In this interview we discuss outcomes in clinical trials and how to improve trials by redefining clinically meaningful outcomes.
Lowell Schnipper, MD
David M. Dilts, PhD
Today, we are discussing outcomes in clinical trials and how to improve trials by re-defining clinically meaningful outcomes. We are speaking with David M. Dilts, PhD, professor in the division of management at Oregon Health & Science University in Portland, Oregon, and Lowell Schnipper, MD, professor in the department of medicine at Harvard Medical School and clinical director of Beth Israel Deaconess Medical Center Cancer Center in Boston.
-Interviewed by Anna Azvolinsky, PhD
Cancer Network: Dr. Schnipper, let’s start with you. Can you describe the way that cancer clinical trials are currently designed?
Dr. Schnipper: Cancer clinical trials have developed over a number of years, actually a number of decades, in an evolutionary way. The major categories of trials are actually conceptualized as phase I, II, and III.
Cancer Network: Could you discuss the typical primary endpoint outcomes for phase III clinical trials in oncology?
Dr. Schnipper: These are either progression-free survival or overall survival, and we all agree that overall survival is the most valuable endpoint but given the fact that patients often, when they progress on a given regimen, are unable to receive another therapy, sometimes crossing over from the standard therapy in a clinical trial to the test therapy, as some trials are designed, it’s quite challenging to identify overall survival as an endpoint. We all agree that it is the gold standard if it’s achievable, but progression-free survival sometimes becomes the endpoint that one chooses because it’s much more readily achievable.
Cancer Network: Dr. Dilts, what are the issues with the ways oncology trials are currently designed?
Dr. Dilts: The situation does not necessarily reside in that there are phase I, II, and III trials-it’s the fact that it takes far too long to actually open a trial and once it’s open, it has modest endpoints that really have difficulty in getting to a really clinically meaningful result. Other issues are the inclusion of all of the newer advances, such as using proteomics and other technologies to design the trial, and using adaptive design which means you can use significantly fewer patients in a phase II or III trial. So, the phases are fine, it’s that it is taking too long and too many patients to get too modest an outcome.
Cancer Network: Dr. Schnipper, you and your coauthors recently published a commentary in the Journal of Clinical Oncology on how to speed up the clinical trial process to get effective treatments into the clinic faster. This came out of an American Society of Clinical Oncology (ASCO) working group analysis. Can you talk about the impetus for this analysis?
Dr. Schnipper: This was an effort led by Lee Ellis who was then the chair of the cancer research committee at ASCO and I was the immediate past chair. Basically, we were motivated by the fact that, just as Dr. Dilts said, clinical trials take too long, and in our opinion, provide relatively modest differences as positive endpoints. So that in the end, the incremental progress we see in cancer-and let’s be clear, there’s been incremental progress, more in some diseases than others-for those of us who care for these patients, it has been painfully slow. To our patients, it has been excruciatingly slow. And so, we were essentially motivated by the fact that the world of cancer biology has given us an enormous opportunity to understand the cancer process better than anybody in any era before us has ever done and to develop very rational strategies to interfere with the aspects of the cancer process.
To us, that means that the future will hold the ability to identify patients who have some degree of high likelihood of responding to a therapy because they are proven to express in their cancer cells a biomarker indicating the target for which a new agent has been designed. If you concentrate your population of patients that have the biomarker against which a new drug is targeted, you will almost certainly be able to get an answer as to the utility or lack thereof of that new treatment with a much smaller study. So intrinsically, this will be a Smarter Trial, and will be done more quickly. We think that the answer will provide a robust change from the standard of care because of the uniqueness of the approach of targeting a biologically relevant pathway that is somehow responsible for the cancer process in a given patient. We decided to tackle this by looking at subgroups. We looked at colon cancer, breast cancer, pancreatic cancer, and non–small-cell lung cancer in the advanced disease setting. That is a scenario in which for many, many patients, who need a treatment for that indication, there are not robust treatments.
And historically, clinical trials interpreted as positive trials are those that have a relatively modest prolongation of disease-free or overall survival, measured in the range of weeks or just a few months. And as a group, we felt that we owed the public to do better. We convened advocates, investigators, and regulators from the US Food and Drug Administration (FDA) to think about this, and we came up with a manuscript that was published in the Journal of Clinical Oncology, suggesting what some would still call relatively modest endpoints but what we think is a set of goals that are a bit more ambitious than those that are operational today.
Cancer Network: Can you highlight some of the short and midterm goals that the group identified?
Dr. Schnipper: We expressed the change that would result from a new therapy in terms of months of overall survival benefit or hazard ratio. We suggest that investigators use a hazard ratio compared to the standard of about 0.75 to 0.8. We argued a lot about which level of a standard we should attempt to achieve, and the group was quite divided. The advocates were very interested in quality of life, almost more than length of life if it meant living with toxicities. Some of the physicians were intent on pushing the bar higher; others said that if the bar is too high, then some of the drugs we now use with good effect in cancer would never have made it to the adjuvant setting because they would never have come through testing in the advanced disease trials. It was an interesting exercise in debating different perspectives, but we came out with what we hope will be a suggestion to industry and investigators to raise the bar higher in this evolving era of targeted therapies.
Cancer Network: Dr. Dilts, can you comment on some of these goals and what would be the positive outcomes and the potential risks of some of these shifts in trial design?
Dr. Dilts: Let me first begin by saying how much I admire the work of the committee. I have done similar exercises, not in the medical domain, and I completely understand the difficulty of dealing with those who have extremely different perceptions on a shared issue and who have very different ways of approaching a problem. I think the fact that the group got together and came up with a consensus is a truly admirable thing. I think that what the group developed is a wonderful call to arms that says, “listen, we like incremental innovation, it’s a good thing to do, but we need to raise the bar, we need to say, let’s try harder than what we are doing.”
One of the things you understand as a management professor is that if everyone says “this is the old bar and this is the new bar,” suddenly people start conforming to the new bar. Let me give you an example. In the cancer realm, one project I did was measuring the length of time it took to open a cooperative group clinical trial. It was well over 800 days to open such a trial. When the National Cancer Institute put together the Operational Efficiency Working Group, they felt that length of time was too long and set the goal that this should take no more than 300 work days, and they have been able to achieve that in under a year. One of the most admirable things that has been happening is people setting the bar that they want to reach. The risks are that other people will stop doing some of the current research being done, regulators will think that a new standard for approval for a new treatment has been reached, or no one will get promoted if they don’t achieve this bar in academia. These kinds of trials are different and they need to be evaluated and approached differently.
Cancer Network: Lastly, are there other examples of a shift in this direction, towards better-designed clinical trials that are already in progress or being designed? Dr. Dilts, let’s start with you.
Dr. Dilts: Absolutely. The BATTLE trial in lung cancer and the I-SPY trial in breast cancer both used an adaptive trial design. So, this is not completely unique in the cancer realm.
Dr. Schnipper: I would second that and say that there are a growing number of trials involving very unique molecules coming out of the biotech and pharmaceutical industry that, once the target of the drug is known, the population of patients being selected for enrollment in the trial constitutes individuals proven to have that particular marker.
Dr. Dilts: There is a concept called fast and frugal failures. In other words, organizations build things to do small pilot tests to very quickly get an answer before ramping up to something like a large phase III trial. And I think that is one of the things that needs to be conceptualized in the cancer arena-how we can get a quick answer of whether or not a biomarker is even close before we go on to a major study.
Cancer Network: Thank you so much to both of you for joining us today.
Dr. Schnipper: Thank you so much for asking us.
Dr. Dilts: Indeed, thank you.