interviewing Dr. Sunil Sheth, who was instrumental in bringing vis AI to Memorial Hermann and has been involved in the project from conception through funding planning and implementation. He's an associate professor of neurology at McGovern Medical School at UTHealth Houston, and director of the Vascular Neurology Program at UTHealth Houston Neurosciences. Thank you for joining us today.

Dr. Sheth: Thank you, Dr. Cross. Great to be here.

Host: So Vis.aidescribes their platform as AI powered intelligence care coordination, and says it enables health systems to connect multidisciplinary care teams earlier to coordinate care and improve outcomes for patients. But let's just back up a second. What are the two types of strokes and why is quick care so necessary?

Dr. Sheth: This is a great question. I think one that everyone should be aware of in the community. There's two types of strokes. One we call ischemic and the other we call hemorrhagic. But to simplify that in one type of stroke, which you call ischemic, the problem is there's not enough blood flow going to certain parts of the brain. Usually it's because of blood vessel has been blocked and that part of the brain that's supposed to get blood, no longer gets that blood and starts to die over time. Every minute that goes by about 2 million brain cells are dying. So every second that you can speed up care the better it is for patients.

The second type of stroke is what we call hemorrhagic. This is when a blood vessel bursts, so rather than blocking the blood vessel, it explodes and blood leaks into the brain or into the spaces around the brain. This is another critical condition where rapid evaluation and rapid treatment will improve patient outcomes. So for both types of strokes, every minute matters, and we use the phrase time is brain frequently for that reason.

Host: Wow. I didn't realize that it was that quick that the cells really just started to deteriorate.

Dr. Sheth: That's the number that we use, that 2 million neurons per minute. It's an estimate that's been out in the literature for some time, but whether it's true or not, it certainly emphasizes the fact that brain is dying very, very quickly.

Host: Absolutely. So once identified, what are the treatment options for patients with these strokes or large vessel occlusions?

Dr. Sheth: So one of the good things is that in the last, about five or seven years, we now have treatments that we can deliver to patients with the worst type of ischemic stroke. So that's what we call these large vessel occlusion strokes, where one of the main blood vessels in the brain gets blocked. So for those patients, we now have Endovascular procedures, which are minimally invasive treatments. We put in small catheters or tubes through the blood vessels of the leg or the wrist.

Navigate them into the brain, grab the blood clot and pull it out. Mechanically remove that offending agent and restore the blood flow to the brain. It's something that we've shown in, now many clinical trials that can dramatically improve patient outcomes. I mean, going from severe disability, you can't speak, can't move, can't walk, can't see. Within, even just hours of the procedures, patients can be sometimes back to normal. So it's been an incredible achievement in the last few years to develop these procedures.

But again, they are very time sensitive. The faster you do them, the faster you can open up that blood vessel on patients that are suffering from stroke, the better the chance they're going to benefit. On top of that, we still have medical treatments that we've been administering for decades now, and primarily those are thrombolysis, so intravenous medications, that go through the body and break up blood clots in the brain. So no procedure involved, just the medication. But again, these are also very time sensitive and so the faster we can administer them, the better the chance of having as little disability as possible from the stroke.

Host: I mean that makes sense. You've got to get blood to the brain and the longer you wait, the more neurons you're going to lose. So what are the time goals that you have? When you see a patient, there must be some sort of gold standard as to where's your window?

Dr. Sheth: Yes. So we have time goals and those are established by a few different, regulatory agencies for when we're in the hospital. So we're in the hospital we try to administer that IV medication within, at least an hour or at say at most an hour of patient arrival. And the same with the thrombectomy procedure, the endovascular procedure, we try to do that within 90 minutes of the patient or even to the hospital. But the reality is that these are really upper limits of where we'd like to be. The faster, the better, and we already have data that if we can administer the thrombolysis agent within an hour of the stroke starting, sometimes we don't even need that in the vascular procedure because that medication will take care of it. And similarly, if we can open up blood vessels within a few hours of them becoming blocked, there's just such a higher percentage of those patients that do well compared to the ones where we treat them, in later time windows.

Host: That makes sense. So then in your experience, I mean, I work at a hospital, I know that there are tons of inefficiencies when a patient comes in where we lose valuable minutes. And in your scenario minutes are brain. So where are some of these inefficiencies in treating patients that are, having an acute stroke and how can technology, specifically this AI technology, the vis ai, help to address them?

Dr. Sheth: Yeah, that's a great question and something we think about regularly. And this was the reason that we brought in this technology because we saw a gap that this could really address, so, using an analogy from cardiology, the same idea as when someone comes to the hospital having a heart attack. It's because one of the blood vessels in the heart got blocked. And they try to take him to the procedure suite and treat them as fast as possible. But what makes our situation a little more complicated is that our diagnostic, modalities are more involved.

There are many more people involved, many different groups involved, and a lot more expertise that has to go into that decision around treatment. Just to give you an example, a patient, will get evaluated by the pre-hospital team, by the emergency medical services crew will come to the hospital. So there'll be an emergency room team that sees that patient. That patient then has to go to imaging, brain imaging, so radiology and radiology techs, and the nurses, in both those areas. Then a neurology specialist will see them, evaluate them for possible thrombolysis.

Then also do an additional imaging study to see if they may need a treatment with the thrombectomy to look at the blood vessels to try and diagnose the large vessel occlusion. After that, we contact the interventionalist and their group. Then we involve anesthesia, and the OR staff and the angio suite staff. So there's many different processes that need to happen. And when they happen in series, like I just outlined, in this, example, it happens very slowly because if you wait for the first step to finish before you start the second, you're going to lose time.

And so, one of the advantages that this platform offered us is that we could move multiple processes in parallel. So, as a part of the application we have a secure chatting and secure image viewing. So on that one platform, all these different groups that I mentioned, can be in one place where they see the images, they can talk about the case and any kind of patient information that needs to be relayed. So the different groups can all be in one place.

Host: So basically you're cutting out the residents. Because having gone through residency, that's basically your job is to try to get everybody coordinated. Look at these images, what do you think? What are we doing? So that makes a lot of sense because, you're right, it's a lot of telephone tag when you are trying to move things quickly at the hospital.

Dr. Sheth: I think we're more modernizing the role of the resident. So rather than having them use outdated pagers and cell phones, to call one after the other, we can use this app where it's all done in parallel.

Host: And that's wonderful that you can actually not only just communicate, but you can literally look at the images, which is so important.

Dr. Sheth: I also do these endovascular procedures, and that was one of the bottlenecks for me is if I'm not sitting next to a computer, which is a substantial portion of my day, I would have to run and find one to log in and look at these images. But now I could be anywhere and I can just view these on my phone. And it's really an amazingone for quality of life. But two, it does accelerate care. There's no question.

Host: That's a wonderful system. And is that proprietary to Memorial Hermann?

Dr. Sheth: So this is a vendor. They contract with a number of different hospitals, throughout the country and maybe throughout the world. I'm not sure to be honest. But it's not unique to Hermann, although what is unique is the way that we implemented it. So rather than implementing it at just a couple of the hospitals that perform these vascular procedures, we implemented it at every single Hermann site throughout the greater Houston area. So that's the hospitals, mainly the comprehensive stroke centers. Also the primary stroke centers, that's 11 of those. In addition, we also put it in the freestanding ERs we call them community care centers.

Host: So then Memorial Hermann has this network of hospitals and care centers. Are all of them able to provide the same care, or do you sometimes need to transfer patients between centers?

Dr. Sheth: We have hospitals throughout the Greater Houston area, and we want to make sure that we have healthcare facilities for the entire population of the city so we can reach them, provide the care that we need. Now, not all of these centers need to be what we call comprehensive stroke centers, where they provide, 24/7 endovascular procedures and neuro ICU. But we do need to have hospitals that can evaluate stroke patients and administer at least the medication, the thrombolysis, intravenously.

So the way we set this up is, again, all these hospitals have access to the same software and use it. So any stroke patient that shows up at any of these Hermann hospitals or freestanding emergency rooms will get evaluated in the same way, using this artificial intelligence software. Then using that same software, we can accelerate the way that patients can get transferred between these hospitals. So suppose a patient who ends up having one of these large vessel occlusion strokes that needs endovascular procedures, if they present to a hospital that doesn't have that capability.

We already know everything about that patient, throughout the system. And that patient can get transferred very rapidly, to the nearest comprehensive stroke centers, that can perform that treatment. And again, these comprehensive stroke centers are distributed throughout the city so that they can reach the population that most effectively.

Host: That makes a lot of sense. So how is Vis.ai advancing care for stroke patients?

Dr. Sheth: So we implemented the software really with patient care in mind. As we went through the time is such a critical feature in how patients do when they're having a stroke. We're also an academic institution here at UTHealth Houston, and so we wanted to study that question in a lot of detail to show that, efforts we were putting in to implementing the software and using it, were effective towards improving patient outcomes in patient care. And we seem to have found that that it is the case.

So in looking how we're doing before and after we implanted the software, we found that at our comprehensive stroke centers where we perform these treatments, we've been able to accelerate the time from when the patient arrives to when we do the procedure, by over 15 minutes. So again, every minute matters, 2 million brain cells a minute. And then at the non-comprehensive stroke centers we call the primary stroke centers, we are also accelerating care.

So, in terms of improving patient care, I think one of the things we've already seen, in just the short time I've implemented it, we started this in April of last year, 2021. And just over a year and a half, we've already seen a substantial reduction in treatment times, which is incredible.

Host: Yeah, when you're talking about needing to do either the treatment within an hour or 90 minutes, as you were saying, I would say 15 minutes and 30 minutes. Those are huge percentages of the time that you're, trying to use as your goal. So that's amazing. And you actually haven't even been using the software that long, so I'm sure it will probably get faster the more you use it.

Dr. Sheth: Absolutely, that's the goal. As, we get more and more used to it, and we start relying on it for communication even more, I think it's going to continue to improve our process.

Host: And in addition to getting patients the treatment faster, have you seen an impact on stroke patient outcomes, like how they're actually doing?

Dr. Sheth: So that is pending. We're running that analysis now and, we'll hopefully be presenting this at a few of our big national meetings. So we have a meeting in November in LA next month where we'll be presenting some of these data. And then another big one, which is the International Stroke Conference in February of 23, and our goal is to have those data ready for you guys then.

Host: I look forward to hearing about it. One last question. Can you explain to our listeners what can comprehensive stroke centers provide versus primary stroke centers in terms of treatment?

Dr. Sheth: So we have, as you described, two types of stroke centers throughout the city. The primary stroke centers, which can receive patients who have stroke, evaluate them very quickly, perform neuro imaging very quickly, and administer the medical treatments, including thrombolysis, the intravenous medication to breakup blood clots, in a very effective manner. The comprehensive stroke centers can do those things and in addition, they have regular neurosurgery coverage and NEUROINTERVENTIONAL coverage.

So the interventional neurologists and neurosurgeons that perform those procedures are also available 24/7. So those, the several, are the differences between the two centers, but either way, both will have access to the same artificial intelligence software and in our case, same physician group, and same expertise.

Host: Well, Dr. Sheth, this has been fascinating. Thank you so much for taking your time, and sharing your expertise with us today. I've learned a lot, and I'm sure our listeners have too, and what you're doing there at Memorial Hermann is phenomenal for patients with stroke. I'm sure that your studies in November are going to show that the outcomes are greatly affected by the reduced wait time and getting the right treatment. So thank you for providing such amazing care to the residents of Houston.

To learn more about stroke outcomes at Memorial Hermann, please visit memorialhermann.org/stroke-outcomes. And that's Memorial Hermann with two n's. That concludes this episode of Advance, the podcast series for Memorial Hermann. If you found this podcast helpful, please share it on your social media channels and be sure to check out the entire podcast library for other topics that might be of interest to you. Please remember to subscribe, rate, and review this podcast. Thank you for listening to this episode of Advance. I'm your host, Dr. Corinn Cross.