Implementation of Early End Points Could Represent New Standard for Breast Cancer Clinical Trials

By relying on early end points in clinical trials for breast cancer such as pathologic complete response (pCR), these studies could better optimize efficacy and limit toxicity, as well as increase the understanding of the biology of responders and non-responders, according to Laura J. Esserman, MD, MBA.

“We should aspire to better. We want better treatments for our patients, and by better, we mean more effective, more targeted, and less toxic. The biggest mistake we can make is standing still. We want to avoid over-treatment and under-treatment, and the only way to do that is with early end points, careful follow-up, and making sure that we do our best to learn from every patient,” she said. Esserman, the 2018 award winner for cancer diagnostics, provided the Giants of Cancer Care® lecture at the 40th Annual Miami Breast Cancer Conference^®.

She went on to discuss the effect early end points could have for clinical trials in breast cancer, how these end points could help usher in more individually tailored treatment options based on an individual patient’s disease and response, and how early end points have been used in the ongoing phase 2 I-SPY 2 trial (NCT01042379). Esserman is the director of the University of California San Francisco (UCSF) Breast Care Center, the Alfred A. de Lorimier Endowed Chair in General Surgery, and a professor of surgery and radiology at UCSF in San Francisco, California.

OncLive^®: How do early end points in clinical trials relate to personalizing treatment for patients with breast cancer?

Esserman: [As we] transition from adjuvant trials to neoadjuvant trials…the identification of early end points has enabled us to tailor treatment in a different way, to [better] understand the biology of breast cancer, and to improve the way we classify tumors to drive the response rates up over the course of the past decade.

The future is about tailoring care, less for the great responders, [but more] for the patients who are not having optimal response. [We must develop] a framework for how to get [more] patients to pCR without standard chemotherapy.

The importance of early end points is they allow you to learn over the course of care. If you operate first and take out a tumor, you are taking out the biomarker. You are removing the tumor, and therefore removing your ability to understand what is happening. The problem with adjuvant therapy is that we took out the tumors, gave every patient the same treatments, and assumed that everyone would respond the same, which is, of course, not true. We know this from the neoadjuvant setting where we leave the tumor intact and discovered that some patients respond well and some people do not.

The early end point means that response to therapy is important. It allows you to pivot if a patient is not having the right response. It turns out that pCR is a great end point for faster growing molecularly high-risk tumors. However, this is not for everyone. You must know for whom that early end point works. For those fast growing, high-risk tumors, [pCR] is an excellent end point.

Why is [pCR] important? It helps you get an early look by 6 months. By the time someone finishes therapy, you have an idea if that is an effective agent for a group of patients and if it is an effective agent for an individual.

Not only does [pCR] help you learn about which agents should be used and for which tumor types, it also lets you learn how to individualize or tailor [treatment] for each patient. Everyone talks about personalized medicine, but the way to personalize it is to come up with that early end point and to be able to pivot both in the face of a great response and in the face of a poor response.

How do you pivot with a great response? You can do less therapy. The next phase of the I-SPY 2 trial is allowing us to de-escalate therapy in the setting of great responses and serially escalate for patients who are not having as optimal responses.

Can you say more about the goals of I-SPY 2 and the patients being evaluated?

Our goal is to get all patients to have better outcomes. By better outcomes, we mean [giving] more effective therapies that are less toxic and to do so in a timelier way to get these therapies to work. It is an ongoing platform trial that gives us the opportunity to continue to add new agents in a series of ways.

In block A, we are looking at nonstandard chemotherapy, meaning targeted agents that have the chance to get people to pCR without standard therapy. We are doing 3 blocks of treatment: block A, block B, and block C. Block A [includes] the novel therapies. If a patient has a great response, they can [proceed] to the operating room. If a patient does not [have a great response], they will get assigned the best treatment for their particular response-predictive subtype as opposed to going right to surgery. We use MRI, functional tumor volume, and biopsy, when appropriate, to try to predict who is going to have a pCR.

Importantly, if a patient does not look like they are having an appropriate response by 6 weeks, we have a strategy where we can escalate early and move on. I believe people are willing to try new things, especially patients who are interested in less toxic therapies. However, you must have a way to say if a [treatment] does not look it is going to [deliver a sufficient] response. We have developed standards and strategies to be able to say that something is not a sufficient response. We have tailored [these strategies] to the type of tumor, and that is important.

Over the past 12 years of running the trial, because we required every patient on the trial to have a biopsy where we could generate molecular profiling, we have been able to recategorize [and develop] a different tumor classification system. It turns out, for example, hormone receptor [HR]–positive [breast cancer] in the high-risk setting is many things. [Approximately] one-third of these [tumors] are more basal-like that have an immune phenotype and respond just as a basal cancer would. Then there are some that are just not very responsive, and you do not want to mix those two. So, we found a better way to sort out how to assign patients different therapies, which is very exciting. This will allow us to move the field forward.

There are a lot of things that we assumed that turned out not to be true. We are letting the data teach us. Overall, we started out at about a 19% to 20% overall CR rate, but by assigning the right patients to the right combinations of therapy, we believe that this [rate] can be at least 60%.

What we also hope is that with some of these agents that are very effective, they do not have to be perfect for everyone, [but we hope] that we can get these drugs to the market. The whole point is, if you can get to [a pCR] with less therapy, that is pretty exciting. From all the patients I take care of, that is what they want: better therapies, less toxicity.

How are you hoping to see the use of early end points further integrated into future clinical trials?

Clinical trials should reflect the way we want practice to be, so our clinical trials must allow us to incorporate these personalized approaches. Trials should look more like care, and care should look more like trials.

In endocrine receptor [ER]–positive patients who have molecularly low-risk but clinically high-risk disease, we are working on testing some of the new and exciting endocrine agents, and what we are looking for is the right early end point, because we need to find it for those slow-growing tumors. You do not want to wait 10 years to find out whether [these patients] have had a recurrence because that kind of knowledge turn is way too slow. So, we are looking for circulating tumor [DNA] in the blood, RNA in the DNA, background enhancement on MRI, for functional tumor volume, pCR, and all of the combinations of features to find out what could be the driver that would tell us if we are on the right track [with treatment].

We have this endocrine optimization pilot, so now anyone who has a stage II or III breast cancer, regardless of their molecular phenotype, has an option to be in the trial. We should learn from every patient we see.

I am also working on the same kind of framework for ductal carcinoma in situ [DCIS]. DCIS should be considered a window of opportunity for prevention. However, if we keep operating on everybody and doing the same thing we have done for the last 3 decades, we cannot expect to have a different result. If we can give endocrine therapy or some of these things in the high-risk setting of invasive cancer, we certainly could do it in the setting of DCIS, which isn’t cancer, but [is rather] a risk factor for getting invasive cancer.

Clinical trials need to take innovative approaches. If we can use [these approaches] in invasive cancer, we could do it in a much lower-risk setting and learn. Everyone [with DCIS] does not want surgery, and many people probably do not need it. We are probably overtreating [DCIS], and we have not figured out the right treatments and the right approach.

That is why early end points are so important. They allow you to learn at a pace that is much faster and much more informative. You put one trial in [motion] and you learn, [then] you have got the next one in mind, and you have a whole pipeline of things that you keep learning and [adjusting] until you figured it out. You are not going to get it right the first time, but over time, you can make substantial progress.