Sarah Psutka: It's a pleasure. Thanks, Tian.
Tian Zhang: I was hoping to highlight your session on frailty here at GU ASCO 2026. Tell me a little bit about what the session was all about.
Sarah Psutka: This was really just a wonderful breakout session, and I'm so grateful to the program committee for ASCO GU that they decided to prioritize an opportunity for us to talk about frailty and geriatric oncology considerations. The cancers that we treat predominantly affect older individuals. They also predominantly affect individuals who have a fairly substantial burden of comorbidity and frailty. And I think the impact of that vulnerability at baseline on treatment decisions and outcomes makes the decision-making for a lot of our folks really complicated. This was a great opportunity. Tullika Garg and Jason Efstathiou were the co-chairs of the session and I was fortunate to participate with a number of different panelists who were highlighting different geriatric considerations that can impact decision-making, treatment receipt, treatment tolerance.
We had Dr. Cort is the pharmacologist who gave a really insightful discussion on the pharmacokinetics of a lot of the therapies we use in our older adults. And Dr. Kanami gave just a total masterclass on sort of considerations in geriatric oncology and how we actually assess frailty, perform geriatric assessments. And the necessary partnership between oncologists and geriatricians and geriatric oncologists and why that specialty is actually something that we should... We definitely need more people who are committed to it and who are doing the great job, the great work that she is doing. And then Hannah Hunter is my... Dr. Hunter is a physical medicine and rehabilitation specialist who I work with really closely at the Hutch and at University of Washington who specializes in developing these personalized prehabilitation programs. And then, I came in and chatted a little bit about sort of how we've gone from problem to solution. Once you identify vulnerability, how do you characterize it? How do you map it? How do you figure out what's driving the frailty?
And then, how do you start to think about moving the needle with these personalized supportive oncology programs? So it was a really rich discussion. We probably could have taken a lot more time than we had, but it was just a lot of engagement from the audience and great questions. And this is an area that I'm pretty passionate about. Spent a lot of time thinking about over the last 10 years or so. And it meant a lot that ASCO GU would prioritize this as a conversation piece here at the meeting in 2026.
Tian Zhang: Absolutely. Well, it's so practical, right? We have so many patients who are in this category of frail, older adults and how should we take care of them in a better way. So that's really important work. And you've actually done some research and done some AI machine learning studies. So tell us a little bit more about that.
Sarah Psutka: This is kind of one of my babies. It's been something that's been in my mind, a problem I've been chewing on for a while. So the short synopsis is started thinking about better ways to understand how you could quantify risk in patients. And we all sort of have this question that we all have to answer every time. Every single time you see a patient, you've got to say, "Fit or not fit. Can I take this patient to surgery? Can I give this patient this chemotherapeutic? Can I give this patient this targeted agent or no?" And what do we base that on? We base that on this kind of gestalt impression of fitness, this rapid calculation that an experienced physician sort of looks at a patient and says, "Yeah, I think I can get you through this treatment," or, "I don't think I can, the risks outweigh the benefit." If we don't do that well, we risk either undertreating or overtreating patients. So that's a really substantial risk. Been thinking about how do you put numbers... How do you put hardcore numbers on everything? Let's get some data. How do you operationalize the eyeball test?
So started a long time ago when I was doing my master's thesis, looking at sarcopenia and thinking about muscle wasting, and then that sort of led me into this concept of frailty, which is more of a functional assessment. It's a multidimensional assessment of vulnerability. That brought me into the world of geriatric oncology. And I feel like I've spent the last like eight years trying to learn how to be a geriatric oncologist and learning from people who are real experts in that area and trying to gain as much knowledge as I can. The comprehensive geriatric assessment is a multi-domain assessment of vulnerability versus fitness. It looks at things like burden of comorbidity, physical robustness, nutrition, but it also includes things like mental health, cognition. To do one of these correctly, it takes two hours and a dedicated geriatrician. All of our guideline bodies recommend that we actually have our older adults go through at least a geriatric screening procedure before they start making decisions about treatments, ASCO, SIOG, EAU, AUA, everybody. None of us do it.
Tian Zhang: Right. Comments.
Sarah Psutka: It's too much time, cost, we don't have enough geriatricians, we don't have access to them. We are not aware that we need to do this. So I started doing some work where I was trying to figure out how to implement CGAs in clinical practice. Long story short, I had a young investigator award from the Bladder Cancer Advocacy Network about seven years ago now, where we prospectively did CGAs on everyone who came into our bladder cancer clinic. We learned that you could do it. It did take some extra time, but we were able to streamline it pretty effectively and we collected a ton of data. And ultimately we have data on this one cohort of 67 patients where we have robust risk assessments. And we learned all kinds of things. We learned that we were underappreciating mental health concerns. There were some pretty substantial identification of previously unidentified dementia in these patients. And then, we learned a lot more about their comorbidities and the impact of those comorbidities on their life and lack of social support. So all kinds of data that as someone who's thinking about taking someone through like a cystectomy, it's really important that I know all of that.
Tian Zhang: Absolutely.
Sarah Psutka: But the biggest problem is actually a methods problem. So what do you do with all that data? And how do I quickly give a handout to my partner who's going to see this patient that says, "I just learned all of this about this patient. Here." And then, what do they do with that? So what's been in my mind that we've been working on is how do you quickly generate sort of a single figure that conveys not only amount of frailty, because just saying frail, yes, no, is really kind of like saying fit or not fit, but that doesn't tell you what's the problem, what's causing the problem, what are the drivers of frailty? And it doesn't give me something to fix. There's nothing modifiable about that. Frailty is modifiable, but only if you know what's causing it.
Tian Zhang: Sure.
Sarah Psutka: So how do you quickly convey the degree of frailty versus resourced? And we talk about vulnerability and resources because it's a little bit more emotionally and neutral language. And then how do you convey what's driving it? So I've been working with actually a machine learning... A team of scientists at Emory led by Anant Madabhushi, our lead scientist on this was Hilmi Al-Shakhshir. The team basically did some very thoughtful application of... They applied a lot of their really robust mathematical modeling skills using principal components analysis to basically take all the data we had, be able to use the different domains to predict frailty in a robust way, and then get at this communication piece and start to kind of map where patients were on this axis that looked at physical vulnerability and psychosocial and emotional vulnerability.
Tian Zhang: And then get to potentially fixing the drivers of therapy.
Sarah Psutka: Well, and we're not there yet. This was really just methods, what we presented here today, but basically showing, yeah, you can actually put this on a plot and you can start to show a physician, "This is where this patient is and here's what's driving their specific risks." So, we've got the math down. And the neat thing about it is because it's a machine-learning model, it's something that's scalable. It could be automated. So then, we've got pretty clear marching orders in terms of next steps in terms of applying this to more robust homogeneous data sets and then trying to get at looking at how we can use these different frailty phenotypes to predict specific recovery trajectories, complication profiles. Maybe even we get to the point where it's a biomarker for efficacy, that's a holy grail, but I'm going to keep... I'm thinking big.
Tian Zhang: You're going, you're chugging.
Sarah Psutka: I'm going for it. And then, a big part of what we do is fixing the problem. So I've gotten very involved in the prehabilitation space. We are currently running what's called the Get Moving Trial, which is sponsored by and funded by the Bladder Cancer Advocacy Network. We're 70% accrued on that trial. It's a perioperative exercise trial. And here, we launched... We showcased our protocol for the first time for what's called the EMPOWER trial, and that's looking at personalized home-based exercise in inactive older adults with non-muscle-invasive bladder cancer. So it's a survivorship population.
Tian Zhang: Yeah. It's fantastic. All the-
Sarah Psutka: Lots of work.
Tian Zhang: ... spectrum of your care from assessing them to understanding what is causing frailty, vulnerability to know some of the more prospective trials to fix it. It's great.
Sarah Psutka: I think hopefully we can do it pragmatically without adding burden to patients, but making it a little bit easier. We've got a great app that our engineering team has to... We work with our sports institute and they had developed this for patients who were trauma victims and patients who were dealing with amputations to help them become more active. And we've been able to work together with my partners, Cindy Lynn, Hannah Hunter, to leverage that into an oncology focused application that's linked with our EMR so we can actually start to really deliver this as part of clinical care. And so, we're validating and trying to understand and evaluate its ability to help people become more active.
Tian Zhang: Fantastic.
Sarah Psutka: We'll see how it goes.
Tian Zhang: Wonderful. Well, thank you so much for joining me.
Sarah Psutka: Thank you so much for letting me talk about this. I appreciate it.