Maurie Markman, MD; Mark G. Kris, MD; John L. Marshall, MD


June 15, 2012

This feature requires the newest version of Flash. You can download it here.


Maurie Markman, MD: Hello. I am Maurie Markman, Senior Vice President for Clinical Affairs at Cancer Treatment Centers of America. I am based in Philadelphia, Pennsylvania. I would like to welcome you to this Medscape Oncology Insights roundtable, coming to you from the 2012 annual meeting of the American Society of Clinical Oncology (ASCO®).

Our topic is the increasing influence that biology plays in the treatment choices we make for our patients. Joining me today are Dr. John Marshall, Professor of Medicine, Georgetown University, and Director of Clinical Research for the Lombardi Comprehensive Cancer Center in Washington, DC, and Dr. Mark Kris, Chief of Thoracic Oncology at Memorial Sloan-Kettering Cancer Center in New York City.

I will start with Mark, and a broad question. Molecular biology and molecular testing in lung cancer: Where are we today, and how do you see the future?

Mutation Testing Standard in Lung Cancer

Mark G. Kris, MD: I still have to pinch myself because of what has happened in the field of lung cancer, both in terms of the improvement in outcomes for our patients, first and foremost, but also because we have become a model for how to treat really tough cancer in 2012. At this year's ASCO® meeting, we have heard more of the same. In the year since the last ASCO®, the National Comprehensive Cancer Network guidelines,[1] the roadmap for treating every patient with this illness, now say that the first stop on the journey is mutation testing to test for epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) in virtually every tumor specimen. That has become a standard of care.

Not only does mutation testing identify which patients to send for treatment with certain drugs that offer a better chance of benefit and less chance of harm, but it also has allowed us to look for additional drivers and other mutations for which we can come up with effective therapies that would be better and less toxic.

At the meeting here at ASCO®, we have heard about a new drug for EGFR-mutated lung cancer, afatinib,[2] and it is the first upfront drug [in lung cancer] that we will have in the United States when it is approved. We have mutations in the ROS gene,[3]which are found in a small percentage of lung cancers, but of importance, these lung cancers are not the ones where we already have mutations. It means that thousands of people every year have an effective, relatively tolerable treatment, with really dramatic effectiveness based on the data that Dr. Shaw[4] showed us yesterday.

This isn't pie in the sky. We don't have to wait 10 years for this treatment. These patients are treated with crizotinib, a drug that is available in every pharmacy in the United States. They found another target, the RET gene.[5] Patients who have mutations in RET don't have any other mutations. Now, thousands more people have another option open to them. Patients with this kind of mutation have drugs available to them today: sunitinib, vandetanib, and sorafenib. All these drugs can be used to fight this cancer. It is an amazing story.

Two things have happened. One, we have entered the generation of multiplex testing. We have seen multiple platforms where patients' tumor tissues (and a reasonable amount of tumor tissue) can be tested for all these driver mutations at the same time, at the start of the illness, so you can immediately get a patient on that treatment. We have seen a number of examples of that.

We also saw that another type of lung cancer -- squamous cell cancer (which had not been included in the testing and targeted treatments) -- will benefit from mutation testing. Information was presented[6] from my institution, where we have started testing these specimens upfront and use that information to either guide people to a targeted trial or guide them to a trial specifically for squamous cell cancer. We have more targets, more people, more availability of testing, and more drugs. It has been an absolutely amazing story.

The Hunt for Colon Cancer Drivers

Dr. Markman: That is fantastic. John, what about gastrointestinal cancers?

John L. Marshall, MD: I wish I had the same story to tell. We have parts of the story, but we certainly haven't gotten it to the same level as our friends in lung cancer. We are very aware of KRAS testing at this point.

In both worlds, the molecular target has come more from the therapy than from the molecular biologists. In colon cancer, for example, there are 50-75 broken genes, but what are the drivers? Which are the important drivers? Of course, our EGFR story is not nearly as successful as yours. We started with using breast cancer rules about receptors being present and finding small amounts of benefit from EGFR antibodies, and only retrospectively we found that it is KRAS that divides patients between responsive and unresponsive.

Then we started to drill down even further to learn what kind of KRAS mutation the patient has. I am discouraged by this discussion because people are saying that if the patient has a 13D mutation, maybe the patient will get a little benefit from the therapy, not nearly the kind of dramatic responses that you [in lung cancer] have seen when you find the right drug for the right patient. We are seeing minor changes, even in the enriched patient population. We are continuing to move this forward, continuing to try to understand who the right patients are. We clearly don't have any sort of black-and-white world.

Even worse is the issue of angiogenesis. The other major biologic in use today in colon cancer is antiangiogenic therapy. At this meeting, 2 new antiangiogenics have come to the table.[7,8] In addition, data are showing that bevacizumab beyond progression has benefit,[9] but the benefit is relatively small: 1-2 months of survival benefit at each stage along the way. The question then becomes one of value. Is the drug worth the cost in an expensive drug world for a relatively small gain? The cry for us is for additional biomarkers and additional understanding of angiogenesis so that we can do some patient selection.

As new drugs come forward, that is going to be a critical element. You set the stage. If you find the right patient and the magnitude of the benefit is high, then it is a no-brainer to get that drug out there. It's easy, it's short, and it's effective. Return on investment is high for all of us who do drug development. The days of large randomized clinical trials with a risk reduction of 0.8 and a survival advantage of 1.5 months are numbered. We saw a lot of that, unfortunately, today in the colon cancer presentations.

For Each Ovarian Cancer Patient, a Disease of Her Own

Dr. Markman: Very interesting. In ovarian cancer, we are even further down the line. We hope for the lung cancer story, or even the colon cancer story. The ovarian cancer story is that there are a tremendous number of mutations and they vary tremendously from cancer to cancer. It is remarkable. In fact, in the high-grade serous cancers (and what you see in all cancers) are p53 abnormalities. If you look from one tumor to the next, there will be 100 mutations in each, none of which is duplicated.

The mutations that seem to be very important in a small subset of patients are the BRCA1 and BRCA2,[10] which are certainly targetable. We are beginning to consider a systems approach to look at the molecular biology of cancers that don't have BRCA1 or BRCA2 but that, in fact, have a profile similar to BRCA. Studies at the meeting began to examine this, but there certainly haven't been any home runs in that area. The pathway that seems to be important right now is antiangiogenesis. Bevacizumab is the major drug, with several positive trials.[11] No overall survival benefit was seen except perhaps in a subset of patients with quite advanced disease, but we have no biological target. It basically would mean treating all patients for long periods of time. There may very well be a benefit. I am not questioning that, but being able to pick out the patients who benefit most is very difficult.

It is a fascinating story for someone like me who has been involved in ovarian cancer research for 30 years. But picking out that particular target is very tough. That is where we are.

Genes, Histopathology Define 5 Colorectal Cancer Types

Dr. Marshall: I want to throw in a bright spot from the colon world. It wasn't all doom and gloom. It's sort of a different way at looking at it. So far, we have sorted colon cancer into microsatellite instability (MSI), microsatellite stability (MSS), KRAS wild-type, and mutated cancers. We are starting to divide it up.

Of interest, some patterns are emerging that couple gene expression and mutations with histopathologic characteristics. A very cool abstract[12]that was presented a little too far off the mainstream looked at colon cancer and was able to divide it into 5 categories. Looking back on big retrospective studies, these categories, both genetically and histopathologically, sorted these patients into very different diseases.

This was a non-therapy-driven analysis. It wasn't based on some drug dividing the populations, but on its innate appearance and genetic profile. Look for that in the future as we develop new medicines and new approaches.

Multiplex Testing: A Tsunami of Information

Dr. Markman: I have one final question. I don't expect a crystal ball here, but as both of you know, the tsunami is reaching us. As Mark pointed out, there is this issue of multiplex approaches that are able to provide genetic and genomic information on tumors, as well as the cost of producing these data. A number of companies either have commercial products or they will shortly, where a patient can choose a test, or go to a doctor and ask for a certain test, and the doctor will order it.Whether it's lung cancer, colon cancer, or whatever kind of cancer, you will basically go in and say, "Please order this test; Doctor, please help me, based upon this information." What's your advice to the practicing oncologist who is going to be increasingly confronted with this world?

Dr. Kris: We always have some information that we don't quite know what to do with. You get the pathology report that [the tumor] is poorly differentiated or well differentiated. How exactly do you use that to choose a therapy? We are not sure, but we somehow manage to do it. We are going to find -- and I'll speak for lung cancer because it's a little easier -- that we will identify more targets and there will be specific treatments for those targets. At the very least, mutation testing tells us what therapies are not helpful. There is more information now about KRAS, actually. If you have a KRAS mutation, your chance of benefit from erlotinib or afatinib up front is very small. At the very least, you will spare patients that side effect. Even a negative piece of information is helpful, such as knowing that you don't have the target.

Multiplex testing is definitely going to happen. These platforms are all over commercial organizations and individual hospitals. The cost for a machine that can do multiplex testing is not huge. There are some in the $50,000 range, and it fits into the practice of a pathology department already, so we are going to see a lot of that. Every tumor is going to be tested with these multiplex platforms. Doctors are going to have the information and make the best of it.

There is going to be a lot to learn, even for lung cancer. What is the role of a p53 deletion in somebody with an EGFR or RAS mutation? It must have some kind of role. It may help us further define a group of patients in whom it is likely or not likely to help, and we are all going to have that ability. I don't think people will be confused. People are going to work through it. Ultimately, it will lead to better treatment.

The other interesting thing about this -- and this has come out at the meeting -- is that you have to delve into the specifics of it. People say "EGFR mutation," but not every EGFR mutation is a sensitizing mutation. That's the first thing. We have to learn more about that.

The second thing is an abstract by Daniel Costa,[13] from the Deaconess Hospital in Boston, who went into these nonsensitizing mutations. When you actually look at the DNA sequence and which amino acids are inserted, there is one mutation that is sensitizing. It's fantastic. It doesn’t pertain to a huge number of people, but it is huge for that patient. With this information, we can precisely define what to do. There will be a learning curve. Even among the negative exon 20 insertions, you might find one that means you are going to have some benefit. We will all be able to do that. That's just an example.

From Spreadsheets to Clinical Decisions

Dr. Marshall: I am a bit nervous. I agree completely with everything you said about the hope and the potential, but there was a study published recently in the New England Journal of Medicine[14] about kidney cancer, in which they took a tumor and did biopsies from 4 quadrants of the same tumor. They performed fairly deep sequencing on these samples and found different profiles within the same tumor.

How much of that variability is garbage, background noise? What are the key drivers? What was good in that paper was that there were some common threads that may in fact be key therapeutic drivers for us. I'm also a bit nervous that much of cancer behavior may be posttranslational. When we are looking simply at a patient or a tumor's DNA, we may be missing some of the important biology that may be in play in that given patient.

You said the key word, and that's "bioinformatics." How do we take these large datasets and convert them into something that we as physicians can find on a report or on a flow sheet and say "this means this" and "this means that"? As a treating physician, that is a fear. To academic oncologists, that's a fear. How do I get this huge spreadsheet and be able to say "you're in trouble" or "you're not," and "chemotherapy will help you" or "chemotherapy won't help you."

Dr. Kris: I see you haven't been watching my Medscape commentaries. Everybody shared your concern after that New England Journal of Medicine article. The truth is, the chronic myelogenous leukemia is not just Abelson (ABL). There are other mutations. The right drug against a driver can do a lot of good. It's the same thing with lung cancer. There are probably hundreds of mutations in those EGFR mutants, yet when you find the driver, it's helpful.

Secondly, the advances in informatics are extraordinary. There are a lot of companies, and I am involved in a project with Sloan-Kettering and IBM to use the Watson technology for this purpose and ultimately to bring it to an iPad in your hand, which can help you sort these things out. This is not pie in the sky. Many people are doing this. ASCO® has a vision to do this as well. It will be here in no time. We are going to get the help that we need, and we do need it. It's going to be here.

Dr. Marshall: We have to link the genetic data with the clinical behavior. There are some major barriers to that in our research worlds. If we could in fact include all patients and all tumors into the haystack and begin diving down on that, the process could be a lot quicker. I agree with you, but I do think there are some real barriers that are getting solved. We aren't quite there yet.

Dr. Markman: I want to thank my colleagues for a most provocative discussion. I want to thank you, our audience, for joining us for Medscape Oncology roundtable. This is Maurie Markman, reporting from ASCO® 2012 in Chicago.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.