Watson Unpacks the Potential Clinical Impact of High-Level MET, HER2, and KRAS Amplifications in NSCLC

Findings indicated that the frequency of driver-positive tumors decreased with increasing copy number gain in non–small cell lung cancer (NSCLC) tumors, and investigators were able to set tissue next-generation sequencing (NGS)–based HER2, KRAS, and MET copy number gain thresholds, according to Alexander Watson, MD, DPhil, FRCPC.1

“This is exploratory work that’s helping build on prior work in the field of amplification and defining what is a meaningful copy number gain. The heterogeneity in the field makes comparing results very difficult because everyone is defining copy number gain a different way, and everyone is defining what threshold defines copy number gain within each mutation group a different way,” Watson explained. “I’m hoping this continues to build the field toward a common understanding of what is an amplified tumor [that] then could be taken into therapeutic investigations and allow us to better understand how effective therapies are in amplified groups. For patients who have these amplified tumors, [I’m hoping we] continue to build better therapies that are able to truly target the genetics of their cancer, which we can see but can’t yet meaningfully leverage to improve outcomes in these tumors.”

Findings from the retrospective study revealed that when driver positivity overlapped with amplification status, same-gene alterations were significantly enriched for HER2, KRAS, and MET. However, BRAF and EGFR mutations were more common in MET-amplified tumors than in HER2- or KRAS-amplified tumors. Furthermore, there was a negative survival association with amplification status independent of driver-positivity status for HER2 and MET; this was not observed for KRAS.

In an interview with OncLive®, Watson detailed findings from the study and the next steps for this research. Watson is an advanced fellow in thoracic oncology and investigational cancer therapeutics at the University of Colorado Denver–Anschutz Medical Campus. In a concurrent interview, he detailed the rationale for the study and how it was conducted, as well as findings showing that oncogene overlap is adequately applied to NGS-based tissue sampling by decreasing frequency and increasing copy number gain.

OncLive: What were the findings from the study once the threshold was defined?

Watson: The threshold for the copy number gain that [we determined to be] meaningfully amplified for KRAS and HER2 was 6, whereas for MET, we used a copy number gain threshold of 4 [because] 4 or more led to a plateau in our frequency. Using that, our next goal was to look at how specific driver mutations coexisted with highly amplified MET, KRAS, or HER2 because you’d guess that overall, we selected a population [of patients] who have a lower frequency of driver mutations, and that’s the principle of mutual exclusivity. But there might be some mutations that are more likely to coexist with an amplified gene. That’s because [with] acquired resistance to therapies, certain gene pathways are more synergistic for cancer growth. For example, EGFR’s reliance on MET as a resistance pathway is well-defined. We frequently combine a MET agent with an EGFR agent if we identify MET amplification or a MET mutation as a mechanism resistance. We can see objective response rates in tumors that didn’t respond anymore to an EGFR TKI responding to the addition of a MET-directed therapy.

[Therefore], we wanted to see which copy number gain genes coexisted with driver mutations that were more traditionally defined. We found that for the MET group, for all 3 genes as a headline, there was frequent coamplification and comutation. We saw same-gene alterations at an increased frequency than would be expected, and this has been previously defined as a phenomenon, but not as generalized and not quantified in the way that this NGS investigation allows us to [be]. [This is] because NGS allows us to simultaneously see mutations and copy number gain in the same cohort in a way that FISH-based amplification is harder to [see with].

We found that, for example, MET amplification was more likely to occur with MET mutations than other driver mutations, and for MET, it was more likely that the high-level amplification occurred with a MET mutation than without. The idea then would be that if you have a mutated gene—for some mutated genes, that cancer is already reliant on that pathway—perhaps further emphasis on that pathway is able to overcome methods of tumor suppression or even targeted therapy. [Although] we don’t have that data for what these cancers were exposed to, we saw for all 3 genes there seemed to be a same pathway mutation and amplification in process. For MET, we did also get a signal that EGFR mutations were more likely to occur with MET-amplified than the other-amplified tumors and BRAF [mutations were, as well]. EGFR is well-defined—we know that MET is a pathway of resistance in EGFR-mutated cancers, but BRAF is less well-defined. BRAF TKIs exist in lung cancer, and they’re approved and used, but the resistance mechanisms to BRAF inhibition or what pathways are synergistic with BRAF [are] less well understood.

What are other key considerations with these findings?

In the future, it would be nice to see a better-annotated data set [examining] whether BRAF-treated cancers are the ones that are popping up with more MET amplification. This is preliminary. The limit of our investigation is that these highly amplified tumors are very rare events. Even with [approximately] 13,000 tumors in our cohort, [a small portion] of them were highly amplified for HER2, KRAS, or MET, which means that if you then subdivide that [small percentage of tumors] into the different mutation subgroups, you start to get limited for power. We would have [liked to] better define what genes were synergistic in a larger cohort of amplification, but these are rare events when you define them stringently, and this is a large cohort. Perhaps focusing on a predefined selection for amplification in the future might allow [for a] better understanding of that landscape.

The big limit to our coexisting mutations analysis is we do not know which treated tumors were [treated] with TKIs and which weren’t. That means that all of the guesses of what a primary resistance mechanism and a secondary resistance mechanism is are speculative. The whole principle of oncogene overlap means it is rare to see primary resistance with 2 active oncogenic pathways in the same tumor. More common is the acquisition of a second pathway in the setting of resistance to targeting of a pathway, but we aren’t able to separate which tumors were treated and which tumors weren’t in our analysis, which limits our ability to understand the implications of the coexisting genes that we’re identifying.

What were the survival findings from the study?

There was a hypothesis-generating survival analysis we added on at the end. We had access to basic insurance claim data, which we could track as the first submitted whenever Caris Life Sciences had a tumor submitted to them for sequencing. We would then be able to track insurance claims for those patients until their last claim, at which point it would be a surrogate for survival, although there are some confounders on what a last claim means. We could track it in a general way, from the time a tumor was submitted until the patient was no longer able to be active on cancer therapies, for whatever reason, and because we focused on advanced tumors in this cohort, the presumption would be that that’s a surrogate of overall survival.

We were then able to look at the impact of copy number thresholds on survival. For each [of the 3] genes [examined], the higher copy numbers of each gene led to a higher HR for death, and when divided by each gene amplified for HER2, KRAS, or MET, there seemed to be negative implications for HER2 and MET with or without driver mutations—it seemed to have a negative prognostic association. For KRAS, that association was present when KRAS amplification coexisted with another driver mutation, but not without.

[However], this is all speculative. We do not have treatment data for these tumors, and one could suspect that if amplifications are acquired over time, perhaps all we’re doing is saying highly amplified tumors are further along in their treatment course, ergo, [patients are] going to have a lower [chance of] survival from sequencing until the last claim. But if we’re looking for other explanations for why a negative survival association exists, we know that amplification of these genes is not yet targetable in general. [For] MET amplification, there are some small cohorts that [appear to show] that applying a TKI has [an] effect on MET-amplified tumors; there’s a response rate, and that response rate seems to track with higher levels of copy number gain, but that’s not understood for KRAS or for HER2 [in NSCLC]. It’s a driver mechanism without an approved therapy right now.

There’s also [an] implication [that] most of these tumors may be acquiring resistance with amplification, and that would therefore also associate with lower survival because that tumor is further on its treatment course [as] it’s a resistant tumor that’s had a first-line therapy or two. [Perhaps] amplifications are more likely to be acquired in certain tumors that have more aggressive phenotypes, such as genomic instability, and that would be speculative. But either way, this furthers [what] has been observed elsewhere, that higher amplification levels seem to be associated with lower survival. This would be a group [of patients for whom] we’d want to see better therapies investigated [and] hopefully approved. [With] these amplified cohorts in particular, these [patients] are not doing as well, whether that’s in a resistance setting or in a primary setting, acknowledging the confounders in our analysis.

What are some of the next steps with this research?

This is a large cohort that has defined oncogene overlap as a potential method to better set our threshold for what is an amplified tumor and what is not. We’d love to see those amplified thresholds taken forward in therapeutic investigations, particularly for targeted therapies against the pathways that are amplified. We’re hoping that this higher threshold defined with oncogene overlap allows a judicious application of a threshold, which defines a cancer that might respond to TKIs vs those that might not, in a way that an arbitrary set threshold may not as well. We want to set a threshold that’s high enough that we’re not including tumors that are just using amplification that’s along for the ride as a passenger mutation, [because] it’s not truly a driving pathway, [and not] too low where we’re ignoring tumors where therapy might be effective.

Prospective investigations, now that HER2 TKIs are approved, [represent] a particular therapeutic gap where we need to understand how HER2-amplified tumors respond to these TKIs, which are coming on the market or finishing early phase clinical trials with promising results. For KRAS, we’re limited by the agents we have, [but] we have KRAS G12C inhibitors. There aren’t data that I’m aware of around the response of tumors that are KRAS G12C amplified, but our investigation included all KRAS-directed agents. In general, the KRAS therapeutic space needs further therapeutic investigations and optimization beyond just amplified tumors.

For MET and HER2, there are agents that are approved. We’d love to see prospective investigations with approved agents in these amplified groups we’ve defined. The annotations are a big limit of our work. It’ll be nice to see future investigations with better-annotated data sets dig into the implications of coexisting amplification and mutations, if that’s truly a mechanism of resistance. The survival implications will come with better annotation as well, understanding what’s a primary mechanism and what’s a secondary. Hopefully, these thresholds and the principal of oncogene overlap will help future clinical trials of targeted agents in the spaces of HER2 amplification, MET amplification, and perhaps, in the future, KRAS amplification. Better annotated data sets will define the associations we’ve identified in the context of the true treatment history of those tumors.

Reference

1. Watson AS, Krause HB, Elliott A, et al. Use of oncogene overlap by tissue-based next-generation sequencing to explore the mutational landscape and survival impact of HER2, KRAS and MET copy-number gain in nonsmall cell lung cancer. Clin Lung Cancer. 2024;25(8):712-722.e1. doi:10.1016/j.cllc.2024.09.001