Shaping the future of cancer research at NCI

Norman E. ‘Ned’ Sharpless became director of the US National Cancer Institute (NCI) in 2017. From April 2019, he served as acting commissioner of the US Food and Drug Administration (FDA), and he returned to the NCI in November. Douglas R. Lowy has been the principal deputy director of the NCI since 2010, having also served as acting director twice, most recently in 2019. Nature Cancer spoke with both last fall to learn more about what makes NCI and its leadership tick.

Ned Sharpless

You have held very different roles as a cancer researcher and medical oncologist, the director of a large NCI-designated Comprehensive Cancer Center at the University of North Carolina (UNC) and later the director of NCI since 2017, a position to which you have returned after serving as acting commissioner of the FDA in 2019. Could you share some of the turning points that led you to follow this career path?

Credit: National Cancer Institute (NCI)

NS: That’s an interesting question. I divide my career into two parts—pre–federal government and federal government—and they are very different. I knew when I was director of the UNC Cancer Center that it probably was not going to be my last job. I was interested in what would come next, but I wasn’t sure what that would be. A defining moment for me was in 2010, when my father died—ironically of melanoma, a disease I had studied. The first new approval of a melanoma drug wasn’t until 2011, so this was right before that landslide of seven new drugs that have been approved for the treatment of melanoma. My father was a veteran, although he never talked about his time in the military, which had been 60 years ago. It never really came up much, other than on Veterans Day, when we would say, “Thanks, Dad.” But at his funeral, someone from the army came and presented my mom with a flag and said, “A grateful nation thanks you for your husband’s service.” I remember this having a very strong impression on me at the time, and for weeks afterwards, I would go around thinking, “What have I done to help my country, and is there some opportunity to serve the US,” because I have this fierce patriotism. So later when the White House called me about taking on the NCI director role, I really didn’t have to be talked into it. I knew after a few seconds that I definitely wanted the job. It was an opportunity to do something I really love—work on cancer research—in an agency that I really respect and admire, the National Institutes of Health (NIH) and the NCI, but it was also something that I could do for my country. So that was what led me to federal service.

How have the different roles you have held shaped your philosophy in leading NCI?

NS: I think it is helpful for leading the NCI to have experience in treating patients. I was a practicing oncologist in academia for many years, so the patient’s voice is constantly in the forefront of my mind. I am interested in solving problems at NCI that will make a real difference in patients’ lives. Of course, NCI is a basic-science organization, and we must continue to support basic cancer research. At the same time, I’m mindful that if we can’t articulate why that basic-science commitment is important to patients, then we are not really doing our job properly.

While my experience as a practicing physician has deeply influenced my philosophy of leading NCI, my time as acting FDA commissioner was also very useful for my job here at NCI. I gained a much better understanding of the federal government in general, as well as a more specific understanding of how agencies like NCI and FDA work together—and can work together better. During my 7 months at FDA, I remember many times thinking, “I wish the science in this area were better developed.” FDA relies on scientific discovery to make regulatory decisions that have profound impact on the lives of Americans. I would think to myself, “Wouldn’t it be nice if an agency like NCI—that really has the infrastructure, talent and resources to answer some of these science questions—could take them on in a very specific way to help FDA’s regulatory decision-making become better?” So I have to admit that understanding the regulatory needs of an agency like FDA has really sharpened my focus at NCI.

Now that you have returned to NCI, what are your priorities, and would you say that your experience at FDA has changed them?

NS: I don’t think my priorities have changed because of my time at FDA. If anything, my FDA experience has given them more of a sense of urgency, as it has sharpened my understanding of how challenging it is to get things across the finish line in the federal government and how much of a constant push it requires. That’s true for both FDA and NCI. I also think I’ve got a very good appreciation now of what advancing cancer science will mean in terms of drug approvals, new therapies and devices for cancer patients at the other end of FDA. It’s added incentive to try and work as efficiently as we can here at NCI to advance our discoveries, but the priorities have remained the same.

When I first started at NCI, I emphasized a commitment to workforce development. I think that the most important thing NCI does is develop the next generation of cancer researchers, and that is a continuing priority for us. That also involves determining where the most promising areas of opportunity are in cancer research and what skills will be required of future researchers. It is incumbent on the NCI to really work on that so that our investment in training and development of the workforce is focused where it needs to be.

I continue to believe that a major challenge in our field has to do with poor data aggregation and data usage. Many novel techniques are being developed for data analysis, such as distributed ledger technologies and machine learning, but how we are going to use those in cancer science is an important question and an area where I’ve always been very focused.

The clinical trials apparatus at NCI has been undergoing fairly dramatic changes over the last decade. We must continue to modernize the clinical-trials infrastructure and make it work better for patients, by giving them increased access to trials. This will in turn improve accrual rates and help us do the kinds of trials that are meaningful for patients in a more rapid and facile way.

As I already mentioned, I think NCI will never stop supporting strong basic science. Job number one for NCI is always to fund and incentivize the foundational science that allows cancer discovery to go forward.

So I completely stand by the four areas of key focus that I articulated when I first started. If anything, my experience leading FDA made me realize that that framework is a really strong one.

Cancer research has become very interdisciplinary. It now involves many different fields, including biology, genetics, engineering, computational science and clinical work. What are the challenges and the opportunities that are generated by this interdisciplinarity, especially for younger investigators? How can we support the younger generations of researchers who are facing this increasing complexity?

NS: Before discussing the challenges, which people tend to focus on, I think it is important to say up front that interdisciplinary science really is a good thing. Working collaboratively is an opportunity to accelerate progress and help patients in a more direct way when it involves individuals with translational expertise. I believe that interdisciplinary, collaborative research is part of the reason our progress against cancer has been so great of late. We sometimes fondly remember the days of one postdoc and one principal investigator writing their paper. Certainly, some terrific basic science got done that way, but when the time came to move that science into the more translational phase, it was hard because the individuals with the relevant expertise weren’t necessarily involved.

I think the interdisciplinary nature of cancer research has created a lot of the enthusiasm that exists in cancer research right now, and I have to say it really is an exciting time to be a cancer researcher. When I looked at the entire portfolio of approvals coming through FDA, I was impressed that cancer is the coolest area. And that’s not my personal bias; that’s just what they’re approving. The progress is not nearly as great in other disease areas.

At the same time, interdisciplinary research does have some challenges. Making sure that credit is apportioned properly is the main challenge. I always worried, for example, about the career ladder for the data scientists, who are absolutely indispensable to a project. They are the people who aggregate and clean up the data, making analysis possible. But they would end up being middle authors on a paper with a long author list. How are those individuals going to get recognized and promoted? Does their academic institution appreciate their efforts? In addition, those individuals may have real opportunities to work in the private sector for companies such as Google or Amazon, so why would they stay in academia doing relatively thankless, underpaid data science when these other more lucrative opportunities exist? It means we have to think of creative ways to incentivize and recognize the achievements of biostatisticians, data scientists and certain other specialists who won’t be the first author on papers but will be vital to the progress of the endeavor. This also applies to young investigators, who may be crucial to these interdisciplinary teams, but, because they are less experienced with this process, they may not get as much credit as they deserve.

Allocating rewards for research progress can be difficult, but it can be solved. NIH, and NCI in particular, have developed funding mechanisms to recognize these scientists, and we have really tried to get this message out to the academic communities. For example, NCI’s Office of Cancer Centers expects cancer centers to have certain core infrastructure capabilities to meet the requirements of a modern cancer center and to be supported through NCI grants. In addition, we have done a lot to protect the status of the early-stage investigator (ESI). The early years are a very challenging time in someone’s career, and we understand that we are asking these junior scientists to work collaboratively as part of large groups while at the same time they need to develop their own careers to advance themselves. So we have, for example, extended the award length for the first research grants, and we have tried to enhance the paylines for ESIs to the extent possible. It’s still a challenging time in their careers, but at NCI we try to make success in science for ESIs as easy as possible.

You mentioned how important big data and data science are in cancer research. How is our ability to generate and analyze big data today changing laboratory-based research?

NS: I’m old enough to remember when you could work on a single gene in your mouse-genetics lab and publish a paper that proposes that a gene is very important in cancer because it has a certain behavior in a knockout mouse or in a cell culture model. In retrospect, many of these papers turned out to be wrong, because they were based on imperfect in vitro and animal-model systems. But now the clinical data are available. The Cancer Genome Atlas has sequenced thousands of patients. The Genomic Data Commons has genomic data from tens of thousands of patients. So it has become fairly easy for molecular biologists to ask whether their gene of interest truly exhibits the clinical behavior they would expect by checking it against large patient data sets they didn’t generate themselves. They may not even need to collaborate, or may work with just one data-science collaborator, because we’ve gone to great lengths to make these data available through web-based formats. So what used to be a long and drawn-out endeavor—finding the gene in paper number one, then characterizing it in paper number two and then having somebody else study the gene’s effect in populations in paper number three—can now all come together in one paper, because these sorts of large data sets that may include clinical outcomes, genomic data and RNA expression profiling are readily available.

Big data is a good thing, but it is not without its problems. First of all, the care and feeding of these large data sets, and making them available to the public, is much more expensive and difficult than I would have imagined before I came to NCI. The data aren’t always applicable in a way that’s straightforward. There can be clinical outcomes from different trials, but merging them together and using them for aggregated data sets can be much more difficult than one would imagine. More and more, we are finding that it is hard to analyze these data sets properly, since they are so large, and often there is a difference of opinion among the data scientists and statisticians as to how to do the analyses. NCI wants to make these data available now, so we are making some choices about how to handle and view them. But those choices are going to be guided by future methodologic work. For example, there is tremendous methodologic work that still needs to occur in terms of how to do single-cell RNA-sequencing analysis. These challenges are real, and as these data are becoming more readily available, they pose a different type of problem for the academic investigator: the problem is not whether the data exist, but, because there are so many data, we have to figure out which data sets to use and how to use them.

What do you think about the application of these methods to clinical practice and the patients today? How close are we in terms of integrating big data, data science and clinical practice?

NS: I think this has already started. FDA has already approved drugs using real-world evidence gleaned from aggregated data from non-clinical trial settings. The palbociclib drug approval, for example, was a label expansion of a breast cancer drug that received expanded approval for treatment of breast cancer in males and was in part based on analysis of real-world evidence. I think we’ll see more and more of that, but it turns out to be harder to do than you might imagine. Real-world clinical data are not gathered in a way that makes them easy to use for practice decisions. We are getting better at that, and there’s certainly a lot of interest from government, academia and industry in trying to get more data out of clinical practice and to collect these data in a way that makes them more useful going forward.

By the way, I think real-world clinical science is a really good space for the NCI to conduct research in. NCI is not tasked with making decisions on what does and doesn’t work the way FDA is, but it is tasked with providing the foundational science that allows these real-world methods to be used. We have a robust grants portfolio supporting investigation on how to collect and aggregate real-world evidence, how to extract features from that data and how to determine which endpoints from such data sets are useful. As I mentioned, it has been surprising to me how hard this problem is, but it’s an area in which we are seeing progress. In the last year or so, we’ve started to see great studies come out, FDA decisions being made and investigators deciding what research to do in the future based on that kind of evidence. I think the era of using real-world data to guide clinical practice and regulatory decisions is upon us.

You mentioned earlier that it is a very exciting time to be a cancer researcher. Which advances in cancer research have you found most exciting over the last few years, and what are the breakthroughs you are looking forward to?

NS: What’s exciting to me about cancer research is the diversity of progress. As I’ve often said, the major cancer discovery in my lifetime is that cancer isn’t one thing, but rather it’s many things. I could say that the last great hurrah of the ‘cancer as a unimodal entity’ idea was the angiogenesis-inhibitor period, when the belief was that one kind of agent could cure everything. There was this sense that all we needed to do was work this out, and we were going to have a silver bullet for all cancers. That didn’t really work, and I’d say a major factor in this was the appreciation of RNA transcriptional diversity in cancer. I think it was very eye-opening for the field when Chuck Perou published a paper back in 2000 stating that breast cancer wasn’t one entity but rather multiple subtypes. What we thought were maybe 10 or 15 unique diseases were really hundreds of diseases, and they each had their own epidemiology and risk factors, and would respond differently to therapies. That was a major advance that radically changed how we study cancer. Instead of giant clinical trials with 2,000 patients per arm, as they are done in cardiology, we now have very important phase II trials with small numbers, like 50 patients, that in some cases have even led to FDA approval of new therapies.

So the ability to find the right patient and use the right approach has been the major development in cancer research in the last 20 years. We’ve now seen very active kinase inhibitors for subsets of lung cancer, marvelously effective immunotherapy checkpoint inhibitors for cancers that present many neoantigens, and CAR-T or cellular immunotherapy for certain kinds of hematologic cancers. Added to that is the persistent efficacy and utility of the old sort of cytotoxic DNA-damaging agents, but now we’re understanding in what rare subsets of patients these drugs work and why. And then we have the very challenging problem of putting all these different treatments—radiation, surgery, chemotherapy and immunotherapy—together and using them in sequence or in parallel in patients. I often have made the remark that if we discovered no new drugs for a decade, we’d still make a lot of progress in cancer during that period, because the clinical-trials community is going to have to figure out how to combine all these modalities in a way that is most effective for patients. By the way, I think that we’re going to see many new therapies approved over the next decade, and then we’re going to have to figure out how to move them into our new regimens in a way that is most efficacious for patients, minimizes toxicity and prolongs survival.

In the modern era, we also have to think about treatment costs. We are keenly aware of the issue of financial toxicity for patients. A drug that’s marvelously effective but that no one can afford is not really progress. Finding ways to ensure that not just some but all patients get access to therapies that are beneficial to them is an important consideration at NCI.

Now that you are back at the NCI, what is your vision for the coming years at the institute?

NS: When I assumed leadership of the cancer center at UNC, it was in great shape, and the same was true when I arrived at the NCI and when I began serving at FDA. So the theme that is a constant around my career is: don’t mess things up; take things that work well, and help them work better. I would argue that the enterprise of cancer research has been very successful and exciting over the last decade, and it’s built on hard-won, intricate basic science that goes back to the many complex studies of the 1970s and 1980s. Now we are starting to see the fruit of those endeavors, and we have a really good understanding of cancer. It’s not perfect, and, as I said, we definitely need further basic-science investigation, but we do have a really good understanding of the heterogeneity and diversity of cancer. So now we have a paradigm that is working and has caught the attention of [people in] the pharmaceutical industry, who have greatly increased their investment in cancer therapies because they feel that the basic science is advanced, and they see cancer as a field where they can develop effective therapeutics.

I just want to keep that really strong momentum going, to make sure that we can move from a great idea to a new therapy or means of prevention as quickly as possible, with the shortest period of experimentation necessary, so that patients don’t have to go through not having access to effective therapies or receiving ineffective therapies. That means making the basic-science infrastructure efficient; using data in a way that we can learn from every patient; doing the smallest, most efficient clinical trials possible to get the information needed; and, once we have put that package together, passing it along to industry and other regulatory agencies so that it can benefit patients.

As we are coming to the close of our conversation, could you share with us what inspires you in your day-to-day work in cancer research?

NS: I spend a lot of time thinking about my first patient in medical school. In med school, they want you to interview a patient so that you can practice what it is like when you’re not in a didactic classroom setting. So they had me put on my little white coat and unpack my new stethoscope that I’d hardly ever used, and I trundled over to the hospital. The very first patient I saw was a really nice guy from eastern North Carolina who was dying of metastatic lung cancer. He presented to the hospital with superior vena cava syndrome, so he was in a tough position. But he was eager to have a medical student interview him. He was a really charming man and told me about his life. I remember I asked him if he smoked and he said, “two packs a day for as long as I can remember.” He was 55 or 60 years old. After interviewing him, I had to present to our attending physician. Keep in mind, the real focus here was on getting a patient’s history and doing a physical examination; it wasn’t on what was going to happen to the patient—that was before we thought about outcomes. As I was reporting about the patient, I remember saying what a nice guy he was and how much I’d enjoyed doing this. The attending physician said, “Yes, it’s too bad he’s going to die soon.” I was shocked. It hadn’t occurred to me that we didn’t have an effective therapy for this patient, and when I realized that, it really bothered me. I felt we should be able to do better. That is something I’ve tried to keep in mind. I have a good sense of what a tick-down on a Kaplan-Meier survival curve means. I can picture the patients whose lives that graph represents, but despite all the great progress and the excitement in cancer research, there are still far too many people suffering and dying from cancer. So, for me, the ability to have that happen as infrequently as possible, or never, is very inspiring. That, I think, is the great aim at NCI, the shared mission. NCI is full of wonderful people who want to make a difference and reduce the amount of suffering due to cancer. It’s a real privilege to work with them.

Douglas R. Lowy

What inspired you to become a researcher and to pursue an academic career? Were you always interested in the life sciences and cancer research?

Credit: Ernie Branson, National Cancer Institute (NCI)

DL: Well, my first choice would have been to be a quarterback in the National Football League, but I was missing two important prerequisites—talent and size. In the long run, cancer research turned out to be a really important area that matched my talents to a much greater degree. My parents were both primary-care physicians, and they talked a fair amount about cancer when I was growing up. When I was a teenager, my mother had melanoma. So I certainly developed an interest in cancer early on.

When the Hammond and Horn report came out in the mid-1950s, linking tobacco consumption to lung cancer and cardiovascular disease, it was a real eye-opener for me because I grew up at a time when tobacco was actually viewed in a positive light. For example, when somebody would hit a home run in a baseball game, they would send cigarettes to veterans’ hospitals. Everybody had a cigarette case in their living room for people to smoke cigarettes when they were socializing. So the Hammond and Horn report really led to major change. My parents talked with me about it at length, so that when the surgeon general’s report linking smoking to cancer and other disease came out about 10 years later, in 1964, it seemed to be coming a bit too late in my opinion, even though I had not yet gone to medical school.

I didn’t become interested in research until I got to medical school. I was actually a non-science major when I was an undergraduate, but the teaching at New York University School of Medicine, where I went, was very oriented toward research. It seemed really interesting, so I got involved in laboratory research while I was there and ended up going to NIH for further training.

Could you share some of the defining moments in your career as an independent investigator?

DL: I was quite discouraged by my initial research experience in medical school because I didn’t think I had accomplished very much. However, my supervisor, Jan Vilcek, thought that I had potential for doing research. Without his encouragement, I doubt that I would have gone in that direction. So his support was really of critical importance and I would say the first defining moment in my career. With his help, I applied to NIH for training, and while I was there, I studied endogenous mouse retroviruses. This was soon after the reverse transcriptase was discovered. I was involved in some experiments on endogenous mouse retroviruses that cause thymic lymphomas in mice and developed a technique for activating them from normal virus-negative cells and then studied their structure and genetics. I showed that the provirus behaved as a Mendelian gene that was part of the mouse genetic chromosomal makeup and localized one of these endogenous proviruses to a particular part of the chromosome. All of this research was very interesting to me, so I decided to become a basic researcher.

The next defining moment was deciding to work on papillomaviruses, which were not well understood. They caused warts, for which there were no really good treatments. I had trained in dermatology, and I wanted to connect my clinical training with my research. My lab developed a cell transformation assay for the bovine papillomavirus, and the molecular genetics of papillomaviruses were worked out using that assay. My lab did it to some degree, but other labs took this assay and used it innovatively to develop molecular genetics.

Another defining moment was a site-visit review several years after I started my lab. By this time, the relationship between the human papillomavirus (HPV) and cervical cancer had been identified, and we had been working on cell transformation by papillomaviruses. The review panel concluded that although we were good, we should use our talents elsewhere, because there were other laboratories ahead of us. We shifted our direction to work on the structural proteins of the papillomavirus, which then led directly to our finding that if you expressed the L1 protein of the virus, it would self-assemble into virus-like particles that induced very high levels of neutralizing antibodies—the cornerstone of preventive vaccines.

So, although neither Dr. Vilcek nor the site-visit panel told me what to do, by suggesting I continue to do research and put my talents in a different project, they did help steer me in a direction that otherwise I might not have gone on my own. That led to our studying the structural viral proteins and making the observations that ultimately resulted in the technology for the HPV vaccines.

Continuing on from this point, given that your work has been successfully translated into the clinic with the HPV vaccines, what are your thoughts on basic versus translational work in cancer research? What can we do to improve translation of basic findings?

DL: My own feeling is that basic research, including basic research that may not have an immediate translational implication, needs to be very strongly supported. High-quality basic research that can help us understand mechanisms of growth control, signaling and various fundamental processes will lead to interventions in the long run.

On the other hand, academic researchers frequently find themselves wanting to pursue findings that might have translational potential. One of the programs that NCI has in the area of cancer treatment is the NCI Experimental Therapeutics (NExT) program. This is specifically designed for academic researchers who want to do translational research but don’t have actual experience doing it. The NExT program is largely carried out by extramural investigators with experience in translating findings to develop therapeutic interventions. When an application is selected, the program will help academic investigators move their findings forward so that they can be translated. This is one way in which we try to overcome the bottleneck of people who have good ideas but don’t necessarily have the right experience for their translation.

Building on this point and the difficulties of research, what do you think are the biggest challenges we are facing in cancer research?

DL: NCI faces several big challenges. The biggest challenge is that we have many more outstanding applications from researchers than we can fund. Although we are funding more awards today than we did a few years ago, we can’t keep up with the interest and enthusiasm there is for cancer research. Our applications have gone up by about 50% over the last 5 years, which is about ten times faster than the rest of the NIH applications. As a consequence, the success rate of our applications has actually gone down, even though we are funding more proposals than we did in the past. The levels are so low that there is no question that it is demoralizing to the research community. So, from the perspective of the investigators, I think this is a very big challenge. At NCI, our top priority is to try to improve the success rate and paylines for investigator-initiated research.

From the point of view of patients, I think that the most acute challenge is for those who have advanced cancer. We need to remember that, of the people who develop cancer today, more will be cured than will succumb to the disease. But, for people with advanced cancer, the problem is that either they will get a response to treatment, but their cancer will then become resistant to the therapy, or the cancer isn’t responsive to treatment right from the start. I am hopeful that putting combinations of treatments together will be an important approach to try to overcome the challenges of primary and acquired resistance.

The third challenge is in the area of cancer prevention and screening. We should recognize that, as important as it is to develop successful interventions for treating cancer, in the long run we are going to benefit more people by preventing cancer and by detecting it early, when it can be treated much more successfully than when it has advanced. So our other big challenge is first to determine how to disseminate more widely the interventions in prevention and screening that we already know work and second to make further advances in prevention and screening, so that we can help even more people. I would say these are three enormous challenges.

We covered the challenges, but cancer is a very active field, as you just described. Which areas in cancer research excite you the most, and what advances are you anticipating in the next few years?

DL: I anticipate that new drug combinations are going to lead to more patients having responses and that the responses will be of longer duration. I think it will be analogous to what happened a number of years ago with the treatment for people with HIV infection. In that case, putting drugs together led to much better outcomes for these patients than treating them with a single agent. We are supporting research on combination therapies, so I expect that we will see that also happen with treatment for cancer.

NCI is funding research in many different areas where we foresee good potential to lead to important and impactful findings—whether it’s at the level of cancer survivorship, treatment, diagnosis or pathogenesis. But I think there is one area where we are going to be surprised a number of times in the next few years: basic research. In basic research, people often make observations that in many ways are unexpected and that take us in new directions. I am pleased to say that we are doing as much as we can to support this type of work. I expect that we will make discoveries that will help us to think about cancer in new and different ways. Ultimately, this will help patients to either not develop cancer or, if they do develop it, to have a better outcome and quality of life than they do today.

How do you view your roles in the NCI leadership?

DL: The cancer research community is a wonderful community to be leading. The cancer advocacy community also plays an absolutely key role, which is critical for all of us to recognize. We also receive a great deal of input from people outside NCI, through advisory boards, for instance. Although ultimately we need to be the ones making the decisions about what to fund, our decision-making is done in as collegial a manner as possible, because we all share the same mission, which is to improve outcomes for all patients and do as much as we possibly can to prevent cancer.

In our final few minutes, could you share some advice for the younger generation of cancer researchers?

DL: I think they should be optimistic about the research opportunities in cancer. NCI and many other organizations strongly support the training and development of young investigators. At NCI, we have a number of different fellowships for graduate students, postdoctoral fellows, people who are transitioning to becoming independent researchers and early-stage investigators. In the last few years, we have tried to give a boost to people applying for their first large grants, because we recognized their relative inexperience in terms of writing grant applications and having only preliminary data, compared to experienced investigators. We believe that training the younger generation is absolutely essential for continuing to make important discoveries. Whether these young cancer researchers are trying to do basic or applied research, they are the people who are going to push the field forward. We need them, and we do everything we can to support them.

Author information

Correspondence to Alexia-Ileana Zaromytidou.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zaromytidou, A. Shaping the future of cancer research at NCI. Nat Cancer 1, 7–11 (2020). https://doi.org/10.1038/s43018-019-0017-7

Download citation