Pathways of Resistance in Metastatic Colorectal Cancer

Transcript:Richard Kim, MD: We know that the EGFR pathway activates the KRAS, the RAS, and the MEK pathways. We know that the KRAS is a predictive marker for EGFR. So that data is very well understood. But most of the patients who are wild-type get an EGFR drug. They respond to some degree. Then, after 4 to 6 months they stop responding. The question is, why do they stop responding? There are many theories and ideas behind the secondary resistance to EGFR drugs. Most of the data that’s currently out there shows that there can be a new KRAS or NRAS mutation that can occur as patients get treated with an EGFR drug. So, that’s probably the first point. Second point is that there can be a KRAS amplification, which can actually confer resistance to the EGFR drug, as well. Third data behind us is that HER2 amplification may play a role in this.

Now, last year at ASCO, there’s some data that if you have HER2/neu positive, but KRAS wild-type, those patients also do not respond to EGFR drugs, in small studies. The HER2 amplification may be a primary resistance, as well as secondary resistance to EGFR drugs. And third but not least, there can be amplification of [the] MEK pathway, which can also confer resistance to an EGFR drug as well.

Daniel G. Haller, MD, FACP, FRCP: Much more is known about EGFR mutations evolution in non—small cell lung cancer. And this is in part because EGFR mutations are the target for tyrosine kinase inhibitors in non–small cell lung cancer. EGFR mutations are not that important in colon cancer. RAS is, but not EGFR mutations. But the whole pathway probably does mutate during treatment, both spontaneously and because of prior therapies. More is known in patients with non–small cell lung cancer, and that’s why the issue of liquid biopsies is much more important in those patients in determining next-line therapies after erlotinib, for example.

But in colon cancer, there has been work looking at mutations and also looking at whether or not one can reuse EGFR agents. I found at least one retrospective study—there are no prospective studies—with about 82 patients coming out of MD Anderson who had been exposed to EGFR agents typically as first-line therapy. And then after second, third, or fourth-line therapies, patients with no other therapies to choose were given either the same drug, cetuximab, or a different drug, panitumumab. And a small number of these patients seemed to respond.

And the two clear markers for the ones who would respond in this particular study were, what was their response to the therapy in the first place? That kind of made sense. And the other one was the duration of their response, the duration from the time they stopped their earlier therapy to the time they restarted it. This has also been seen for chemotherapeutic drugs. For example, people have been re-challenged with oxaliplatin or irinotecan after a vacation, and have been seen to show response. There seems to be changes in all of these pathways that can occur that can not only cause worse outcomes, but may actually restore sensitivity both to standard chemotherapeutic drugs and to biologics.

Charles S. Fuchs, MD: One of the great controversies in the management of metastatic colorectal cancer is which biologics to use and when. The simple answer is that if you have a KRAS- or all RAS-mutated patients, they shouldn’t get an EGFR antibody ever, and that’s simple. But if it’s RAS wild-type, and all-RAS wild-type, we have choices, which essentially are bevacizumab or the EGFR antibodies, cetuximab, and panitumumab. And what we really don’t know is, how do we best sequence it? I don’t presume to know the answer. In the US, you’ll see a lot more frontline bevacizumab use. In Europe, you’ll see a lot more EGFR antibody use. Is anybody right or wrong there? No, because what we know from CALGB 80405, a large US study, is that whether you start with one antibody or the other, you do relatively the same. And moreover, we’re getting survivals in the range of 30 months, which, as a median, is inadequate—but it is three times better than what we were doing 10 years ago.

These antibodies clearly have value, but the truth is, I don’t know the best way to do it. Anyone who presumes to tell you that this is the way to do it first-line, second-line, they’re guessing. We’ve got to stop guessing, and we need more data that actually fundamentally understands what a patient’s individual tumor is, how it’s made up, and which drug is going to work better for that patient.

Transcript Edited for Clarity