Next-Generation KRAS Inhibitors in NSCLC: A Pipeline So Active It's Hard to Keep Up

Introduction: From Undruggable to Overcrowded

KRAS was considered undruggable for decades. The protein's smooth, shallow surface offered no obvious pocket for small molecule binding. Its extremely high affinity for GTP/GDP made competitive inhibition impractical. Its lack of suitable allosteric sites seemed to foreclose most conventional drug design approaches. The first approved KRAS inhibitors, sotorasib and adagrasib, broke through those barriers by exploiting a covalent binding opportunity specific to the G12C mutation in its inactive (GDP-bound, off) state. The results were modest — better than docetaxel, but not by much, and resistance emerged quickly.

In 2026, the picture looks dramatically different. The G12C space has at least six or seven next-generation inhibitors in active clinical development, several showing more than double the response rates of the approved drugs. On-state inhibitors and protein degraders are tackling KRAS mutated NSCLC from entirely new angles. G12D, the most common KRAS mutation across all solid tumors, now has its first clinical data with some promising results. Pan-RAS inhibitors capable of blocking multiple RAS mutations and wild-type RAS simultaneously have reached phase 2. The field has gone from a single approved target in one disease to a proliferating landscape that is genuinely difficult to track.

KRAS mutations occur in approximately 20 to 30% of non-small cell lung cancers, predominantly in lung adenocarcinoma. G12C is the most common subtype in NSCLC, followed by G12D. Standard of care has not yet changed — sotorasib and adagrasib remain the only FDA-approved options, both in the second line — but that will not be the case for long.

The Approved Drugs: What We Have and Their Limitations

Sotorasib (CodeBreak 200): Compared to docetaxel in second-line KRAS G12C NSCLC, sotorasib produced a median PFS of 5.6 months versus 4.5 months. No significant overall survival difference was observed. Diarrhea at grade 3 occurred in 12% of patients; 27% required dose holds, and 11% discontinued.

Adagrasib (KRYSTAL-12): PFS 5.5 versus 3.8 months versus docetaxel. A notable feature was a 24% intracranial response rate. Nausea is the dominant patient-reported complaint — seen in 34% of patients, though mostly grade 1 to 2. Both drugs target the G12C mutation in its inactive GDP-bound (off) state, and both face a common resistance challenge: on-target mutations at the G12C binding pocket, off-target upregulation of wild-type RAS and other RAS family members, epithelial-to-mesenchymal transition, and histologic transformation. The same mechanisms that create resistance to one drug create resistance to the other, limiting sequential use.

Next-Generation G12C Off-State Inhibitors: Substantially More Active

Several next-generation G12C inhibitors share the same off-state mechanism as sotorasib and adagrasib but offer significantly greater potency and selectivity.

Olomorasib: A phase 1/2 study combining olomarasib with pembrolizumab showed an overall response rate of approximately 77%, with approximately 75% of responses in the first-line setting. Tolerability appears better than the current approved drugs: the dose reduction rate was 20% and discontinuation was 6.5%. Diarrhea remains the most common treatment-related adverse event. The SUNRAY-2 phase 3 trial is now enrolling patients with both resectable and unresectable KRAS G12C NSCLC.

Divarasib: Has shown ORR of 55% range in phase 1 data, with PFS ranging from 9 to 14 months, and discontinuation rates of approximately 5%. Phase 3 trials combining looking at divarasib and pembrolizumab compared with pembrolizumab alone or in combination with chemotherapy. 

**Garsorasib—**Pooled phase 1 and 2 study showing median PFS of 9.1 months and median OS of 14.2 months. 

MK-1084: Second-line data show PFS of approximately 8.3 months in MK-1084 monotherapy and ORR 38%. The combination with pembrolizumab in the frontline setting is particularly striking: 77% ORR in patients with PD-L1 TPS ≥1%, and a median PFS of 25 months in that population. Even in patients with any PD-L1 expression combined with chemotherapy, PFS reached approximately 10 months.

Glecirasib: Phase 2 data showing median OS of 13.6 months and PFS of 8.2 months with ORR of 47.9% with grade 3-4 treatment related adverse events of 38.7% with 5% discontinuation rate. A phase 1 and 2 study studying glecirasib with a Shp-2 inhibitor sitneprotafib has also been studied showing ORR 71% in untreated KRAS G12C NSCLC and 49% in patients previously treated with systemic therapy but naïve to KRAS G12C inhibitors.

Elisrasib (most recently reported at AACR): 60% ORR and PFS of 9 months in patients without prior G12C inhibitor exposure; approximately 33% ORR in G12C inhibitor-pretreated patients. An important finding in this study was the performance in patients with STK11 and KEAP1 co-mutations, which are among the most immunotherapy-resistant molecular contexts in NSCLC. Elisrasib showed a 55% frontline ORR in this co-mutated population, suggesting that improved KRAS G12C inhibition could benefit this subgroup of patients whose tumors are resistant to immunotherapy.

The On-State Revolution: RAS(ON) Inhibitors

The approved G12C inhibitors only work on the inactive (off/GDP-bound) form of KRAS. This creates an inherent limitation: the on-state (GTP-bound, active) form of KRAS is inaccessible, and resistance can emerge when cells keep KRAS in the active state or when downstream signaling is maintained through other means.

RAS(ON) inhibitors take a fundamentally different approach. Rather than competing for a binding pocket on inactive KRAS, they form a tri-complex with KRAS and a chaperone protein called cyclophilin A, locking KRAS in a conformation that cannot signal downstream regardless of its nucleotide state. This is categorically different pharmacology.

Elironrasib (Revolution Medicine, G12C-specific RAS(ON)): In previously treated patients, RMC-6291 produced a 42% overall response rate. At ESMO last year, data in treatment-naive patients showed a 56% ORR, 10-month median PFS, and a 12-month overall survival rate of 65%. An ongoing study is evaluating elironrasib and pembrolizumab.

BBO-8520 (BridgeBio): This agent is distinctive in that it inhibits KRAS in both the on and off states. In monotherapy, 17 patients showed a 65% ORR with a 68% 6-month PFS rate. In combination with pembrolizumab, 8 patients showed a 62.5% ORR. Grade 3 or higher diarrhea was 7% in the combination arm; grade 3 liver enzyme elevations were 7%. The waterfall plot shows deep responses. Very early data, but directionally compelling for a dual on/off inhibitor.

KRAS G12D: The First Clinical Data

KRAS G12D is biologically distinct from G12C. It lacks the reactive cysteine residue that enables covalent G12C inhibitor binding, making the drug design challenge more difficult. Preclinical data suggest that G12D tumors grow more rapidly and create a more immunosuppressive microenvironment than G12C tumors. G12D is more common in never-smokers and, across all solid tumors, is the most prevalent KRAS mutation subtype.

Zolondorasib (RMC-9805, Revolution Medicine, G12D on-inhibitor): Phase 1 data presented at AACR showed a 52% ORR and a 73% 12-month overall survival rate. Grade 3 toxicity was 13%, with only 3% grade 3 diarrhea. These are modest response rates but meaningful proof of concept for a mutation that had no targeted options until recently.

G12D targeted protein degrader (NEJM publication): Setidegrasib is a PROTAC that recruits VHL E3 ligase to degrade the KRAS G12D protein itself, rather than inhibiting its function. In a combined NSCLC and pancreatic cancer cohort, the NSCLC cohort of 45 patients receiving 600 mg IV weekly showed a 36% ORR, median PFS of 8 months, and continued activity in second and third-line settings (40% ORR, PFS 11 months). Infusion reactions occurred in approximately 60% of patients, but no grade 3 infusion reactions were seen and no patient discontinued due to infusion reactions.

Pan-RAS Inhibitors: Targeting Multiple RAS Variants

Beyond mutation-specific approaches, pan-RAS inhibitors targeting multiple KRAS mutations and wild-type RAS simultaneously are advancing.

Daraxonrasib, Revolution Medicine: A non-covalent protein tri-complex inhibitor with activity across KRAS, HRAS, and NRAS at codons 12, 13, and 61. First-in-human NSCLC data presented at ELCC showed a 38% ORR in second-line patients, median PFS approximately 10 months, and overall survival approximately 17 months. The most significant toxicity is rash, but prophylactic antibiotics and topical agents reduce the incidence of grade 3 or higher rash to essentially 0%. Phase 3 studies versus docetaxel are enrolling, along with studies combining daraxonrasib with ivonescimab and with other RAS-directed agents.

Where the Field Is Going

The standard-of-care question is not fully resolved. Sotorasib and adagrasib remain the only approved second-line G12C options, and frontline use of KRAS inhibitors is not yet standard. But the pipeline will change this within two to three years.

The key open questions are: Should frontline KRAS inhibition use an IO combination, a chemotherapy combination, or the KRAS inhibitor alone? Does PD-L1 status modify the benefit of KRAS inhibitor plus IO combinations? Should on-state and off-state inhibitors be used sequentially, or can they be combined? How do SHP2 inhibitors and other downstream co-targeting strategies fit in?

The only certainty is that the KRAS treatment landscape will look dramatically different within the next few years. For patients with KRAS-mutated NSCLC today, referral to a center with active phase 1 and phase 2 KRAS trials is the most important recommendation alongside the current approved options.

For Patients

If you have been diagnosed with non-small cell lung cancer and told that your tumor has a KRAS mutation, the therapeutic landscape is changing rapidly. KRAS was considered untargetable for decades; it now has two FDA-approved drugs for KRAS G12C mutation and many more in development that show substantially higher activity. Ask your oncologist specifically which KRAS mutation your tumor has (G12C, G12D, or another variant), since different drugs target different mutations. Ask about clinical trial options — many of the most promising next-generation agents are currently available only through trials, and a center with active KRAS-focused research may offer options beyond the current standard.

Key Takeaways

• Sotorasib and adagrasib, both second-line G12C off-state inhibitors, produce modest PFS benefits (5-6 months vs ~4 months for docetaxel) with relevant toxicity; resistance emerges through both on-target and off-target mechanisms.

• Next-generation off-state G12C inhibitors (olomarasib, MK-1084, elisrasib, others) show 50-77% ORR in early data, substantially exceeding approved agents; several phase 3 trials are enrolling.

• Elisrasib shows 55% frontline ORR in STK11/KEAP1 co-mutated patients, suggesting improved G12C inhibition may partially overcome IO-resistant contexts.

• RAS(ON) inhibitors target active-state KRAS through a tri-complex mechanism, bypassing the limitation of off-state-only inhibition; some examples include elironrasib (RMC-6291), daraxonrasib (RMC-6236), GFH547, BI-3706674, YL-17231

• KRAS G12D now has first clinical data: Setidgrasib (22% ORR) and a protein degrader (36% ORR, published in NEJM); both represent proof-of-concept for a historically untargetable mutation.

• Pan-RAS inhibitors (daraxonrasib) show 38% ORR in second-line NSCLC across multiple RAS mutations; rash is manageable with prophylaxis.

• Standard of care has not yet changed; clinical trial referral is the most important action for eligible patients.

References

1. CodeBreak 200. Sotorasib vs docetaxel in second-line KRAS G12C NSCLC. New England Journal of Medicine, 2023.

2. KRYSTAL-12. Adagrasib vs docetaxel in second-line KRAS G12C NSCLC. New England Journal of Medicine, 2024.

3. Olomarasib plus pembrolizumab in KRAS G12C NSCLC. Phase 1/2 data, ASCO 2025.

4. MK-1084 in KRAS G12C NSCLC. Phase 1/2 data, ASCO 2025.

5. Eli-rasif in KRAS G12C NSCLC including STK11/KEAP1 co-mutated. Phase 1 data, AACR 2026.

6. RM-018 (RAS(ON) G12C inhibitor). Phase 1 data, ESMO 2025; AACR 2026.

7. KRAS G12D protein degrader in NSCLC and pancreatic cancer. New England Journal of Medicine, 2025.

8. Zolbetuximab (pan-RAS inhibitor) in KRAS-mutant solid tumors. ELCC 2025.

9. Hsu RC. "Next Generation KRAS Inhibitors." Presented at the 2026 San Francisco Lung Cancer Conference, San Francisco, CA, May 2026.

About the Author

Robert Chun-Hao Hsu, MD, is a medical oncologist at the University of Southern California Norris Comprehensive Cancer Center in Los Angeles, specializing in thoracic oncology and head and neck oncology. His research encompasses studying new biomarkers with potential clinical impact, biomarker-driven treatment selection, and reducing cancer disparities and improving access to cancer care.