Treating Advanced Gastroesophageal Cancers - Episode 2

Genomic Analysis in Gastric Cancer

Transcript:Johanna Bendell, MD: The molecular biology of gastroesophageal cancers is becoming more and more important, and we keep hearing about this TCGA analysis. Now, they’ve done it across multiple types of cancer, but certainly the gastric cancer data have been particularly poignant. Yelena, you know a ton about this. Can you tell us about some of the molecular profiling that we’ve been doing?

Yelena Janjigian, MD: Absolutely. So, the gastric TCGA is part of the effort by The Cancer Genome Atlas group to really characterize the molecular subtypes of solid tumors: breast cancer, lung cancer, colon cancer, etc. In 2014, gastric cancer joined the ranks of the molecularly sophisticated diseases, which is an enormous step forward, and ignited interest in the research of drug development and so on. It was really a seminal development for our field. Since 2014, other groups have published other classifications, including the Asian Cancer Consortium and, most recently, the Cambridge Group in Nature Genetics. So, there are different ways to characterize these tumors, particularly as we learn more about their clinical response.

I would say the TCGA classification is now still the most encompassing and well understood. The way that we understand these diseases, phenotypically, we always knew that gastric cancer is not one disease. If you treat enough of these patients, you see the patterns, and it’s pretty logical to now see that, for example, within esophageal adenocarcinoma, these tumors are more similar to the proximal GE junction tumors and are completely different from esophageal squamous cell cancers. These esophageal squamous and adenocarcinomas should not ever be evaluated in the same trial for either the preoperative or metastatic setting, while the esophageal adenocarcinomas could potentially be studied with the same biomarker-driven trials as a GE-junction cancer, for example, of adenocarcinomas.

In short, there are four classifications for the gastric cancer subtypes. The subtype that we probably most commonly see in clinic, at least in the Western clinic and as Ian alluded to, are the tumors on the rise. There’s an epidemic of these tumors in my clinic; for example, the chromosomally instable or the GE-junction subtype. These are receptor tyrosine kinase-driven tumors. These are the EGFR- and HER2-amplified tumors. More commonly, they are p53-mutant, and this imparts potentially more aggressive tumor biology and resistance to treatment. Interestingly, these are the subtypes of tumors that we commonly see in the Western world, while, perhaps in Asia, they’re more distal tumors that are more commonly microsatellite unstable—that’s the MSI subtype. And they’re characterized by mis-match repair protein deficiency. So, this is not your typical Lynch syndrome germline patient with a high risk of passing on these alterations to their family and children. These are somatic problems that happen in the tumor and are unique to the tumor and, again, could be targeted in a certain way.

Of the other two subtypes, one is Epstein-Barr virus—driven gastric cancer—also a very molecular homogenous subtype. Eighty percent have them have PI3 kinase mutations, and these are PD-L1 overexpressers. So, again, with the advent of immune-oncology, it’s a very interesting subtype to target. And the last, and probably the most clinically and biologically challenging subtype, is this silent, genomically stable subtype which is most commonly related to a problem with the cell adhesion proteins. So, e-cadherin, claudin, fusions happen in these tumors. And, as Manish alluded to, epidemiology of these subtypes are different. They are the same across the world, and, again, they’re a very important subtype to target.

Johanna Bendell, MD: Yes. Just an example, and you’ve been doing a lot of work with this, is looking at these subtypes in relation to their response to immunotherapies. Like you were talking about the microsatellite unstable group, and the patients that overexpress PD-L1 and PD-L2 might be the ones that would benefit the most from immunotherapy—so, super exciting. And, actually, to keep you on that point, we talk a lot about molecular profiling of tumors for patients in our clinics. What are some of the things that we’re looking for in gastric cancer patients, both as standards to help us direct treatments now, but also things we might want to check to potentially direct a patient toward a clinical trial or give them some prognostic information? What are some of the things that we look at and look for?

Yelena Janjigian, MD: Absolutely. I want to make a distinction between patients with metastatic versus curable disease. It’s clear that the majority of our patients, even when they have locally advanced stage III, stage II disease, which is the most common presentation for nonmetastatic tumors, they’re at a high risk for recurrence. And the tumor, even after surgery, may come back. At that time, it would still be very useful to have the information that we’re looking for. For metastatic tumors, it is clear that the only really validated biomarker, and the most important biomarker to routinely reflexively check, is HER2. Generally, we begin with HER2 immunohistochemistry. It’s a quick and easily done test even in the community setting. And for tumors that are equivocal or uncertain, such as those of immunohistochemistry of 2+, we do fluorescence in situ hybridization (FISH). At a minimum, you have to do immunohistochemistry for HER2 and FISH, and then you could stop there.

But the majority of our patients now, for gastric cancer, at least, are becoming more and more sophisticated. They’re reading these articles. They are more driven and seek out second-, third-line studies. So, I think the future of gastric cancer characterization is probably next-generation sequencing. Because, in this sense, you can do a one-stop test where you can use unstained slides of paraffin-embedded tissue, do the HER2, and already know the MSI and EBV status—perhaps know the claudin transformation status and so on—instead of having to go from one test to another and then realizing you ran out of tissue and you can no longer do the next test. But that’s probably a few years still away, so, in short, the only proven test right now is HER2.

Transcript Edited for Clarity