TOPICS > Science

Extended Interview: Janet Woodcock Discusses Cancer Biomarkers

March 28, 2007 at 4:30 PM EDT

SUSAN DENTZER: What do biomarkers mean to FDA in terms of the future of drug discovery, diagnostics, therapeutics for patients, particularly patients suffering from cancer?

DR. WOODCOCK: Well, cancer is probably the most promising field right now for biomarkers, and from FDA’s point of view, I think biomarkers are the future of medical therapy, both for diagnostic purposes as well as for cancer therapeutics. And it’s going to be a very challenging road ahead in getting this new science of biomarkers into medical products the doctors and patients can actually use, but that is the future we need to move toward.

SUSAN DENTZER: And why is it the future?

DR. WOODCOCK: Because these new markers get to the mechanism of cancer. They get to the actual causes in the cancer cells of what’s gone wrong, why they’ve turned into cancer cells. They provide a lens, a new way of looking at people so we could determine they have cancer earlier, diagnose earlier, as well as pick the right treatment.

SUSAN DENTZER: Let’s talk about FDA’s Critical Path Initiative, what it is, and how biomarkers fit into it.

DR. WOODCOCK: The Critical Path Initiative is the track that a new innovation, an idea, scientific idea has to follow from the laboratory, to actually get in the hands of doctors and patients. It has to go through a whole series of steps that we call the critical path. And the Critical Path Initiative is attempt to bring more science and pull those innovations along in a rigorous, safe, and useful way and move them faster into the hands of doctors and patients.

SUSAN DENTZER: And the role that biomarkers will play in that is what?

DR. WOODCOCK: Biomarkers are extremely important in the critical path because at every step of the way we need new biomarkers to predict whether or not a product is going to be safe enough, whether or not it’s going to be effective, who it should be used in, and how to manufacture it properly. And biomarkers play a role in all of these critical steps that have to be taken.

SUSAN DENTZER: And they play a role in two senses, both in the diagnostic sense and in the therapeutic sense. Would you elaborate a bit on that?

DR. WOODCOCK: Yes. Well, biomarkers can be turned into diagnostic tests, of course, and those tests can be used to either detect, for example in cancer, detect a tumor much earlier than the kind of tests we use now, or they can be used to direct therapy. Or they can be used to say you are a patient who is at high risk for a side effect, and we shouldn’t give you this therapy. So biomarkers can really help us make treatments safer.

SUSAN DENTZER: So they really are, as you say, just opening up a whole new world of understanding the disease and how to treat it.

DR. WOODCOCK: That’s right. And it becomes – because all the science that we’ve invested in over the last 30 years, all of the medical advances people read about in the paper, they are now at the point where we can actually use them, turn them into real diagnostic tests that doctors can use on patients, looking at the proteins in their body, looking at the genes in their cancer cells and how they’ve mutated, looking at images of their cancer in ways we’ve never been able to do before, to tell us all kinds of new information.

Teaming with other agencies

SUSAN DENTZER: Now, obviously this is of such import that you all are teaming with NCI and with CMS to move this process along as fast as possible.

DR. WOODCOCK: We think this is so urgent that amongst the federal sector -- the Food and Drug Administration, the National Cancer Institute, and the Center for Medicare and Medicaid Services, or CMS, that we've teamed up together to make sure from the innovation side, the science side, which is NCI, the medical product development side, which is FDA, and then CMS, the reimbursement, that we can have a smooth process for biomarkers. We call this the Oncology Biomarker Qualification Initiative.

SUSAN DENTZER: And the CMS end of it is to make sure that at the end of the line Medicare is going to pay for these tests?

DR. WOODCOCK: Yes. We need to make sure that not only are the tests developed into medical products, but there's enough information about them and about their value that CMS is able to reimburse for them.

SUSAN DENTZER: Now, only nine of 1261, or some numbers close to that, of proteins that have been identified as having links to cancer have so far been approved by FDA. Why is that?

DR. WOODCOCK: Only a fraction of the potential biomarkers that have been identified in cancer have been approved in tests by the FDA, and there are a lot of reasons for that. It relates to all the barriers to getting the scientific discoveries actually translated into medical products.

First of all, not all associations are true predictors, and so a lot of scientific work has to go on clinical trials, on patient volunteers and so forth to determine whether or not these proteins really predict accurately enough some event in people.

The second thing is a diagnostic company has to actually create a test, a kit, and configure it in a way that's reliable and precise so you, if you're a patient, if you get results, and your doctor gets results, you can rely on those results because those results are going to make a life altering decision for you, maybe say you might have cancer, or you should take this chemotherapy and not that therapy. And these are very important pieces of information that need to be right.

So there's a barrier there in gathering that information, manufacturing that kit, and getting enough reliability of the test that it can actually be used out in practice.

From idea to reality

SUSAN DENTZER: And this whole process of so-called validating these biomarkers could take large trials over many years. What's validation and why is it so hard?

DR. WOODCOCK: The process of taking a scientific idea or a discovery of an association and moving that to where you can rely upon it and understand it is called validation, or we prefer to call it biomarker qualification. But whatever you call it, it's a very difficult process because you have to gather up enough information about the performance of the test in a real world situation to know you can rely on it.

You have to do clinical trials, not only in patients who may have the disorder, but in people who don't have a disorder because you have to see if you test them, does the test still tell you something wrong? All right, and you have to see in people who have the disorder, for example, how many of them does the test indicate are positive. It might have not a high enough response. And so, you know, would have too many false negatives.

So there's a lot of work that has to be done. It might be expensive, it might take a fair amount of time to complete. And it's very difficult for the small businesses often that are formed around these diagnostic tests to do that type of work.

SUSAN DENTZER: And because this is a long disease process often in cancer, you might have to follow people for years, correct?

DR. WOODCOCK: Yes. The situation is somewhat different between a diagnostic test that's a screening test for cancer and a test that might predict which treatment you should use. Some of the screening tests for cancer are the most difficult, and if you think about it, you can see why.

First of all, it may take a long time for the actual cancer to develop. So you have to wait a long time after you did the test to determine how predictive the test actually was.

The second issue is, it's very difficult, obviously, to get a test that you're going to rely upon maybe to do surgery, or to say you don't have to have further follow-up. That test has to have very high characteristics, very good characteristics, and it's hard to get tests that have that good a prediction.

SUSAN DENTZER: So given all of this, when could the American public expect that biomarkers are going to be a routine tool in cancer diagnoses, cancer therapy - they already are but really it's tumbling off of everybody's lips. "Oh, I had a biomarker test. It showed that my breast cancer is X versus Y." When is that going to be a reality?

DR. WOODCOCK: Well, when you think about it, we use biomarkers now. We use X-rays, MRI's, we look at pap tests, and look at the cells. What we want is a new generation of biomarkers that use the molecular characteristics, genetic characteristics, advance functional imaging, things like that, and that's coming. We'll probably see it soonest, along with therapeutics as we're doing now.

SUSAN DENTZER: So we will see in the next few years diagnostic tests that say should you take this targeted cancer therapy or that therapy, or not?

DR. WOODCOCK: The screening tests are going to take longer for the reasons we have discussed. It just takes longer to validate them, and there's a very high bar for a screening test because you don't want to send people to the operating room based on a false result.

A 'high bar'

SUSAN DENTZER: Just to get in general terms at this notion of labs being able to do these tests properly, let's say something to that effect, that there is - it's probably not likely these are going to be little kits that you buy in the supermarket or on a drugstore shelf to diagnose yourself for cancer. You're probably going to get them in your doctor's office. They'll be sent to a lab.


SUSAN DENTZER: How will we know that the labs are getting the right results, because we have a history of not getting the right results, even in pap tests, from the labs.

DR. WOODCOCK: Right. Well, there is going to be - there's a very high bar all the way from getting the specimen out of the patient, processing the specimen correctly, performing a test correctly, and in interpreting the results, right. And all of this is going to have to happen correctly.

Now, I think we as a society can do this. If you think about airline travel, okay, we send all these passengers - we can do these high tech things safely and correctly, but it's going to be challenging, and we can't minimize the difficulties because the consequences for an individual patient, if any of those steps are performed incorrectly, those consequences are very serious.

If you have a blood sugar test done or a cholesterol test done, and it come back abnormal, the first thing your doctor is going to do is repeat the test. But in this case, in these types of tests, you know, you may not have a second or third chance. You can't just have more blood drawn and have this done again and again.

So we're going to have to get this right at every step of the way, including the laboratory performance.

A new generation of imaging

SUSAN DENTZER: And then finally, we have begun a biomarker trial, the FDG PET trial, which is an imaging linked trial. Let's explain what that trial is.

DR. WOODCOCK: As a part of our partnership with the National Cancer Institute, FDA and NCI have worked to start an advanced biomarker trial. This trial is an FDG PET, which is a kind of functional imaging it's called, and in cancer it looks at how the cancer tumor takes up glucose and uses it.

So it's how active your tumor is basically, and we're looking in a disease called non-Hodgkin's Lymphoma, which is a type of cancer, and we're trying to see - we think that potentially FDG-PET can tell doctors more definitively if a patient has had a complete response to the treatment. And obviously, this is very important.

This is important in drug development because you can tell us earlier if new treatments were better. And it's important out in the clinic because if you've had a complete response, you may not need additional therapy. And right now, in many, many patients, we don't know whether or not they've had a complete response because some tumor residual is still there according to X-ray. But we don't know if there's lost cells inside or not, and a scan like this, which takes up glucose and so forth can tell us whether that tumor is active, or whether there's just inactive mass left.

SUSAN DENTZER: And in this instance we'll be able to test the activity, if you will, of the biomarker, not the way we do with many other proteins, by giving it essentially an antibody really to test. We'll be doing this by watching the action in the cell, watching it with a PET scan. Correct?

DR. WOODCOCK: Yes. In this case, because this is called functional imaging, we can actually look at a picture of the tumor and watch it take up glucose and see how much of the glucose it took up, and what its functional activity is like. And I can tell you, there's a whole generation of new kinds of imaging coming along that will be new probes into the behavior of tumors and their biology, and how well they're responding to treatment. And this is an extremely exciting field. But just like these other diagnostics, it's going to take a lot of work to make it precise, reliable, and something we can count on.

SUSAN DENTZER: And just to close on this, if the tumor is taking up glucose rapidly, is that good or is that a negative?

DR. WOODCOCK: If the tumor is taking up glucose, it means the tumor is in there and it's - it's active, and that's bad as far as that means the tumor hasn't completely gone away. For example, we've seen some cases with chemotherapy where the patient is given the chemotherapy and they've had a PET scan first that shows the tumor is very hot, taking up a lot of the glucose, and then a few days later do another PET scan and the tumor is silent. It's not working.

That's very good. That's what we want to happen with chemotherapy. We want to shut that tumor right down. But we have to validate how this works, and whether or not it's adequately predictive before we can make treatment decisions based on these PET scan results.

SUSAN DENTZER: And how long will this trial have to go on, and how many patients will have to be involved before you have the answer?

DR. WOODCOCK: There are several hundred patients that will be enrolled in this trial, and the trial has to go on several years to enroll that many patients. However, for any individual patient it will take about six months or so to have the course of treatment, and have the last PET scan, and then they have to be followed for quite some time to see if they relapse, and that will give us the information about whether that final PET scan actually predicted what happened to them many months later.

SUSAN DENTZER: So taking a guess about at what point you'll have enough information statistically to make a judgment call on this - five years?

DR. WOODCOCK: It's going to be many years. Yes, it may be five years before we have all the results in.