
Clinical Utility of KEAP1 and STK11 in NSCLC
Clinicians discuss using KEAP1/STK11 mutations to guide CTLA-4 therapy escalation in NSCLC, aiming to overcome PD-1 resistance and improve 5-year survival goals.
While PD-L1 expression remains the traditional cornerstone for directing immunotherapy in non–small cell lung cancer (NSCLC), the emergence of genomic "resistance markers" is complicating and refining the treatment landscape. Among the most challenging are mutations in KEAP1 and STK11, which are frequently associated with a "cold" tumor microenvironment and a lack of sustained benefit from PD-1 monotherapy.
In a Case-Based Roundtable discussion, Martin Dietrich, MD, PhD, led a panel of experts to examine how these genomic signatures influence real-world clinical decision-making. The conversation centers on a pivotal question: when these mutations appear on a next-generation sequencing (NGS) report, does it provide enough evidence to escalate treatment to a CTLA-4 combination?
Martin Dietrich, MD, PhD: Who is using KEAP1 or STK11 as resistance markers to PD-1 therapy? When you see this on your NGS report, does that make you think toward a CTLA-4 combination?
Ryan Griffin, MD: Now I’m looking at it. That’s more of a high-risk prognostic factor, and I’m thinking about using a CTLA-4–targeted agent.
Dietrich: For me, pragmatically, I think there’s so much room for improvement that every time I find another reason to escalate to a more intense therapy, I probably will, simply because the majority of patients will not live until 5 years, even with our best therapy. I try to take all of this into account as a whole equation of factors for treatment.
Some of the pros and cons that people have mentioned with tremelimumab [Imjudo] is that we only have 5 doses as opposed to ongoing ipilimumab [Yervoy]. In reality, they’re both self-limiting. CTLA-4 has a dose dependency that is very different from PD-1, where most patients eventually will run into some treatment-limiting adverse effect, and very few patients end up taking ipilimumab on an ongoing basis.
Samip Master, MD: Deciding between the 2, I would probably use ipilimumab/nivolumab and chemotherapy because of familiarity and also, the long-term data is better.
Dietrich: What factors would make you think to escalate to a CTLA-4 combination?
Master: Upfront in a new patient who is PD-L1–negative and has a robust, good performance status, I think they can tolerate the [escalated] regimen.
Stephanie Cimo, MD: [Dr Dietrich], I’m interested to hear your thoughts on KEAP1 and STK11, because I do feel like that influences me to add the CTLA-4 inhibition. What is your practice with that?
Dietrich: The problem is—and I think that’s why the guidelines are so shy in incorporating them—is we have a very consistent signal that when patients have STK11 or KEAP1 mutations, we don’t see a PD-1 impact independent of PD-L1 level. So, they’re really turning the tumor microenvironment cold. We have…data that CTLA-4 dose seem to be distinctly different and may have an immune inflammatory impact that may overcome at least some of that [resistance]. [STK11 and KEAP1] are still prognostically negatively, but predictively, there’s a big difference.
I use STK11 and KEAP1 in the context of all the other factors that are coming in as reason for escalation. It’s sometimes difficult because the reports don’t feature them prominently. They’re obviously not companion diagnostics, so then you have to dig through the reports. But if you’re sensitizing yourself to looking at your NGS reports, you’ll see them left and right. They’re certainly not rare mutations by any means… Sometimes we see them as a compounding presentation.
I use PD-L1 as the starting point histology and then STK11, KEAP1 and obviously the oncogenic drivers… I tell my patients that we don’t have the most solid prospective data for these 2 markers, but if it was me, I would go for a combination… So I try to really look at the full setup of mutations.
I also use, sort of as a prognostic bonus, the tumor mutational burden assessment. If patients have high tumor mutational burden, I think that this is an additional plus. If they’re low, certainly it’s more concerning for less impact. It does correlate a little bit with smoking status.





























