Where Are KRAS G12C Inhibitors Heading? Sotorasib, Adagrasib, and the 5 Axes of Resistance in 2026 | KRAS Frontier Vol.1

Key Points

• Sotorasib (Lumakras / Amgen) and adagrasib (Krazati / Mirati→BMS), approved in 2021-22, were the first clinical drugs to directly target the long-“undruggable” KRAS G12C mutation. Approved initially in second-line non-small-cell lung cancer (NSCLC), they have since expanded into colorectal, pancreatic, and biliary cancers.
• By 2026, single-agent objective response rates (ORR) sit at ~30-40% in NSCLC with median PFS 5-7 months. Early acquired resistance is the main clinical bottleneck. Mechanisms: (1) KRAS switch II pocket mutations (R68S, H95D/Q/R, Y96C); (2) EGFR / RTK reactivation; (3) bypass via NRAS, BRAF, MET, etc.
• Clinical strategy has pivoted to combinations: (a) KRYSTAL-7 (adagrasib + pembrolizumab, ORR 63% in PD-L1-high); (b) EGFR antibody combos (FDA-approved in CRC for both drugs); (c) SHP2 / SOS1 inhibitors; (d) chemotherapy.
• 2026 priorities: (1) front-line (1L) advancement; (2) neoadjuvant / adjuvant trials; (3) resistance sequencing; (4) tailored strategies in non-lung cancers (CRC, PDAC, biliary).

Introduction — When “Undruggable” Cracked

KRAS, identified in the 1980s, is mutated in ~25-30% of all human cancers (~30% of NSCLC, ~45% of colorectal, ~90% of pancreatic). For decades it was deemed “undruggable”: extremely high GTP/GDP affinity made competitive inhibition impractical, the protein surface lacked deep hydrophobic pockets, and high intracellular GTP outcompetes most candidates.

In 2013-14, Kevan Shokat’s UCSF lab showed that KRAS G12C exposes a cysteine residue inside the “switch II pocket” amenable to a covalent acrylamide warhead. Amgen (sotorasib / AMG-510) and Mirati (adagrasib / MRTX849) translated this into clinical drugs, with FDA approvals in 2021 and 2022.

This article reviews the 2026 status of both drugs, resistance mechanisms, and combination strategies.

Body

1. Clinical Profiles

Table 1: Approved KRAS G12C inhibitors

Item	Sotorasib (Lumakras)	Adagrasib (Krazati)
Sponsor	Amgen	Mirati (BMS acquired 2024)
FDA approval	May 2021 (NSCLC 2L); 2024 (CRC + panitumumab)	Dec 2022 (NSCLC 2L); 2024 (CRC + cetuximab)
Dose	960 mg QD	600 mg BID
NSCLC 2L ORR	~37-41%	~43%
NSCLC 2L median PFS	~6.8 months	~6.5 months
NSCLC 2L median OS	~12.5 months	~12.6 months
CRC monotherapy ORR	~10%	~22%
Notable AEs	Diarrhea, transaminase elevation, nausea	Diarrhea, nausea, vomiting, QT prolongation
CYP3A interactions	Mild	Strong (QT cautions)

Comparative analyses (PMC 2024) suggest similar efficacy with sotorasib showing a more favorable safety profile; adagrasib has better CNS penetration, suggesting potential advantage in NSCLC brain metastases.

2. Resistance — Switch II Mutations and RTK Reactivation

2021-25 clinical and translational research has crystallized the major acquired resistance mechanisms:

(A) On-target resistance:

• Switch II pocket mutations: R68S, H95D/Q/R, Y96C — block sotorasib/adagrasib binding.
• Secondary KRAS mutations: additional changes (G13D, Q61H) acquired in the same cell.
• KRAS amplification: increasing target levels to overwhelm drug.

(B) Off-target / bypass:

• EGFR / RTK reactivation — the most frequent and important. KRAS-pathway inhibition releases negative feedback, upregulating EGFR and reactivating RAS-MAPK via bypass.
• NRAS, HRAS, BRAF acquired mutations: WT RAS paralogs or downstream BRAF compensate.
• MET amplification, FGFR3 fusion: alternative RTKs.
• Epithelial-mesenchymal transition (EMT): phenotypic plasticity-mediated escape.

(C) Population dynamics:

• Drug-tolerant persisters: a reversible “dormant” state that resumes proliferation after treatment interruption.
• Tumor heterogeneity: clonal restructuring, minor-clone selection.

3. The CRC Single-Agent Failure — Context Dependence

While NSCLC sees 30-40% ORR, monotherapy in CRC achieves only ~10%, because:

• EGFR reactivates immediately upon KRAS G12C inhibition in CRC (more pronounced than NSCLC).
• Concurrent KRAS G12C + EGFR blockade is required.
• 2024 NEJM CodeBreaK 300: sotorasib + panitumumab achieved ORR 26%, PFS 5.6 months vs sotorasib alone (ORR 9%, PFS 4 months).
• KRYSTAL-1 expansion (adagrasib + cetuximab): ORR 46%.
• FDA approved both as combinations in CRC in 2024.

The takeaway: “the same KRAS G12C mutation produces different resistance landscapes by tumor type”.

4. Front-Line Advance — KRYSTAL-7

NSCLC 1L progress centers on immune checkpoint inhibitor (ICI) combinations:

KRYSTAL-7 (Phase 2, BMS):

• Treatment-naïve advanced NSCLC (KRAS G12C+) given adagrasib + pembrolizumab.
• PD-L1 TPS ≥50%: confirmed ORR 63%; PD-L1 TPS 1-49%: 49%; PD-L1 <1%: 36%.
• This exceeds historical ICI monotherapy and ICI + chemo benchmarks.
• Toxicity manageable, but adagrasib + ICI hepatic toxicity warrants monitoring.

Sotorasib + ICI is in parallel Phase 2-3 development with similarly promising readouts.

5. Other Combinations — SHP2, SOS1, Chemotherapy

• SHP2 inhibitors: SHP2 (PTPN11) is the RTK adapter. RMC-4630 (Revolution Medicines), BBP-398, JAB-3068 in development.
• SOS1 inhibitors: SOS1 is the KRAS GEF. BI 1701963 (Boehringer Ingelheim) and similar.
• Chemotherapy combinations: cisplatin + pemetrexed in CodeBreaK 202, etc.
• MEK inhibitors (trametinib): downstream MEK blockade — toxicity management is the challenge.
• Radiation: radiosensitization, oligometastatic settings.

6. Biomarkers — ctDNA and Molecular Screening

• NGS: comprehensive tumor profiling (KRAS, NRAS, BRAF, EGFR, MET, ALK, ROS1, etc.).
• ctDNA / liquid biopsy: detection at treatment initiation and serial resistance monitoring.
• PD-L1 IHC: ICI combination triage.
• MSI / TMB: especially for CRC.
• Resistance profiling: post-progression rebiopsy / liquid biopsy to identify switch II mutations or bypass pathways and guide next-line choice.

7. Adaptation Beyond Lung and Colon

• PDAC: ~90% KRAS-mutant, but only ~1-2% are G12C (predominantly G12D/G12V). G12C inhibitor activity in PDAC is limited.
• Biliary tract: small-percent G12C, basket trials.
• Endometrial, ovarian, esophageal: limited cases, identified via NGS, aggregated into basket trials.

8. Safety — Hepatic, QT, and Drug-Drug Interactions

• Sotorasib: AST/ALT elevation, exacerbated by ICI combinations. Hold/reduce protocols required.
• Adagrasib: QT prolongation, strong CYP3A interactions. ECG monitoring and concomitant medication review essential.
• GI adverse events: common, mostly manageable, occasional dose modification.
• CRC + EGFR antibody: dermatologic toxicity (acneiform rash, paronychia) management is critical.

My Thoughts and Outlook

KRAS G12C inhibitors opened the “undruggable” door, but single-agent activity has limits and combinations are the next battlefield. By 2026, three frameworks have crystallized: NSCLC 1L with ICI, CRC with anti-EGFR antibodies, and PDAC with chemotherapy backbones.

First, NSCLC ICI combinations like KRYSTAL-7 may redefine first-line standards. PD-L1-high ORR 63% likely surpasses pembro + chemo (~50-55%). Safety, optimal sequencing, and subgroup stratification still need work.

Second, CRC absolutely requires anti-EGFR partnership, illustrating context dependence — a universal lesson likely to apply to G12D, G12V, and pan-KRAS programs.

Third, resistance sequencing remains unsettled as of 2026: SHP2/SOS1 combos, next-gen G12C inhibitors, and immunotherapy switches are all candidates without a definitive winner.

Fourth, the G12D, G12V, and pan-KRAS frontier — including Revolution Medicines’ zoldonrasib and daraxonrasib — opens a new chapter on top of the G12C framework, covered in Vol.2.

Beginner’s Perspective

“KRAS” is one of the most important cancer-related genes. About 25-30% of all cancers carry mutations in it. For a long time it was considered “undruggable” — meaning no medicine could target it directly.

In 2021, that finally changed: a drug called sotorasib (Lumakras) got FDA approval to target a specific KRAS mutation called G12C. Adagrasib (Krazati) followed in 2022. These were historic moments — cancers thought to be unreachable with targeted therapy could finally be attacked.

The catch: resistance shows up fast. Within months, cancer cells learn ways around the drug. So researchers are now combining these drugs with others. In one important trial (KRYSTAL-7), the combination with an immunotherapy drug shrank tumors in 63% of patients with high PD-L1 expression.

Vol.2 looks at the next frontier — G12D and pan-KRAS drugs that are reshaping pancreatic cancer treatment. Vol.3 maps the full competitive landscape.

Science Writer’s View

Sotorasib (Lumakras) and adagrasib (Krazati) broke the KRAS undruggability paradigm with FDA approvals in 2021-22. NSCLC 2L ORRs run 30-40% with median PFS 5-7 months. Resistance emerges quickly: switch II pocket mutations (R68S, H95, Y96C) and EGFR/RTK reactivation dominate. CRC monotherapy is weak (~10% ORR), so anti-EGFR antibody partnerships (panitumumab, cetuximab) became required — established in CodeBreaK 300 and KRYSTAL-1 expansion, with FDA combo approvals in 2024. 1L strategy hinges on ICI combos: KRYSTAL-7 PD-L1 ≥50% ORR 63% may shift the standard. SHP2, SOS1, chemo, MEK, and radiation combinations are progressing in Phase 2-3. ctDNA / liquid biopsy enables longitudinal resistance monitoring central to precision implementation. G12D and pan-KRAS expansion is detailed in Vol.2.

Expert Perspective

Sotorasib (AMG-510, Amgen, FDA 2021/05/28) and adagrasib (MRTX849, Mirati→BMS, FDA 2022/12/12) achieve switch II covalent inhibition by acrylamide warhead binding to the KRAS G12C cysteine (Shokat lab 2013-14). NSCLC 2L single-agent ORR: 37.1% (CodeBreaK 100, sotorasib) and 42.9% (KRYSTAL-1, adagrasib); median PFS 5-7 months; OS 12-13 months. CRC requires anti-EGFR antibody co-administration due to EGFR feedback reactivation under KRAS G12C inhibition — confirmed in CodeBreaK 300 (NEJM 2024, sotorasib + panitumumab; ORR 26%, PFS 5.6 months) and KRYSTAL-1 expansion (adagrasib + cetuximab; ORR 46%); FDA combo approvals 2024. KRYSTAL-7 (adagrasib + pembrolizumab, 1L NSCLC): PD-L1 TPS ≥50% confirmed ORR 63%, exceeding pembro + chemo SOC ORR ~50-55%. Major resistance mechanisms: (1) on-target switch II mutations (R68S, H95D/Q/R, Y96C; Awad et al., NEJM 2021); (2) EGFR / RTK / SHP2 reactivation; (3) MET amplification; (4) NRAS/BRAF/FGFR3 bypass; (5) drug-tolerant persister phenotype; (6) EMT. SHP2 (RMC-4630, BBP-398, JAB-3068), SOS1 (BI 1701963), MEK (trametinib), and chemo combinations are in Phase 2-3. Liquid biopsy / ctDNA (Guardant360 CDx, etc.) implement longitudinal molecular surveillance. G12D (zoldonrasib RMC-9805), pan-RAS (daraxonrasib RMC-6236), and KRAS degraders are Vol.2 topics.