Johanna C. Bendell, MD: We’ve been hearing a lot lately about doing molecular profiling or next-generation sequencing for patients with both colon cancer and other types of advanced cancers. I personally am a proponent of doing next-generation sequencing for patients with metastatic colorectal cancer because you can learn a lot of things. First of all, it’s standard of care to know the expanded RAS analysis and BRAF analysis, but we also need to learn about microsatellite stability or microsatellite instability [MSI], which can be done on next-generation sequencing or other molecular platforms. But now we’re starting to find subtypes of patients, including HER2-positive patients and other types of mutations, where we may have other targeted therapies or clinical trials available for those patients to see if other targeted therapies may work.
Tanios Bekaii-Saab, MD: When we’re looking at all the GI [gastrointestinal] cancers, for example, and we test MSI status and tumor mutational burden, they seem to be somewhat overlapping. In colorectal cancer, they overlap pretty nicely, almost a 90% or 95% overlap. What that tells us is [that] MSI high is essentially just a surrogate market for tumor mutational burden.
MSI high’s presence just tells us that the tumor is loaded with a mutation. It doesn’t necessarily predict for who may or may not respond to immune therapy. Tumor mutational burden just tells us how heavy the tumor mutational burden is. The higher it is in a tumor, the more likely the tumor will respond to immune therapy agents such as PD-1 or PD-L1 inhibitors and their combinations.
It appears that the tumor mutational burden is higher on the left side in the presence of MSI high versus on the right side, for whatever reason. It’s a bit unclear, although these are preliminary studies. But overall, it is safe to say that MSI high is a good surrogate marker for tumor mutational burden. Now, that said, not every patient with a high tumor mutational burden will be MSI high or vice versa. That’s very important to keep in mind. We’re relying more and more and more on tumor mutational burden, or TMB, to define who would be the best candidate for immune therapy. It just tells you…how loaded this tumor is in terms of mutations and how much more likely it will be to respond to immune therapy agents.
We’re not where we should be right now. Think about the landscape in first-line metastatic colorectal cancer today. Most patients in the United States will be treated with chemotherapy plus bevacizumab. Some will switch the bevacizumab for an EGFR inhibitor. That’s based on essentially all the data we have without the presence of a good predictive marker. For bevacizumab, we don’t have a predictive marker. We essentially treat everyone to benefit many, but not all.
With the EGFR inhibitors, we’re starting to learn a little bit more about who should not get them. We don’t know how to select for those who should get it. We have a negative predictive biomarker but not a positive one. The positive markers are the ideal because they do select better. We know that, for example, in the presence of a RAS mutation, of a BRAF mutation, or of a HER2 amplification, those patients will not benefit from an EGFR inhibitor.
We know, as we understand a little bit more about the molecular origins, on the right side there is no benefit from an EGFR inhibitor, at least in the first-line and maybe in the second-line setting. We have some of these predictive biomarkers. We’re learning that HER2 amplifications may predict for a response to HER2-targeted therapies, and many are under development. We also understand that MSI-high or tumor mutational burden—high tumors will respond to PD-1 or PD-L1 inhibitors. Now we know that. We have few predictive biomarkers.
For the majority of the patients though, we don’t know. We just have no idea why bevacizumab works or why it doesn’t or why other VEGF inhibitors work or not. We don’t know why chemotherapy works or not in some patients. We don’t know why regorafenib works in some but not in others. We don’t know why TAS-102 works. We certainly continue to explore the intricacies of other molecular or genetic drivers of the disease so we can target them better.
We’re getting better overall. Ten years ago, we didn’t have almost any predictive biomarkers other than the KRAS gene. That was it. That was the only 1 we could use. Everything else was just unselected. So, we’re doing better, but we’re far from where we should be. We’re on track to understand a little bit more about additional predictive biomarkers that will break this disease down into multiple pieces.
Johanna C. Bendell, MD: As we’re doing more next-generation sequencing of patients with advanced cancers, we’re learning about different subsets of these patients and potential treatment options for them. For example, we know that for patients who have RAS mutations, we shouldn’t be using anti-EGFR antibodies because these probably won’t work. For patients with BRAF mutations, we probably want to use some sort of combination of an EGFR inhibitor and a BRAF inhibitor. In fact, NCCN [National Comprehensive Cancer Network] guidelines recommend second-line use of cetuximab plus irinotecan plus dabrafenib for patients with BRAF-mutated colon cancer. We also know in the first-line setting for these patients that a combination of FOLFOXIRI [folinic acid/fluorouracil/oxaliplatin/irinotecan] plus bevacizumab may be more beneficial for them.
For patients with HER2-positive disease, which accounts for about 5% of patients, there is some suggestion that these patients may not respond to EGFR inhibitors. But possibly more importantly, we’ve seen some preliminary data that suggest they may respond to combinations of anti-HER2 agents like trastuzumab plus lapatinib, or trastuzumab plus pertuzumab.
We’re also very excited about NTRK fusions. We’ve seen these in multiple types of cancers and in colon cancer, it probably accounts for about 2% of patients. But we now have NTRK-targeted therapy for these patients that might be beneficial to them.
Transcript Edited for Clarity