Esteva Explains State of Genomic Testing in Breast Cancer

Francisco J. Esteva, MD, PhD

Physicians must continue to develop an understanding of breast cancer tissue and plasma to personalize medicine and deliver the optimal treatment to their patients, says Francisco J. Esteva, MD, PhD, who spoke on genomic testing at the 16th Annual International Congress on the Future of Breast Cancer® (East).

OncLive: What are the clinical implications of genomic testing in breast cancer?

In an interview during the meeting, Esteva, a professor in the Department of Medicine at NYU Langone Medical Center, discussed the clinical implications of genomic testing for patients with metastatic breast cancer.Esteva: There are different technologies that are being used to profile both primary and metastatic breast cancer tumors including next-generation sequencing, circulating-free DNA in the plasma, and other technologies looking at the microenvironment. In my presentation, I just focused on next-generation sequencing and liquid biopsies.

The most important area right now in terms of research is the profiling of metastatic tumors. In breast cancer, there are not as many mutations in cancer-related genes as we have seen in other tumor types, such as melanoma or others where the mutational load is not as big.

In The Cancer Genome Atlas study published in Nature a few years ago, they profiled 500 tumors and found that around 10% of all tumors had more than 6 mutations. Most tumors either had no mutations identified in the top-40 cancer-related genes or just had very few mutations. I don’t think we have the whole story about cancer drivers using genomic sequencing today. It is important information, but it remains a research tool to direct patients to clinical trials targeting those mutations.

These are all major challenges that we need to deal with. If we find a specific mutation, whether that mutation is a driver of the cancer or whether it is a passenger—meaning it is not driving the tumor phenotype—is still unclear. There are several examples where this technology may be useful in the future. We have found mutations in HER2 that may predict a response to neratinib (Nerlynx). For example, in one of the studies presented by Dr David Hyman at Memorial Sloan Kettering Cancer Center at the 2017 AACR Annual Meeting, findings demonstrated a promising approach to sequencing tumors.

However, in that study, researchers looked at breast cancer and other types of cancers. In mostly breast cancer with HER2 mutations without HER2 genome implication, they saw responses to neratinib. Not all mutations are the same, and not all mutations have the similar results in all tumor types as we may have thought.

PI3K is another interesting gene where mutations have been associated with a better prognosis in breast cancer. The question is, “Can we use them to predict response to PI3K inhibitors or mTOR inhibitors?” To date, the data are inconclusive.

Another thought is the estrogen receptor (ER) mutations that may predict responses to different types of endocrine therapies, such as fulvestrant (Faslodex) compared with aromatase inhibitors. There are studies also looking at fulvestrant versus CDK 4/6 inhibitors, such as palbociclib (Ibrance), suggesting that they do not predict [responses with] CDK 4/6 inhibitors, but they may be predictive of response to some of the basic endocrine therapies.

What are the current guidelines for genetic testing?

That research applies more to the liquid biopsies where we have been able to find mutations in circulating-free DNA in the plasma, which may be predictive but it is still an area of investigation. In patients with metastatic breast cancer, the biomarkers we always test for are ER, HER2, and perhaps BRCA. Now, since we are more interested in identifying BRCA mutations, it is standard of care—or at least recommended by national guidelines—to sequence the BRCA1/2 genes in patients with triple-negative breast cancer in which the ER, progesterone receptor, and HER2 tests are negative.

Next-generation sequencing, meaning sequencing an entire tumor, is not recommended as a standard of care yet. It is done whenever possible, and especially at the major academic institutions, with the goal being to direct those patients to clinical trials—single-institution trials, big institutions, and also the National Cancer Institute (NCI) MATCH trial, for example. This is where there are more than 20 arms depending on the mutations identified, so patients may be directed to those trials.

How are these genomic tests personalized for patients?

The NCI-MATCH trial screened about 10,000 patients with metastatic cancer of many types and they closed the screening step. Now, they are only allowing patients to enter those trials if sequencing is done by different laboratories. For patients with metastatic breast cancer, those tests are relatively easily accessible and available. There are many companies now where you can send them the tissue or the plasma, and they can look for mutations in the tissue or mutations in circulating-free DNA in the plasma. The question is, “What do you do with that information to be able to personalize these therapies?” That goes into the area of clinical utilities. Technology is ahead of the clinical utility of personalized medicine. We can identify mutations in tumors or in the plasma, but what we do with that information is not clear. That is why it is important to continue to use clinical trials.

In clinical trials available today, patients with ER-positive breast cancer were randomized to endocrine therapy with or without PI3K inhibitors or mTOR inhibitors. In some trials, particularly with the mTOR inhibitor everolimus (Afinitor), for example, PI3K mutations were prognostic, but not very predictive of response to the mTOR inhibitor. With PI3K inhibitors, there is a study that showed that those patients may respond better to a PI3K inhibitor in addition to the endocrine therapy. However, that inhibitor was too toxic. Therefore, we are now looking at more selective inhibitors, such as PI3K inhibitors.

What is the main takeaway message for community oncologists?

Is there anything else you would like to add?

We need to continue to test those. I don’t think we are at the point to be able to personalize and use this information in the clinic—to select one therapy versus the other.The main takeaway message for community oncologists is to understand the limitations of these technologies that are available. We can sequence tumors, look for mutations in specific genes in the plasma, and do a very deep sequencing of ER and PI3K. However, it is not clear that those mutations are driver mutations. My hope is that those patients, whenever possible, should be referred for clinical trials. Over the last 15 years or so, we have been focused on protein coding genes. We even tend to focus on larger genomic testing, but each human cell has 20,000 genes—making it a large and complex system. There are thousands more noncoding genes, which we don’t know much about, and that is another area where we are trained to focus on if they are expressed, what they regulate, and what they do.

Additionally, we have the microenvironment, which relates to the immune system and many other things that are not directly in the cancer cell. We need to be able to understand all of this to truly develop personalized medicine in the future.