New ARPA‑H effort aims to change how doctors understand and treat critical illness in real time

Terry Gerton All right, well, ARPA-H has issued a new solicitation for the CIRCLE program. Before I invite you to tell me what CIRCLE is, I want to first talk about the medical conditions that it’s trying to address. What is happening in ICUs?

Yoram Vodovotz What is happening is, we call it critical illness. And a lot of people don’t really understand that term because it doesn’t seem to convey a lot of meaning. To make it maybe perhaps a little easier for the audience, think of sepsis, which is the one very core reason that the patients end up in critical illness in an ICU. So there you have an infection. The infection kind of gets out of control. The immune system revs up and drives damage to organs. And then that process requires support for those organs because they just don’t function properly. So for example, you need support for your lungs via mechanical ventilation or support for your kidneys via dialysis or other forms of support. And this happens in the Intensive Care Unit. And then unfortunately, well over 4 and a half million people in America have this happen to them through a variety of causes, and then anywhere from, depending on the specific situation, anywhere from 10 to 30% or more of those people can succumb to that process and die. And then importantly, even if you do sort of okay and you get through that ICU stay, you experience elevated morbidity and mortality for months to years after that ICU state. So sepsis is like the kind of prototypical cause that people think of when they think of a critical illness. But if you have a severe injury in a car crash, for example, and you survive that injury, you will end up in a state of critical illness as well because the support that’s needed will land you in an ICU. And then the process that’s very similar outwardly to what happens in sepsis will happen there. And it can also happen for people that, for an example, have a complication of heart failure or some other type of organ failure due to chronic disease. That also brings them in. And this was what was confusing for clinicians over the years, is that how all these different, seemingly very different causes could end up in a syndrome that has some very similar manifestations. And that, I think, was the big kind of a-ha moment quite some years back, that really, this is not just about organs that fail, it’s about why is it that these organs are failing? What is the process that’s driving that? And that’s what we’re trying to address in CIRCLE.

Terry Gerton So explain to us then how CIRCLE would drive a different approach to treatment or therapy and recovery.

Yoram Vodovotz So let’s take kind of a step back as to how that pathology sort of evolves and it evolves very rapidly. You have that entry point. So you have an infection or you have a severe injury or you some rapid complication, a severe complication of an underlying chronic disease. You have this very strong inflammatory and immune response that drives organ dysfunction. And most people view that as a kind of linear process. You know, one leads to the other, which leads to other. But what we started to understand about 25 years ago is that it’s really a feed forward process. You can think of it as inflammation to organ dysfunction to more inflammation. And this revs up very rapidly, very non-linearly, which kind of practically means that by the time physicians are getting a signal for an organ that is heading towards failure and they begin to treat that organ, or a second or a third or a fourth organ is already on the way to failure, which means that inevitably they are late to that situation and they’re aware of it. This is not something that they kind of don’t understand. They just don’t have a means of being able to predict that well enough in advance and to treat it well enough in advance. So the ICU was a huge step forward for medicine when it sort of came about in the roughly 1960s. Because it brought together all of the necessary pieces to care for someone intensively, as the name says, in one place. And of course it was a huge improvement over what there was before, and so then it was tremendous advance. But then really it was left up to iterative kind of evolutionary improvement that has sort of plateaued out. And the clinicians are trying to just get earlier and earlier indications that organs are failing and then try to treat those failing organs. But what I just described to you as the underlying process means that just simply knowing when an organ is going to fail and then trying to rush to treat it is insufficient because you never address the mechanism that leads it to fail, right? And they have very limited tools for doing that. They’re aware that the runaway inflammatory response is involved, but they have very, very few tools and they have very few appropriate diagnostics. And what we’re trying to do in CIRCLE is kind of break that log jam, sort of cut through that problem. And the way we are thinking of doing that is to substitute what is now a vicious cycle of this kind of inflammation to organ dysfunction to more inflammation, with what I’d like to call a virtuous cycle of measurement, modeling, and modulation. And what that means is, think if you could have a virtual version of the patient, and this is a virtual computational model, but it’s essentially a virtual patient, it’s not just a bunch of data that you sort of try to use machine learning on. It is an actual virtual patient. It’s an abstraction of a real patient. But imagine that you have, whether it’s through equations or computer programming, you have the relevant immune biology, the relevant organ physiology that plays a role in critical illness. So now then you feed that with data. That are appropriate for the scale and scope of that digital twin, as we’d called it. And now these data are being fed in regularly and this model is updating and it’s making rapid predictions about the state of the patient. So you think of it, that’s great, that’s a diagnostic, it’s wonderful. But really it’s more than a diagnostic because it’s also predicting for you, the clinician, therapy options. It says, based on what I’m seeing, and the evolving trajectory of this patient, you may want to consider option one, two, or three, or even a combination of those options for therapy. And so that’s the first two parts of the program. And that’s already a major advance, but the key part, of course, is testing that that’s true. So the third part that closes the loop is the validation part, the modulation part, whether it’s experimentally or using selected kind of clinical studies. The idea is to take those inferred control points that come from the digital twin model that is fed by the data from the patient and testing them during the research and development phase of the program to make sure that they’re actually correct. And then that is, so that’s the core structure of the program. And I can talk more about some of the ancillary parts of the programs that are also really important for solving other related problems in the field to make all of this into a reality.

Terry Gerton I’m speaking with Dr. Yoram Vodovotz. He’s the program manager for the Resilient Systems Office at ARPA-H. Well, let’s take that a little bit farther. What steps are you building on now in both the regulatory side and the program development side so that a tool like CIRCLE could move from the lab actually into the ICU room?

Yoram Vodovotz That’s a fantastic question. We’ve been working diligently with potential partners within the federal government to unlock data sets that would hopefully help during the R&D phase to help build these models and to calibrate them and to validate them, and also to expand their reach. And we’re also working, in fact, we just got off of a discussion with the FDA on the whole digital twin side of it, because that’s an emerging technology and I think our program may be the first one that will really, really stress test the digital twin paradigm because of the ICU use case. So, digital twins are being used in various industries outside of health and also in some cases in the health care setting but, usually, at least in the healthcare setting, the time horizon for the predictions is a bit further out in time. In the ICU, you’re looking to get insights within minutes to hours to a day at most, because that’s kind of the decision time frame that makes the difference. Trajectories that are set in motion, correctly or incorrectly, in the beginning of that ICU time period, kind of get set in stone and require ever more heroic efforts the further out in time you go. And we’re not talking about further out in time like weeks. We’re talking about further out in time like a few days. And really, that’s kind of how all complex systems are. They are very dependent on those initial conditions. They’re very set in motion and modulated by their feedbacks. And once they reach that kind of terminal velocity, where they’re really going, they’re very, very difficult to retarget, to reorient. And so the idea is that you want to be able to get actionable information and act on it as soon as possible. And that’s what CIRCLE is trying to bring to bear.