Endpoints in Cancer Trials: Putting Patients First

In the debate over patient-centric considerations in cancer care, one main focus has been the choice of proper endpoints for clinical trials. As the CEO of a cancer therapeutics company with multiple ongoing trials in a wide range of cancers, as well as a cancer survivor myself, I have found myself drawn into this discussion and have developed a core argument as well as some complementary insights. Before I share these, let me offer a bit of context. The debate on endpoints is embedded in the context of a larger discussion on the value of patient-centered care, which has now made it to center stage in discussions of quality. Writing in Annals of Family Medicine, Ronald Epstein and Richard Street note that patient-centered care is based on deep respect for patients as unique living beings and the obligation to care for them on their terms. As they observe, there have been concerns that patient-centered care, with its focus on individual needs, might be at odds with an evidence-based approach, which tends to focus on populations; but proponents of evidence-based medicine now accept that a good outcome must be defined in terms of what is meaningful and valuable to the individual patient.[1] As noted by Christopher M. Booth in Clinical Cancer Research, although patient outcomes have improved as a result of clinical trials since the 1970s, in the contemporary era, patients still define a useful therapy as one that improves the quality, or increases the quantity, of survival. Despite considerable change in the treatment and outcome of patients with cancer, it is critical that investigators, clinicians and policy makers not lose sight of these fundamental patient-centric outcomes.[2] Increasing recognition of the importance of patient-centric considerations in healthcare came with the creation of the Patient-Centered Outcomes Research Institute (PCORI). Congress authorized the establishment of PCORI in the Patient Protection and Affordable Care Act of 2010. According to its website, PCORI’s mission is to help people make informed healthcare decisions, and improve healthcare delivery and outcomes, by producing and promoting high-integrity, evidence-based information that comes from research guided by patients, caregivers and the broader healthcare community.[3] In a nutshell, I believe that patient-centric decisions should be the primary consideration in clinical trial design for cancer therapy; we should be able to design clinical studies that accommodate what is best for the individual patient. This leads us to the evaluation of endpoints. Currently, the three endpoints that the research community typically evaluates in clinical studies for cancer therapy are: response rate (RR) (i.e. tumor shrinkage), progression-free survival (PFS) and overall survival (OS). I believe that clinical trials should focus on RR as a primary endpoint and OS as a follow-on endpoint. If industry changed to RR and the FDA accepted that for registration, there might ultimately be more drugs that can be approved. The adoption of RR as the sole primary endpoint would allow cancer trials to run with smaller enrollments, since the number of patients needed to show tumor burden changes in a randomized study is much smaller than required in trials where PFS or OS are the primary endpoints. This would ultimately result in the development of products whose developers would not be making OS claims initially. Instead, they would claim the product kills more tumor than the standard of care (SoC) when added to SoC. The reduction of tumor burden in a safe manner should be a registration endpoint. Having put forth this basic position, it bears elaboration in several respects. For example, to understand why patient-centric decisions have traditionally been overlooked in cancer trials, it’s important to note that two fundamentally opposed viewpoints are in play: those of doctors and patients. Doctors are frequently consulting scans to see tumors and if the patient has progressed or not so that they can plan treatment courses. But these are typically not the patient’s concerns; patients care about whether a tumor has disappeared, not whether it is static or shrinking. Instead of being stressed about tumor progression, they are much more concerned about lifespan. When translated into the regulatory environment, these concerns are muddied. The U.S. regulatory environment is very much focused on OS and PFS and less on RR, while in Europe there has been a move toward PFS as an endpoint, and in Japan there is little emphasis on efficacy with a much stronger emphasis on safety. Regulators would do well to listen to patients’ concerns when evaluating trial design, so that trials are focused on killing tumors, maximizing lifespan and ensuring patient safety. It is common in oncology to talk about how patients will essentially endure any kind of therapy in order to live longer, but this is not true; patients place high priority on their safety. Response rates would allow developers to prove that drugs mechanically do what they’re supposed to do-kill tumors and are safe. The OS endpoint can come later, after the drug is on the market. This endpoint should play a role in pricing, not whether or not it can be commercialized at all. Following on this line of thought, the biotech industry has pushed very hard to go to PFS as an endpoint, because it allows companies to move through the development cycle more quickly-if you assume there is a formula or ratio that says X days of PFS equates to X days of OS. While that may be accurate for older chemotherapies, it is important to recognize that PFS is not directly related to OS in any context involving the immune system, including the development of new immunotherapies. So to use PFS as an endpoint is neither scientifically nor technically warranted. Think of PROVENGE® for metastatic prostate cancer, which exhibited relatively little change in PFS but gave patients three to six extra months of lifespan-which is what patients wanted. We can also ask whether the choice of endpoints, patient-centric or otherwise, is driven in some degree by investors with a stake in publicly traded biotechs. It is true that in some cases, endpoints are tailored to what investors are looking for; namely, a relatively quick return on investment. For most of a biotech company’s history, it is in pre-product approval mode. However, it is typical for a company-especially if is a new first-in-class or second-in-class new area in the field-to spend 15 to 20 years before the first product approval. If an investor values a company on the likelihood of an FDA approval, which in our current environment is OS, then that’s an awfully long time to wait. So there is constant pressure to continue releasing data and other information to help investors evaluate a company’s potential. As you can imagine, RR endpoints can provide meaningful information that can satisfy patients and can sometimes be obtained in weeks or months. What sorts of consequences would the use of RR as the primary endpoint have? It would not particularly affect the claims that could be made for new drugs, but it would affect their reimbursement. This is a reflection of the fact that certain clinical studies are run to obtain approvals to treat a certain cancer, while other clinical studies are run to justify reimbursement. As far as regulatory agencies are concerned, OS is currently the gold standard. A drug developer must realize that it is unlikely to obtain approval for a drug based on RR alone; the agency will likely want RR linked to OS at some point. It could be worthwhile for an agency to give the developer a limited approval while OS is being calculated. And what if a sea change actually did occur, with the biotech industry and regulatory agencies both dedicated to patient-centric endpoints? The result might very well be the approval of greater numbers of drugs. A logical consequence of the move to RR as an endpoint would be the acceleration of clinical trials and the quicker approval of products. At the same time, companies would not be sacrificing safety considerations, because safety data is obtained at about the same time as tumor response data is obtained. RR and safety endpoints ought to be sufficient for product approval, apart from OS. The arguments that I have been making in favor of more patient-centric trials are valid for all different types of cancer, as well as metastatic and non-metastatic forms. For those who wonder why this debate is a timely one, it is due to the fact that cutting-edge therapies are holding out the promise of dramatically increasing lifespan over what was previously possible. In such a climate of innovation, the potential to influence patients’ interest in trials is profound. Patients who know that a treatment stands to benefit them directly are much more likely to be psychologically invested in the outcome of a trial and more motivated to stick with the treatment regimen-a benefit for them and researchers alike. There has never been a more exciting time in cancer research. We owe it to cancer patients to keep their interests in mind as we design the trials that will ultimately push back the frontiers of the field. Brad Thompson, Ph.D. is president & CEO of Oncolytics Biotech Inc., a Calgary-based biotechnology company focused on the development of oncolytic viruses as potential cancer therapeutics. [1] Epstein RM, Street, Jr, RL. The values and value of patient-centered care. Ann Fam Med. 2011;9:100-103. [2] Booth CM. Evaluating patient-centered outcomes in the randomized controlled trial and beyond: informing the future with lessons from the past. Clin Cancer Res. 2010 Dec 15;16(24):5963-5971. [3] http://www.pcori.org/about-us