A key decision in the design of clinical trials in oncology involves whether to select progression free survival (PFS) or overall survival (OS) as the primary endpoint of a clinical trial. Both are important to patients and clinicians, yet their impact on various features of a sponsor's clinical development and market access strategy move beyond the rudiments of traditional trial design. Indeed the evaluation of tradeoffs when choosing from these endpoints requires careful consideration and expertise to channel statistical strategy towards financial success and industry advantage.
Most sponsors use PFS when trying to secure regulatory approval, and OS for market approval and market access. Building clinical trial designs that optimize across both has historically proven challenging, but new technology like Cytel's Solara has also paved the way for new methods to find these potent designs.
PFS & OS: The Basics
PFS and OS both measure important benefits to a patient. PFS measures the length of time after receiving a treatment, where the disease does not progress. OS by contrast measures the time between receiving therapy and succumbing to illness. Both are important to patients and clinicians for determining the right therapy for a given diagnosis.
Progression Free Survival
PFS is the endpoint that usually generates a shorter trial. When OS is not required by regulators, using PFS as an endpoint can accelerate trials and thereby reduce sponsor costs. Regulators are likely to approve of trial designs with PFS as the primary endpoint. Since many sponsors are concerned with the time it would take to measure overall survival and know FDA and other regulators will generally accept clinical trial designs that are built around PFS, they often choose to use this as the primary endpoint.
This has led many to criticize PFS as being the endpoint of choice that is measurable and which expedites a clinical trial, but which does not necessarily reveal what is in the best interest of patients. As such many sponsors aim to balance their trial designs by optimizing for both PFS and OS in the choice of trial.
OS measures the amount of time between receiving a therapy and succumbing to an illness. Essentially the best medicines for patients may well be those that maximize overall survival, but the better the new therapy the longer it will take to accumulate the events necessary to measure OS. As a number of healthcare systems and payers require information on OS before a new therapy can be brought to market, those who choose PFS for accelerated timelines also need to collect information on OS.
This means that many sponsors who choose to use PFS for regulatory submission, will still need a clinical trial the optimizes over both endpoints. Although PFS can supply evidence for regulatory acceptance, there is still an additional hurdle. This sometimes leads to the complex situation where a sponsor has achieved regulatory approval but still needs to generate evidence for market approval and access. Arguably, the best time to do this is during an interventional study that will supply evidence for OS while using PFS as a primary endpoint. Yet this often leaves a sponsor in a bind when pursuing an effective clinical strategy.
Alternatively a sponsor might choose OS as the endpoint of concern with PFS as a secondary endpoint, but this requires overpowering for PFS. This is because PFS and OS are obviously correlated, and powering for OS means that unnecessary resources are used to achieve regulatory submission that could have been achieved had PFS been the primary endpoint.
The Technological Solution
Up until now technology has made optimizing over both endpoints a difficult undertaking. A statistician would generally have to optimize over one endpoint (say PFS) then optimize a design for the other, and thereafter tinker with both until a trial design is discovered that satisfied numerous objectives. The time and process were less than ideal, and ensuring that better designs are unavailable would cut into even more clinical development hours. Furthermore, PFS and OS are correlated -- a high likelihood of PFS would generally mean a high likelihood of OS. Therefore, examining both in isolation rather than at once affects the precision of the forecast.
Cytel Solara has a Dual Endpoints feature which enables sponsors to optimize over both endpoints, thereby aligning a clear path between regulatory and market strategy. Solara's powerful engines scan millions of simulations to pinpoint designs that reflect sponsor circumstances and satisfy preferences like speed, power and cost of a clinical trial. Yet they also determine how well each design performs over both endpoints.
The Statistical Benefit
The central question then becomes whether a Win Strategy for a sponsor involves choosing one or both endpoints. This process, when done in Solara, also helps solve for an integral challenge of controlling Type I error. When assessing tradeoffs between various performance characteristics, Solara can replace the win strategy for power to offer clearer insights into what is ideal for a project team. (Contact us below to learn more about this approach).
The Strategic Opportunity
The complex designs generated by Solara provides a framework for sponsor leadership to streamline regulatory and market access strategy. The Dual Endpoints function assesses designs for one or both endpoints. This opens several possibilities to sponsors. A sponsor can have an interim look that analyzes PFS and continue the trial for OS; alternatively it can have one interim analysis to see if significance in either has been achieved, and make a tactical decision based on this knowledge about whether to file for regulatory submission and when to begin the process of market access.
Further Solara demonstrates the tradeoff between these various approaches in terms of a clinical trial's other performance characteristics like speed and probability of success. The Pareto Optimality feature ensures sponsors can quickly see which designs are optimized, and a number of scoring algorithms and visuals help sponsors to detect how to rank various priorities.
About the Author of Blog*:
Dr. Esha Senchaudhuri is a research and communications specialist committed to helping scholars and scientists translate their research findings to public and private sector audiences. She received a doctorate from the London School of Economics in philosophy, and is a former early-career fellow at the American Academy of Arts and Sciences. She has taught medical ethics at the Harvard School of Public Health (TH Chan School) and sits on the Steering Committee of the Society of Women in Philosophy's Eastern Division, which is responsible for awarding the Distinguished Woman Philosopher Award.
*Esha thanks Jim Bolognese for support with the content of this blog.