E46: Dr. Charles Roberts on the Next Era of Pediatric Cancer Care

Dr. Charles Roberts, director of the St. Jude Comprehensive Cancer Center, joins Timothy Chou to discuss how pediatric cancer research is moving beyond nonspecific toxic therapies toward targeted treatments, AI-driven discovery, immunotherapy, RNA therapeutics, and gene editing. He also explains why childhood cancer is biologically distinct from adult cancer, how St. Jude is working to improve survival globally, and why the future of pediatric medicine depends on rethinking both science and access.

This episode is brought to you by BevelCloud—powering distributed AI in healthcare and driving the Pediatric Moonshot forward. Learn more at BevelCloud.ai."

Transcript

SPEAKER_00 0:03

Good morning, good afternoon, good evening, everyone to another edition of the Pediatric Moonshot Podcast series. I'm really pleased to be joined today by Dr. Dr. Charlie Roberts. Charlie got his bachelor's degree at the University of Wisconsin, beautiful campus. His PhD and MD degrees at Wash U. He did a residency and fellowship at Boston Children's. From there, he spent 15 years at the Dana Farber Cancer Institute, where he was the Deputy Chief Scientific Director. In 2015, he went to St. Jude's, where he's an EVP and director of St. Jude's Comprehensive Cancer Center, the only NCI-designated cancer center devoted to children. Furthermore, he is the co-founder of the Pediatric Cancer Dependency Project, which has screened over 200 cancer cell lines to identify actionable targets unique to children's tumors. Welcome, Charlie.

SPEAKER_01 1:09

Thanks so much for having me, Tim.

SPEAKER_00 1:13

I gotta start with why pediatric cancer as a career?

SPEAKER_01 1:18

Yeah, that's a great question. I come from a family of scientists, but not medical scientists. And I just had real curiosity for science as I went through high school and in college. And then suddenly I had this sort of sense if I'm gonna do research, I want to have it be relevant to the human condition. And so I started doing cancer research. Then in medical school, I thought for sure I would be doing adult um oncology mixed with research. And then I hit the pediatric wards, and it was just an amazing experience seeing young children fighting for their lives and at the same time having the spirit of kids. And almost immediately I switched into pediatrics. And it's just been a fascinating and motivating career of discovery science to understand childhood cancer, you know, thinking about translation and how do we do better in treatment? And so I've I couldn't be happier with my career choice.

SPEAKER_00 2:24

Well, and uh, well, the universe thanks you for doing that. So uh could you uh I don't know that everybody, maybe everybody's familiar with St. Jude's, but I think it occupies a rather unique position in children's hospitals in the world. Could you just tell us a little bit about St. Jude's?

SPEAKER_01 2:42

Sure. Um, I think lots of people know about St. Jude, but they don't necessarily really understand um the St. Jude mission or the full breadth of what goes on. And so St. Jude's mission is to advance cures in means of prevention for pediatric catastrophic disease through research and treatment. Um, and so you'll notice the word research is right there, and we really have the motto of finding cures and saving children. We're motivated by our founder Danny Thomas's vision that no child should die in the dawn of life. And so ultimately, the goal of St. Jude is really to treat the untreatable and cure the uncurable. Um, and as soon as we make advances to share that information out there so that others can adopt it and move on to the next problem that is um preventing um success or harming children. And right from its founding, um, it was an institution that is open to all and no family ever pays for treatment, travel, housing, or food. Um, and so it is designed to make the exceptional care at St. Jude available to all. And almost all of our patients are on research protocols of one type or another that take the best current therapy that we have and are trying to look for what can do even better. Um, we've been heavily focused on childhood cancer, but are increasingly, as technologies coming along, moving into other diseases. And so we span the spectrum of fundamental discovery science from structural biology, chemical biology, computational biology, imaging sciences, all the way through translation and to a large cadre of clinical trials, most of which are investigator initiated here. So it really is a unique engine that is trying to constantly focus on pushing the envelope to do better, um, including into survivorship and increasingly major efforts around the world to improve the outcome for kids in in low and middle-income countries as well.

SPEAKER_00 5:13

Actually, could you just for a second talk about that? How, because I do, I mean, maybe as we all do, we think of St. Jude's as a US thing. So maybe talk a little about the international efforts.

SPEAKER_01 5:24

Sure. So St. Jude really has um uh been focused in the US and growing its impact. And then it really began dawning on us um really a couple of decades ago, as we had trainees come from other countries and go back, and we had what was called a twinning program. So that you know, a number of hospitals um uh around the world that were twinned with St. Jude. Um, and then about 10 or 11 years ago, just really the realization that that was not a scalable model to have impact around the globe, and also the recognition that your zip code, if you will, meaning where you're born on the planet, really is probably the biggest determinant of what your outcome is going to be for um diagnosis with childhood cancer. In the US and developed countries, we're now above 80%, and in some cases above 85% in survival for children with childhood cancer. Low and middle-income countries, it's below 20%. Um so, I mean, those are things that we already know how to do, yet the outcome is dramatically worse. And so we launched what we called um St. Jude Global, that is now developing regional networks around the world where we are teaching people how to care for kids with cancer. We've partnered with the World Health Organization in UNICEF. And so we have both a bottom-up approach of meeting doctors and hospitals and starting those systems, and a top-down approach through World Health of getting together with finance ministers of countries. Um, and for instance, a major initiative once we learned that access to drugs, to the chemotherapy drugs that are needed to cure cancer is one of the greatest limiters. Um, we started a $250 million initiative funded by St. Jude to create a buying network of the drugs that are needed to cure childhood cancer into an international consortium. World Health gets the finance ministers in, UNICEF does the drug distribution. We've already launched in five countries. Um, and so we're seeing the impact of now being able to get drugs to people in low and middle-income countries. And so just one example of how we're thinking about the global problems and starting to then move also into sickle cell disease and beginning to think about neurologic disease as well.

SPEAKER_00 8:08

And you know, you you you I've heard that statistic at 80% of the kids are curable with cancer. So uh given that amazing statistic, so what are what are you focused on now? What's the next chapter of the story?

SPEAKER_01 8:23

Yeah, so I I didn't quite say cured, what I did say survival. Now, what okay, okay. No, but it's actually a very interesting point, Tim. Back in the 60s and 70s, when we first started getting kids into remission, five years out meant you made it that far, you were cured. And that's true. You make it five years out, and you're now quite unlikely to have um cancer come back. Um, but what we have found is that the impact of the diseases um of the treatment, sorry, the impact of the treatments that we are using are causing lifelong health effects and shortened life expectancy for a number of survivors. And so that is really a key focus. And so you asked about the next chapter. I think in many ways, childhood cancer is often cited as a field that really exemplifies the power of scientific research to have on the human condition and improve outcomes. And that's um unquestionably the case. When St. Jude, just to pick a time, opened its doors in the night in 1962. Survival rates for childhood cancer were about 10%, um, and we're now up to 85%. And that was through not only at St. Jude, but many other places, the application of scientific discovery. Um But here we are in 2026. Um and these are incredible achievements, and yet cancer remains the number one cause of death by disease in the United States and developed countries. Um and the therapies that we've been using um are essentially nonspecific toxins that cause health damage um lifelong. And so I think it's really interesting to think about why historically were we so successful in childhood cancer compared to, say, sickle cell disease or inherited gene defects, other things like that, other pediatric diseases. And it's really because cancer was a disease that you didn't need to really understand the mechanism and you didn't really need to understand how the drugs worked. It turns out that deploying non-specific toxins damage cancer cells more than the healthy cells because cancer cells can't stop and arrest to repair, whereas the normal cells can. And so we succeeded, the whole field succeeded by using quite damaging toxic medications. But that has now been maximized. Um, there's going to be little additional mileage out of it, and yet we've got the number one cause of death by disease. Um, and so going forward, um uh really focused in, I think, four key areas. The first is finding vulnerabilities, and we can talk more about that, but now we have the ability to find targets that will kill cancer cells and have minimal effect on normal cells. The second is therapeutic technologies, so getting away from these nonspecific toxins. Um, we have the ability of AI to design small molecules, new classes of chemistry that allow us to degrade targets, to drug the undruggable. We can manipulate individual genes with RNA therapeutics, we can use the immune system, we can use gene editing. All of these things open the door. Now we had that success. The third is really more diseases. We were so successful in cancer with nonspecific um therapies, which did almost nothing for any other field of medicine. But now, as we move into being able to manipulate individual proteins and genes, that opens up to many more diseases. And finally, then is global impact of as we make these discoveries, we really need to bring them um to the globe, uh, to everyone everywhere to enhance access. So those are really the four things that we're focused on looking forward.

SPEAKER_00 12:36

Wow, wow. You you you started touching on AI. So, what do you see the role of AI in accelerating diagnosis and treatment of pediatric cancer? How do you see that playing out?

SPEAKER_01 12:50

Yeah, it's such, you know, I mentioned St. Jude is really about research and treatment, and so research is really core to what we do. And um at St. Jude, AI is just about everywhere. We have launched, we have data scientists all across the campus, but we've launched a major initiative in data science that's um adding um recruiting 130 more data scientists on campus. And so, in some ways, it might actually be faster to list areas that were not using AI than to tell you about where we where we are. Um, but with that in mind, I can give you a few examples. So if we think about these areas looking forward, the first is vulnerabilities. And so in order to have a therapeutic impact that will um hurt the cancer cells and not normal, we need to understand those vulnerabilities. And so St. Jude has a major collaboration with the Broad Institute, the cancer dependency map there, Dana Farber Cancer Institute, um, to bring the power of the cancer dependency map into what we're you you mentioned, the pediatric cancer dependency accelerator. And so we are searching every single gene in the genome across thousands of cancer cells, over 200 pediatric cancer cell lines, um, and looking for what will predict the vulnerability, what are those vulnerabilities, what will predict them. Can we actually use AI models to then extrapolate from the cell lines that are robustly characterized to the data we get from patients coming out so that we can really mechanistically understand and make solid predictions about what drugs can be used. In terms of therapeutic technologies, um we used to have to just screen across all drug chemicals out there to see what would happen. Now that when we find a specific target, for instance, Christoph Gorgul in structural biology here uses AI to virtually screen over 50 billion molecules in silico, um, that looks like it's going to dramatically accelerate our ability to go from vulnerability to a chemical or a drug that can bind that, including AI-mediated drug design. We're using it for CAR T cells, so the immune system, using a patient's own immune system to attack cancer. One of the challenges there is that cancers can become resistant by losing the target. And so we've used AI to engineer CAR T cells that recognize two different targets on AML cells, so it's much harder for the cancer to get away because mutating one target won't do enough. I know on your podcast you've been interested in diagnostic technologies, AI and radiology, particularly around childhood cancer. There are some things that can port over from the adult studies, but some things that don't work well. And so we're we're doing it there. And then finally, just to mention, um, as we now expand to other diseases and think about RNA therapeutics, what are the rules that will determine which RNA therapeutics are going to work well? Um and it turns out that there are very complex rules to doing that. And so AI there. And so, you know, that um just a small spectrum of things that we're thinking about, but but all of that really um critical part of our data science going forward. Um maybe I'll just mention one more imaging science. Um we have a new department of imaging science, and so some of the larger we used to think of genome data as being big data. Now genome data is trivial compared to imaging data, um, looking at cells from you know angstrom level all the way um uh on up through, you know, cryo EM and all of those things, and so understanding um medicine links and so um critical uses there in terms of understanding the the biology and vulnerabilities.

SPEAKER_00 16:56

You you touched on this, but I'm gonna uh ask you it, you know, um as is often said in pediatrics, uh kids are not just little adults. Talk a little bit about, I mean, there's a lot of research work going on in adult cancer, right? I mean, how much of that translates, how or said differently, how unique is pediatric cancer?

SPEAKER_01 17:18

Great question. So we now know through um the Pediatric Cancer Genome Project um launched at St. Jude and other sequencing endeavors that 55%, 55% of the gene mutations that drive childhood cancer are not found in adult cancer. So, on the one hand, 45% are, and so we can learn from the adults there and use drugs that target those coming down to kids, but the majority of the mutations are unique to childhood cancer. And so what it really means is from a therapeutic standpoint, the old approach of just trickle down of chemotherapy, because remember, we were talking about nonspecific drugs that are generally toxic, you can deploy them to lots of different types of cancer and they'll have similar effects. But as we're now moving into an era of targeted therapies, uh what's going on in adults is not going to be relevant in most cases to childhood cancer. And so that really changes the paradigm of how we have to think about this disease. It has implications for economic models. Um, but then as we go into, for instance, radiology that we were talking about, because we have different types of tumors, some of the algorithms and things work fine, a tumor versus not, but because the pediatric ones can look different, are occurring in developing tissues that have different um uh densities and appearances, we also need to adapt and learn how to do things even on the imaging side.

SPEAKER_00 18:52

Yeah. Could you you talked about economics a little bit? I mean, you know, we uh we like to say kids are our future, but we don't really spend that way. I mean, the amount of money spent in research in adult cancer versus pediatric cancer, I have no idea. What what do you see the future of funding in this area?

SPEAKER_01 19:14

Um, yeah, funding is is one thing um that we think uh about and you know how do we get um focused investment um from funding agencies? And I I think overall I don't see that as the biggest challenge right now because we we are reasonably successful at getting uh money from NIH in terms of grants. The real challenge going forward is as we start developing targeted therapies that will kill children's cancer but not damage kids' lifelong, the market size goes way down.

SPEAKER_00 19:55

Yeah.

SPEAKER_01 19:55

So if you are using a nonspecific drug that will help childhood cancer and adult cancer, breast cancer, colon cancer, lung cancer, no problem. Economic models are there. As you're talking about a drug that can save a child specifically and not cause damage, that market size is quite small. And the current economic models um really do not provide a pathway forward for those. And so that's one of the things as medicine now advances so that we can be targeted. This is a major problem that that um is yet to be solved, and it's not just the US. We're talking about lots of places, and it's not just childhood cancer. If one thinks about inherited neurologic defects and other things that need um really bespoke therapies, this is a major challenge going forward. And so, you know, are there going to be tax incentives or other things that can be done that would truly incentivize companies uh to go after the kids who have the brightest future ahead of them and uh decades of life ahead? So that that's a major challenge to be thought about.

SPEAKER_00 21:10

Yeah. And what does it mean for I mean, I you think probably think a lot about this more than I do, but I mean, if you're starting to create a drug for two kids in the world or whatever, how do you do clinical trial? How does that whole process work?

SPEAKER_01 21:26

Right. So so this is um, you know, we're we're often not we will actually think about N of one um kids that are there. We're thinking about smaller and smaller cohorts. Um, but one has to rethink about the science of it. And many of the regulatory things are set up to be, you know, for patients or drugs where you're going to be treating 100,000 people and you can think about control arms and other things. But let's even push where you were talking about instead of two patients, imagine one patient. Um You know, the only reason you would be doing it is because it's there's been somebody in some other gene uh in the same gene, but it's known to be sort of a fatal, serious disease. But now this mutation is different than anybody else has. Certainly, as scientists and and um people interested in health, we can't just say, oh, you're an N of one, sorry, you're gonna die. And so now the model becomes well, we sort of know what the historical outcome of this disease is as a fatal illness. It is now not only it is ethically required or compelling to take your best shot at fixing that, and that's something that parents absolutely will get behind. And so designing, for instance, antisense oligonucleotides that can um target an individual gene, and then you monitor how the patient is doing over time. And at St. Jude, for instance, uh Richard Finkel here, there was a mother who had lost a child at seven months of age to spinal muscular atrophy, and she got pregnant again and reached out at 25 weeks of pregnancy, and Dr. Finkel and the St. Jude team raced and ultimately got FDA approval to treat the mom with an SMA-altering drug while she was pregnant in order to treat the infant. Um, that infant was born and doing incredibly well. The sibling died uh at seven months of age. I think the toddler is somewhere near three years of age now and doing significantly better. And so there's an example of essentially how you can intervene even when it's a single patient.

SPEAKER_00 23:53

Yeah, wow, amazing. Let me um totally fascinating conversation. Let me kind of wrap it up by uh asking you it's uh it's 10 years from now, we're having this conversation. Uh, where do you think we're gonna be?

SPEAKER_01 24:08

You know, it's an interesting question. You know, advances happen so rapidly that, you know, discoveries can transform the way forward. So it makes it um challenging to accurately predict the future. Um, with that said, um, I think some of the things that are most likely to happen are really reflecting what I think is likely to be a transformation and how we target cancer. We're now gonna, 10 years from now, have uh many more understanding of better targets that we can go after to get away from the toxicities. We're gonna have um a much better portfolio of therapeutic technologies that get us away from toxic chemo and into many, you know, whether it be small molecules, degraders, RNA therapeutics, immune system, antibody, many more available for targeted things. Um so the immune system um will be leveraged much more. I think the um uh spread to other diseases and more and more inherited diseases. And then um really hopefully the goal um the cure rates for childhood cancer in low and middle-income countries have really started to rise in a substantial way because we're we're making these things available um to kids not only across the entire US but um across the world.

SPEAKER_00 25:37

Yeah, well, hey, uh again, thank you for taking the time. Uh thanks for the amazing work you're doing uh and joining our podcast this this week.

SPEAKER_01 25:48

Well, thanks for the invitation. It's uh it's always um energizes me to talk about how we can do better for for children that are facing catastrophic diseases.

SPEAKER_00 25:58

Well, thank you again, Dr. Roberts. Thanks, Tim.