Adagrasib Elicits Responses in KRAS G12C–Mutant NSCLC

Adagrasib, a KRAS G12C inhibitor, showed early responses in the phase 1b portion of the phase 1/2 KRYSTAL-1 multi-cohort trial (NCT03785249), which studied patients with KRAS G12C–mutated non–small cell lung cancer (NSCLC) who harbored active and untreated brain metastases, according to Joshua K. Sabari, MD.

In the phase 1b cohort of KRYSTAL-1, the overall response rate (ORR) of 19 evaluable patients with active, untreated central nervous system (CNS) metastases was 32% (n = 6; 95% CI, 12.6%-56.6%), with objective intracranial responses observed in patients with target and non-target lesions. Sabari presented these data at the 2022 ASCO Annual Meeting.1

“KRAS has been undruggable for many years, and we’ve had many failed trials in this disease,” Sabari said. “[Adagrasib and sotorasib (Lumakras)] are the first generation of KRAS G12C inhibitors. It’s an exciting time.” Adagrasib has a long half-life, 23 hours, which differentiates it from other KRAS G12C inhibitors, Sabari added.

In an interview with OncLive^®, Sabari, an assistant professor in the Department of Medicine at NYU Grossman School of Medicine and the medical director of Kimmel Pavilion 12 at NYU Langone Health, discussed the significance of adagrasib as an agent with favorable CNS penetration. He also emphasized the importance of performing next-generation sequencing (NGS) in all patients with NSCLC to detect KRAS G12C mutations.

OncLive^®: Could you provide some background on adagrasib and explain what makes this agent unique from others available in this patient population?

Sabari: Adagrasib, formerly known as MRTX849, is a KRAS G12C inhibitor. It has dose-dependent pharmacokinetics. It also has a long half-life, 23 hours, which differentiates it from other KRAS G12C inhibitors.

What has been previously observed with this agent regarding CNS penetration in treated, stable CNS metastases?

KRAS G12C mutations are quite common, occurring in 10%−12% of patients with NSCLC. In that population, upward of 40% of patients have brain metastases at presentation. This is an unmet need for our patients. The prognosis is poor; the median overall survival [OS] is about 5 months.

It’s important for us to study agents that have activity in the brain. Adagrasib has an interesting property in that it inhibits the P-glycoprotein efflux pump. This is a transporter that usually pumps out drugs from the brain, and adagrasib directly inhibits it.

When we studied this drug preclinically in mouse models, we were able to see nice CNS concentration. In the clinic, when we administered this drug in patients at a regular dose of 600 mg [twice daily], which is the recommended phase 2 dose, we were able to do lumbar punctures in 2 patients in a phase 1b cohort specifically for patients with active and untreated brain metastases. The average Kp,uu unbound partition coefficient ratio was 0.47, suggesting that this drug has very good CNS penetration.

We also have seen clinical regression in patients. The intracranial ORR in the phase 1b cohort was 32%, which is similar to what we’ve seen extracranially. The CNS is going to be an active area of ongoing development of agents. In other targeted alterations such as EGFR and ALK, having drugs with significant CNS activity is critical. It’s similar in the KRAS space.

Can you expand on the design of the phase 1b portion of the KRYSTAL-1 study and the overall trial objectives?

The KRYSTAL-1 study is a phase 1/2 [study including] dose escalation followed by expansion. We initially looked at just safety with the recommended phase 2 dose to understand the maximum tolerated dose. We rapidly escalated to 600 mg twice a day, and then we went into the expansion cohort, where all patients were treated at 600 mg twice a day.

Data [from cohort A of the phase 2 portion of that study] were presented at ASCO and were quite impressive. We saw a 43% objective response rate, a median progression-free survival [PFS] of about 6.5 months, and a median OS of 12.6 months.

This [cohort included] an additional 25 patients enrolled at the recommended phase 2 dose. These patients had active and untreated brain metastases. KRYSTAL-1 is the first study to look at patients with [with KRAS-G12C mutant NSCLC] and active and untreated brain metastases. We defined these patients as those who were clinically asymptomatic, were not on high doses of steroids, and were not on anti-epileptics. We excluded patients who had brainstem lesions, but we did allow patients who had leptomeningeal disease, if they were clinically asymptomatic.

An efficacy objective of this cohort study was to look at the intracranial response rate by modified [response assessment in neuro-oncology (RANO)] brain metastasis criteria. We also looked at extracranial response by RECIST v1.1 criteria. We wanted to look at safety and, if feasible, we also wanted to look at the CSF concentrations of the drug. We were able to do that in 2 patients.

What is important to know about patients with untreated CNS metastases?

We know from prior presentations of adagrasib that this drug was associated with a 33% intracranial response rate in patients with stable, treated brain metastases. With sotorasib, formerly known as AMG 510, we saw prior presentations of about a 13% response rate in patients with stable treated brain metastases. [Some] patients had prior radiation to the [brain] lesion, so it was difficult to identify whether the prior radiation or [adagrasib] caused the regression or response.

In the phase 1b cohort of adagrasib in the KRYSTAL-1 study, we specifically looked at patients who had active and untreated brain metastases. The response rate by the modified RANO brain metastasis criteria by a blinded independent review committee, was 32%. This was similar to the rate we saw extracranially, which was 37%, with an 88% concordance in responses in the brain, as well as responses in the chest and extracranially. A key [detail about this study] is that no significant, new CNS-specific adverse effects [AEs] were identified with this agent.

What has this study shown in terms of durability of responses?

The median intracranial OS was not yet reached in this study, which is exciting, because the median OS for patients with NSCLC with brain metastases is poor. At [a median follow-up of 6.6 months], the median intracranial PFS was 4.2 months. This is in line with what we would expect from this patient population.

We also saw patients who had prolonged responses to therapy. One of the patients who I personally treated had a durable long-term response to treatment of over 12 months. We’re seeing responses in the clinic, and we’re also seeing durability of those responses. Further data is needed to confirm these findings.

How did intracranial responses differ in those with target lesions vs those with non-target lesions?

We enrolled 25 patients onto the study, and [19] patients had RECIST v1.1–measurable disease. Of those, [15] had target and non-target lesions, and 4 patients had non-target lesions only. The response rates were similar [across these 2 groups]. The intracranial disease control rate was also similar across both cohorts.

When developing a prospective study measuring active and untreated CNS metastases, it’s important [to include patients with] target lesions. Moving forward, [trials with] adagrasib, as well as with other agents being studied in the brain, should have strict criteria for enrollment like we did.

Could you expand on the relationship between systemic and intracranial disease control?

Current KRAS G12C inhibitors are not as effective as some of the EGFR inhibitors that we see. With EGFR inhibitors, we’re seeing response rates in the 80% range. With ALK inhibitors, we’re also seeing response rates in the 80% range. In KRAS G12C–mutant disease, we’re seeing responses in the 35%-45% range, so we still have a long way to go.

The intracranial response of 32% that we saw with adagrasib [in the KRYSTAL-1 study] appears low, but it’s similar to what we would expect from an extracranial response. If a patient has a KRAS G12C mutation and is responding extracranially in the chest, they’re likely also going to respond in the brain. However, if they’re not responding in the chest, they will probably not derive benefit in the brain.

We do need to do more; we need to understand these agents better. We need to find better combination strategies in later lines without sacrificing or forgetting about the need for CNS activity.

What is important to know about the safety profile of adagrasib?

[In KRYSTAL-1], patients were treated in a fasted state at 600 mg twice a day with the capsule formulation [of adagrasib]. We did see higher rates of nausea, abdominal discomfort, and emesis. I’m hoping that with the formulation now as a tablet, these AEs will be less frequent, particularly in the fed state.

In clinical practice, most patients had these gastrointestinal AEs with adagrasib in the first 2 to 4 weeks of therapy, and then they significantly improved over time, either with some anti-emetics, or as the patient became more used to these AEs.

What do these data mean for patients with KRAS G12C–mutated disease who have untreated CNS metastases, and how do you see this agent fitting into the treatment paradigm?

For patients with KRAS G12C–mutant NSCLC, the current FDA-approved frontline regimens are chemotherapy and immunotherapy or immunotherapy alone. Chemotherapy does not have great CNS activity. Immunotherapy, particularly pembrolizumab [Keytruda], which is often used in the frontline setting, does have CNS activity.

There is an ongoing cohort in the phase 1 expansion study with sotorasib for patients with active and untreated CNS metastases. We’re looking forward to seeing that data. Additionally, we’ve shown that adagrasib does have CNS activity, but it’s not yet FDA approved. It has a PDUFA date of December 14, 2022.

An approval could change the potential landscape of this agent. Preclinically, the agent is robust because of its dose-dependent pharmacokinetics and its long half-life. Because of the CNS activity, particularly in a patient with active untreated CNS metastases, I would prefer to use adagrasib over sotorasib. However, according to the data that we currently have, if a patient is symptomatic, has vasogenic edema, has a centimeter or centimeter-and-a-half–sized lesion, or required steroids, I would recommend radiating that lesion prior to considering a systemic therapy.

What are the next steps with adagrasib and sotorasib?

The next step for these agents is to confirm their activity in a phase 3 study. Both sotorasib and adagrasib are being compared with docetaxel in the second line. [These agents] work in the second-line setting.

From there, the next logical step is asking: What if we move them to the frontline setting? Would patients respond longer and better? Both agents are now being studied in combination with a PD-1 inhibitor, pembrolizumab, and I look forward to seeing more results from these studies in the frontline setting, because there’s a potential role for these agents there.

Lastly, in the later lines in patients who’ve had progression on KRAS G12C inhibitors, the resistance landscape is diverse. Thinking about combination strategies, like potential combinations of upstream inhibitors such as SHIP2 inhibitors or SOS1 inhibitors or potential downstream inhibitors, might be an effective strategy for targeting some of those resistant clones.

What main message about KRAS G12C mutations in NSCLC would you like to impart to colleagues?

Test your patients in the frontline setting. You need to molecularly profile your patients. If you’re not obtaining NGS, you can’t identify the best possible treatments for your patients.

In the past, we always profiled our young, never-smoking patients, but KRAS G12C occurs almost unanimously in patients who have smoked in the past. It’s no longer appropriate to select who you sequence or who you obtain NGS for; it needs to be done in all patients.

KRAS is an oncogene and a driver alteration. If we target it and successfully inhibit it, we are benefitting and improving our patients’ quality of life and longevity. Thinking about the future, many inhibitors are being developed that may be even more potent than the covalent inhibitors that we have now in G12C.

Reference

Sabari JK, Spira AI, Heist RS, et al. Activity of adagrasib (MRTX849) in patients with KRASG12C-mutated NSCLC and active, untreated CNS metastases in the KRYSTAL-1 trial. J Clin Oncol. 2022;40(suppl 17):LBA9009 doi:10.1200/JCO.2022.40.17_suppl.LBA9009