Michael J. Pishvaian, MD, discusses expanded testing for biomarkers, using liquid biopsies to predict response to treatment, and the importance of identifying increasingly specific subsets of GI cancers to make treatment decisions.
Michael J. Pishvaian, MD, PhD
Molecular biomarkers are gaining significance in the gastrointestinal (GI) cancer realm outside of HER2 overexpression, including BRCA, NTRK, PALB2, and microsatellite instability (MSI). Now, researchers are exploring how they can best be targeted—and how to optimally detect them.
The surge of molecular testing has led to the development of targeted therapies against some of these actionable biomarkers. An example of this, he adds, includes the HER2-directed drug trastuzumab (Herceptin) to treat patients with HER2-positive colorectal cancer (CRC).
In an interview during the 2018 OncLive® State of the Science Summit™ on Gastrointestinal Cancers, Michael J. Pishvaian, MD, an oncologist at Georgetown-Lombardi Comprehensive Cancer Center, discussed expanded testing for biomarkers, using liquid biopsies to predict response to treatment, and the importance of identifying increasingly specific subsets of GI cancers to make treatment decisions.Pishvaian: What we are looking for are predictive biomarkers that lead us toward specific therapies. The term “actionability” is often used to describe a biomarker that predicts full responsiveness, or at least a high responsiveness to a certain therapy. Sometimes that therapy is approved and used in the cancer that is being treated, and sometimes it’s approved and used in a different cancer. Ultimately, that biomarker seems to have promise in the cancer you are treating.
We spoke about turning broad profiling into specific actionable biomarkers and using that to shape care. There are several ongoing and emerging examples of that in the world of GI cancers.
I do have a couple of things to [focus on] with regard to liquid biopsies. I'm a little biased because we looked at liquid biopsies, at least in pancreatic cancer, and we found what a lot of other people have found. Liquid biopsies have tremendous potential, but because they're not very sensitive tests, you need to know what you're looking for. If you know the mutation you're looking for, it’s a great test to use because it’s very specific. However, if you need to look broadly, it’s more difficult because the technology does not allow you to use this broad profiling into the tumor cell. At least in the world of GI cancers, I still prefer to look in the tumor if I can— and in the blood if I have no other choice.There's a great deal of adoption of liquid biopsies because they're so easy to do. The problem is that we need to couple that with a lot of education. If you don't find what you're looking for—a negative answer—a lack of identification doesn't necessarily mean it’s not there. I have to go to my lung cancer colleagues to give an example of this. In lung cancer today, there are 3 mutations that have to be tested for in treatment. If a liquid biopsy is done and shows a mutation, they don't have to do any further testing. If no mutation is found, it might be a result of lack of sensitivity and they’d have to look in the tumor. If we find a KRAS mutation in colon cancer in the blood, we're done; however, if we don't see one, it still might be present. If there's tumor available, there's no good time to do a liquid biopsy in GI cancers. If there's absolutely no opportunity to get to the patient's tumor because they have metastatic disease, then we'll certainly work around it. There is a separate discussion about liquid biopsies and some intriguing emerging data about tracking response to therapy. In KRAS mutations, you can have very high levels of KRAS-mutated DNA in the blood and that level can plummet when treated effectively. It may actually become an early marker of response to therapy even before you get that first test scan. That is definitely a very useful way to use liquid biopsies, but we need good data and studies to prove that.
Working by analogy, about 10 years ago in the world of Hodgkin lymphoma, there were some data that PET scans can be a better early indicator of response to therapy. However, it took a 1000-patient study to prove that PET scans can be used after just 2 cycles of therapy. We're going to need to go a long way before we start using liquid biopsies to direct therapy for KRAS- and other mutational statuses dependent on changes in the blood levels of the mutation.There was a large study that proved definitively that this overexpression exists in about 5% of colon cancers and, when appropriately targeted with the right drug, work very well in the HER2-positive CRCs. There are other GI subsets emerging, most notably BRCA- and PALB2-mutated pancreatic cancers. They seem to respond really well to platinum-based PARP inhibitors.
Then, there are very small subgroups of tumors such as NTRK, which are very rare, but when they're present are most common in colon cancer and sarcomas. When a tumor has an NTRK mutation, the targeted drug works very well. MSI is present in a fairly small percentage of patients, but the tumors at high levels respond really well to immunotherapy. There's many, and a lot of the clinical trials are becoming smarter in their design. They're only looking for activity in the small group of patients who have that biomarker. Rather than testing 100% of patients and figuring out after the fact whether the biomarker is present, they're looking for that biomarker first and proving whether the drug will help.You need to be testing particularly the patients with advanced cancer. This isn’t as important for the early-stage curable patients because the tools we have work really well. The patients with advanced cancer are where we need to find better options. There might be a couple of therapies that work for a little while but overall, the patient is going to progress. In pancreatic cancer, there is the opportunity of looking to see if that patient's cancer happens to be one of the small group of patients who have a BRCA or NTRK mutation.