November 2012, Vol 1, No 5

← Back to Issue

Cultivating Personalized Medicine Clinical Acumen in the Management of Breast Cancer:

An Interview with Edith Perez, MD

Edith Perez, MD

Interview with the Innovators

If any cancer can be said to have launched personalized medicine into orbit and establish it as the standard for all cancer therapy to aspire toward, it is breast cancer. The discovery of the estrogen receptor (ER) and HER2 biomarkers has astonished healthcare professionals and the public alike and raised forever the bar of cancer care. What conditions have arisen to make breast cancer the envy of other cancer researchers, and how is this field lev­eraging its discoveries to unlock new opportunities for researchers and clinicians alike? And most important, how can today’s practicing oncologists put into practice the life­saving techniques of breast cancer personalized medicine, ally them with the conventional medicine that still dominates the field, and cultivate a base of knowledge of this burgeoning, data-heavy field of medical science?

To capture the essence of the breast cancer treatment revolution, we interview Dr Edith Perez. Translational research is the soul of Dr Perez’s work. Asked what she sees as the vision statement of personalized medicine in breast cancer, she mentions not the advances in science but the lives that will be saved through a new understanding and application of its science. Asked about the “competing” model of predictive modeling versus biomarkers, she instantly embraces it as a welcome and helpful adjunct to biomarkers. This same inclusiveness extends to wellness-based healthcare, to patient-reported outcomes, to the link between the clinical, business, and policy sectors of healthcare in making breast cancer research get traction and stay on course.

And what is that course? Without hesitation, she advances the translational research goal: a new molecular classification of breast cancer that identifies the relevance of types of breast cancer to drive treatment decisions. For Dr Perez, it’s all about treatment. As one of the leading researchers in the area of cancer research that other cancer experts look to as a model of success, we sought her out to give not only an empirical view of the principles and drivers of breast cancer success and future research goals, but also to help define the issues helpful to all cancers. Above all, she applauds the principle of personalized medicine in oncology: to cultivate an understanding by practicing oncologists of the biological basis of cancer and the pharmacodynamic basis of the drugs – biologicals and conventional alike – used to combat it. Asked what the practicing oncologist must glean from the informational overload of this new science to keep up, her advice was immediate and precise: “The practicing oncologist needs to understand genes, the regulation of genes, protein expression, and then metabolomics.”

We turn now to a tour of personalized medicine for the practicing oncologist by Dr Edith Perez, who describes the strategy and tactics that she and her fellow colleagues are employing to transform what was once a universal killer into a manageable, nonlethal condition.

PMO How would you define or otherwise describe personalized medicine and its role in the clinical management of breast cancer?

Dr Perez It’s our ability to identify markers of the tumor of the host that impact pathogenesis so that we can target them to optimize treatments for the individual patients.

PMO How does predictive modeling assist in this pro­cess, compared with the more specific matter of
biomarkers?

Dr Perez Predictive modeling can be really fascinating and provide us with some guide of the path that we want to follow. The same applies to preclinical modeling, utilizing cell lines or newer models. In the area of predictive medicine for patients, or personalized medicine, we have realized that all of these models really need to be validated with a robust number of patients who have been followed for an adequate time, along with having appropriate tumor specimens to test these biomarkers.

We can chase a lot of ideas with theoretical concepts and preclinical models, and even studies in a small number of patients. But in the area of oncology, certainly in the area of breast cancer, we have come to a realization that these tumors are heterogeneous. There are many genes and proteins that impact the pathogenesis. Thus, we need a fairly large number of patients with appropriate molecular testing to reach definite conclusions.

I’m very enthusiastic, though, about the way the field is moving, but at the same time I’m concerned about the volume of reports that are published in abstract form or peer-reviewed manuscripts every month in this field that may not have a robust number of patients to reach definite conclusions.

PMO Yes, and that’s what you’re looking for: an enriched patient population.

Dr Perez Yes, and we’ll get to that, but another area that is tremendously important in terms of interpretation of the literature is that it has become easier for journals to prominently publicize and publish material that reflects a possible relationship between a biomarker and patient outcome.
It remains difficult to print data of negative associations.

For the needs of patients, reports of negative associations are as important, so that researchers don’t have to keep repeating the same work, as well as acknowledge the work that is done by not only the patients volunteering tumor specimens but also the investigators who have done the calculation work.

I would like to ask for a call to analyze and publicize the data, even if they are negative, because that is the way to advance the field.

PMO Yes, everyone from the public to the research community is getting this skewed, overly positive perspective on the degree of success – and the ease of success – that’s present in personalized medicine. Without sufficient recognition of the research failures, we will lose the impetus to continue pursuing research areas that are going to overcome these setbacks.

Dr Perez Yes, but I wouldn’t actually use the word failure. When we have hypotheses, it’s our job as scientists and translational investigators to validate the hypothesis or not, and so our obligation is to be able to publish, to report on the information we have provided. That’s the way I look at the field.

But I am very enthusiastic about the way we’re moving because we are all learning about better ways to do the analysis. We have started to look more carefully at the way the specimens are collected and stored, the need to validate any laboratory testing that we do with appropriate controls, and potentially even have more than 1 pathologist look at stains or results of gene analysis. And also how carefully we have to analyze the data from these markers in a blinded way in the context of patient outcomes, so that the results that we obtain truly could relate with patient performance.

PMO You’re seeing basically a success in changing not just prognostic expectations that personalized medicine provides, but even changing the culture of medicine regarding prevention, diagnosis, and treatment?

Dr Perez Well stated. Let me phrase the situation in this way. It is clear that we have made advances over the past 30 to 40 years, but at this time I think we are in a position for the advances to be of much greater magnitude than what we have been able to accomplish.

Second, the 2 basic stories that are most positive in the setting of breast cancer are following the only 2 targets that we currently use for therapeutic decisions, which include the ER and HER2. At the same time we know that we have to move beyond these 2 markers in the context of 24,000 genes, validation of the expression of these genes, then posttranslational modification approaches.

PMO What is the degree of the proliferation, if you will, of personalized care in breast cancer compared with other cancers, and what would be the reason for the differences in the degree of personalized medicine proliferation across these cancers?

Dr Perez I think there’s a wide interest in the field of personalized medicine for all malignancies. The 1 issue in the setting of breast cancer that makes it unique is composed of 2 things. Number 1, breast cancer is a common disease, and we have had some good therapies that have allowed many patients to survive for longer times than other tumor types, which has then led to an advocacy coming from the patient’s standpoint to get researchers to do research in the setting of breast cancer.

Also a positive for breast cancer has been the fact that some of these therapies that we have found to be beneficial really have impacted the lives of so many, so essentially we have used the low-hanging fruit. Some things have worked very well, so we have a lot of impetus to keep moving in that direction.

But clearly at the same time I’m really very happy that this field of personalized medicine has started to pan out in the setting of some lung cancers and also some other cancers, especially melanoma, where we’re starting to see some really fascinating data. But the issue is going to be, are these new data going to impact the 5-year survival of those diseases, which remain way below 20%.

So although the data with the EGFR inhibitors for lung cancer, for the ALK inhibitors for lung cancer, and the data for the BRAF inhibition in the setting of melanoma, look very interesting, we need a little bit longer follow-up, whereas in breast cancer, we know that targeting ER, targeting HER2, truly impacts the long-term outcome of patients in addition to having an impact on short-term response or progression-free survival.

Initially it seems like the group of individuals involved in breast cancer is quite cohesive at the global level, but this is happening also with other tumor types. But we have had long-term meetings where research data are discussed on an ongoing basis so that we don’t duplicate work, and at the same time we collaborate with each other to make important data available.


PMO What would be some of the important examples of personalized medicine algorithms of this type as they’re generally understood? And next, what are some examples of such algorithms that have been discovered by researchers but have failed to be adopted by oncologists expeditiously, and how can we expedite the uptake of that kind of progress for breast cancer patients?

Dr Perez First of all, one of the issues that we have to deal with, and I deal with this all the time, is how to convey the message of the work that is currently being done, as well as the complexity of the work being done, to the physicians who are mainly involved in practice, or to physicians or to philanthropists who we absolutely need for support of some of this innovative work. We have to distill the technical jargon to understandable concepts that can be conveyed to people because they have to be part of the team as we try to solve this puzzle of breast cancer.

PMO Yes, including the investors.

Dr Perez Absolutely. Then we as scientists also have to keep in mind that the honesty related to our research findings is something that cannot be underestimated. We have to be judicious in the way we report our own data and just absolutely always, always tell the truth, not make more than what we have, not make less than what we have. We’ll always be 100% honest and accurate related to the information we share with others.

We cannot compromise on this issue because we may send confusing messages to all of these stakeholders as well as all the researchers.

In terms of personalized medicine algorithms, I think probably the most important thing I can say is that we have to work with our pathologists to be able to collect the tumor specimens in a consistent way, and for pathologists to follow algorithms and recommendations that have been generated by expert panels related to exposure of tissues to various elements so that the tissue remains appropriate for testing of these markers.

At the same time it is very important to realize that at least from the treatment standpoint, depending on the data that we get from these molecular profiles, the clinical condition of the patient is also very important. So we don’t manage patients only based on laboratory results. We manage patients based on the prior therapy they may have had, the patient’s overall condition, and organ function. We have to put all the pieces together, but I hope that in the next few years we’re going to learn so much about the molecular profile of these tumors that we’ll be able to be a lot smarter related to putting together molecular profiles with patient profiles to come up with the best treatments.

When we think about personalized medicine and molecular markers, there are so many different things that we can look at. We can look at whole genome. We can look at exomes. We can look at RNA. We can look at proteins. We can look at microRNAs, RNAs that modulate function of genes. So this is a very vast field that speaks to the complexity of what we need to elucidate.

In the area of microRNAs, there are hundreds of RNAs that have now been identified, and RNAs appear to modulate the function of about a third of our genes.

PMO Given the limitations of personalized medicine, empirical science continues to play a predominant role in cancer care and in research, and we won’t be abandoning this approach anytime soon. Because of the need for personalized medicine and empirical medicine to coexist, we’d welcome your comments to put personalized medicine into an operational perspective.

Dr Perez I think in order to provide personalized medicine there will always be a portion of empiricism, because I don’t know if we are going to be able to devise a treatment that will target all molecular abnormalities that may be driving a tumor for each individual patient.

I think what we’re going to be able to do is subclassify patients based on pathways, and I hope that will give us enough information to be able to have maybe 10 types of breast cancer and guide therapy based on that. Ideally, if a tumor is found to have 3 abnormalities and we can find 3 drugs that address those abnormalities, then obviously we will be able to provide individualized care to each patient. But I think the way things are going, each breast cancer may have many more than 3 abnormalities, so we may not be able to treat a patient with 20 drugs.

A realistic approach means that, while most of what we do now is empiric, the balance will change. But we’re going to have to be realistic related to whether we will address every target in the tumor. Think about this: It’s more than just being able to identify the abnormalities in the tumor with genomic analysis. We need to identify whether they are drivers of tumor growth or metastatic potential.

PMO Is there any strategic guidance that you might offer practicing oncologists to give appropriate emphasis to personalized medicine in order to integrate it into their clinical strategic process?

Dr Perez I think the guidance is going to be that we are going to need more access to more specimens in the evolution of the disease process to be able to manage patients in an optimal way. Right now we’ve been practically relying on the limited molecular testing that we can do on the original tumor and expect that that will provide us the guide on how to manage patients when they develop relapse, when they go into a metastatic setting, when they develop further progression, so I think that has to be the guide.

We may need to be much more attentive and consistent in our recommendations to obtain biopsies so that we can do this molecular analysis at the different stages of the disease process. Then people like me and others will figure out the best tools with which to evaluate the molecular changes. But without tissue that can be collated with patient outcome, we will not be able to advance as quickly as we should.

PMO How would you advise oncologists so that they could develop a sophistication in their understanding of personalized medicine techniques to distinguish real innovation from just novelty?

Dr Perez Very good point. I think it’s going to be up to all of us. I think it’s our responsibility to educate physicians, because the terminology is different from what they learned in medical school. So we need different specimens to do different types of gene sequencing. We need to help them understand the difference between doing an exome analysis versus a whole genome sequencing and the data that we can get. There’s a realization that we’ll need to work together to educate practicing oncologists, and it will be done


PMO Can you describe breast cancer patient subgroup risk stratification as part of the personalized medicine revolution? What are the different categories of breast cancer prognostic expression signature sets?

Dr Perez They’re currently being defined. That’s actually 1 of the challenges in the field. Different people are using different technologies to come up with signatures, and the genes from one signature do not correlate with the genes found with the other mechanisms, so these signatures seem to be nonconsistent, in my opinion. That’s because the platforms used to develop the signatures are different. So I think as we get more and more into understanding the exome, and we get to understand the whole genome sequencing, this will get better, because right now we just have a smatter of signatures and no real clear-cut relation between one versus the other. To complicate matters even more, not many studies have been done in which a specific set of tumors has been tested using different techniques to see if the signatures that are derived are the same or different.

PMO How do they reflect tumor status?

Dr Perez Well, right now we and others are working on redeveloping signatures for what’s ER-positive versus ER-negative breast cancer. I think we will eventually develop signatures for patients with HER2-positive breast cancer that will have a likelihood of responding to trastuzumab or lapatinib or other agents.

Right now the status of the signature is that no one can tell me whether it’s a primary tumor or a metastatic site. That’s not the area that we are going to be pursuing. We’re just going to take a tumor wherever it is, try to understand the biology so that we can offer the best therapy for our patients.


PMO Drilling down 1 step further. These signature sets, how do they help delineate outcome?

Dr Perez When we find mutations in the signature sets that are important for the pathophysiology of the breast cancer, then we’ll be able to really devise therapeutic strategies using combinations of agents, and that will be directly correlated with improved patient outcome. At least that’s the way I envision this.

PMO Historically, how reliable are they?

Dr Perez Well, we haven’t had the tools in the past. Right now the best we can do when we have a patient with advanced cancer is to choose any of the drugs that have been approved by regulatory agencies. Each drug has approximately a 20% to 30% activity, and that’s just not good enough. I hope that we’ll get to the point where we’ll be able to do the analysis of the tumors and tell the patient, “hey, your tumor has an 80% chance of benefiting.” Even with that I’m going to be very happy.

PMO Are these codified in algorithms by the National Comprehensive Cancer Network or American Society of Clinical Oncology?

Dr Perez Not yet, because there are so many signatures that are being published, based on what I consider to be evolving technologies, that they have not been validated by others enough to really have big names yet.

PMO What are the specific breast cancer patient types – and tumor types? Is the revelation of personalized medicine processes in breast cancer leading to subspecialization based on these patient types and tumor types?

Dr Perez This will happen, but right now they’re really not validated enough for me to tell you that there are going to be 4 subtypes of breast cancer, or 10. I think this is going to require a little bit more work.

Right now we’re still using ER/PR [progesterone receptor] and HER2 to make the decisions for patients. Everything else is to be validated.

PMO In what way would this profiling aid the oncologist in risk stratification, in establishing prognostic expectations, and in identifying personalized medicine diagnostic tests and treatment?

Dr Perez Just to give you an example of how this is going to be revolutionized: Right now when a new patient is diagnosed with breast cancer, we do surgery and take the tumor out, and then we give adjuvant therapy, or we may decide to give neoadjuvant therapy. But there are essentially about 5 drugs that are used where I think there may be 40 abnormalities in the breast cancer that may be important to drive pathogenesis. I think in the future we’re going to have many more agents that are really directed toward their normal targets. Instead of having 1 general recommendation for adjuvant ther-apy, we may have 10. Because right now everybody’s treated the same way, bottom line. Everybody gets chemotherapy, that is anthracycline, taxane-based chemotherapy. That’s it. That is the backbone of treatment for patients with cancer, even with early-stage breast cancer, and we need to be much more specific than you’re seeing, you know, 2 classes of drugs that are most effective.

PMO If you could, please discuss the discovery and advancement of the approved biomarkers for breast cancer, and the investigational ones.

Dr Perez In terms of a routine management of patients with breast cancer, there are only 3 markers that are done: ER, PR, and HER2. And it’s not even HER2/neu. HER2/neu refers to the HER2 gene, so we prefer to call it HER2 because HER2 addresses evaluation of the protein or the gene, and we can do either for a therapeutic decision. We don’t have to do both.

EGFR has been looked at, and it has not panned out yet as a marker to drive therapy in patients with breast cancer. Regarding p53, certainly we know it’s expressed as abnormal in a majority of tumors, but it’s not tumor-specific enough to guide therapy, and people have tried to look at p53 in the context of anthracyclines, but again nothing specific.

So we just wonder when looking at AKT, looking at PA3 kinase, looking at IGF, are those markers going to be important, but right now we don’t know. MET may be a very important marker in breast cancer. Protein kinase D1 may be important, but again, all these things are a part of research right now. Nodule may be important. But we need to figure out what those markers mean in the context of the physiology of this disease.

PMO Why are these attracting the most interest and research?

Dr Perez One of the reasons is that companies have drugs that target a particular marker, because the other markers that we haven’t talked a lot about today are the microRNAs, which may control 30% of our genome, and theoretically they could be extremely important. But more work needs to be done in that area.

Another area certainly is phosphorylated proteins, but that is very difficult to test for because phosphorylation occurs, and in 1 minute it’s gone, so you cannot assay for the impact of phosphorylated proteins in the context of tumors. We can do that very well in cell lines, but that’s very different than in humans, so I think that’s going to be a very challenging area to conquer.

The other biomarker of interest has been circulating tumor cells. Before we can really explore this more, we need to first understand if the genomic markers in circulating tumor cells are equivalent to the genomic markers of the tumors that are invading different organs. And second, can we assay for a variety of molecular markers in circulating tumor cells when we may have just 5 cells per mL? When is the technology going to be good enough that we can take 2 cells and do enough biomarkers in those 2 cells to be able to drive a better understanding of breast cancer?

If we can avoid having to do invasive biopsies and just assay the circulating tumor cells, that would be great. But before we do this we again need to address those 2 issues: are they the same cells, or is there something different about the cells that decide to shed and be in the circulation compared with the tumors that are invading the liver and causing death?

PMO What would you regard as landmark research in personalized medicine in breast cancer and the clinical basis for its significance?

Dr Perez The term “personalized medicine” is a very broad term, but essentially the most important biomarker in addition to ER really has been HER2. HER2 was identified, and then we developed a strategy in the metastatic setting to target HER2, clearly improving survival. And then we were able to take that marker and use these agents in the adjuvant setting, and we decreased relapse by 50%, and we are improving survival by 38%. So now by using this biomarker, this is part of personalized medicine. We’ve been able to change the natural history of HER2-positive breast cancer from a very aggressive type of breast cancer to one that is very manageable and potentially curable.

PMO Are there any particular clinical, business, or policy factors impeding progress in breast cancer, especially involving personalized medicine.

Dr Perez It’s really time and money, because we need to put enough resources not only to have sequencing machines but to have the patient resources to have enough tumor specimens to not only put the specimens in a machine but also to have the bioinformatics personnel to put the story together. There has to be a focused and concerted effort to devote enough resources to this area in a timely fashion to really get the answers.

PMO Are there particular areas of consensus and particular areas of controversy concerning biomarkers in particular and personalized medicine in general involving breast cancer that you see as standing out, very obvious ones?

Dr Perez The controversy is, in my opinion, that some people feel a lot of gene work has been done and “it hasn’t translated into anything yet.” Some people have become naysayers because they’re not as involved as some of us are, trying to really make specific advances on this.

I think that is the main controversy. This is a new field, and people don’t understand it. Breast cancer is a genetic disease, and we really need to unravel the genetics of this disease, what drives gene function, what is the product of those genes, so that we can realize this dream, going back to the dream of personalized medicine for our patients with breast cancer.

PMO In your research priorities, are you going to focus in 1 area?

Dr Perez Absolutely, although our interests are very vast. I hope that our work will eventually lead to a new molecular classification of breast cancer and we’ll identify the relevance of types of breast cancer to drive treatment decisions. Along those lines, this finding of fusion genes in breast cancer may be very important. But first of all, we need to identify what the redundant fusion genes are, and number 2, do they have any associated protein products, and can they be utilized for the development of new treatments.

PMO That gives us a perfect place to end our discussion today. Thank you for your time.

Dr Perez My pleasure.

Dr Perez is Deputy Director, Mayo Clinic Cancer Center in Florida, and director of the Breast Program and the Serene M. and Frances C. Durling Professor of Medicine at Mayo Medical School. She is a cancer specialist and an internationally known translational researcher at Mayo Clinic. Her roles extend nationally, including chairing the Breast Committee for the North Central Cancer Treatment Group, as well as other positions within the American Association for Cancer Research, the American Society of Clinical Oncology, and the National Cancer Institute. Dr Perez has developed and is involved in a wide range of clinical trials exploring the use of new therapeutic agents for the treatment and prevention of breast cancer. She also has developed studies to evaluate the role of genetic biomarkers in the development, aggressiveness, and therapeutic efficacy for breast cancer.

Uncategorized - November 19, 2012

EGFR Mutations, Not KRAS Mutations, May Be Predictive for Sorafenib Response in Advanced NSCLC

Treatment with sorafenib did not improve overall survival (OS) in patients with heavily pretreated advanced non–small cell lung cancer (NSCLC) in the overall analysis of the MISSION trial reported at the 2012 ESMO Congress. A post hoc companion biomarker study of MISSION presented at the same meeting suggested that EGFR [ Read More ]

The Last Word - November 19, 2012

A Personal Introduction and Invitation to Join My Journey Through the World of Personalized Medicine

Welcome to our new column, The Last Word. At the outset, let me comment on the irony of the title, for this is not a forum that presumes to have the “last word” on any aspects of personalized medicine. The column will serve as the concluding editorial in each issue [ Read More ]