John Marshall, MD: I want to take what you just reviewed apart, because a lot of us are playing catch-up on understanding these molecular reports. You just listed a lot of things that are in the BRCA family—the HRD [homologous recombination deficiency] family. Mike Pishvaian, MD, one of my partners and friends, did a nice look using the Know Your Tumor through Pan Can [Pancreatic Cancer Action Network]. You’d mentioned RAS before. If you look at the group that’s not RAS-mutated, a lot of these BRCAs and MSIs [microsatellite instability] are clustered in that 10% or so that aren’t.

I keep wondering if one day we’ll have some sort of, “Well, you’re RAS. You’re over here,” and if you’re in that 10%, we’re looking for other targets. I think you quoted his number. If you looked at, depending on how you lined up, what was HRD—or what you thought was a real, actionable mutation—it was something around 24% of all pancreas cancers, which had something that we thought was a legitimate actionable mutation. Paul, some other data came out around HRD and the like at the meeting. Do you want to give us an update on where that stands?

Paul Oberstein, MD: There are a bunch of data. The data from Know Your Tumor—which is a great initiative to pull together, this kind of genetic data—did show that there’s a pretty high number of patients who have these DDR [DNA damage repair] deficiency mutations, but it didn’t seem in his data as if it really mattered so much, unless they were treated with a certain kind of chemotherapy called platinum therapy.

John Marshall, MD: Right.

Paul Oberstein, MD: There have been other studies, including, at this meeting, a data set from Memorial Sloan Kettering Cancer Center.

John Marshall, MD: It was predictive.

Paul Oberstein, MD: It was just predictive of response.

John Marshall, MD: Not prognostic.

Paul Oberstein, MD: Not prognostic without treatment.

John Marshall, MD: That was the first demonstration using that panel.

Paul Oberstein, MD: Correct, and it was hundreds of patients. It was over 500 patients.

John Marshall, MD: You mentioned Foundation Medicine before KRAS, and they all are guilty of this. If they find any one of those, they might pull down a platinum recommendation. This was at least, on some level, a little validation that it’s not too far off, but you were talking about the power of each one of those individuals versus the collective.

Paul Oberstein, MD: Memorial Sloan Kettering Cancer Center released data at this meeting. There are over 400 patients with pancreatic cancer in whom they looked at both the germline and somatic mutations. What they found, again—with the caveat that this is just in their center—are almost the same findings. The germline mutation patients did better when they got a platinum, but the somatic mutation patients did not seem to do better. It raised a question. There was a trend, so it wasn’t clear where to go with that. But that may not matter if it’s somatic, meaning if it’s just in the tumor. It may matter only if it’s germline.

John Marshall, MD: Yes. I do also want to comment that many of the molecular profiling companies are working to develop their assays, so that when you do the somatic, you’re kind of also getting the germline. We’re not really there yet, because when you do germline testing, you get much broader coverage, so you don’t miss stuff. Whereas with the somatics, they looked at hot spots, if you will. The more we’re seeing whole exome sequencing, the more they may overlap. For now, I just wanted to reiterate that these do measure different things, and one is not necessarily concordant with the other. Certainly there are more targets to come. Pancreas cancer is not dead to targets. Ed, some other stuff is coming out—the COMPASS trial. Give us an update on what that showed.

Edward Kim, MD: The investigator should be applauded for a very aggressive approach capturing what we’re talking about. There is a need to try to profile our patients, and they attempted to robustly profile over 150 patients—looking to patients who had not been treated. They attempted to take tumor tissue, isolate it, profile it for whole genome sequencing as well as RNA, and return that information as quickly as possible. The median time to return data was about 35 days, which is quite fast.

What I think is still one of the issues is, is that fast enough to make decisions in real time? One of the things that came out of that study was that they could profile patients into a Moffitt Cancer Center classification of either classical or basal. Classical tends to do better, and basal does poorer. About a quarter of patients are basal-like. There was an identification of a marker—a surrogate that you could test alone, something called GATA6, which is a transcription factor that’s involved in differentiation. If you have high expression of GATA6, that clearly goes along with the classical subtype. So instead of doing this whole sequencing, you could cut down the timeline and test just for GATA6, potentially.

John Marshall, MD: One of the questions that we all get asked is, why aren’t we moving the bar faster in pancreas cancer? What you’re getting at, Ed, is that we’re starting to subclassify. Does anybody think that this is really the way we’re going to have to go with this disease and find its different groups to make headway? Shubham?

Shubham Pant, MD: Let me take a contrary view on this subject. I do target therapy 50% of my time, so I keep on looking at this. Again, everyone wants it to be like lung cancer, but it’s not lung, right? There’s a lot of pancreatic cancer.

John Marshall, MD: Who would have ever thought we would say that?

Shubham Pant, MD: I was joking with somebody. I had a presentation recently on newer therapeutics and early case drug development in pancreatic cancer, and 1 of my slides was on a glioblastoma patient on TGF beta therapy, in whom I actually got a response. I showed that saying, when I was doing my fellowship, which was—I’m dating myself—sometime back, we used to give platinum: carbo [carboplatin]-Taxol. It was the easiest question on our boards; carbo-Taxol for lung cancer.

It used to be a fight between lung and pancreatic, saying, “This is a tough disease.” And that’s been settled. I said, glioblastoma is really where we’re looking to make advances and everything, as far as a median survival is concerned. For our audience—for community oncologists—when you look at the bigger picture: 1) when you look at the germline testing, I think everybody should do it. That’s 5%. You can pick that up and give them platinum first, or we’ll see what happens with the olaparib story as it further matures, along with other PARP [poly ADP-ribose polymerase] inhibitors. With the somatic testing, you still have to have access to clinical trials. You do a somatic testing, and you find something. If you don’t have access to clinical trials, if you’re in a small town somewhere, then it’s harder to.

John Marshall, MD: To be fair, you still have your rare stuff. You have your MSIs. You have an occasional NRK. And if you don’t look, you won’t find it.

Shubham Pant, MD: Right.

John Marshall, MD: There is some motive for wanting to do that with approved drugs.

Shubham Pant, MD: The other thing is, I think for the MSI—yes, you can do a simple IHC [immunohistochemistry]—you don’t need to do a whole profiling. I do think people should be profiled. I’m not saying that. I’m saying, to break that down like a little pie chart into 5% or 2%, it’s much harder in pancreatic cancer. I think we really need to focus on the KRAS-driven tumors, because when the RAS is turned on, it’s very hard. The pancreatic cancer microenvironment is very tough to overcome, so I think that’s where we have to focus our further research efforts.

John Marshall, MD: I jokingly have called 2019 the year of the RAS…just because I do think there are some emerging therapies in RAS. Fingers crossed that we will find that. That’s obviously a major breakthrough if we could start to drive against RAS.

Paul Oberstein, MD: When you compare this with lung cancer, they have a pie chart that looks nice.

John Marshall, MD: Well, they didn’t always have a pie chart.

Paul Oberstein, MD: They didn’t always have a pie chart; they’ve discovered more. We have a pie chart too. It’s just dominated by RAS.

John Marshall, MD: Yes.

Paul Oberstein, MD: I think it’s not that we don’t know the mutations; we do. We just need effective therapies. Now, I don’t think it’s going to be one. I wish it would be. There’s going to be RAS. There better be, or else we’re in trouble.

Allyson Ocean, MD: Also, I want to point out that the numbers traditionally have included the scope of everyone diagnosed with pancreatic cancer. There’s a decent percentage of patients who never get offered any treatment at all, and they’re included in that. If we’re treating more patients and also doing genetic testing to uncover more mutations, I think that the survival rates will increase in a meaningful way in the next 20 years or so. We really have to make sure that everybody gets offered treatment.

John Marshall, MD: I think that’s such an amazing point, because 5 years ago, what we were just saying is, “Quit being therapeutic nihilists. Treat them; we have new medicines.” Now what we’re arguing is the next step of that. It’s worth looking, because you don’t know what we’re going to find until we start looking.

Shubham Pant, MD: I completely agree with Allyson on that. I think it’s all of the above. We need to be using chemotherapy, and targeted therapy when it’s appropriate, and then be looking for ways to target the RAS pathways. It’s all of the above that will drive our science, so I think we should be looking at the bigger forest, or looking at the whole picture to say, “How can we target these different ways? And when we cannot, what’s the chemotherapy backbone that we can use in these patients?”

Transcript Edited for Clarity