Zev A. Wainberg, MD: Gastric cancers and gastroesophageal junction adenocarcinomas are risk stratified similar to other cancers with respect to stage. Obviously, more advanced stages have a poorer prognosis. There are very few data on the molecular characterization as it relates to providing accurate stage prognostication, but there is some suggestion in the literature that patients of Asian descent have perhaps a better prognosis when they have gastric cancer, versus patients of American or European descent, and gastroesophageal junction adenocarcinoma, which is staged similarly to other cancers with respect to stage and performance status. Those are the main prognostic criteria.
Alan P. Venook, MD: The importance of molecular features in gastric cancer, stomach cancer, and GE juncture cancer are limited. I would say that the most well established and probably the best-understood TCGA, Tumor Cancer Genome Atlas, analysis of any cancer has been of gastric cancer, where we’ve learned a lot about the causative agents. But in terms of the specific molecular features, the dominant issue is HER2 overexpression or amplification. Amplification of HER2 predicts potential benefit from inhibitors of that molecule, with trastuzumab for example.
Another factor, MSI, microsatellite instability, is of course relevant in every cancer now that we have checkpoint inhibitors. That is a rare finding, but that would be worth evaluating. Otherwise, there are features like molecular mutations in RAS or in MET. We really don’t understand the ramifications, and they’re not informative in terms of how to manage patients.
There are a couple of familial syndromes that are associated with gastric cancer. Lynch syndrome is one. The one that is the least common is from the inherited gastric cancer families, which is the E-Cadherin gene mutation, CDH1. That’s rare. It has actually been described at Stanford largely. People with that mutation are almost certain to develop gastric cancer, diffused gastric cancer, over the course of their lifetime. But it’s rare. If you go looking for it, the chances are you won’t find it. We might look for it only as a genetic marker for patients felt to be at familial risk. In general, we don’t look for that as a happening other than inside of kindred family.
Other features that we talk about, other heritable risks, are being MSI high, which is part of Lynch syndrome. As for other heritable cancers, perhaps BRCA mutations may occasionally be associated with gastric cancer, but those are rare. In terms of the average patient, I would not even look for CDH1 unless there’s a strong family history.
Zev A. Wainberg, MD: I think with the recent approval of immunotherapy drugs in patients with microsatellite instability, regardless of underlying primary tumor, there are good data to suggest that we should be testing all patients with any GI cancer for mismatch repair proteins, which can be tested by immunohistochemistry or, for microsatellite instability, by PCR. So, that’s a recent update I think we should look at virtually for all GI tumors, recognizing that it’s very infrequent and less common in gastric cancer than it is in colorectal cancer, but still worth checking.
The data for PD-L1 being a prognostic marker are very mixed. It’s not clear yet in the literature that PD-L1 represents a prognostic marker for gastric cancer. There’s some suggestion based on a recent publication that it may be, but I think it’s too early to conclusively call PD-L1 a prognostic biomarker in this disease.
The TCGA subsets have divided gastric cancer into 4 groups, one of which is an EBV (Epstein-Barr virus) subset, and the EBV subset represents a small group of patients with gastric cancer. That may have prognostic implications as well, but it’s also not clear that that group necessarily represents a worse prognosis.
Alan P. Venook, MD: There’s always a good question about the value of doing markers that are unsubstantiated or not necessarily proven to be a benefit. A good example of that would be circulating tumor cells. We would assume that if you find circulating tumor cells, that would be bad. Some people have talked about using them to monitor the effectiveness of ongoing therapy or following surgery, to see whether you’ve cured a patient or not.
In gastric cancer, I’m not convinced that there’s good evidence to support the use of circulating tumor cells. There is some evidence to look at circulating cell-free DNA. For example, with patients who are HER2 amplified, at least if you lose the genetic evidence of that in the blood, it may signify the disease has changed stripes and lost that amplification. But in general, we would not recommend following circulating tumor cells other than in a research environment.
Transcript Edited for Clarity