A Conversation With
A Conversation with Dr. Patricia Steeg on Redesigning Clinical Trials to Test Therapies that Could Prevent Cancer Metastasis
How big a problem is metastatic breast cancer, and would your proposal apply to other cancers besides breast cancer?
This year, nearly 40,000 women will die of breast cancer, and metastatic disease is the largest contributor to these deaths. Although the annual number of breast cancer deaths is declining, it's not declining by much.
The proposal I've put forth should apply to a number of different cancers, particularly those where the majority of patients are diagnosed before they have full-blown metastatic disease, or if they have limited, treatable metastatic disease. One could imagine applying this to prostate, bladder, and colon cancers.
Why is treating metastatic cancer so challenging? Are there any effective, approved treatments?
First off, metastasis research is tough to do. There is no in vitro, or lab bench, model for metastasis, so you have to use animals and the experiments are long and complex. Most of our animal model data show that we can prevent the formation of metastases. Taking those preclinical findings and validating them in a clinical trial is, currently, just about impossible.
We have a few adjuvant treatments in breast cancer that can partially prevent metastasis when given early after the initial surgery and radiation therapy. These include cytotoxic chemotherapies, which kill cells; estrogen receptor antagonists; and some molecularly targeted therapies, like the HER2 inhibitor trastuzumab [Herceptin].
But nothing cures in the metastatic setting, and up to 30 percent of women with early breast cancer will suffer distant recurrence and ultimately die of their disease, so there's a lot of room for improvement.
How do cancer metastases differ biologically from the original tumor?
When you compare metastases to the patient's primary tumor, many gene and protein expression patterns can be the same, but others differ. In addition, there are more mutations in a metastatic tumor than in the primary tumor, and one metastasis may have mutations that another metastasis in the same patient does not. So, our old approach of just looking at the patient's primary tumor may not give us all of the information, or the best information, that we need for "precision medicine," or personalized medicine.
What makes current clinical trials poorly designed to test drugs that could prevent metastasis?
My lab and others have identified a number of compounds that will prevent the formation of metastases in mice. These compounds are not inherently cytotoxic—they don't kill tumor cells. Nor were they designed to be synergistic with traditional chemotherapy drugs. But those are the two criteria by which most compounds [are judged] in clinical trials.
Phase I and phase II clinical trials are done in patients with advanced, refractory metastatic cancer—patients who have had many therapies, and these therapies have failed them. In the phase II trials, we ask, does the drug shrink the metastases? And then, if it does, the compound will keep progressing through the clinical trial process, and it will be tested against the current standard of care in a phase III trial to see if it will produce responses in patients with metastatic disease.
But a drug that prevents metastasis may not shrink a large, refractory metastatic tumor. It has a different mechanism of action that is not being picked up by the clinical trial system. That is why I've proposed that we take a detour in that path of clinical trials, either before or just after phase II trials, to do what I call phase II randomized metastasis-prevention trials.
How should such trials be designed?
After we see that a drug is safe, we have a hint of activity, and we know that it can be combined with the standard of care, let's try looking for metastasis prevention. Let's take patients who either have no detectable metastases but are at high risk of disease progression, or patients who have a limited numbers of metastases and have been treated with standard therapy. Those patients would receive the standard of care and be randomly assigned to receive the placebo or the potential metastasis preventive in a clinical trial.
For patients who don't have metastases, the most important endpoint would be time to development of the first metastasis. For patients who do have metastases and have been treated, the endpoint would be time to development of a new metastasis. I think that this would give these drugs a fair chance to show metastasis-preventive activity. If the endpoint is only shrinkage of an existing, bulky metastatic lesion, these drugs are going to fail.
What will it take to change the process and implement your proposal?
This is doable; it's just going to take collaboration. Part of the reason I wrote this perspective is to get a conversation going. Oncologists are going to have to work closely with the molecular biologists, who can tell them what compounds to test and in what circumstances. And they're going to have to work very closely with the FDA [Food and Drug Administration] and get guidance on the appropriate trial design. Then they're going to have to work closely with patients and advocates, so that patients understand that this is a new paradigm—it's untested but it holds promise—and will work with us to get these trials moving.
Beyond publishing the perspective in Nature, how are you working with others to implement these ideas?
I'm part of an ad hoc group of government and academic researchers who have been meeting monthly to discuss new trial designs, new approaches, and new drugs for brain metastases of breast cancer. I've heard many of the concepts that I've laid out in this editorial in these conversations.
It's extraordinarily productive when preclinical molecular biologists can get feedback from a wide variety of clinicians. [That feedback] shapes our work when we go back to the bench and design the next experiment. The clinicians hear about our mouse results and take them with the appropriate grain of salt, but seriously consider them. And we have to do our mouse experiments in a way that can be translated into a clinical trial. It's a valuable two-way conversation.
For this proposal to spread and be accepted, the FDA is going to have to issue some guidance. I don't think a drug company is going to spend money, or that clinicians are going to risk part of their careers, to bring something forward until there is a path to approval.
Here sits the preclinical research that says drug X and drug Y will meaningfully prevent metastasis. But these drugs are failing in the clinical trial system, and they are never getting to the patient in the appropriate setting. [As a result], the drug company loses the investment in drug development, the research scientist loses a clinical hypothesis, and the patients continue to lose their lives.
—Interviewed by Elia Ben-Ari