Expert Highlights Current Advancements and Questions in Breast Cancer Treatment

Kornelia Polyak, MD, PhD

Kornelia Polyak, MD, PhD

One exciting change in the treatment and management of patients with breast cancer in recent years has been the advent of liquid biopsy testing. Liquid biopsies using circulating tumor DNA (ctDNA) testing, which is often utilized in clinical trials, could potentially help oncologists detect cancer to make diagnoses and monitor progression and responses in a patient undergoing treatment.

“[A] very interesting area is applying liquid biopsy at earlier stages to detect both for primary diagnosis, which is potentially very relevant clinically, and for earlier detection of potential metastatic progression of the patients,” Kornelia Polyak, MD, PhD, says. “That's a very promising and exciting area of research.”

Many questions and barriers still exist before the true potential of liquid biopsies can be uncovered and used, not just in an academic center setting, but also for oncologists at community centers. Unlike researchers at these centers, a community oncologist secluded to a rural area in the United States does not typically have access to liquid biopsy testing or clinical trials, notes Polyak, a professor of medicine at Harvard Medical School and Dana-Farber Cancer Institute.

In an interview withTargeted Oncology, Polyak discusses several exciting advancements in the field of breast cancer, including immunotherapy, PARP inhibitors, and liquid biopsies.

TARGETED ONCOLOGY:What are some of the current areas of research that you are excited by?

Polyak:One area that is very interesting and clinically relevant is combination therapies. There are several clinical trials of combination therapies going on. Some of them are more rationally designed than others. For example, Alan Ashworth, PhD, [recently presented] about how to design synthetic lethal interactions when you have 2 agents that will block different pathways or that are taking advantage of a particular mutation that sensitizes the cells to a particular therapy. That's one area that is very interesting. For PARP inhibitors in BRCA-deficient tumors, it has been clinically proven that if you have a genetic deficiency that predisposes the cells to a particular synthetic lethal interaction, in that case the DNA repair effect can effectively and selectively kill those kinds of cancers. The question is, how do we find additional therapies like that?

Also, a very exciting area is in immunotherapy. Particularly, in immunotherapy, how do we rationally combine agents with immunotherapeutic agents when we [often] don't even know what the effects are of a particular therapy on the immune microenvironment. There's a lot of combination trials going on and a lot research in terms of more rationally designed trials because combination therapy is more effective for controlling disease.

Another exciting area is how to monitor disease progression and the response to treatment. We know that traditional matters, such as the imaging of tumors, may not be sensitive enough to monitor changes or actually detect response to treatment. One of the very interesting areas is liquid biopsy, where you can draw blood and assess mutations that are present in the blood, like with ctDNA testing. Particularly, metastatic patients could be more sensitive because ctDNA may give you a broader overview of the mutations present. In metastatic patients, it's not possible to sample every single lesion that they may have, but the ctDNA could detect mutations that could be in one of their lesions and not the others. [We recently presented a poster on] findingBRCAmutations in ctDNA that aren’t found in a tumor. The question is whether the ctDNA is going to be more sensitive for monitoring disease progression and responses.

TARGETED ONCOLOGY:The use of PARP inhibition in patients with breast cancer has generated some excitement. Do you think that PARP inhibitors will be used similarly and in similar settings in breast cancer as it's been used in ovarian cancer?

Polyak:PARP inhibitors have proven efficacy in patients withBRCA1andBRCA2mutations, either germline or somatic mutations in the tumors. This is only a subset of patients with breast cancer. The question is, can you somehow induce BRCAness phenotypes in a tumor that don’t have a mutation inBRCA1orBRCA2, and would that make them able to respond to PARP inhibitors?

A [recent] poster from my group discussed a combination of some epigenetic agents with PARP inhibitors. They observed synergy because if you are targeting some epigenetic regulators ofBRCA1orBRCA2regulators, like bromodomain proteins, when combining that with PARP inhibitors they also observed synergy. [Investigators are also looking into] using a similar approach [in prostate cancer] to induce PARP BRCAness even in wide-type patients. Whether that will be as effective as theBRCA1mutation carriers remains to be seen.

The other interesting observation is that there are many genes in the DNA repair pathway besidesBRCA1andBRCA2, like for exampleATM, which is one of the upstream regulators of this DNA checkpoint pathway. Breast tumors that arise inATMmutation carriers are very different fromBRCA1andBRCA2carriers, maybe even suggesting that there may be different mutations in the same pathway that may not have the same effect because PARP inhibitors don't work in those, whereas they do work inBRCA1andBRCA2. That just shows that we need to understand more about what makes a particular tumor DNA repair defective or not and produces that phenotype.

TARGETED ONCOLOGY:What is currently known about immunotherapy in breast cancer, and what are researchers still trying to learn?

Polyak:Breast cancer has been one of the more disappointing tumor types in terms of success with immunotherapy in the trials that have been conducted. Only a subset of patients seems to respond, most of those limited to triple negative. Of patients with triple-negative disease, only about 20% had a response to immune checkpoint inhibitors alone. Some of those responses were not as durable as what's been observed in, say, melanoma. The question is whether it's because we are not treating with the right drug, or if breast tumors are just immunologically more inert.

The combination of immunotherapeutic agents with chemotherapy and, for example, PARP inhibitors, seems to be promising. In ovarian cancer, there is a trial going on and it will be initiated in breast cancer as well to test these combinations. Many other agents [are being studied in combination] with immunotherapies. The question is that we don't know that it's going to be more effective or not. Some data with the chemotherapy and radiation combination shows that they are more effective than if you are using immunotherapy alone, whether those are going to be durable responses or not—that again we have to see. Triple-negative breast tumors have a lot of leukocytes in the tumors. Many tumors have more leukocytes than cancer cells, so the question is, how can we activate them, or maybe the tumor escapes control in different ways so that checkpoint inhibitors alone may not be efficient. It is a very intense area of research.

TARGETED ONCOLOGY:Why has there been less success with immunotherapy in breast cancer tumors compared with lung or melanoma? Why is the potential with immunotherapy better when we combine therapies?

Polyak:One of the potential reasons why there's less response [in breast cancer] is because there's not as many mutations. Melanoma and lung cancer are both caused by carcinogens like the sun and smoking, so they have many more antigens. The presence of antigens has been correlated with response in some trials, but it's not always a perfect predictor. That's certainly one of the reasons that you don't have any antigens, and that could be that maybe mismatch-repair deficient tumors of any cancer type will respond to checkpoint inhibitors. Triple-negative breast cancer has mutations, but may not have as many antigens as other cancer types. We recently published a paper showing that in breast cancer, the tumors go through this in situ phase [when the tumor] grows in the duct—ductal carcinoma in situ—and then, when they become invasive, there seems to be a big selection and a drop in immune active cells in the tumor. Even though you have more T cells and more leukocytes in an invasive tumor, they seem to be less active compared to the early stage tumor. This could mean that applying immunotherapy at earlier stages could potentially be more beneficial, or increasing the mutation—as in chemotherapy and radiation—you could activate the immune response that way. Those are the mechanisms that people are working on now.

TARGETED ONCOLOGY:Can you talk about some of the excitement in liquid biopsies? How could this change treatment?

Polyak:Liquid biopsies can be applied at different stages of the disease and for different purposes. What has been the most well characterized is in the more advanced stage of the disease where you can follow the mutations present in the tumor and in the blood. ctDNA has been a very sensitive way of detecting that and there have been studies where you can see the fluctuation of the particular mutation that you are targeting so that you see the patient response. In several studies, the response in the ctDNA was observed several months before the imaging actually showed the response to the treatment, so they seem to be able to catch response and, at the same time, potential recurrence or failure to respond earlier (potentially by a few months) than traditional imaging technologies, so that seems to be very promising. Of course, it requires that you know the mutations in the tumor, which is not always the case that you have mutations you can track. The other very interesting area is applying liquid biopsy at earlier stages to detect both for primary diagnosis, which is potentially very relevant clinically, and for earlier detection of potential metastatic progression of the patients. That's a very promising and exciting area of research. 

TARGETED ONCOLOGY:What impact will that have on outcomes?

Polyak:It will have an impact when you have something to offer those patients, so a prognostic marker is always useful if you can act on it. Obviously, if you would test somebody who you would see a recurrence in or who has a higher risk of reoccurrence, then you could apply another type of therapy at an earlier setting and not wait until they have increased tumor burden that you would see in imaging. That could increase the efficacy of response. Of course, metastatic disease is very difficult to cure so that would certainly have an impact but may not be that you would cure somebody. Whereas, if you could detect patients diagnosed at an early stage, that could have an even bigger effect on morbidity and mortality because we know that earlier diagnosis always leads to better outcomes just because we can control tumors better. Particularly in breast cancer where we have the option of adjuvant therapy, if you could predict potentially who is likely to recur who would give adjuvant therapy at the same time who may not need adjuvant therapy because you know that's not without side effects and some patients may not recur, so that could be beneficial to apply in those settings, in earlier disease, and at the same time monitoring advanced disease.

TARGETED ONCOLOGY:So where are liquid biopsies currently and where do we need to go?

Polyak:Most institutions are using liquid biopsy, but it's not part of the standard of practice yet. I think that incorporating it into the standard of practice would require a bit more understanding of what you really are detecting and how the heterogeneity of the tumor may influence an either positive or negative biopsy. Also, [we need to] follow patients overtime to see if somebody has a particular mutation in a tumor and if you have a false negative in a liquid biopsy, is it really a false negative or will repeated testing get different results? It still requires more research in terms of understanding how to apply it. Many institutions like Dana-Farber also use it for patients because they are sequencing the patient's tumor, and at the same time, detecting the ctDNA.

TARGETED ONCOLOGY:What advice do you have for a community oncologist who is treating patients in a more rural area?

Polyak:That's an interesting question because many of these technologies and tests are only available in larger academic centers like UCSF, Dana-Farber, MD Anderson, or Memorial Sloan Kettering. Even the clinical trials are only available in those centers. I think we have to see what the best way to treat people is, because 1 option is that some diagnostic testing and designing treatment occurs mostly in a centralized manner, and then administering the treatment could occur in the community practice in rural areas. That requires that patients have to travel to those centers, or at least, the sample coming from the patient has to be available for the center. Many of the major centers, including Dana-Farber, have been reaching out to community hospitals in 2 ways: 1) to have doctors from these centers go to the community hospitals and see patients there, so that way even if you are at a community hospital [rather than at an academic center], then you could see the same doctor personally, and 2) have virtual treatment to have an interaction with the oncologist at major centers.

For example, Maine is one of the most rural areas in the United States, and you have these areas that have larger centers, but they could be several hundred miles away from where the patients live. For daily treatments, you don't want people to travel there, but at the same time, you want them to be treated and diagnosed with the latest technologies. Many of those centers are offering genetic testing or are testing the tumor profile and assessing what clinical trials for which the patient could be eligible. Then they work with the local community oncologist to make it possible for that patient to enroll because enrollment in clinical trials is very low outside of the major centers, which means that some of those patients may not get the latest potentially helpful therapy that is available.