COMMENTARY

Molecular Profiling Is Becoming Key to Successful GI Cancer Treatment

John L. Marshall, MD

Disclosures

February 07, 2019

Editorial Collaboration

Medscape &

This transcript has been edited for clarity.

Hey, everybody. This is John Marshall for Medscape. We just returned from what I call "GI ASCO," the Gastrointestinal Cancers Symposium out in San Francisco. One of the most exciting things that happened out there was that we had a little bitty earthquake. I thought the guy above me fell out of his bed. Turned out it was a little bitty earthquake. It rocked my world and was kind of fun.

GI ASCO is always a great meeting. It is always good to get together with your colleagues and update each other. I wish ASCO would stop presenting abstracts at the big meeting. I think it would be much better to present everything at the disease-specific meetings, such as this one, and then take to ASCO only the presentations we all really need to see.

Today I want to cover some of the highlights from GI ASCO 2019. I believe a few things are quite interesting and fascinating. Across all of the GI tract cancers, one of the major messages out there is to do molecular profiling. You need to understand the molecular subtype of your cancers before you go forward [with treatment].

First, let's look at gastric cancer. Janjigian and colleagues[1] found that if a tumor is HER2 positive by immunohistochemistry and PDL-1 positive, using a checkpoint inhibitor, trastuzumab, and chemotherapy achieves a whopping > 80% response rate in frontline metastatic gastric cancer. You will have patients who fit into that category, and this clinical trial suggests that this approach is [worth trying]. This was primarily in patients with esophagogastric adenocarcinoma. We'll have to see how it plays out with the other cancers.

The other important immunotherapy trial in gastric cancer compared pembrolizumab versus chemotherapy in second line for gastroesophageal junction adenocarcinoma. Participants had a combined positive score greater than 10, so that's a higher bar for the biomarker. This was the KEYNOTE-181 study,[2] and essentially the study was positive, with an improved outcome—overall survival included. I believe that this will result in more solidification, if you will, of using checkpoint inhibitors in that space, in second line in the right patient with the right profile.

Next, David Ilson[3] presented a nice follow-up study to the use of TAS 102 in patients with gastric cancer. He asked: What if you don't have a stomach—does the drug still work? And he showed quite nicely that having a stomach is not required to [benefit from] that oral agent.

I want to shift gears now to pancreas cancer. To me, the biggest story about pancreas cancer this time around is that, indeed, there are molecular subtypes. So it is worth doing genetic profiling. My partner and good friend, Mike Pishvaian,[4] presented the Know Your Tumor data, a nationwide molecular-profiling analysis that showed that about 25% of pancreas cancers will have some actionable abnormality. One of the lead abnormalities that got some traction at this meeting was the homologous repair deficiency, or "BRCAness," as you may know it.

There does appear to be some benefit to platinum-based therapy if you have one of those mutations. So if you're trying to choose between FOLFIRI and oxaliplatin/gemcitabine/Abraxane, then BRCAness may be a good tool to help you decide. But more important, we see more and more clinical trials incorporating PARP inhibitors in that space. So I believe that in pancreas cancer, that was a big deal.

In bile duct tumors with BRAF mutations, the combination of dabrafenib and trametinib in BRAF V600E mutated cholangiocarcinomas resulted in a 41% overall response rate (ORR).[5] This was a nice study, part of the ROAR basket study; it was the cholangio arm, if you will, of the larger basket study. Again, that subgroup had a 41% ORR. I believe that what we're seeing is that for BRAF mutations, in multiple diseases, using a doublet in this case is effective.

Switching to colon cancer, there wasn't a ton [of new information]. Honestly, I'm disappointed and I feel personally responsible for that on some level. But I do think one of the key messages we are getting in colorectal cancer is the role of circulating tumor DNA in making predictive prognostic decisions. We know that this matters. We're looking at it in adjuvant therapy and we're looking at it in other settings. Do you have persistence of circulating DNA? Is it a marker for needing adjuvant therapy? Is it a marker of sensitivity to adjuvant therapy? Is it a marker for emergence of resistance in certain settings? A lot of nice data were presented in that space and it is worth following.

In immunotherapy for colon cancer, I don't believe we moved the bar very much. An important big study[6] that tried a vaccine after metastasectomy was a dud. It didn't work and no one really knows why.

Another study kind of scared me: a phase 2 randomized study[7] that looked at the dual checkpoint inhibitors durvalumab and tremelimumab versus best supportive care in microsatellite stable patients. There was a trend toward improvement in overall survival. I get nervous that people are going to say, "That's enough; go ahead and go forward with that therapy." But I don't think that's what this will show in the end. So that trial needs to be repeated. It is not justification to throw [this combination] at everyone with colon cancer in a refractory setting. So, a note of caution there.

So, a little earthquake, some big data, some little data, and practice-changing stuff from GI ASCO 2019. From Washington, DC, I'm John Marshall for Medscape.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....