Prognostic and Predictive Markers in Colorectal Cancer

Transcript:Daniel G. Haller, MD: Personalized medicine is a phrase that’s been used much more in recent years. Although having been in practice for 40 years I think we’ve always practiced personalized medicine, which means taking all the factors you know about a patient and their tumor and then making decisions with patients about what is the best therapy for them. There rarely is one therapy that is best for everyone.

The use of biologic markers obviously is extremely important today. There are two types of markers. One are the prognostic markers, and we are awash with prognostic markers. Every year at ASCO, thousands of abstracts are submitted about markers that tell us how patients will do independent of their treatment. Unfortunately, they don’t tell you which drugs to use, they only tell you how a patient might do. This is particularly true, for example, in adjuvant therapy where prognostic markers and recurrent scores are extremely important in telling people who are at high or low risk of having recurrence or dying of their disease.

In metastatic colon cancer, we look more toward predictive markers. Predictive markers tell us how patients will do in regard to a certain treatment. So, you develop a target, which is now very easy to do with next generation sequencing, where you can look for thousands of mutations at one time, some of which are actionable and some of which are not, and then try to pick a drug that might best work for that patient. So we’re awash with prognostic markers and in colorectal cancer in particular, we’re short on the number of predictive markers that we can actually use, and one of them obviously is RAS, KRAS and NRAS testing.

What is known about predictive and prognostic markers? Let’s do prognostic first. Probably the strongest prognostic marker, biologically, that we have in adjuvant therapy is microsatellite instability. In the 15% or so of patients who have MSI-high tumors, especially with stage II disease, these people do extraordinarily well, and they very rarely need adjuvant therapy. So, this is part of Oncotype DX. If the patient has an MSI-high tumor, there’s no point in getting a recurrent score, because none of those patients need adjuvant therapy.

Now, it turns out, that as we get down to metastatic disease, very few patients have MSI-high tumors, because they don’t relapse. So, there’s an irony there. MSI-high tumors are very rare in metastatic disease for the very reason that these patients, when they have surgery with stage II disease, rarely recur. Patients with MSI-high tumors in metastatic disease only represent about 5% of the total patient population. So, that’s a small number. It’s roughly equivalent to the number of people who have ALK in non—small cell lung cancer. Using drugs like checkpoint inhibitors seems to be restricted only to those patients with MSI-high tumors.

Charles S. Fuchs, MD, MPH: This is a really exciting time in colorectal cancer, because we know so much more about the genome in colorectal cancers, and we have technology that allows us to do this in real-time in clinical samples. And specifically, the simple rule is, we know that KRAS and RAS, in general, is an important predictor of whether you respond to an EGFR antibody. So, obviously every patient needs to get all-RAS testing; KRAS, NRAS, and all the appropriate areas. But, beyond that, we know that we need to know the BRAF status, because there’s evidence that you would have a worse prognosis if you’re BRAF mutated.

Moreover, we recognize that there are BRAF-directed strategies. We need to know microsatellite instability. We used to know that, because for patients who were getting adjuvant therapy, that informs prognosis. MSI-high is associated with a better outcome. But more recently, we know that MSI-high patients with stage IV disease are more likely to benefit from a PD-1 or a PD-L1 antibody immunotherapy. So those things are critical.

But beyond that, I think the advent of next-generation sequencing, really learning the entire genome, is going to inform prognosis. It’s going to inform prediction, how we know how to give the current therapies, and it’s going to give us new targets for therapies. So, I would try to advocate for this extensive testing for clinical care, and ultimately, future drug development. We need to get patients to undergo that testing. And it’s available in real-time, and there are a number of resources through your own pathology lab or companies to do that.

Beyond next-generation sequencing is the advent of other targets. Most interesting and provocative is HER2, clearly relevant to breast cancer. We know that trastuzumab is a drug for HER2-positive breast cancer. We know it’s important in gastric cancer, but recently, we learned that HER2 over expressed in colon cancer, in a small study, benefitted from a combination of trastuzumab and lapatinib. It’s too early to know whether this should be routine, but we may find that beyond all the other testing, we may want to know the HER2 status. Beyond that, I suspect we will realize a variety of other targets, be it PI3K or other genes, p53, that there are drugs being tested. They’re not proven yet, but I bet in the next five years, we will want to know the status of those mutations because we’ll have available therapies.

Transcript Edited for Clarity