Recently, Ravi Parikh, MD, MPP, an Assistant Professor of Medical Ethics and Health Policy and Medicine at the University of Pennsylvania, Perelman School of Medicine, released a research letter titled, "Exposure to US Cancer Drugs With Lack of Confirmed Benefit After US Food and Drug Administration Accelerated Approval." In this interview with OncData, he further discusses his research and gives a thought-provoking analysis of what these findings could mean for the future of the accelerated approval program, as well as how the multidisciplinary team could better advocate and let their voices be heard for more quality clinical trials.

Oncology Data Advisor: Welcome to Oncology Data Advisor, an expert resource for the multidisciplinary cancer team. Today, I am joined by Dr. Ravi Parikh, an Assistant Professor in the Department of Medical Ethics and Health Policy and Medicine at the University of Pennsylvania, to talk about his research letter regarding lack of benefit from accelerated approvals and negative confirmatory trial data.

Thank you for joining us today, Dr. Parikh. Would you like to introduce yourself and tell us a bit about your interests?

Ravi Parikh, MD, MPP: Sure. I'm Ravi Parikh. I'm a Medical Oncologist and a Health Policy and Clinical Researcher at the University of Pennsylvania. A lot of my research focuses on applications of real-world data to understand oncology prescribing patterns and particularly how we can encourage high-value utilization of drugs and discourage low-value utilization of drugs in clinical practice.

Oncology Data Advisor: Incredible. Again, great to have you here. So, we are here today to talk about your research regarding the lack of benefit of accelerated approvals from the FDA. Would you like to give us a brief overview of your research?

Dr. Parikh: Sure. So, the FDA accelerated approval program, which began many years ago, is intended to speed up access to novel therapies that show promise in non-randomized, largely phase 2 trials and accelerate market access to those drugs prior to them undergoing standard of care confirmatory phase 3 testing in comparison to a standard of care. This program has been a good thing; it's been associated with the entry of drugs into the market earlier, and usually, those drugs are confirmed to have benefit in phase 3 testing. So, it really is earlier access to drugs that can be beneficial to patients—and I've seen this in my own clinical practice.

One of the issues that's come up with the accelerated approval program is that about a quarter of indications that are approved in the oncology setting are subsequently withdrawn, meaning that they don't show benefit in confirmatory phase 3 trials. This is, in some ways, a problem, because it exposes patients to drugs that may not be beneficial over standard of care and may, in fact, be either harmful in comparison to a standard of care, or pose adverse events and pose cost to the health care system. Trying to redesign the accelerated approval program so that it can still maintain its benefits of speeding up access to effective drugs, while figuring out ways to discourage access to ineffective drugs, is a big policy interest of the FDA, of federal legislators, and to our research group.

Oncology Data Advisor: Incredible. Yeah, it's definitely such an important topic to bring up. So, thank you for your research and your team's research. Next question I had was, I'm wondering, what was the genesis of your study and why you felt it was important to bring awareness to the situation?

Dr. Parikh: One of the criticisms around drugs that are withdrawn after receiving accelerated approval is that those indications have the potential to expose patients to drugs that are subsequently ineffective. But we don't really have a good sense of how many patients, and what percentage of patients, are exposed to an ultimately ineffective or withdrawn accelerated approval drug. There have been some studies out there that try to estimate cost of accelerated approval drugs and try to estimate other aspects, but none have really been able to estimate the proportion of patients who are exposed. That's because they don't have granular enough data to do so. So, we partnered with a real-world data company, Flatiron Health, to actually be able to identify patients who would qualify for those accelerated approval indications, specifically those who would've qualified due to line of therapy settings or due to certain biomarker status.

We were able to estimate the proportion of patients who were exposed at a much more accurate level than previous studies have. And arguably, I think we've come up with the highest level of data that shows what percentage of patients are truly exposed to a subsequently withdrawn drug. What we found is that, when we studied five indications of accelerated approval of drugs that were subsequently withdrawn in the oncology setting, around a quarter of patients—26% of those who were eligible for the indication—were exposed to it during the time between when the drug was approved and when it was subsequently withdrawn.

Furthermore, there is this sort of interval after accelerated approval where these withdrawn drugs ultimately may get some negative confirmatory trials. Before the publication of those negative confirmatory trials, that number of patients that were as exposed was almost a third, 33%, as opposed to 25%. So, that period between when the accelerated approval occurs and when the confirmatory phase 3 trial is reported is a very high-leverage window, because it's an evidence-free zone or it's a point where we're not having the full formal phase 3 evidence. And it tends to be a time when a lot more patients are exposed to these ultimately ineffective drugs, compared to a standard of care.

Oncology Data Advisor: Were there any limitations you experienced when conducting this research?

Dr. Parikh: Yes, so I would say that one, since we're looking at real-world practice, there may be some reasons why patients got these drugs, which were largely immunotherapies, that had nothing to do with the accelerated approval indication. Maybe, for example, a patient had Lynch syndrome or maybe, for example, they had other biomarker-based studies that qualified them to use the immunotherapy. So, we weren't able to estimate that, because sometimes our biomarker data was a little sparse for patients. There may be some reasons why this 26% number may be a little bit of an overestimate, although we don't expect that it's a significant overestimate. So, that's one potential limitation. The other limitation is, "Were you only using data within the Flatiron Health Network?" So, although that covers about 25% of all community oncology practices in the United States, it is possible that these findings may not be generalizable to practices that are outside the Flatiron Health Network. And that's something that I think will be an important limitation for this use.

Oncology Data Advisor: Speaking on the patient's perspective, how do you explain the negative results to the patients and potential changes in their treatment regimens going forward? And do you change the treatment regimen?

Dr. Parikh: In my practice, when I get a patient for whom we had previously discussed using a drug that was approved in the accelerator approval program and then there's been a subsequent indication change, what I normally will do—what I've done in that scenario—is explain the rationale behind the indication change and what new data has come about that's changed our thinking for it. And then we engage in shared decision making with the patient to figure out whether they want to use this drug that may not be better than standard of care. I think there are a couple of important caveats here. One is that when a patient has already started with a line of therapy for an accelerated approval drug that's subsequently withdrawn, it's rare that I will take them off that drug, because usually, I want to figure out whether these drugs are having any response in this particular patient.

And that gets to a broader point, that even if these drugs "fail" confirmatory phase 3 testing, that doesn't mean that the accelerator approval drugs are ineffective. They still may be effective for a certain proportion of individuals. It's just that when you run your phase 3 trial of all comers with one drug versus all comers with another drug, there's no benefit that's found in that general comparison. That doesn't mean that there aren't select patients for whom these accelerated approval drugs can be quite effective. So, I'll never interrupt a course of therapy, but for a patient for whom I was previously thinking, if I get new data that suggests that the accelerated approval drug doesn't work as well, then I will likely change my recommendation and go back to the standard of care or go back to another drug that has shown benefit in the phase 3 setting.

That is normally how I'll practice. I think the other thing to sort of realize too is that just because the drug fails confirmatory phase 3 testing, what it usually means, to me, is that there's been some evidence of benefit shown in phase 2 that's not born out in phase 3, but we need to be doing a better job in identifying better biomarkers and to be identifying the patients who may benefit from these accelerated approval drugs in the future, rather than entirely casting that drug aside.

Oncology Data Advisor: Going off what you were talking about, from the clinician's perspective, do you and your team have any potential strategies for minimizing exposure to accelerated approval therapies with lack of benefit and negative confirmatory trial data?

Dr. Parikh: I think that really forces us to step into the role of the FDA and what they do, because once a drug receives their approval, we're oftentimes obligated to sort of at least explain what the treatment options are, including that accelerated approval drug, as a clinician. I think, really, the change ought to occur upstream at the FDA level primarily. So, what could the FDA do? Well, they just recently passed an act called FDORA, F-D-O-R-A, that tries to address some of these limitations of the accelerated approval program. Namely, they propose that drugs should be already undergoing confirmatory trial testing at the time of accelerated approval, rather than waiting to start a confirmatory phase 3 trial. That could minimize the amount of exposure to a drug that may potentially be withdrawn later, because it speeds up the conduct of the phase 3 trial. It sort of is a contingency for having begun that trial before that drug received accelerated approval, which is really important.

I think the second thing that the FDA could do is use harder end points in the phase 2 setting to justify accelerated approval. Right now, oftentimes, we use surrogate end points based on phase 2 testing to grant accelerated approval, but using things like overall survival or other surrogate end points and comparing them to historical controls, or even running a simultaneous historical control in the real-world setting during phase 2 testing, may be promising strategies to get a stronger sense of efficacy of these drugs prior to accelerated approval. That's at least on the FDA side. On the clinical side, I think that we can do a better job with explaining to patients what the data is actually showing, rather than just saying, "This drug is effective and might save your life."

We might say, "This drug was approved based on an end point that is not overall survival. Maybe it was approved based on progression-free survival or response rate. So, I don't know if that drug is actually going to improve your personal survival. It improves some aspects that aren't survival, but it's tough to know whether it's going to improve your overall survival. However, these are the reasons why I think it may be beneficial for you or why it might pose less side effects than the other types of therapy for you." And then we leave it to the patient to decide, in light of all of that information, what they may want to do. So, I think we could do a better job of understanding the evidence so that we're presenting the evidence in favor of the drug correctly, rather than just saying "All of these drugs are going to save people's lives."

Oncology Data Advisor: Brilliant. And you just started talking on what I was going to ask you next, but I'm going to ask you anyway, just in case it brings up further thought. How do you feel clinicians could better recognize these lack of benefit trials and advocate for more thoroughly investigated approvals?

Dr. Parikh: Yeah, it's a great question. I would say two things. First, clinicians could recognize potential limitations on the data, just by looking at the end point that's cited in the phase 2 trial that's justifying the accelerated approval and seeing whether that end point is really meaningful or not and whether it's in line with what you might think. Just because a hazard ratio is a certain level or not, that doesn't necessarily mean that the drug is all that much more effective than the current standard of care. I think that clinicians can also play a strong role and have a strong voice in talking to the FDA about how they ought to be changing some of their tactics: sending letters, participating in open panels, and really making their voice heard, that either you disagree with the surrogate end points or you think that the accelerator approval program has a lot of benefits. Either one of those sides is a totally legitimate argument, but I think we need physicians to be stronger advocates to the regulators, who actually can make change in these types of programs.

Oncology Data Advisor: Definitely. And final question I have for you is, out of all this research, are there any plans for further research? Or are you planning on doing anything else regarding the issue in the upcoming future?

Dr. Parikh: Yeah, it's a great question. So, I would say that there are three lines of research that we're thinking about. First, when it comes to uptake and effectiveness of drugs receiving an accelerated approval, we're hoping to expand the list set to all oncology indications, rather than just the five that we study, to see if our findings are truly generalizable. The second thing that we're hoping to do is to get a sense of metrics outside of just clinical efficacy of these accelerated approval drugs. One that we've thought is adverse events. A very high proportion of immunotherapies are approved by accelerated approval, but accelerated approval isn't usually based on adverse event adjudication. Trying to get a sense of real-world adverse events associated with FDA accelerated approvals compared to current standards of care, I think, is really important. And coming up with better data streams, so that we can actually collect adverse events in a more structured fashion, is important.

Maybe the last thing I'll say is that the FDA has just released a bunch of legislation pertinent to this area, and I think there's a really rich position to have some sort of policy evaluation analysis to help enable the FDA to see whether their programs are working or not, in the short and in the intermediate term. Pursuing some of that experimentation and planning for some of that, given that the FDA has released recent legislation, I think is really important for characterizing how effective that legislation is likely to be.

Oncology Data Advisor: Definitely. That was all I had for you today, so thank you so much. This is such brilliant research; so important to talk about and discuss. Thank you so much, Dr. Parikh, for being here today and for your time.

Dr. Parikh: Great. Thanks very much, and thanks for thinking of me.

About Dr. Parikh

Ravi Parikh, MD, MPP, is an Assistant Professor of Medical Ethics and Health Policy and Medicine and the Associate Director of the Penn Center for Cancer Care Innovation at the University of Pennsylvania, Perelman School of Medicine. Dr. Parikh's research and passion revolves around delivery system reform and informatics, including the use of health technology to improve routine patient care, quality of life and survivorship care in oncology, and payment reform for advanced illnesses. His research, advocacy, and legislation can be found in multiple academic journals, and his legislative recommendations were recognized and earned commendation from the Massachusetts Speaker of the House.

For More Information

Transcript edited for clarity. Any views expressed above are the speaker's own and do not necessarily reflect those of Oncology Data Advisor.