Editor’s note: Dr. Hudis hosts the ASCO in Action Podcast, which focuses on policy and practice issues affecting providers and patients. An excerpt of a recent episode is shared below; it has been edited for length and clarity. Listen to the full podcast online or through iTunes or Google Play.
I am really excited to have as my guest ASCO's current president, Bruce E. Johnson, MD, FASCO. Dr. Johnson is the chief clinical research officer and institute physician at the Dana-Farber Cancer Institute. He is a professor of medicine at the Harvard Medical School, and he is probably best known to many of you as the director of the Dana-Farber Cancer Center lung cancer program.
Our conversation today is going to focus on a passion of his, precision medicine. This is, in fact, the theme that he selected for his year as ASCO president and you'll hear more of that at our upcoming Annual Meeting. In January, ASCO held an event focused on precision medicine at the National Press Club. This was part of the rollout of our new State of Cancer Care in America event series, and Dr. Johnson was our host. We focused on extending the reach of precision medicine and addressing some of the challenges that exist in making it even more available to more people with cancer.
CH: Let's begin with a level set. What is precision medicine? How do you define it?
BJ: I'm going to use the National Institutes of Health (NIH) definition and then expand on this a little bit. If you take a look on the NIH website, it says that precision medicine is defined as an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person. And I'm going to adopt this from one of the things you say, Cliff: precision medicine is the right drug for the right patient at the right time.
CH: What are some of the concrete, tangible ways in which therapy, for example, has changed with this evolution? What are the differences in the therapeutic options that patients and practitioners need to consider?
BJ: We gave similar therapies to patients with lung cancer and had very similar outcomes for about two decades in the 1990s and 2000s. Then there was a discovery of a genomic change in lung cancer, EGFR mutations, and the realization that you could treat it with inhibitors of the epidermal growth factor receptor. These ended up working for about twice as long, and we've learned that the patients who get this treatment live about 2.5 years longer. Since that initial discovery in 2004, we’ve found three additional genomic changes that have been able to help predict for the efficacy of targeted agents: ALK rearrangements, which are about 5% of lung cancers, and then two that are relatively rare, ROS1 rearrangements present in 1%, and BRAF mutations (a mutation commonly present in melanoma, where it was first developed), also in 1% of patients with lung cancer.
One that you've been very involved with is an observation made in the 1980s that subsets of breast cancers have multiple copies of the HER2 gene present. In about 20% of breast cancers, one can administer agents like trastuzumab directed against that potential target.
An antibody directed against the epidermal growth factor receptor is used for treating colon cancer. One of the things they learned from studying the genomic changes in colon cancer is that those patients who have a mutation in one of the most common oncogenes present in all cancers, a KRAS mutation, do not have any benefit from getting the antibodies directed against the epidermal growth factor receptor (panitumumab and cetuximab). Precision medicine means that you can use KRAS as a negative selection factor: those patients don't benefit, and you can avoid those side effects.
CH: How does the responsible clinician and the entire clinical care team keep up with this? What do you suggest they do to make sure that they're not missing opportunities, for example, to test appropriately, but also to avoid overtesting and obviously wasting some resources? What do you recommend?
BJ: One of the things that we want people to keep in mind is that there are common tumors in which, at the current time, there aren't any specific precision medicine therapeutic activities that one can target. For instance, in pancreatic cancer, although it has lots of KRAS mutations, we don't have an effective agent for it. You don't particularly need to test this. For breast cancers, there aren't particular lesions that one can target other than HER2 amplification, so you can do an isolated test.
The way we encourage people to keep up on this is first, follow the guidelines. NCCN guidelines give a lot of information. Our ASCO guidelines also give information. Also, one can also contact the academic folks who follow this and seek second opinions of places where they not only have the capability to do the testing but also have a broad portfolio where physicians can refer people to these specific tests. The other thing I would add is ASCO's Targeted Agent and Profiling Utilization Registry (TAPUR) Study, which allows both academic as well as community oncology practices to access these targeted treatments.
CH: Part of the reason that TAPUR is needed, and I think one of the drivers of some level of head-scratching on the part of clinicians, is that while they may need one or two or three genes tested, the clinical experience mostly is that they end up receiving the results of very long, multipanel gene tests. They get many, many results, some of which are relevant and some of which might not be. Can you explain why it is that they receive these long genomic reports when they are really just looking for one or two results?
BJ: When you test a solid tumor, typically these large panels will test between 200 and 500 or 600 genes. Within those panels, you'll get somewhere between 5 and up to 20 genomic changes. And at any given time, when we discuss things that are actionable, there will be trials for these.
For instance, KRAS is one of the really common ones, and there are a lot of clinical trials for KRAS. However, it's not clear that there's any effective therapy. You end up getting this report that says the patient has this KRAS gene. It's actionable by referring them on this clinical trial testing agents directed against KRAS, but it doesn't give you enough information to know that there's not an effective therapy. You probably wouldn't go to great lengths to send a patient off for a KRAS trial at some distant place. Nor would you go to the effort to open the trial until there's some sort of clinical evidence that it's somewhat effective.
This is one place where clinical decision support can be very helpful. One of the things that we think is important for the development of guidelines is keeping that information up to date about what's actionable and what’s useful. Things can be actionable—and one of the things that they consider actionable is referral for a clinical trial—but what you really want to know about is the agent against that particular driver. Is there enough information that you could say that the person is likely to benefit from the therapeutic intervention?
CH: Is it fair to say that a large number of the genes are there because, essentially, once you've done a few, it doesn't cost any more to do these large panels? So clinicians should recognize that they're not paying more but they're getting many more results than anybody thinks that they need in these multipanel gene tests right now. Is that a fair assessment?
BJ: That's a fair assessment. We began systematically testing our patients starting in 2004 for EGFR mutations. One of the things that is problematic is that it's very difficult, once you've done one test, to go back—to do another test if there's a new indication. Our experience was going from EFGR mutation testing in 2004 to adding ALK rearrangement testing in 2008. And to have to go back and find the tissue from the original biopsy, send it off to another diagnostic company, is a real challenge.
The second thing that we've learned is that kind of the magic number is that once you get to around four or five different genomic tests, the cost of doing one of these large panels is about the same—and when I say large panels, I'm talking about one that generates results from hundreds of genes. When you factor in the cost, it's about the same to do the panel of several hundred genes as it is to do the three or four that are indicated. And that's really what drove us.
In lung cancer, one of the things that we anticipate is that there are two additional changes where we think there are likely to be FDA-approved drugs sometime in the next 1 to 3 years: exon 14 skip mutations in a gene called MET, as well as NTRK rearrangements. Having those tests available, if there's an approved drug you can put patients on, would be useful both in academics as well as in community practice.
CH: I know that precision medicine fills you with hope and has not only helped to define your career but has allowed you to help define an advancing standard of care for the disease that you focus on. But even with that said, what part of this keeps you up at night? What is the thing that makes you most worried about either the delivery or the advancement of this area of work?
BJ: The part that concerns me the most is being able to disseminate the information in a meaningful way to the people at the point of care who are taking care of these patients. We've seen a gradual evolution from having more general oncologists to being more specialized, and we did it decades ago at the academic centers. We've seen the movement from community-based practices to hospital-based practices and seen the size of the practices grow. One of the things that we hope is that, as it grows, people become familiar with the diseases that they treat, but to know what's meaningful. And when I say meaningful, to know that a specific genomic change can be effectively targeted.
For instance, in my own practice, we've seen the length of time a given drug works, and the most dramatic example is with ALK rearrangement, where the initial treatment worked for 10 months, and now it looks like it's going to work for somewhere around 3 years or even longer in some of the patients. That really transforms the care for that group, and we would like that information be disseminated rapidly to the point of care for the people who commonly see and treat these patients.
It's one thing to be able to manage the common diseases where the genomic changes get a large degree of publicity, but there are other things—the unusual things, like Merkel cell tumors, where some checkpoint inhibitors look like they may work in a subset. We have to have that information disseminated so people either know that there is a treatment where there wasn't one previously available, or know where to send patients to get those treatments that have been developed for these unusual tumors.