Prevention and early detection studies frequently require engaging physicians from specialties other than oncologists, said Peter O’Dwyer and Mitchell Schnall, co-chairs of ECOG-ACRIN Cancer Research Group.
The logistical problems this creates can be observed in the case of the ECOG-ACRIN study EA2185, which compares two regimens for monitoring pancreatic cysts.
“The individuals who would put patients on this study are not medical oncologists, who are the drivers of most of the therapeutic studies, and they are often in specialties outside the traditional cancer research specialties, in this case, gastroenterology and surgical specialties, but, to a certain extent, primary care,” O’Dwyer said to The Cancer Letter.
“And we’re trying to figure out what are the best structures to allow this to happen? How much of it needs to occur in CTEP versus in the Division of Cancer Prevention or elsewhere, and could we streamline it in such a way as to have something like a ‘registration-lite’ for individuals who are not going to be treating patients, who don’t need to have their pharmacy information in there, for example, and other aspects that are required by our current registration procedures, so simplifying this, making it easier, ultimately for primary care.
“Our focus should be on primary care physicians, because, if large screening trials are going to ever be implemented, we’re going to need the patients—they’re not even patients—the subjects to be screened at the level of their local doctor. So, that’s a big part of the way that we’re thinking about how to engage more people in this process.”
Another challenge would be to figure out how money flows through institutions, Schnall said.
One man’s cure is another man’s overtreatment.
Mitchell Schnall
“Institutions are becoming very complex these days,” Schnall said. “We have to be more agile in how we engage these systems, because the assumption that we can send dollars to the cancer center and they can figure it out, are more complex in this complex environment. And so, we also need the agility to be able to engage the right constituency within systems that are going to be participating in these trials, through contracts and other.
“And again, we’ve been talking to NCI about this, with the same groups, as part of this sort of rethinking, because it is a barrier when you’re trying to engage a group who is part of a system, who has very arms-length relationships to the cancer center—through the cancer center. We then have to sort of negotiate these internal relationships rather than just say, if they’re going to enter the patients, let’s just directly contract to them and get it over with.”
A video of the conversation is available here.
O’Dwyer and Schnall spoke with Paul Goldberg, editor and publisher of The Cancer Letter.
Paul Goldberg: Thank you for agreeing to meet with me, gentlemen. Last March, an ad hoc committee formed by NCI presented a set of recommendations on cancer screening trials. How would you boil down the message of that report? Or maybe there’s another way of asking it: What are the key issues surrounding screening and prevention trials in the publicly funded system?
Peter J. O’Dwyer: Mitch, do you want to start?
Mitchell D. Schnall: Sure. To me, among the most important things that the committee concluded was that the [TMIST breast cancer screening] trial was important, and to me that’s a statement not just of that trial, but of the importance of screening, prevention, to the armamentarium in the fight against cancer. There’s a tendency, naturally, to be focused on patients who are suffering and focused on therapeutics, which are critically important.
But these trials, I think, are equally important in preventing severe disease. And so, I think first and foremost, the statement by the committee that this is an important trial and we should put effort into concluding it was, I think, a very important statement for them to make, and much appreciated. I think much of the other comments made, from my perspective, were fairly high-level comments focusing on the need to do everything to promote accrual and participation in the trial and execution of the trial, as per the protocol, which is frankly a generic comment on any of the trials that we do. I don’t know, Peter, if you want to expand on some of that.
PJO: I think highlighting the importance of screening and prevention was a big part of the main message of the report, and I also liked the recommendation that they look into the framework of the design of these studies, feasibility, and the operations supporting them. I think they put their finger on a difference between the screening and prevention studies and the therapeutic studies, and recognize that they may need to be done a little bit differently.
They also suggested a portfolio analysis, which has implications of prioritization and so on. We have relatively few screening trials in the whole system, so maybe that’s not an urgent requirement right now, but it certainly, with the development of new technology, is going to become so.
What’s ECOG-ACRIN’s history and interest in screening and prevention trials? By the way, we should probably just back up because the report we’re talking about focused on TMIST, where I’ve been talking broader.
PJO: Yes. I should probably address that in terms of ECOG’s history. So, ECOG has participated with all of the cooperative groups in the development of prevention trials, which probably had their peak in the late ‘80s, early ‘90s, particularly, and we had two major contributions.
One was through [Waun] Ki Hong in a number of ECOG institutions in head and neck cancer, the role of 13-cis-retinoic acid in prevention of second primary tumors, and that was a positive study, and I don’t think that it ever wound up in a huge intergroup study, such as, for example, the breast cancer studies, or the prostate cancer prevention trials mounted by NSABP and SWOG.
And we participated in the design of all of those and were involved in those, and our part of that effort, if you like, was for lung cancer prevention, and it was with the selenium trial [E5207], and that was led by Dan Karp, who, as you probably know, came out of Ki Hong’s group also.
That didn’t turn out the way we wanted it, but it was a very successful study in that it accrued very well, and accrued quickly. It had correlative endpoints that were led by Steve Belinsky, and, of course, he was at a SWOG institution at the time, so it highlighted cross-system collaborations, as there were in the other studies, too.
So, that kind of was the history of our involvement in that, and since the end of those studies, until TMIST came along really… well, no, actually, the other screening studies. Mitch, I should hand over to you to talk about the screening studies of imaging in NLST [National Lung Screening Trial] and so on.
MDS: Right, and we on the ACRIN side, we have a rich history in this space; right? Really starting with our first DMIST trial, the digital mammography trial, which interestingly enough, if you look at the proliferation of digital mammography—and it wasn’t a given that it was equivalent even to standard film-screen mammography—there were a lot of concerns. Digital mammography proliferation in the U.S really came concurrent with the release of the results of the DMIST trial, and followed with multiple other screening trials.
We did a CT colonography trial showing that it’s relatively effective for centimeter polyps. We did the NLST trial, which obviously had major impact. Since the merger, we’ve actually done some additional trials.
We did the trial that recently published—the AB-MRI trial for abbreviated breast MR, demonstrating that for patients with radiographically dense breasts seeking supplemental screening, that there was a two-and-a-half fold increase in cancer detection rate by adding an abbreviated breast MR.
So, this screening focus has been part of our history and focus on the ACRIN side for a long time that we brought into, into ECOG-ACRIN together, and sort of culminating both with this TMIST trial as well as the view going forward of an approach to screening that will be much more combined biomarker imaging-based, and the need to sort of integrate blood based biomarkers and imaging approaches together into more effective screening programs, which is something that we’re interested in pursuing.
Finally, the other thing that we’re interested in is trying to think collaboratively between screening and prevention—for a couple of rationales. One is mechanisms used in prevention might also speak to mechanisms used for screening, particularly for use in molecular-based screening approaches. The second is that the screening component of a prevention trial could represent a way to measure endpoints and serve as a sort of a marker for the effectiveness of the prevention. Peter, any other thoughts?
PJO: Well, yeah, you bring up an important nexus, and I think that my predecessor in this role, Bob Comis and Mitch saw the potential of this—and kudos to both because this is really where [we’re interested in] bringing together different biomarkers—in particular imaging biomarkers because they’re non-invasive—and clinical markers for trials of various kinds. I think we’re really at a remarkable moment when therapeutic advances in genomics and targeted therapies and immunotherapy are set to influence strategies for prevention and screening.
And our belief is that these interventions may have an even greater impact in that setting, in prevention rather than in treating established cancer. So, this is the time really for a big focus on this area, and I think that probably underlies the CTAC committee’s sense of the whole field.
Right, this is not the time to abandon it just because the costs are high. Could we establish the context for another trial, which is really what I’m writing about, which is the EA2185 pancreatic cyst surveillance trial, and screening and surveillance trials require participation of medical specialists other than oncologists. So, what obstacles at NCI may prevent other medical specialty groups from getting involved? I mean, the radiologists have done pretty well, but who else is out there that you need?
PJO: Mitch? Do you want me to take a stab at that first?
MDS: No, go for it. I’ll pitch in.
PJO: You know, this study is a multi-level study.
I think that it’s really unique in its screening and surveillance focus, because it has a primary imaging entry. It addresses the biology of cancer risk. It has correlative studies that’ll come around behind the positives on this study and analyze various aspects of risk in the context of this screening.
So, it has value at a number of levels. I think the context that you bring out, and that is that the individuals who would put patients on this study are not medical oncologists, who are the drivers of most of the therapeutic studies, and they are often in specialties outside the traditional cancer research specialties, in this case, gastroenterology and surgical specialties, but also maybe less for this, but to a certain extent primary care.
And all of these different components of the medical system need to be considered in how the structures are put together to make these trials happen. Historically, the therapeutic approaches have driven those structures, and really a one-size-fits-all to bring in other specialties may not work as well as it could with a different approach.
This is kind of an unintended consequence of the history of where we are in this type of cancer research, and so we’ve been in discussions with, particularly with Phil Castle, Ned Sharpless, and Worta McCaskill-Stevens, but also on the CTEP [Cancer Therapy Evaluation Program] side, which regulates the registration of investigators, with Jim Doroshow and Meg Mooney.
And we’re trying to figure out what are the best structures to allow this to happen? How much of it needs to occur in CTEP versus in the Division of Cancer Prevention or elsewhere, and could we streamline it in such a way as to have something like a “registration-lite” for individuals who are not going to be treating patients, who don’t need to have their pharmacy information in there, for example, and other aspects that are required by our current registration procedures, so simplifying this, making it easier, ultimately for primary care.
Our focus should be on primary care physicians, because if large screening trials are going to ever be implemented, we’re going to need the patients—they’re not even patients—the subjects to be screened at the level of their local doctor. So, that’s a big part of the way that we’re thinking about how to engage more people in this process.
MDS: There’s another corollary to what Peter said, and that is, in addition to the NCTN structure, which assumes broad participation across the whole portfolio of trials, it also sort of has a single point of contact into a health system or into an institution, and institutions are becoming very complex these days; right? I mean, I’m in the University of Pennsylvania Health System.
The University of Pennsylvania Health System is six hospitals, and we’ve got the academic radiology group, we’ve got the private group, and same thing in multiple specialties.
And so, we have to be more agile in how we engage these systems because the assumption that we can send dollars to the cancer center and they can figure it out, are more complex in this complex environment. And so we also need the agility to be able to engage the right constituency within systems that are going to be participating in these trials, through contracts and other.
And again, we’ve been talking to NCI about this, with the same groups, as part of this sort of rethinking, because it is a barrier when you’re trying to engage a group who is part of a system, who has very arms-length relationships to the cancer center—through the cancer center. We then have to sort of negotiate these internal relationships rather than just say, if they’re going to enter the patients, let’s just directly contract to them and get it over with.
Well, let’s say I’m a primary care physician. What would be my obstacles in participating in, say, any of the trials? I don’t want to choose one, because I’ll probably choose the wrong one.
PJO: Under current structures, you mean?
MDS: Yeah. I mean, you’d have to be registered in the NCTN, which basically means you’ve got to be qualified to do an NCTN study, including—because I go through this as a radiologist—identify the pharmacy they’re going to send the chemotherapy agents to.
PJO: The IRB.
MDS: Right. Deal with a ton of paperwork to show that I’m capable of being able to treat patients on NCTN trials. Then I would also have to figure out how I’m going to work with my cancer center, so that if there’s accrual [that] happens and the dollars will go to the cancer center, that I actually get my costs reimbursed, for me doing that trial.
So, you can imagine if I’m a primary care doc, this sort of bar that holds. And what it means to us as a group. It means that we’ve got to put a ton of resources into that primary care doc to educate them on what it is to be in the NCTN, to walk them through a bunch of paperwork, most of which are irrelevant to the trial they’re going to go on that they’ve never seen before, heard before, and really are scared of, to work and try to create a relationship between them and the cancer center that would keep them whole for their costs and participating in the trial. It really creates challenges to getting these folks on these trials.
Are NCORPs going to have an easier time or a harder time? Community docs, non-academic docs.
MDS: I don’t think it much matters. What I say is that, the very large institution… I think it has to do, to a large extent, with internally the structure of your organization and how you operate, but if you’re an arms-length specialist, in terms of participating, normally in cancer trials, you’re going to be, it doesn’t matter whether you’re in a big or small place, you’re going to suffer from these concerns.
PJO: Ultimately, it’s really important not to reinvent the wheel here. It’s taken us 50 years to get to this point where we have an established nationwide system of cancer research that really works, and so tinkering with the existing system rather than trying to stand up a different but parallel system, and I think that it’s logical, seems to us the best way forward. And I think we’ve embarked on that over the last year or so, and we’ll see some results, I think.
MDS: That’s worked in the past, right? Before the formal NCTN, there was a little more latitude in how people could get rostered for studies, and there were special rosters developed. We at ACRIN had the latitude to be able to engage sites this way. I know speaking with some of our colleagues in the NCTN, the legacy trials, like the STAR [Study of Tamoxifen and Raloxifene] trial and some of the other breast cancer prevention trials, they were able to separately roster people that way. So, in legacy before the formality of the NCTN, there was some of that flexibility, which was effective.
So, the recommendations sought by NCI prior to this focused on TMIST. How would you apply those recommendations globally to EA2185? We touched on some of this, but maybe there’s something we forgot.
MDS: I mean, I think again, we touched on a lot of this. I think a key would be to be able to create a separate, if you will, “registration-lite” program, and to be able to allow direct engagement, in this case [with] gastroenterologists and the GI surgeons that are going to primarily be seeing these patients, and that way we could greatly expand the participant pool and be closer to where these patients are.
And what kind of changes are you making to your studies? I guess we’re getting back to TMIST, that’s probably where it mostly needs to happen. Where is it now? I haven’t looked at it in about six, seven months, so I probably missed a bunch.
MDS: Well, I mean, TMIST is accruing quite well now, we’ve done some things based on recommendations to the committee and frankly things we were doing ahead of time, and I think you and I may have talked about that.
We started to work on some redesigns and have negotiated some of those out with NCI to decrease the total sample size a bit. The work that I described in getting sort of the non-standard NCTN sites accruing was work that we’ve been doing all along in TMIST, but you can imagine, each site getting them up and running because of the effort that it takes, was quite a bit delayed. But now we’ve got, we’re solidly accruing between 2,000 and 2,500 patients a month. I don’t remember the latest, but I think we’re over [56,000] patients in at this point, and so we’ve been on solid footing on the accrual in TMIST in the last six to 12 months.
So, at this rate, you would be expecting results when, roughly?
MDS: So, we’re projecting about two years ‘til we finish accrual, so by now probably about 18 or 19 months that we’d be finished and then we’ve got follow up after that. So we’d be expecting results in about, with all the follow up (it’s a funny design, because it’s a design that you could end at any time) but something of the order of four years of follow-up, so probably about six years from now.
What about European data? Can you pool it with Europe or not? Or is that still being discussed?
MDS: It’s being discussed. There’s a lot of nuance about this design—because it’s not diagnostic-based; it’s outcome-based, which I think is becoming even more and more important as we think about a lot of the challenges that the next generations of screening are going to create and then how important it is for us to be outcomes-based.
As we become better and better at finding disease, we want to find a disease that matters. So, we’re looking and we’ll consider anything, but at this point I don’t know that we’ve found some slam-dunk data that we can pool.
Is there anything we forgot?
PJO: Well, I would say that in terms of the other changes that we’re making to our studies,[it] is [an] all-in approach, particularly for studies that are going to be challenging or complex. We develop study teams, we tune the composition of those study teams to what skillsets are required to manage them.
I know Diane [Dragaud] is on the call. We have a big focus on communication, and I think we’re getting better at communicating with our sites and with potential participants, and that’s a lot of resource. But we think it’s worthwhile and it’s something that what we’re doing for EA2185.
Oh, that’s interesting. What do you need to do there?
PJO: What we need to do is to get to a critical mass of participants. Within the existing participants, we’re looking at novel methods of alerting people to the existence of the pancreatic cyst—because it can be missed, and because there are criteria for how long the cyst has been there that make a patient eligible or ineligible and what kind of other tests they might have had. So, some of this is actually amenable to an algorithmic approach within particular systems, and that’s the approach we’re developing.
We think that in that way, I don’t know if [PI] David [Weinberg] mentioned it, that the potential exists for EA2185 to be a kind of a laboratory environment for what may work in the direction of accruing patients to screening studies. Can we be smarter about it? And that’s something we’re going to try and work out within that trial.
MDS: Leveraging the EHR to identify candidates in ways that may be valuable, because we’re looking through broad populations.
The question, the research question is really fascinating to think about, because, is this stuff benign, or is this going to kill you?
PJO: It’s a complex design and those are exactly the issues. And that’s why it’s valuable.
Our focus should be on primary care physicians, because if large screening trials are going to ever be implemented, we’re going to need the patients—they’re not even patients—the subjects to be screened at the level of their local doctor.
Peter O’Dwyer
MDS: And the challenge, and I can tell you, because I do this for a living, is that the imaging studies are becoming so good that, inadvertently, we’re finding pancreatic cysts are incredibly common, and we’ve got to get our arms around figuring out how to manage these. Because every time somebody’s getting a workup because of hematuria, we’re looking at their kidneys. Every time somebody’s getting a workup because of GI distress or because their biliary tract is being worked up, right, we’re seeing the pancreas.
And there’s almost an epidemic of pancreatic cystic lesions, and frankly, this is emblematic. There’s corollaries in other areas as well, where imaging is so good, there’s a proliferation; of small renal cell carcinomas, for example. And it’s not because there’s suddenly a proliferation, it’s because we’re seeing stuff that we’ve never seen before and trying to figure out how we should be caring for these patients is absolutely crucial.
Yeah. You’re talking about Whipples, I mean that’s not [trivial].
MDS: That’s exactly my point. One man’s cure is another man’s overtreatment; right?
Well, thank you so much for talking with me and I look forward to writing about this.
