Exploring the intersection of artificial intelligence and next-generation cancer therapies, Lisa Hatfield, who brings the perspective of living with multiple myeloma, speaks with Dr. Krina K. Patel of The University of Texas MD Anderson Cancer Center about how AI may accelerate the development and safety of CAR T-cell therapy.
Dr. Patel discusses how AI is being used to improve diagnostic precision, identify better therapeutic targets, and better understand resistance mechanisms, with the goal of delivering more personalized and durable treatment strategies. She also shares how patient participation in research is essential to advancing AI-driven innovation and offers an [ACT]IVATION tip to encourage patient engagement in shaping the future of CAR T therapy.
Related Resources
Transcript
Lisa Hatfield:
Dr. Patel, you’ve seen CAR T evolve from highly customized early generation therapies to more sophisticated platforms. Right now, there’s a lot of interest in how artificial intelligence, or AI, could help accelerate CAR T development while improving safety.
How is AI being used or explored to improve CAR T design, such as identifying better targets or reducing safety risks, and what could that mean for patients in terms of faster access to next-generation therapies?
Dr. Krina K. Patel:
Yeah, this is a great question, and I’m really excited about AI, because I’m hoping that, you know, in the last 10 years, how quickly we moved from myeloma is not immunogenic, and we’re never going to have immunotherapies to now all of our therapies becoming immunotherapy for myeloma. This will help that even faster with the next generation of treatment.
So I think, one place AI could really help us is that’s already helping a lot of people is really in pathology and radiology in medicine, and so, you know, learning to say who has active myeloma even faster, maybe before, you know, our biomarkers tell us, or for patients who have non-secretory disease, right?
We can’t follow their markers, but being able to look at imaging at a much better way to say, okay, this really looks like myeloma in the bone marrow, or this bone lesion really looks like myeloma, and not just you know, something else benign. Same thing with pathology, being able to tell are these plasma cells that look benign actually malignant? You know, we have certain testing we do right now.
But I think AI would be able to help us even more to say, who from MGUS to smoldering is going to have myeloma sooner, right? So those are places where I think, in terms of diagnostics, they will help us with CAR T itself. I think the bigger issue is going to be figuring out which targets and then, really, do we need to give therapies together? So, do we need to attack two different targets at once, or should we be doing this sequentially?
I don’t think we have enough data yet for even AI to help us, but it’s, we’re working on it with all the data we are collecting from all these different trials of patients who have had, you know, prior bispecifics or prior CAR Ts and different targets.
So again, in that resistance mechanism, what is happening when someone gets an anti-CD38? They get lenalidomide (Revlimid), they get, you know, bortezomib (Velcade), then they get a CAR T, or they got a bispecific. This is one target GPRC5D, this is another target, BCMA, now there’s FCRH5. What happens to those targets over time for different patients and different sequencing? So again, I think it’s not just for CAR T, but it’s about, you know, how does that myeloma go through? How do we kill every last myeloma cell so it never comes back? But when it does come back, what’s happening, right? So how can we be smarter about it in the way we attack it? Because today, it really is just about how my patient’s doing, what do I have available, and what seems to be the next one.
We do it based on which one looks like it has the most efficacy, right, which one has the best response rates, the longest progression-free survival, and you know, safety profile that’s okay for my patient. But I’m hoping that there’s a smarter way we can do that for different patients with their different myelomas, right? And that’s what I’m hoping that AI will eventually help us with.
Dr. Krina K. Patel:
My [ACT]IVATION tip would be, you know, the more our patients participate in research, the more it helps us actually get to where we need to for AI to help us, too.