You are here
Molecular classification of gastric cancer
Prof. Adam Bass
Molecular classification of gastric cancer
Professor Adam Bass
Department of Medical Oncology, Dana-Faber Institute, Boston, USA
Q. Hello. We are here in Barcelona at the World Congress of GI Cancer. Today we heard a wonderful key note lecture from Professor Adam Bass from the Dana-Farber Institute in Boston, USA. You know him as the lead author of the TCGA, The Cancer Genome Atlas classification of gastric cancer.
Adam, it is great to have you here for an interview. Do you want to tell us some short sentences what the TCGA classification is about, what you did and what you found out?
A. Sure. The Cancer Genome Atlas project was a big effort through the US National Institute of Health to collect a large number of cancers from different cancer types and then do a systematic analysis from different molecular platforms, gene expression mutations, somatic copy number and so forth and try to do it all on the same group of cancers and try to then analyse and find what kind of patterns and groups of tumours you can find as well as different features of distinct cancers.
Within the stomach cancer project, when we did all of these different, more detailed analyses, we found some clear patterns that emerged over and over again that essentially there were different groups of cancers that would share features across multiple platforms, meaning that there were certain groups of cancers that might have had certain features in their DNA mutations and they would also share features in copy number and also share features in their DNA methylation. From that we got the idea that increasingly there are really distinct subgroups of gastric cancer, just like there are subgroups of other tumours, breast cancer, where people think of triple negative, HER2 positive, ER positive.
We also identified there were some clearly different subgroups of gastric cancers. We developed a classification putting them into four little groups, one being EBV or Epstein-Barr virus positive cancers and one being microsatellite instable cancers and the last two being ones that here have what are we call CIN or chromosomal instability, a lot of very abhorrent copy number or gains and losses in chromosomes and the last, what we call genome-stable, was mostly composed of diffuse gastric cancers.
In each of the four groups we saw some different molecular features, different kinds of oncogenes that were active, so we hope that this classification will help us better stratify patients for trials and, even within these four groups, there will be subgroupings based on the more distinct features, for example with the chromosomally unstable cancers, there are some that have amplification of HER2, some amplify the KRAS oncogene or amplify EGFR or MET and so forth, so it is not that all CIN tumours will get the same kind of therapy, but you would be looking at those tumours for a target that may be amplified for example.
Q. In the current time point, this new classification is not so much informed clinical practice yet, but it is stratification, it is a pathway for future clinical research; is that correct?
A. Yes. That is a good way to think about it. There are a few different questions; one might be when we think of the conventional empiric treatment we use just now, be it empiric chemotherapy in the palliative setting or some of the up-front chemo, chemoradiation for potential in curative patients, so some questions might be if you take the patients and stratify them by these groups do you identify that some of these therapies may be more or less effective in particular subgroups? The other way, especially in the setting of looking at the systemic therapies, will be whether these different subgroups may be more or less likely to respond to particular therapies that exist already or are being developed?
One of the most exciting places where we are hopeful that this classification approach will be useful is in the implementation of new immunotherapies. We have already seen in colon cancer how the tumours that are microsatellite unstable are more likely to respond to some of the PD-1 pathway agents, and that makes sense being that the microsatellite unstable tumours have lots of mutations and therefore have lots of new antigens or neoantigens that might ?signal the immune system.
We have seen now that some of the markers that may suggest benefit to immunotherapy such as expression of PD-L1 are higher ?upper-side tumours and even higher on Epstein-Barr virus cancers, so especially with the development of immunotherapy I think, considering these different subtypes will be quite important.
Q. Very interesting. Apart from immunotherapy do you see a chance that your analysis will lead to targeted treatments, like we have seen in lung cancer for instance? There will be drug mutations or targets that can be hit by one drug or do you see the future different for gastric cancer?
A. I think there are lessons to be learned from all cancer, but I think unfortunately it is not going to be as straightforward to develop these targeted therapies for gastroesophageal as it is for lung cancers. The gastroesophageal tumours are more genomically complicated and they have, especially compared to the lung cancers that do well to monotherapy such as EGFR therapy, so my increasing thinking is that we are less likely to see the dramatic responses to monotherapy just by hitting the right kinase in the right patient. Just because we might not see monotherapy, just like we have seen that lapatinib and afatinib haven’t been dramatically beneficial, especially in single agent, in HER2 positive cancers, that doesn’t mean that hitting the target isn’t part of what is going to be effective therapy. I think it means we have to really go back to the laboratory and look at patient samples more carefully, including after you get the therapy, to say “Okay, you have this cancer cell in this organ that is very fast growing cancer, very fast growing” – we are looking at baseline, thinking the lining of your intestines are turning over constantly compared to say, a lung where the cells are more hanging out, so to speak, so I think that when you hit the one key oncogene in these tumours there are going to be ways that they adapt, that they turn on other pathways and compensate. To develop combination effective targeted therapy we should know first what target to hit and then what the cancer does next so we are ready to develop effective new combination therapies. In general I think we can develop target therapies, but it won’t be like the lung cancer model where you either get the EGFR or the ALK inhibitor or the ROS, I think monotherapy is not going to work nearly as well in gastroesophageal.
Q. The last question: do you think the tumour heterogeneity is a particular issue in gastric cancer?
A. That is something we are increasingly looking at now. We have seen some very exciting new evidence pointing to that being a problem with targeted therapy. The recent study that was published out of the United Kingdom with David Cunningham’s group looking at the target FGFR2 was very telling and that was a trial looking at a particular tyrosine kinase inhibitor, FGFR2, and what they saw was that the patients in whom the target FGFR2 was found throughout the tumour was very homogeneous. Those are the patients who benefitted from the inhibitor, whereas the patient where the target was spotty, it was just a part of the tumour, they were not benefitting as well. That is an area of very exciting possibilities now and we now have increasing opportunity to look not just at the tumour, but at circulating DNA which could be a proxy for more globally the state of the tumour, including the metastatic lymphocyte, I think that has potential to help guide targeted therapy.
Q. Thank you very much for giving us so many insights. We are looking forward to the further research of Adam Bass and to his future papers. Thank you for listening.
A. Thank you.