The Evolution of Endpoints

At the American Society of Clinical Oncology (ASCO) meeting, we celebrate our collective efforts in the fight against cancer. We are making real progress. People with many types of cancer are living longer than ever before. But this success has created a paradox when it comes to clinical trials.

Overall survival, the “gold standard” measure of how well cancer medicines are working, has become increasingly difficult to assess within clinical trials. The reason is mathematical. As people with cancer live longer, the probability of measuring an overall survival benefit within the timeframe and size of a typical clinical trial drops dramatically (see slideshow below).

That doesn’t mean new medicines aren’t improving survival. But it does mean that we might be missing something important if we rely only on overall survival to understand the effectiveness of a medicine.

If we want to keep pace with the tremendous advances we’re seeing against cancer and get new medicines to people who need them more quickly, we need to expand our approach to endpoints, the measurements used in clinical trials.

Meet the Endpoints

There's an ever-expanding variety of endpoints used in clinical trials. Like people, they each have a different job to do. Below are some of the most common endpoints used in trials of cancer medicines, as well as some newer potential endpoints on the rise.

The Heavyweight

Overall Survival (OS)

OS is the historical "gold standard." It's a measure of how long people with cancer live. As people in a typical trial live longer, it becomes more difficult to measure survival.

The Scout

Progression-Free Survival (PFS)

Like a scout checking out an athlete's 100 yard dash, PFS measures time. Specifically, the time until disease worsening or death. Related endpoints include relapse-free survival (RFS), event-free survival (EFS), and time-to-progression (TTP).

The Sleuth

Response Rate (RR)

One of the classic endpoints, defined as the percentage of people whose cancer has decreased by a defined amount. This is commonly based on a 30% decrease in the size of solid tumors viewed on medical scans, like a sleuth examining evidence through a magnifying glass.

The Investigator

Minimal Residual Disease (MRD)

A high tech spin on traditional detective work. MRD is an endpoint that uses newer, highly sensitive technologies to search for traces of certain blood cancers where traditional tests may have come up clean.

The Surgeon

Pathological Complete Response (pCR)

An assessment of the efficacy of a medicine given prior to surgery, defined as no cancer detectable at the time of surgery.

The Enthusiast

Quality of Life (QoL)

QoL is a complex concept involving many aspects of a person's satisfactions in life, ranging from the ability to work to the experience of certain side effects. Unlike other endpoints, it's based on what patients report.

The Disruptor

Immune Related Response Criteria (irRC)

Like a disruptive innovator in the tech world, irRC is shaking things up in the world of endpoints. It's a fundamentally different way of assessing anti-tumor efficacy that accounts for the unique response patterns seen with immunotherapies.

That’s no easy task. We need to continually and carefully balance scientific rigor and speed in developing medicines. The U.S. Food and Drug Administration (FDA) has already made great strides in expediting the regulatory process while retaining this rigor. Noteworthy examples include programs like breakthrough therapy designation and approvals based on tumor shrinkage (response rates) instead of survival in lung, skin, and breast cancers. Even bigger changes could be the horizon.

Despite the challenges, I'm optimistic about several areas where I already see regulatory science advancing.

Refining Surrogates

As survival becomes more difficult to measure, “surrogate endpoints” are one possible solution. These are measures that may predict a clinical benefit (like overall survival) but can be evaluated much more quickly.

Although the use of surrogate endpoints may allow us to get medicines to people who need them faster, they are not a guarantee. Each endpoint must be thoroughly vetted in well-designed clinical trials that also include established endpoints. Even after a surrogate endpoint is validated in one disease, much work is required to translate this progress across other medicines and cancer types.

A blueprint for success can be found in the history of progression-free survival (PFS), an endpoint that measures the time until disease worsening or death. As shown in the slideshow above, PFS was practically unheard of 30 years ago. Over time, however, it was increasingly adopted in clinical trial designs as a surrogate for survival. Based on extensive validation, PFS is now a widely-accepted surrogate endpoint and can be the basis for FDA approval in oncology, provided the magnitude of PFS benefit is considered clinically meaningful and the overall benefit-risk profile is acceptable.

Other potential surrogates like minimal residual disease (MRD) or pathological complete response (pCR) may follow this path in the future. I’m especially excited about these measures because they reflect the complete removal of the disease—a goal that is finally within reach for some types of cancer.

However, the true test of a medicine does not hinge upon any single endpoint. As we know from our experience with pCR in early breast cancer—it’s about the totality of data. Every medicine comes with risks that we need to balance against the potential benefits, so extensive safety testing is crucial.

“The federal government, academia, patient advocacy groups and private sector research-driven companies are all essential parts of our nation’s innovation ecosystem. Successful collaboration among all sectors of the ecosystem will deliver new innovative medicines with greater access for improved patient outcomes, reduced burden on healthcare and increased societal benefit. Advancing regulatory policy for surrogate endpoints in cancer trials is just one example how we can work together so that cancer patients can access new, effective medicines faster than ever before.”
— Celgene

Rethinking the Rules

In some cases, breakthrough science forces us to take a more radical approach and completely change the way we assess medicines. When that happens, we need to work together to rethink the rules.

This is happening right now in response to a new class of medicines called immunotherapies, which are designed to unleash the body’s own immune system to fight cancer. Because the immune system is inherently dynamic, immunotherapies can elicit fundamentally differently response patterns that our historic response criteria can’t capture. That’s why a new set of criteria has been proposed for evaluating whether these medicines are working.

Another area where we may have to rethink the rules is pediatric oncology. As my colleague Raphaël Rousseau wrote, we can’t simply extend the rules from clinical trials in adults to trials in children. The way we assess their responses to treatment must account for their unique biological differences.

Refocusing Our Goals

Despite progress in assessing anti-cancer efficacy, our ability to measure how a patient actually feels still lags behind. This is a deeply important consideration that I believe should be a primary goal in developing new medicines.

The endpoints used to assess quality of life, known as patient reported outcomes (PROs), are collected through patient surveys. PROs cover wide-ranging aspects of a person's well-being, from symptoms experienced to the ability to perform daily activities. However, these measurements can be subjective, making it challenging to establish reporting standards. Yet, when the quality of one’s life changes so significantly upon a cancer diagnosis, it’s a goal too important to ignore.

Despite progress in assessing anti-cancer efficacy, our ability to measure how a patient actually feels still lags behind.

Historically, the emphasis of cancer treatment was to keep a person alive, with PROs being relegated to the back seat (if they were considered at all). However, as treatments improve, PRO differences (“comparative tolerability”) could become a critical factor guiding treatment decisions. In fact, one medicine was approved with a PRO endpoint just within the last couple of years.

I believe it’s increasingly important to standardize PRO metrics, ensure their inclusion in clinical trials, and ultimately get them into medicine labels (so doctors and patients better understand how a medicine might make the person feel).

Redefining Success

The evolution of cancer treatments is necessitating an almost continuous rethinking of how we define success. I’m encouraged by the progress we’re already making through surrogate endpoints, regulatory pathways like breakthrough therapy designation, and increasing focus on a patient’s quality of life. But this is just the beginning.

We need to continue to build our scientific understanding of endpoints, and the regulatory science that supports their use, in order to keep pace with the promising advances in clinical science. In this way, the potential of new approaches can quickly be realized as new medicines. After all, at the end of the regulatory path for a medicine is the beginning of an important journey for the person who is waiting for it.