May 26, 2023
The National Center for Health Research (NCHR) appreciates the opportunity to provide public comment on the Food and Drug Administration (FDA) draft guidance on clinical trial considerations to support accelerated approval of oncology therapeutics.
NCHR is a nonprofit think tank that conducts, analyzes, and scrutinizes research on a range of health issues, with particular focus on which prevention strategies and treatments are most effective for which patients and consumers. We do not accept funding from companies that make products that are the subject of our work, so we have no conflicts of interest.
Oncology drugs account for more than 60% of all accelerated drugs.1 There is reason for concern about the benefits of these drugs; research indicates that as of 2019, only 20% of cancer drug indications approved through the accelerated approval pathway from 1992-2017 had demonstrated improvements in patients’ overall survival (OS) based on their confirmatory trial data.2 Our Center reviewed published studies for 18 cancer drugs that had been approved based on surrogate endpoints at least 4-8 years earlier, but had subsequently not been proven to improve OS. Our analysis of post-market studies regarding their impact on quality of life found that only one showed better quality of life than the control group and two showed worse quality of life.3 In a more recent analysis of all drugs granted accelerated approval, there was a significant delay in confirmatory trials demonstrating a clinically meaningful benefit to patients. Of the 100 accelerated approval confirmatory trials completed or due between 2012 and 2021, more than half were past their expected completion deadline set by the FDA.4
The FDA draft guidance document contains numerous recommendations that would greatly improve the quality of research submitted in support of accelerated approval for oncology drugs. For example, there is clear evidence that a randomized controlled trial (RCT) is greatly superior to a single-armed trial. We therefore agree that an RCT is the preferred approach to support accelerated approval applications and we agree with the description of the weaknesses of single-armed trials specified in pages 2-3 of the Guidance. The evidence is clear that studies that depend on only a historical control group for comparison are subject to bias and confounding variables and therefore lack the scientific rigor and validity to support approval. The control arm should always evaluate the best available therapy if there is one, not a placebo or any other treatment.
We agree with the draft guidance that the preferred plan is a single randomized trial design that should serve as the basis for accelerated and regular approval. The FDA guidelines for two randomized controlled clinical trials is reasonable but the probability that the second study will be completed as planned is greatly reduced for a second study that enrolls patients later than the first study.
We agree with the guidance document that endpoints should be determined early and discussed with FDA to decide whether a different endpoint would be more appropriate prior to the start of a clinical trial. However, we have seen numerous recent examples where the sponsor rejects the FDA’s strongly recommended endpoint and experts express concern about the quality of the data, and yet the FDA subsequently grants accelerated approval despite the clear concerns of FDA scientists and statisticians. This sends the wrong message because it creates an incentive for companies to ignore FDA advice. Moreover, research has shown that nearly 20% of oncology clinical trials have changed the designated endpoint midway through the study; these post-hoc decisions are clearly subject to bias and uncertainty when interpreting the results.5
We are concerned about the vague wording in the guidance encouraging drug developers to have confirmatory trials “well underway, if not fully enrolled,” prior to receiving accelerated approval. The wording in section 506C was recently revised to state that the FDA “may require, as appropriate, a study or studies to be underway prior to approval, or within a specified time period after the date of approval.” This gives the FDA the authority to require that confirmatory trials be fully enrolled prior to approval, which would ensure that they can be completed in a timely manner and as designed. As FDA has often acknowledged, when a new oncology drug is approved, clinical trial participants are likely to drop out of the study to ensure that they are treated with the newly approved drug, and not the placebo or older treatment. For that reason, the FDA should explicitly specify that studies need to have a majority of patients enrolled prior to receiving accelerated approval, with a timeline for completion that will ensure that approval will not result in a drop out of patients that would jeopardize the appropriate completion of the study. FDA should never accept a definition of “underway” that includes merely having research staff in place, with protocols and methods developed, since that would not adequately ensure that the study can be completed in a timely manner or with adequate recruitment or retention.
Another major concern regarding the draft guidance is the statement that progression free survival (PFS) is a valid “longer-term clinical endpoint to verify clinical benefit.” PFS may provide short-term hope to patients, but is not a meaningful measure of patients’ health or functioning and does not have clear implications for overall survival.6 As is well established in clinical trials, PFS in the short-term often does not predict OS;6 in fact, initial tumor shrinkage or stability can be followed by rapid tumor growth, or a patient may suffer or die due to the toxicity of the drug or other adverse events from the treatment, rather than the cancer itself.
We generally support most of the views expressed in the draft guidance in the section on Single Randomized Controlled Trials and the potential shortcomings that need to be addressed in Single-arm trials submitted for accelerated approval. For example, we agree that response rates to a treatment, however defined, do not necessarily predict clinical benefit and that the durability of a response may need to be studied for considerably longer than 6 months. However, we are concerned that the wording in these sections is often vague and the tone seems less like regulatory guidance, but rather as friendly suggestions. And, we disagree with the suggestion on page 4 that a confirmatory study that is based on a different indication may be acceptable evidence. The example of studying the same drug on patients who have not yet received any treatment for an indication for refractory disease is worrisome, because the risks and benefits for patients with refractory disease could be significantly different.
In addition, we urge the FDA to delineate their plans to quickly remove accelerated approval if the results of a confirmatory trial are negative. The guidance should warn that automatic withdrawal of an accelerated approval is likely when the confirmatory trial does not have a meaningful clinical benefit.
We understand that an FDA guidance document is not enforced, but we are concerned that the FDA’s draft guidance and repeated focus on flexibility in approving cancer drugs will continue to send the message that FDA will allow large loopholes in scientific evidence going forward. To ensure that the guidance is followed as rigorously as possible, the FDA should explicitly advise that single-armed trials will not be considered acceptable, with the exception of specifically described rare circumstances. For example, the “urgent need” for better treatments should not be adequate justification for accelerated approval when there is no clear evidence that the benefits outweigh the risks of a new drug compared to existing unproven treatments or even to the existing standard of care. In other words, when there are already approved treatments that lack confirmation of meaningful clinical benefits, the FDA should not add to uncertainty by granting accelerated approval to a new drug based on a single-armed trial. Similarly, if the existing standard of care is sometimes (or even rarely) effective, the inevitably questionable results of a single-arm trial would not justify accelerated approval.
1. Schaefer, N. (2022). Oncology has highest number of FDA accelerated approvals. Clinical Trials Arena. https://www.clinicaltrialsarena.com/comment/oncology-fda-approvals/
2. Gyawali, B., Hey, S., & Kesselheim, A. (2019). Assessment of the Clinical Benefit of Cancer Drugs Receiving Accelerated Approval. JAMA internal medicine, 179(7), 906–913. https://doi.org/10.1001/jamainternmed.2019.0462
3. Rupp, T., Zuckerman, D. (2017). Quality of Life, Overall Survival, and Costs of Cancer Drugs Approved Based on Surrogate Endpoints. JAMA Intern Med. 177(2):276–277. doi:10.1001/jamainternmed.2016.7761
4. Deshmukh, A., Kesselheim, A., & Rome, B. (2023). Timing of Confirmatory Trials for Drugs Granted Accelerated Approval Based on Surrogate Measures From 2012 to 2021. JAMA health forum, 4(3), e230217. https://doi.org/10.1001/jamahealthforum.2023.0217
5. Florez, M., Jaoude, J., Patel, R., et al. (2023). Incidence of Primary End Point Changes Among Active Cancer Phase 3 Randomized Clinical Trials. JAMA Netw Open. 6(5):e2313819. doi:10.1001/jamanetworkopen.2023.13819
6. Fleming, T. R., Rothmann, M. D., & Lu, H. L. (2009). Issues in using progression-free survival when evaluating oncology products. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 27(17), 2874–2880. https://doi.org/10.1200/JCO.2008.20.4107