NSCLC Paradigm Continues to Push Forward With Emerging Targets, but Predictive Biomarkers Are Limited

As targeted therapies emerge for EGFR exon 20 insertion mutations and KRAS G12C mutations, the composition of a patient’s genomic profile continues to be an important clinical tool for treatment selection in non–small cell lung cancer (NSCLC), said Nagashree Seetharamu, MD, MBBS, who added that although biomarkers, such as PD-L1, tumor mutational burden (TMB), and microsatellite instability (MSI) can provide prognostic value, they are not perfect tools to predict response.

“The targets [for NSCLC] are evolving,” said Seetharamu. “We are no longer stuck in the zone where we are just checking for EGFR, ROS1, and ALK [mutations]. We need to have comprehensive genomic testing with broad molecular panels. Many times, the roles of tissue-based testing and blood-based testing are complementary, and [we should be] testing at every point of progression.”

“Multiple drugs are in development, and the portfolio for lung cancer is changing. The landscape is changing so rapidly that it is extremely important for us to make sure we keep updating our knowledge,” Seetharamu added.

In an interview with OncLive® during an Institutional Perspectives in Cancer webinar on lung cancer, Seetharamu, an assistant professor at the Donald and Barbara Zucker School of Medicine at Hofstra/Northwell and a medical oncologist in the Department of Medical Hematology and Medical Oncology at Northwell Health, discussed emerging targets in NSCLC and the limitations of current biomarkers.

OncLive®: How is the ongoing ZENITH20 trial (NCT03318939) contributing to our understanding of treating patients with EGFR exon 20 insertion mutations?

Seetharamu: The trial design [of ZENITH20] is quite smart. [Investigators] are including patients with EGFR exon 20 insertions, as well as HER2 insertions, so the number of cohorts [included] is appropriate.

We’ve known that EGFR exon 20 insertions are very different from all EGFR mutations. [EGFR exon 20 insertions] are smart molecules that try to duck into the C helix pocket [structure].

[ZENITH20] also brought to attention that the drug poziotinib is not that easy [for patients] to tolerate. [The study] led us to understand the class adverse effects of the TKI, and that perhaps dosing [of poziotinib] needs to be modified further. That was the take-home message [of ZENITH20]. Poziotinib is effective in this population, so if we can tweak the dosing, perhaps it may end up being a very good option for patients.

Mobocertinib was granted a breakthrough therapy designation from the FDA in April 2020. What clinical implications could an approval for mobocertinib have for patients with EGFR exon 20 insertion mutations?

We are desperately waiting for something to be available for [patients with EGFR exon 20 insertion mutations]. Many of the population-based studies have cited [the incidence of] EGFR 20 insertions to be somewhere [between] 3% and 10% [of lung adenocarcinomas]. It is a wide range, but the more we look [for these mutations], the more we will find. With the robust comprehensive genomic testing that many centers are adapting to today, the number is much higher. Is it the luck of the draw? I don’t know. I have a lot of patients [with EGFR exon 20 insertion mutations] that I am referring to clinical trials or [enrolling to] the expanded access protocol [for mobocertinib] that we are fortunate to have for our patients. An approval in this setting would be huge.

Once we have something for EGFR exon 20 insertions, we will be comfortable with the frontline [setting], but we know that is a finite space. We are clocking [disease progression] scan-to-scan; the patients are anxious, we are anxious, and we know that at some point in time, patients will have disease progression.

How do treatment decisions change when a patient with NSCLC experiences disease progression?

A big question is: What’s next? I don’t know if we have the answers. Amivantamab or the bispecific antibodies may be answers [to this question].

In practice currently, when patients progress, we assess them individually to see whether it is oligoprogression or systemic progression. If it is oligoprogression, the answer is easy. We treat the progressive spots and continue [the patient] on treatment. We can watch those patients with slow progression as well. Patients who progress rapidly move on to chemotherapy or a clinical trial [if one is] available. However, for the most part, [patients with rapidly progressing disease represent] a big unmet need. Combining chemotherapy or VEGF inhibitors with TKIs is something that we do in practice. I typically do not include a checkpoint inhibitor until I have no other options because I do not think [checkpoint inhibitors] are very effective in this particular patient population.

How are you navigating the available TKIs for patients with MET exon 14–altered NSCLC?

We had none, and now we have plenty [of MET TKIs]. We used to use crizotinib [Xalkori] when capmatinib [Tabrecta] or tepotinib [Tepmetko] were not available. Given the better tolerability profile and efficacy of capmatinib and tepotinib [compared with crizotinib], crizotinib is not my go-to [agent].

Between [capmatinib and tepotinib], it is a toss-up. Capmatinib came first, so [using that agent] would be my knee-jerk reaction. I’ve treated more patients [with capmatinib vs tepotinib], so I’m comfortable [using capmatinib].

One of our first-year fellows presented a case of a patient who had mild progression and was not tolerating capmatinib; [however, the adverse effects (AEs) reported] were class effects. I’m not sure switching to tepotinib would be considered for this particular patient, but in those situations, having multiple drugs is helpful.

How is the field pushing forward in the development of KRAS G12C inhibitors?

The randomized CodeBreak 200 trial [NCT04303780] is ongoing. That study [could] answer a lot of questions, but as noted by multiple investigators, this is also a finite [space] where a lot of patients quickly develop resistance.

Typically, patients [with previously treated KRAS G12C-mutant NSCLC] also respond to checkpoint inhibitors, unlike patients with EGFR mutations. Combination strategies might be more appropriate in [the KRAS G12C-mutant population vs the EGFR-mutant population].

In terms of updates, [the CodeBreak 100 trial] is moving along with sotorasib. [Adagrasib] is also moving along].

What still represents an unmet need are the patients with KRAS mutations [other than G12C]. Some indirect efficacy has been demonstrated with [sotorasib and adagrasib], but a lot of unmet need still [exists] in that specific population. [Targeting KRAS G12C] was an impossible nut to crack just a couple of years ago. Now, we’ve made a dent, and I’m hopeful that the future will change. It will be great to have multiple options [to target KRAS mutations] as well.

What are some of the limitations with current biomarkers in NSCLC? What are your thoughts on the utility of biomarkers overall?

No biomarker is perfect. We are used to using PD-L1, TMB, and MSI; those are common biomarkers. Another biomarker that has been brought up is tumor-infiltrating lymphocytes. Some peripheral blood markers, including peripheral blood TMB, [which has been put forth] as a better biomarker compared with tissue-based [TMB], is another biomarker that I’m aware of. Some radio-genomics and bioinformatics [work] is also ongoing.

Ultimately, it’s helpful to know when [a therapy] is actually hurtful for the patient, meaning when it is going to cause significant AEs vs when it is going to be beneficial. I do not want to exclude a patient [from receiving a therapy] based on biomarkers when they might actually have a benefit. I’m sure every oncologist, myself included, has had a patient that checks zero on every box but still benefits from [a therapy]. We’ve all also had patients with a PD-L1 score of 100 and high TMB who don’t respond [to an immune therapy]. Biomarkers are imperfect.

I’m not [as interested in using] predictive biomarkers to exclude patients, but I am interested in finding biomarkers that will [predict which] patients will have more AEs.

How could investigational biomarkers affect treatment decisions in NSCLC?

In every single category [of NSCLC treatment], a lot of progress [is being made]. Other biomarkers, including STK11, could have negative predictive value for checkpoint inhibitors. However, again, biomarkers are not perfect, so just because I see an STK11 [mutation], does not mean I’m going to prescribe [a certain] medication that a patient is unlikely to benefit from. I don’t think [STK11] has that kind of [predictive] strength. However, when I see an STK11 mutation with a KRAS G12C mutation, and when the TKIs become available, I know that KRAS G12C has the upper hand and that these patients [will] benefit [from KRAS inhibitors].

We can put [biomarkers] into the algorithm and present them to the patient; however, I would hate to take away the hope [of a potentially effective treatment] based on one, little biomarker. A lot of progress is ongoing [regarding] new combinations and new targets.

What unanswered questions need to be addressed in NSCLC?

The most [significant] unmet need is what to do after a patient progresses and how to sustain responses. My biggest problem is treating patients with disease in sanctuary sites, such as liver disease, bone disease, and central nervous system disease. [In these cases], entire-systemic disease is controlled, but [the disease] continues to progress [in these sanctuary sites].

I’m often at a loss as to how to treat these patients. Based on clinical experiences in small studies, a lot of times I add VEGF inhibitors, but there is nothing specific [to treat these patients]. As we are separating lung cancer, looking at these individual patient populations separately rather than grouping them all together is something we should start doing in the future.

What advice do you have for fellow clinicians who are making treatment decisions in the rapidly evolving landscape of NSCLC?

Even today, we get referrals for patients who have been started on TKIs for EGFR overexpression despite that clearly not being the right [treatment choice]. It is hard to keep up with the science, so we need to attend meetings, connect with a group of providers, and update each other. That is probably the only way to [ensure patients are getting optimal care].