As artificial intelligence becomes an indispensable tool in cancer research, drug sponsors and researchers should focus on building and utilizing rich real-world evidence databases that can be used to complement clinical trials.

“We have blood markers now. We have tissue markers,” said Roy Herbst, deputy director of Yale Cancer Center and Smilow Cancer Hospital, assistant dean for translational research and Ensign Professor of Medicine and professor of pharmacology at Yale School of Medicine. “We have to take a study with pembrolizumab, a study with durvalumab, a study with atezolizumab, all different drugs, but similar enough that they need careful investigation, and the data needs to be evaluated and pulled together.

“We’re collecting all those data, but we’ve got to use the tissues in a rational way. In fact, if we don’t associate these biomarkers with patients who are responding or have long-term stable disease or progress, we’ll never figure out what to do.

“And then, we can design the prospective trials. Then, we have a smarter Lung-MAP. We have a smarter Pragmatica, because we’ve learned and adjusted based on our data.”

Herbst, who is also the chief of medical oncology at Yale Cancer Center and Smilow Cancer Hospital, was elected to the Friends board of directors Jan. 8. He is vice chair for developmental therapeutics at SWOG Cancer Research Network and was the original steering committee chair and principal investigator of Lung-MAP (The Cancer Letter, Nov. 15, 2013).

“What Roy brings to the table, beyond the science and his clinical trial expertise, is he never forgets what we’re doing this for. The science is important. The processes and the speed are important, but ultimately making life better for patients is really what we are about,” said Ellen Sigal, chair and founder of Friends.

“We’re all incredibly interested in the role that AI and real-world evidence can play in advancing research. We have these partnerships in both of these areas, and I know that Roy will be very involved in them.”

In 2024, Friends will be investing in ongoing work on digital pathology and ctDNA, pragmatic clinical trial design, and validation of real-world endpoints for drug development and label expansion (The Cancer Letter, July 7, July 14, 2023).

On Feb. 1, Friends announced a new research partnership, Digital and Computational Pathology Tool Harmonization (Digital PATH) Project. This project will identify factors that may contribute to variability in biomarker assessment across computational pathology platforms, propose areas for alignment in the field, and provide insights for shaping regulatory processes. The organization’s 2023 scientific report is available on the Friends website.

Pragmatica-Lung, a streamlined phase III trial evaluating the efficacy of ramucirumab plus pembrolizumab in non-small cell lung cancer, was the result of an unprecedented collaboration between FDA, NCI, industry, research groups, and Friends. Launched over a year ago, the trial has to date accrued about 300 patients—nearly half of the 700 target—since it started enrollment in April 2023 (The Cancer Letter, Pragmatica-Lung).

“With a group like Friends, through the meetings, the policy papers, the advocacy, hopefully this becomes even more prevalent in other tumor types,” Herbst said. “Right now, we’ve been called by people who are researching melanoma and other tumors, wanting to do very similar things. Our goal is to change the standard of care and to get drugs to patients with cancer as soon as possible.”

Over the past six years, FDA has been working with cancer informatics companies, drug manufacturers, and academic researchers in a collaboration led by Friends to develop and validate real-world endpoints—routine care analogues to traditional clinical trial metrics—and determine whether they can be used to study treatment effectiveness and inform future drug approvals based on real-world evidence:

• In 2018, Pilot 1.0 concluded that it was possible to identify a high level of shared characteristics across varying data sets—demonstrating that it is feasible to extract data about specific patient populations from disparate sources of data.
• In 2019, in the early stages of Pilot 2.0, the data groups developed a set of common definitions for real-world endpoints—including real-world overall survival and progression-free survival—and other non-traditional endpoints that have the potential to serve as proxies for conventional endpoints (The Cancer Letter, Nov. 22, 2019).
• In 2020, Pilot 2.0 demonstrated that the magnitude of benefit obtained from immuno-oncology treatment regimens in the real world—i.e. measured via real-world overall survival and time to treatment discontinuation—is directionally similar to outcomes in a conventional clinical trial (The Cancer Letter, Sept. 25, 2020).
• In 2023, the rw-Response Pilot study provided evidence that tumor response rates can be assessed across real-world datasets (The Cancer Letter, Oct. 6, 2023).

“I think the FDA is very firm on this,” Sigal said. “We really do need this data in the real world to see what happens to patients that were not on the trials.”

Oncology isn’t currently maximizing the utility of real-world data, Herbst said.

“We’re not going to learn it solely through clinical trials. Trials are too hard to accrue to, they’re too expensive, and not everyone is eligible to be on a trial,” Herbst said. “In fact, I’d rather see patients being treated on-label so that we can get them the drugs. But, we have to continue to learn and monitor. And I think that’s something that groups like Friends and others can help with.”

Herbst and Sigal spoke with Paul Goldberg, editor and publisher, and Matthew Ong, senior editor of The Cancer Letter.

Paul Goldberg: What projects and directions in oncology are the both of you most excited about? What are you planning to do in the coming year?

Ellen Sigal: We’ve seen remarkable progress on many fronts in cancer research, but additional policies and programs can help ensure that new discoveries reach patients as broadly and efficiently as possible.  This has led us as an organization to pursue a diverse portfolio of initiatives.  These include projects related to real-world evidence, diagnostic test optimization, use of digital pathology tools, cell and gene therapy development, and efficiency in clinical trials.

Roy Herbst: First of all, I just want to add how delighted I am to join the board of Friends of Cancer Research. I’ve been working with Ellen for many years, and we will talk about some of those projects, but I think as we move forward, we’re in an era of precision medicine, but we don’t yet routinely use it in our daily practice.

I think one of the areas that I’d like to focus on and help with is access—to make sure all patients get access to care, get access to precision therapy, meaning we profile their tumors to help better guide the new targeted therapies and immunotherapies. Some of the work Friends has already done has suggested to us new biomarkers such as ctDNA that we can use to follow patients, but we need to move to the next level and understand why patients are resistant to immunotherapies and find the right drugs or combinations to give them more effective therapy.

There’s so much we can do to advance key research and policy initiatives through Friends and their public-private partnerships by bringing groups together, that I’m so excited to be working even more closely with this team.

ES: There continues to be collective opportunities to improve clinical trials.  In the past several years we’ve been part of efforts to expand eligibility criteria to make trials more accessible and develop strategies to optimize dose determination.  These have helped improve trial designs, but we first had the opportunity to work with Roy regarding innovative approaches to operationalize clinical trials, and some of those initiatives are still continuing today.

RH: Yes, Ellen called me about 11 years ago and asked me to join a project that she was presenting at the Friends Annual Meeting to construct a “master protocol” for lung cancer. At the time, it was still early in the era of the EGFR inhibitors and mutations in lung cancer—work that I was closely involved with our BATTLE trial with Waun Ki Hong at MD Anderson, and ALK and ROS1 and the potential for other targeted therapies.

But the question was, how could we test some of these agents in large enough clinical trial populations, because many of these alterations impact fewer than 10% of the population. So, we first discussed how we could bring a group together—clinicians, physician-scientists, statisticians, regulators and the community—and we formed a panel and that was the beginning of the Lung-MAP, the Lung Master Protocol.

And then, of course, the thing that I think is extraordinary is, we actually did it, and today more than 4,000 patients have been accrued to Lung-MAP, proving we can find patients, we can profile their tumors, and then put them on the right study. I was honored to be the founding PI of this study and to lead the steering committee for more than a decade.

ES: The unique part of Lung-MAP is that it truly is a public-private partnership between the National Cancer Institute, all of the NCTN cooperative groups, the Food and Drug Administration, the Foundation for NIH, Friends, patient advocates, dozens of pharmaceutical companies, and diagnostic testing through Foundation Medicine—and the point is to get things done that you can’t do it with just one stakeholder. And by partnering, we were able to expedite these master protocols and implement the necessary culture shifts across the board that otherwise would be very difficult.

RH: Lung-MAP has now led to a subsequent collaboration called Pragmatica.  As part of a recent Lung-MAP sub-study, the combination of ramucirumab, a VEGFR2 inhibitor, and pembrolizumab, an anti-PD-1 agent, together showed promise in patients with refractory lung cancer who had already had immunotherapy. And with this trial, maybe we can tell the story, Ellen, about what we did when we had that positive result and it was presented at ASCO.

I remember calling Ellen late one evening and we both said, we can’t stop here. We need to figure out how to get to the next step as quickly as possible and to get new therapies to patients! This is the evolution of the Pragmatica Lung trial, working closely with our FDA colleagues who approached us with this new concept. We then immediately took it to the NCI and our NCTN team-and of course the pharmaceutical companies.

We have a real issue with the complexity of clinical trials and getting patients on trials. Despite everything we’re doing, only about 5% of eligible patients are going on clinical trials, and of course fewer in underserved populations. So, we hoped the Lung-MAP effort would help with that. It also helped with trial diversity and getting drugs to the point of care. We saw that there were more patients from underrepresented minorities that went on Lung-MAP than standard cooperative group trials.

But now, with Pragmatica, we are promoting the idea of doing a trial that is more simple since both the drugs are already available individually as standard of care. The only unique thing is combining them. We prepared a trial document of only six to 10 pages as opposed to 40-50 pages, by having fewer eligibility criteria and importantly many fewer endpoints.

Of course, we have to make sure each patient is able to tolerate the agents but the goal is to collect only the essential data.  Frankly, in this trial we want to as quickly as possible confirm the survival endpoint-everything else is secondary and won’t matter if the study does not meet the survival endpoint.

We’re comparing to a dealer’s choice control option. The physician can pick whatever they want as their standard of care, and we’re only collecting survival data because that’s really what’s going to matter here. If we see a positive survival benefit, that option will go forward, and it will be available to more patients. If we don’t, we’ll move on to the next option.

ES: I think part of the innovation comes from leadership also. This was something that Rick Pazdur and Harpreet Singh brought to us, and they were very anxious to implement a more streamline approach to this study that minimized data collection, expanded patient eligibility, and would be available at a large number of non-traditional sites.

They wanted to understand if the initial overall survival would really hold up. And we immediately got in touch with NCI leadership, Doug Lowy and Jim Doroshow, and said, “Look, this is really important.” Jim was and Doug recognized what an important opportunity this was.

We then went to the trial leadership and the companies. I think the companies were on board. They were a little curious. This isn’t the type of thing that they normally do, but the fact that leadership was very involved from FDA, both Rick and Harpreet, made a huge difference. Roy and I were really on top of it and we got this through very, very quickly.

But I don’t think you can do anything that really changes culture unless you are on top of it and you have leadership, passion, and you stay focused on it to make sure that you’re going to execute. And I think that was there from the very beginning with Pragmatica. And I hope with Roy’s help and with Friends, we’re going to do a lot more of that.

RH: Yes, I’m looking forward to that. I haven’t told you this, Ellen, but we had our Pragmatica monthly call yesterday—the trial opened in March of last year. So far, we have about 10 months of accrual and the trial already has 300 patients, which is pretty darn good. It’s really accruing quickly, which is an early validation of this new concept.

It is also available at many locations bringing these drugs to the point of care and serving a much more diverse population.  It’s almost at the halfway point and has enrolled almost twice the number of Black or African American patients than we would see in a standard study. So, while early, many of the trial goals are being met.

Will the combination pan out? I don’t know. We’re doing a hypothesis-based trial based on a phase II trial, but is the paradigm working? So far, so good. And this is a true public-private partnership—on our monthly call we had individuals from major pharmaceutical companies, representatives from all the cooperative groups, statisticians, and the NCI. So, this is a wonderful new model.

With a group like Friends, through the meetings, the policy papers, the advocacy, hopefully this becomes even more prevalent in other tumor types. Right now, we’ve been called by people who are researching melanoma and other tumors, wanting to do very similar things. Our goal is to change the standard of care and to get drugs to patients with cancer as soon as possible.

I was just in clinic yesterday, I care for patients who have lung cancer—our new immunotherapy is amazing—I’ve personally seen so much change over the last 25 years, since I was a fellow at Dana-Farber. Back then when someone had lung cancer, we offered cytotoxic chemotherapy and few patients lived for even a year. But now the results with immunotherapy are phenomenal, but still, only 15-20% of patients see long term success from their treatment.  We are seeing similar results with targeted therapy but again resistance is the main problem- so guided precision medicine clinical trials are a must.

We have come so far but there is still so much to learn—we need to do much better. We must develop drugs that are less toxic. We need to know which agents to use and how long to treat, and we have to make sure all patients have access to these newer treatments. These are all issues that I’m thrilled that Friends covers at its many panels and conferences. Now I guess, like Ellen, I’ll be at all of them!

Matthew Ong: What challenges, politically and scientifically, do you anticipate in the short term? How will they impact your efforts to accelerate cancer research and drug development?

ES: Well, I think one of the biggest problems we’re facing now is declining budgets. And that makes it very difficult, because the basis for any discovery is strong support for basic science and translational research. Second, is the overall view of science by many policy makers.

Support for NIH/NCI and research for many years was an area for bipartisan agreement and a source of pride.  Recently though, there has been a growth of misinformation that has, for some, led to a mistrust of science and reluctance to support our research and public health agencies.  This is something that as a community we need to work together to inform and repair as quickly as possible.

Despite these macro challenges, technological advancements will help facilitate progress in research. For example, we’re all incredibly interested in the role that AI and real-world evidence can play in advancing research. We have these partnerships in both of these areas, and I know that Roy will be very involved in them. We’re looking toward the future for digital pathology tools, specifically related to a new collaboration that we just launched to develop strategies for validation of digital pathology tools used to aid patient selection in cancer research and care.

RH: I would agree. I’m very excited about that because here at Yale, Chen Liu, our chair of pathology and David Rimm are leading a large effort in digital pathology. So, to extend the efforts that we have here at Yale to a nationwide or worldwide effort, working with Friends, is so exciting to me.

But I’ll give you two operational challenges to clinical trials that we’re working to address. The first is through Lung-MAP, with 4,000 plus patients over 10 years, we needed to find companies that had drugs that were ready to bring into this study where we could find the right biomarker and treat the patient. And I’ll tell you, finding the right drugs at the right time and doing it quickly is a challenge. And we can’t get agents onto trial fast enough. Patients, unfortunately still die with lung cancer; there are over 150,000 deaths a year in the US alone. There are many patients that are, even in this era of immunotherapy and targeted therapy, still not doing well.

So, we have to find these new agents, and entice companies to work with us. We need to be quick; we must have endpoints that matter and that can quickly get drugs approved. There’s the challenge for us as a community.  Our public-private partnerships need to come together to achieve that goal. We’re having a retreat in a couple of months in Washington to all get together to determine, ‘What’s the most exciting science that we can bring quickly into the Lung-MAP, so we can impact patient care in the most effective way?’

PG:  What about real-world evidence and real-world endpoints? How far are we from seeing them really being used?

RH: I think they’re already being used. I know that they help with approvals in other countries. I’ve seen pharma doing that. Certainly, once a drug gets approved, there’s very little data. Dr. Rob Califf talked about this at one of Ellen’s meetings recently—the drug’s approved and it’s effective, but we still don’t know how long to use it for.

We’re doing better with dosing, but still there are questions and the setting that the drugs should be used. So, continuing to follow up with real-world evidence gives us much more information regarding how to use cancer drugs.

We’re not maximizing what we can learn and we’re not going to learn it all through clinical trials. Trials are too hard to accrue to and they’re too expensive. Not everyone can be on a trial. In fact, I’m happy to see patients being treated on label so that we can get them the drugs. But, we have to continue to learn from patients on treatment and monitor them. And I think that’s something that groups like Friends and others can help with through many of their projects, meetings, and other endeavors.

ES: Yes, Roy brought up a very important point. So, we finish a trial and a drug gets approved, and that’s a success. Following that, it’s important to be able to see what’s happening in the real world and get additional data to extend our understanding, optimize use if needed, and be able to examine effectiveness in broader patient populations. That’s something you can do effectively with real-world evidence.

RH: I wonder if we could do better in this country. Other countries, I think, do it better. We need to create a database where we capture all the data on patients that receive treatment. In some places, we have access to insurance status and Medicare and Medicaid data, but we really need a way that we can track patients who are being treated.

If we took advantage of the large numbers of patients on treatment and looked at a drug like pembrolizumab, for example, one of the drugs in Pragmatica, we would know more about biomarkers. We would know more about length of treatment. We would know more about re-treatment. We would know more about how to treat toxicities.

We’re not maximizing what we can learn and we’re not going to learn it solely through clinical trials. Trials are too hard to accrue to, they’re too expensive, and not everyone is eligible to be on a trial. In fact, I’d rather see patients being treated on-label so that we can get them the drugs. But, we have to continue to learn and monitor. And I think that’s something that groups like Friends and others can help with.

ES: Yes, I think the FDA is very firm on this, and they should be, that once this is approved, we really do need this data in the real world to see what happens to patients that were not on the trials.

Our goal from the very beginning of Friends has not changed, and that’s to really make life better for cancer patients and to get to innovation and to get better treatment. We started with advocacy, which is still important in funding, but we shifted to the regulatory and to innovation. And that’s where we hope to really get new treatments that are better. And we can detect earlier and help patients at an earlier stage where we know that they will do better.

MO: I’ve been covering Friends’ work on real-world evidence for years now, and there seems to be an ever-growing need for databases that combine biological and clinical data—including for studying and validating real-world endpoints. Should industry—both data companies and drug sponsors—be investing more in these datasets?

RH: I’m not speaking for Friends because I haven’t been to a board meeting yet, so I’m speaking from my own experience. As a clinical trialist who has led a large number of studies and has devoted my career to biomarkers, we have to learn from our patients who have been treated on studies and the tissues collected for biomarker analysis.

Now, the problem is not hypothesis-based because we collect the tissue, but we’ve got to then use it. We have blood markers now. We have tissue markers. We have to take a study with pembrolizumab, a study with durvalumab, a study with atezolizumab, all different drugs, but similar enough that they need careful investigation, and the data needs to be evaluated and pulled together.

We’re collecting all those data, but we’ve got to use the tissues in a rational way. In fact, if we don’t associate these biomarkers with patients who are responding or have long-term stable disease or progress, we’ll never figure out what to do.

And then, we can design the prospective trials. Then, we have a smarter Lung-MAP. We have a smarter Pragmatica, because we’ve learned and adjusted based on our data. This is likely our next step and will need involvement from even more companies and groups. Companies like Caris, Tempus, and Foundation Medicine. Everyone has IP, proprietary knowledge, but we need to pull resources together for our patients.

We have to figure out how to develop these biomarkers. I can’t stand the fact that I give immunotherapy to patients with PD-L1 zero. So, there must be something else going on that immunotherapy and chemotherapy is working. Why four or five years in don’t we understand why yet? And so, we need to be smarter, and we’ll be smarter by collaboration, by partnership.

PG:  To wrap up, what are you most excited for as you expand ways that you’re able to work together?

ES: Well, I think also what Roy brings to the table, beyond the science and his clinical trial expertise, is he never forgets what we’re doing this for. The science is important. The processes and the speed are important, but ultimately making life better for patients is really what we are about.

And Roy, as a physician, we’ve seen over and over again, that whenever someone needs help, whenever there is a question, he will 24/7 be available. So, that humanity is what we’re looking for too. We’re looking for good science and leadership, but we’re also lucky to have his humanity.

RH: When I trained many years ago, Emil “Tom” Frei was my mentor, and he always said, “Bring your best therapies earlier.” So, I think what we’re finding now is to get more impact, we’re bringing immunotherapies, targeted therapies earlier in many diseases.  But this is going to take new tools, new approaches, and perhaps even new policies.

What Ellen’s taught me, and this is why I love working with her, it’s about the patient. We need to do what’s best for the patient and to help the patient. All of us have dealt with friends, family, maybe ourselves who’ve had cancer.

But oftentimes, you need to have a common broker to instigate success. Someone has to pull it all together. And I think Friends really serves an amazing role doing that, and I’m excited to be part of the group at an even higher level.

https://cancerletter.com/conversation-with-the-cancer-letter/20240209_4/