Exclusive Interview: NCI Director Talks Big Data, Clinical Trials, the Cancer Research Workforce—and Why He Lives to Conquer Cancer

Subscribe to the podcast through iTunes and Google Play. 

 

Dr. Clifford A. Hudis: Welcome to this ASCO in Action podcast. This is ASCO's podcast series where we explore policy and practice issues that impact oncologists, the entire cancer care delivery team, and the individuals we care for-- people with cancer. My name is Clifford Hudis, and I'm the CEO of ASCO as well as the host of the ASCO in Action podcast series. For today's podcast, I am delighted to have as my guest Dr. Ned Sharpless, the director of the National Cancer Institute.

The NCI is the largest funder of cancer research in the world, and it has helped to drive many of the major prevention and treatment advances we've seen over the past 50 years. This includes things like HPV vaccination and the identification of the link between HER2 status and breast cancer outcomes and treatment, as well as new discoveries that have dramatically improved outcomes for childhood cancer.

Dr. Sharpless, welcome, and thank you for joining me today. Now, we really have a whole lot to discuss, but before we get to our planned topics, I have to jump ahead and start with the president's State of the Union address, when President Trump mentioned that he wants to see $500 million appropriated for childhood cancers over the next decade.

Can you talk a little bit about how you expect that, specifically, to play out? What will the NCI be able to do with those new specified funds for pediatric research?

Dr. Ned Sharpless: Sure. I think childhood cancer-- childhood cancer is an area where the National Cancer Institute has had a long interest and a robust portfolio of research. And I think it is an area where we've made some progress, in terms of mortality, over the last few decades. But you have to say two things about childhood cancer. While progress has been good, and we're making-- more kids are surviving cancer therapy today than ever-- there's still a long way to go. Too many kids dying of cancer in the United States, and even the kids that we're able to cure have these significant lifelong survivorship challenges, in some cases.

So the therapy that is curative may leave patients with side effects of surgery and chemotherapy and radiation for the rest of their lives. So better treatments for kids and less toxic treatments for kids are what we are really looking for. And with that amount of money, I think a good-- the thing that it appears to me that one could do to most quickly move the needle in childhood cancer-- which, as you know, is a collection of less common cancers, even rare cancers-- is really a more intentional effort at aggregating and using and linking clinical data with molecular data and other sorts of patient data, so that we can really learn from every child with cancer in the United States, so that we can really figure out what's working in certain populations and then disseminate that information as rapidly as possible-- without having, in all cases, to rely on slower clinical trial structures that are challenged for certain populations where accrual can be difficult.

So I think that is the vision for the president's initiative, is to, with additional funding, allow for very aggressive, intentional, and organized data linkages and data aggregation so that we can learn from every trial and therefore treat every child's cancer in a better, more effective way.

Dr. Clifford A. Hudis: You know, I think that's great. And that actually provides two different segues-- one I'm going to pick up right now, and one I want to come back to. The first is about data-- big data, specifically. We'll come back to that. The second is about the way the pediatric oncology community for years has really led in designing studies that could accrue the majority of children diagnosed with various specific diseases.

And that leads me, that idea of eligibility and the structure of research, to ask about the way that you're thinking about modernizing clinical trials. This is something I know you wrote about in JAMA Viewpoint in the last couple of months. You addressed financial pressures, the need to increase overall rates of accrual to the trials, especially representing patients from underserved populations. Can you expand a little bit on that effort and what kind of progress you see as possible in the coming months and years?

Dr. Ned Sharpless: Yeah, I think everything we do successfully in cancer today is in some ways the results of a clinical trial. And this is clearly one of the most important things the NCI does, in terms of moving basic science into patient care through experimental clinical trials. And it's an area where we-- frankly, a lot's changed in the last couple of decades. When I was a wee fellow, the clinical trials apparatus was very different from the way it is 20 years later today. And we need to make sure that we modernize the clinical trials process to keep up with the changes in our understanding of the biology and the new kinds of therapy we have for cancer.

So that brings up a bunch of items that are areas where the NCI is really doing a lot of things. So, for example, one of the first problems I noticed when coming to the National Cancer Institute was that the clinical trials infrastructure, the big networks that we have for doing these kinds of trials, were under-resourced, that they had a funding problem. And they were becoming non-competitive with the trials sponsored by industry.

And this showed itself in many ways, in accrual fees for patients, or the wait times to get the trial open, or the slow accrual once the trial was open. And so they were laboring under a number of problems. And so we decided we had to invest in the Clinical Trials Network and have been doing that and will be continuing to doing that in a number of ways-- through direct funding to attempt something like the National Clinical Trials Network or the NCORP, for example, the NCORP organization, but also by additional funding for biobanks and data aggregation initiatives, targeted clinical trials, et cetera.

I think we've also-- there are some structural problems with the clinical trials that you alluded to. For example, eligibility criteria, I think, hadn't really kept pace with modern clinical trials. And I think ASCO and other groups have played a really important leadership role in identifying what are good eligibility criteria and which ones are not as necessary anymore.

And then, do we have to have the same criteria in all the trials, and be more thoughtful about how those are used as a way to enhance accrual, because often we have a-- superfluous eligibility criteria can limit accrual. And increasing accrual by a variety of measures is really important. And we've thought a lot about how to do this through novel ways of clinical [? house ?] matching.

I think one of the more successful efforts we've had in clinical trials accrual recently has been the MATCH trial, the NCI MATCH trial, which was able to accrue 6,000 patients at 1,100 sites in the United States, filling a targeted accrual two years ahead of schedule. It's the fastest-accruing trial in the history of the NCI.

And I think one of the things MATCH teaches you is that if you have an interesting trial that's written in a nimble way that is open in the community-- that patients don't have to drive six hours to a cancer center, but can go to a local NCORP site, for example-- then those trials will accrue. We can accrue quickly, and we can accrue underserved populations, and we can accrue rare cancers. And that framework is more nimble than, say, the large phase III randomized trial run only at cancer centers that we had 10 years ago.

There is still a role for large, randomized, phase III trials. The NCI is not backing away from that, or where we will support those. But I think, as we discussed in the JAMA piece, we really have to be thoughtful about where the NCI needs to be involved with those kinds of trials, compared to which of those should be supported by industry, for example.

Dr. Clifford A. Hudis: It sounds like you're alluding to something I think you and I discussed even when you first got into your current role, which is the identification of those trials that industry should run, essentially, itself, and those trials that the NCI should support as complementary to industry trials. Can you expand a little bit on how you see that distinction and where you draw that line?

Dr. Ned Sharpless: Yeah, the thing to know about clinical trials in oncology in the United States right now is most are actually paid for by industry. There's a huge pharmaceutical industry spend on clinical trials, and from my point of view, that's great. The fact that industry is paying for trials to develop therapies for cancer patients-- that's less money the NCI has to spend on those same questions. So we think that's a wonderful development and healthy for cancer research.

But if that's the way it's going to be, then the NCI has to ask itself-- for the precious moneys that we have to spend on clinical trials, we need to use those in a way that's maximally effective and, in particular, not duplicative with what industry sponsors are doing. It's important to say, we do a lot of work with industry. So it's not just us either-or.

Many of our trials, through these agreement processes called CRADAs, allow us to do trials with pharma sponsors and use their compounds in our trials. And that's a real boon to our research effort, as well. But there are certain kinds of trials that are very important where we really want to know the answer, but they're a bad fit for what industry is going to fund.

For example, a de-escalation trial-- that's a trial where there's a standard of care that's pretty good, but the therapy is toxic. And so we'd like to see if we can get the same good outcome in a population using less aggressive therapy. A very important example of this was the TAILORx trial recently, where we showed that based on a genetic risk score, an RNA-based risk score of the breast cancer, women with estrogen receptor positive breast cancer-- many of them could forego cytotoxic chemotherapy and just take anti-hormonal agents and have the same good outcome in terms of their long-term survival.

So that's a trial that is not going to be industry-led, for a variety of reasons. But I think it is the kind of question that's really important for patients. It's important, also, to say that de-escalation trials are hard to do. They require a lot of thought. They don't always work. And so they require these comprehensive thoughtfulness and infrastructure that the National Clinical Trials Network can provide.

So an additional example is these multi-modality trials we have, where maybe two different agents come from two different pharmaceutical companies, and then there's some surgery and some radiation. There's very complex, multi-integrated care. And those can be very hard for a single sponsor to run, but, again, can be a very good fit for the NCI. And there are many other examples like this.

But I think the real question we have to ask is, if our budget is limited and finite, what are the trials that the NCI really should do and lead on?

Dr. Clifford A. Hudis: Yeah. And I think one of the points there is you need to conduct-- we need to conduct-- trials as efficiently as possible, getting the most so-called bang for the buck. You alluded to the fact that the NCI, along with ASCO, has been working on making trials essentially more efficient by making them more representative of the actual cancer population we end up treating.

And a specific area of focus for us at ASCO, in this collaboration and also in our TAPUR trial, has been driving the eligibility age down below 18. My understanding is that this is something that you're adopting as a recommendation across the NCI, as well. I guess my question is, how broad and how quickly do you expect to see this implemented?

Dr. Ned Sharpless: We have a number of efforts related to these barriers to accrual. You mentioned age as one of them and other sorts of exclusion criteria. And we've looked deeply and thought about this sort of care across the continuum of life-- both age limits on the less than 18 side, but also at the greater than 65-year-old side, where we see, often, eligibility criteria structured around a maximum age that don't often make a lot of sense.

So that is one of several topics that we are addressing. As you know, we have a variety of networks and programs, and we fund a variety of kinds of trials. Some are led predominantly by the academic institution. Some are led through NCI networks. And so we are rolling out these policies, not in a one shot fits all way, but across these networks at different scales. They often require scientific buy-in from the other participants, and you know how that process works.

I think this is an area, fortunately, where there is a lot of buy-in, where we're not having lengthy debates about whether or not we should do this. Really, the question is how we operationalize it and make it happen as quickly as possible.

Dr. Clifford A. Hudis: That's great. And you know how strongly supportive we are, on lots of levels, for this effort and the related ones, in terms of barriers to accrual. I want to pivot, though, back to something that you introduced earlier about the big data. And my understanding is, in the annual plan and your bypass budget for 2020, you specifically called out the need to harness big data to speed up all of our work across the cancer research enterprise.

And there are many companies, organizations-- we ourselves at ASCO have CancerLinQ-- that are involved in trying to collect data, share it, analyze it, and advance science and clinical care. But what exactly do you see as the NCI's role in facilitating this, and what do you think is our biggest challenge going forward?

Dr. Ned Sharpless: Yeah, it's an interesting topic. I think the-- it's maybe two things to say off the top about big data in cancer research. The first is the NCI already has one very important example of how big data can transform a field, and that's The Cancer Genome Atlas, which later became the Genomic Data Commons. This is petabytes of genomic data that we make available in the cloud now to any researcher, basically, who is interested in cancer.

And that set of data has led to thousands of papers and just a fundamental reorganization of how we think about cancer biology in many ways. And it's been a huge success, I would argue, and well worth the investment of the NCI to do it. And the data has been used in ways we never envisioned. We never thought of some of the papers and applications that would come out of the analysis of the Cancer Genome Atlas, for example.

But the problem, then, one quickly sees, is that while that data set is great, it's limited. It doesn't have the clinical data, it doesn't have radiology and histology, it doesn't have-- we don't really have a way of binning big epidemiologic cohort data, for example. So the GDC, the TCGA, the Genomic Data Commons, proves how useful these kinds of data aggregation efforts can be, but also makes very clear what the shortcomings of our modern efforts are.

The second thing to say is that this is a problem where the NCI is well-poised to be a leader, right? There are a number of issues around data sharing and data aggregation that really benefit from a Switzerland-like federal entity, a non-conflicted, dispassionate entity like the NCI that just wants to create the data structure in a way that's maximally beneficial for everyone, so that there are-- this is an area where the imprimatur of the federal government really allows us to play a role that would be hard for other groups to take on directly. And so I think this is a reason why so many groups have been looking to the NCI for leadership on this topic.

So what are the challenges to big data? Well, I think that one challenge that has been spoken about a lot publicly is this issue of data hoarding by scientists and physicians and people who have these sets of data they don't want to share for academic competitive reasons. That is a problem. I'm not going to say that doesn't exist. But I don't actually think that's the biggest problem.

I think a bigger problem around data sharing is just it turns out to be really hard to do. And by hard, I mean expensive. It turns out to be-- these various data sets were not created, initially, with the intent of sharing them. They're often in different formats. They're often governed by different kinds of data use agreements, which are governed by the consent form that the patient signed to have their data included.

And so linking them can be both very technically difficult, from just a computer science point of view, and can also provide a lot of administrative and logistical hassles from the data sharing, data use agreement point of view. And so each one of these things is just something the NCI has got to work through, or someone like the NCI-- is figuring out how to link disparate data sets, how to get the right kind of data abstracted from charts that we want, how to develop the right work force to study big data with big data analytics, and then that is a big problem.

So there are a number of areas where the NCI can address the challenges. And I think we'll make progress. I mean, the good news is that we understand these problems. This is not like we need to-- there's some fundamental problem of biology that we need to figure out. The bad news is that the problems are weedy, complex, and many, many layered, and require us working through them.

But that's what we can do. We have support from the government for this. The moonshot had a lot of funding for data initiatives, which we've been employing to get these structures going. And now the Childhood Cancer Data Initiative, for example, I think could really-- that's a nice demonstration project, if you will, because it's the right size. Childhood cancer is about 16,000 cases a year.

And so I think we can show what this radical data sharing, if you will, this data liberation project can do-- you know, that population and how useful it could be to larger groups of patients like lung cancer, breast cancer, things like that. So I think that these are the kinds of things the NCI can do with help from other federal agencies and academic partners and groups like ASCO.

This is certainly not an area where we plan to go it alone. There are a lot of stakeholders and a lot of great ideas. And I think that by organizing and convening these initiatives, we'll make progress.

Dr. Clifford A. Hudis: Well, I really, first of all, appreciate your calling out the fact that data hoarding in isolation is not the single biggest problem, because I think that's a frequently-cited limit. And I agree with you that it's less of an issue than all of the other ones that you highlighted. In that regard, I understand that you just announced a new office. I think it's the Office of Data Sharing? Can you expand on or explain how that relates to these challenges and what it's going to, hopefully, accomplish for us?

Dr. Ned Sharpless: Sure. The Office of Data Sharing is something within our Center for Bioinformatics and Information Technology. It's getting stood up now. It's been around for about a year, even less than that. It has a new leader and a few FTs, and it has a number of jobs intended for it. I mean, there are a number of ways that we would like the Office of Data sharing to-- a number of problems that we think that the ODS can help serve with the external community in terms of data sharing, like these issues around consent and data privacy that I mentioned.

But right now, an intense focus of that office, because it's something we really need to solve, are related, really, to the issue of accepting data and allowing access to NCI data at present. So we have this complex structure whereby academic investigators can give data sets to the NCI. That's harder than it sounds, because we have to make sure the data are of good quality and they're properly consented, and we understand the data usage agreements and that kind of stuff. And then we have a means to allow access to those data to accredentialed investigators. And there are a bunch of issues with that that are more complicated than you and I would want to go into right now.

But I think that's consuming a lot of the bandwidth at that office right now, is the problems around, for example, the dbGaP entity, whereby different investigators give data to the NCI and the rest of the NIH. That has caused a bit of a bottleneck, and so we're trying to work through some of those issues.

One thing, for example, that I think the ODS can do and is doing already is this sort of concierge-like function. For people who have large, valuable data sets that they'd like to give to the NCI, we should be able to take those data sets as quickly as possible. But something that's happened in the transmission of those data is that we've realized the quality isn't quite what we wanted or the format isn't exactly right, and so we have these questions, and they go back to the investigator. And there's this sort of cyclical loop that can take months and really substantially delay the process.

And so the ODS is jumping in there early on and intervening on that loop and making sure the data are the right format and the right quality at the time of initial submission, so that we don't have this back and forth that wastes a lot of time. So I think those data access and data transmission issues are a prime focus for the office right now, although it has a much larger mission as it gets stood up.

Dr. Clifford A. Hudis: Yeah, a little bit like CENTRA that Rich Schilsky runs for us here at ASCO, in terms of access. But at any rate, I want to take the remaining time we have, and maybe this is a speed round on the cancer research workforce. So a couple of quick questions, perhaps-- first of all, has the Cancer Moonshot Initiative had an impact directly on the kinds of awards that you're making available to researchers? And if so, how do you think that might evolve in the next couple of years?

Dr. Ned Sharpless: I think the moonshot, as you know, was intended to focus on these 10 areas identified by a blue-ribbon panel that were thought to be ripe for clinical translation, just about ready to go into clinic and to benefit patients in a very direct, immediate way. So the moonshot per se didn't include funds for things like really hardcore basic science or training, although certainly moonshot moneys are being used to some extent in both those areas, as necessary, as part of these translational efforts.

So I think that what the moonshot has done-- it's done a couple of things. So first of all, that most of the awards granted by the moonshot mechanisms are more these-- are not the traditional R01, but are more of these consortia and network grants. And I think we've built a lot of infrastructure for research efforts, say, in immuno-oncology or in pediatric cancer or in survivorship. And those networks will both-- well, they will live on beyond the moonshot in some cases, I'm sure.

And those networks will provide integrated research efforts, but also some training opportunities. So most of those include junior scientists and junior clinical investigators, and so there will be some opportunity for the moonshot both to drive the scientific area of study and also provide some training opportunity for the new people coming up.

Dr. Clifford A. Hudis: Well, speaking of junior and new, I listened to your conference call, I guess, about a week or two ago talking about the pay line. Can you expand on your plans to support young investigators right now, given the always-present constraints in funding?

Dr. Ned Sharpless: Right. This is a particular problem for the National Cancer Institute, because we've seen this relatively-- there's no other word than "massive" influx in the number of applications for the so-called R01 grants, the independent investigator-initiated award at the NCI. And this is-- our award number is something up like 60% over the last nine years or so.

So this rapid increase-- which is, in most ways, a very good thing. I mean, that says that new scientists are coming to our field with new ideas and new ways to treat cancer, and the NCI can pick among these many applications and fund the very best ones. But it has this pernicious bad effect for the academic investigator community, and that is that their individual chances of getting a grant are lower.

If paylines are really the number of funded awards divided by the number of applications, and the denominator goes up faster than the numerator-- both are going up, but the denominator goes up faster-- then the paylines are going to go down. And we think this is particularly a problem for junior scientists, because established scientists have seen paylines come and go and funding realities change.

But new scientists aren't as used to the life of the independent researcher and, we think, are most likely to either leave science or move out of cancer research to another area of science. And we'll have to try and minimize that from happening, to the extent possible. So one of the things we've done at the behest, in fact, of 21st Century Cures, which included language asking the NCI in the United States to do this, was really focused on these so-called early stage investigator, the ESI.

So the ESI is faculty. That's someone who's gotten a job, generally in an academic institution, and is now writing their first R01 grant, their first independent scientist grant. And we've done a few things for this population. One thing that's really important is we give them a special payline. We give them, effectively, a higher chance of getting funding.

So if, say, paylines are on the order of 8% now for all Comer grants, for ESIs they'll be more like 14%, right? So a significant-- or 12%, in that range. So, significantly higher than what the general community is. I want to point out, also, that paylines are lower than the actual success rates of the NCI, which is a better number. The reason success rates are higher is because we do fund a lot of grants outside of the score. It's a little bit of inside baseball. But generally, if you write a grant to the NCI, your chance of getting it is more like 12%. And if you're an early stage investigator, it's more like 16%.

Dr. Clifford A. Hudis: Thanks, Ned. To switch gears a bit, I know you've worked with the NCI throughout your career. But now you've been at the Institute's helm for nearly a year and a half. Has your understanding of the NCI and its role in cancer research changed or evolved in this newest assignment?

Dr. Ned Sharpless: I think it has to be said that I was an NCI watcher my entire research career, and I thought I knew the National Cancer Institute and the National Institutes of Health pretty well-- as well as one can know these organizations from the external perspective. But since starting at the NCI, I've really learned that this amazing organization is much larger than even I realized, and that the scale and scope of the NCI is truly both awe-inspiring and, in some ways, daunting.

I had a series of meetings as I started as NCI director where I would learn about these sprawling comprehensive cancer prevention and control efforts or new areas of basic research or clinical trials. And I just really had had no idea that the NCI was involved in some of these activities. So it was very illuminating.

In some ways, it's thrilling, the things the NCI is doing. But I think it also made very clear to me another thing that I think I knew at some level, but didn't really appreciate the full scale of this until becoming NCI director, and that's the issue of-- although the NCI is huge and has this great reach and comprehensive nature, we are limited in scale. Our resources are finite, and the NCI, therefore, is really forced to make these difficult choices about which areas of cancer research to fund and how best to address our mission of reducing cancer suffering.

So I think I was surprised both by the scale and scope of the NCI, but also by the fact that, despite how big the NCI is, it still has significant limitations on what it's able to do and has to make these difficult choices.

Dr. Clifford A. Hudis: ASCO recently launched the "I lived to conquer cancer" awareness campaign that spotlights federally-funded cancer researchers and the patients who inspire them. I want to close out our conversation today by asking you, why do you live to conquer cancer?

Dr. Ned Sharpless: Yeah, I think like just about everybody in the United States, my life has been personally touched by cancer. I've had friends and family members get cancer, and my father even died from cancer. Both of my sisters are cancer survivors. So I think I have a real personal stake-- like everyone in the United States, almost-- in seeing the reduction of cancer suffering and conquering cancer, if you will.

I also find the problem fascinating from an academic point of view. I was drawn to cancer research because I found the biological questions of cancer research so fascinating. So I live to conquer cancer from this intellectual point of view, as well.

And lastly, I have the experience of being a doctor, of being a medical oncologist taking care of patients with cancer. And I've had the frustrating experience of having patients not do well who I thought, I wish we could have done more for-- as well as the experience of taking someone who has a pretty terrible cancer but yet driving it into remission with therapy and then watching that person effectively survive the disease and become cured of it over years. And that is so special and so thrilling to be a part of that as a physician.

So I live to cure cancer because it's personally touched my life, because I am a scientist who is fascinated by the biology of cancer, and as a doctor I've had the experience of helping people survive their cancer. And once you do that once, you just want to do that over and over again.

Dr. Clifford A. Hudis: That's really great, Ned. It's fascinating to hear why progress against cancer is personally so important to you. And I'm sure all of our listeners enjoy hearing that, as well. I want to thank you again for joining me for this ASCO in Action podcast and for all the work you do at the NCI and across the entire cancer care community.

Well, thank you for having me. As you know, one of NCI's most important partners in this effort against cancer is really ASCO. And so it's great to speak to you today. And thanks for all the things that you guys do for patients with cancer.

Again, thanks to all of you for listening today. Those of you who want to follow Dr. Sharpless on Twitter, he's @NCIDirector. And you can always follow me @CliffordHudis, as well as ASCO @cancer. If you do that, you can stay connected to our work, of course, on social media. You can also go to the NCI's website, which is NCI.gov. With that, again, I want to thank Dr. Sharpless for joining me today. And thanks to all of you for tuning in.