Video
Role for IO Therapy in Advanced NSCLC With Concomitant KRAS and KEAP1/STK11 Mutations
Author(s):
Key opinion leaders in the field of non–small cell lung cancer management discuss the role of immunotherapy in patients with concomitant KRAS and KEAP1/STK11 mutations.
Transcript:
Balazs Halmos, MD, MS: For the patient we heard about before, I have an update on their NGS [next-generation sequencing] report. This patient has a TMB [tumor mutational burden] of 14 [mutations per megabase], a KRAS G12C mutation, and an STK mutation…. The patient’s TPS [tumor proportion score] is 0. What do STK11 and KRAS mean? How do you look at the literature to make decisions about these molecular markers?
Joshua Reuss, MD: We’ve learned a lot going beyond surface-level KRAS and other driver mutations and into alterations that connote immunotherapy resistance…. There are different data that are primarily retrospective. Data out of the Boston group about not just the STK11 and KEAP1 but also the cooccurrence with KRAS are important in influencing immunotherapy resistance. These alterations are important in the biology of this disease, though we need more data.
The most enticing and thought-provoking data come from the subgroup from POSEIDON, which is a complex study to describe in terms of design. This subgroup’s data is from patients who have STK11 with KEAP1 alterations. They suggest there’s something important about the CTLA4 drug in this patient population. It might be important in proving benefit. Keep in mind, these patients do poorly regardless, unfortunately. These patients tend to do poorer than those who don’t have the STK11 and KEAP1 alterations.
Back to our patient, I’d consider a POSEIDON treatment. At the same time, I’m hesitant to make broad, sweeping declarations off a subgroup of 30 to 40 patients. Our tried-and-true in adenocarcinoma that’s PD-L1 negative is the chemoimmunotherapy KEYNOTE-189 regimen. Until I see large prospective data with that as our standard rather than a chemotherapy-alone approach, I’m not inclined to make large changes to my practice.
Balazs Halmos, MD, MS: The [University of Texas] MD Anderson [Cancer Center] pooled experiences that suggested STK11 and KEAP1 may be poor prognostic factors but possibly also negative predictive factors for immunotherapy efficacy. Are you looking at the NGS? Does that impact your decision when you’re facing a patient?
Martin Dietrich, MD, PhD: Initially, I didn’t because it was inconsequential. It was a prognostic marker based on the retrospective review, and it wasn’t reassuring. Obviously, those data were greatly concerning for not only STK11 and KEAP1 but also other comutations: TP53, ARID1A, and others. There wasn’t much to be done about it. One of the messages of POSEIDEN is that the concern that there’s a big outcome difference is equalized. I don’t understand how STK11 and KEAP1 influence immune responses in the tumor marker environment, but the data consistently suggest that they have an impact. Many mutations on NGS reports list variants of undetermined significance. We don’t know what a pathogenic STK11 is, so it’s not clear what a variant of undetermined significance would mean. If it’s seen, it would be further down the line of intensification.
For the patient with a PD-L1 level of 0%, that would be a main biomarker. For combination immunotherapies with anti–PD-1, PD-L1, and CTLA4, the first biomarker is PD-L1 level if I have KEAP1 and STK11 as additional markers. That’s especially true in the context of KRAS mutations. The impact outside KRAS is much less clear. There has to be an additional intensification or a clinical trial to serve this patient population in the best way possible. This makes it very challenging. We’ve had some basic implementation difficulties with NGS. By going into this addendum review of large NGS reports, we’re making this even more impractical. Unfortunately, genetics in the context of immunotherapy are highly complex. There isn’t going to be a simple answer to a very complex problem.
Balazs Halmos, MD, MS: Kristen, are you looking for these alterations in the reports? If they’re important, should everybody be obtaining them? If we need to start treatment quickly, how would they do so? Do clinicians have the luxury of waiting until these reports come back in time? Tell us about your experience and approach.
Kristen A. Marrone, MD: You definitely need to have molecular data back to make a first-line therapeutic decision. Not everybody would give a heavy smoker with a BRAF mutation single-agent I/O up front, regardless of PD-L1, so you want to have those data available before you use an I/O-containing regimen. If the reports don’t come back in time or you have insufficient tissues, and you’re be waiting for a plasma test that will take another 10 days, I usually start with chemotherapy alone if a patient is symptomatic. I think about adding in the I/O once I have those data back….
When I see a KRAS mutation, I fall back to PD-L1 to help me determine my course of action. The only way STK11 and KEAP1 change my treatment decision-making is for rare, comutated tumors that have a high PD-L1. This is because when we look at the data, they’re less likely to have PD-L1. I wouldn’t feel comfortable giving that patient single-agent PD-1. I don’t always reach for ipilimumab when I see a KRAS/KEAP1. It’s something I consider. I consider it more often for my fit patients, but it’s not the only 1 I’d reach for. We should explain to a patient that as we learn more about these molecular results, we also learn about when the treatment isn’t going to work. Ultimately, we look forward to a time when we’re having clinical trials in this space, similar to when we were first thinking about different EGFR mutations. Hopefully we’ll have more prospective data in the near future about what to start these patients with.
Balazs Halmos, MD, MS: That’s exactly right. We need more data, both specifically targeting these alterations and also understanding the benefit of combination I/O for this particular subset. We’re hearing that these studies are being planned. Hopefully they’ll get done in a reasonable time. But right now, we have those little subset analyses. They seem intriguing in terms of improved outcomes for those very key patient subset.
Transcript edited for clarity.
