Gene

KRAS

Kirsten Rat Sarcoma viral oncogene

PredictivePrognosticDiagnosticStrong evidenceLast reviewed: July 12, 2026

Overview

KRAS encodes a membrane-associated small GTPase that acts as a molecular switch downstream of receptor tyrosine kinases. Activating variants reduce GTP hydrolysis or favor the active state, causing persistent signaling through MAPK and PI3K pathways. [K1-K3]

KRAS testing has two major clinical uses. In metastatic colorectal cancer, any activating RAS alteration generally predicts lack of benefit from anti-EGFR antibody monotherapy or standard anti-EGFR combinations. In NSCLC and colorectal cancer, KRAS p.G12C can identify eligibility for direct KRAS G12C inhibitors under defined labels. [K2-K6]

At a glance

Biomarker type
Oncogenic driver gene; allele-specific predictive and resistance biomarker
Primary roles
Molecular classification; negative predictor for anti-EGFR therapy in CRC; direct-target eligibility for selected variants
Core specimens
Tumor tissue, cytology, and plasma cell-free DNA
Core methods
NGS, allele-specific PCR, digital PCR
Critical caveat
KRAS variants are not one disease: G12C, G12D, G12V, G13D, Q61, and other alleles have different biology and evidence

Biological function

KRAS cycles between GDP-bound inactive and GTP-bound active forms. Guanine-nucleotide exchange factors activate it, while intrinsic and GAP-assisted GTP hydrolysis switches it off. Active KRAS signals through RAF-MEK-ERK, PI3K-AKT, RAL-GDS, and other effectors to regulate proliferation, survival, metabolism, motility, and differentiation. [K1]

Oncogenic substitutions cluster at codons 12, 13, and 61 and impair normal switching. The amino-acid change affects nucleotide cycling, effector preference, tissue distribution, and drug susceptibility. Therefore, the exact allele must be reported. [K1, K7-K11]

Associated cancers

KRAS alterations are common drivers in pancreatic ductal adenocarcinoma, colorectal adenocarcinoma, and lung adenocarcinoma. They also occur in low-grade serous ovarian cancer, endometrial cancer, biliary cancers, appendiceal tumors, and other malignancies. [K1, K3]

The dominant alleles differ by cancer: G12C is a major actionable subtype in NSCLC and a smaller colorectal subgroup, while G12D, G12V, and G12R are prominent in pancreatic cancer. Clinical evidence must be allele and disease specific. [K1, K7-K11]

Diagnostic role

KRAS is usually not a stand-alone diagnostic test for malignancy. Histopathology establishes the diagnosis. KRAS contributes to molecular subclassification and can support the interpretation of selected lesions when combined with morphology and other markers. [K1-K3]

In metastatic colorectal cancer and advanced non-squamous NSCLC, KRAS is part of standard or broad molecular profiling because it changes treatment selection. Its presence may reduce the likelihood of another mutually exclusive dominant driver, but co-alterations still occur. [K2-K3]

Prognostic role

KRAS-mutant cancers can have different outcomes from KRAS-wild-type cancers, but the effect depends on cancer, stage, allele, co-mutations, treatment, and resectability. Some NSCLC co-mutation patterns, such as alterations in STK11 or KEAP1, can be more informative than KRAS alone. [K7-K8]

In colorectal cancer, KRAS can be associated with adverse features or recurrence in some settings, but it should not be used alone for an individual prognosis. Its validated negative-predictive role for anti-EGFR therapy is clearer. [K2-K3]

Predictive role

Colorectal cancer: activating KRAS or NRAS variants predict lack of benefit from standard anti-EGFR antibody therapy. Extended RAS testing includes specified exons and codons in both KRAS and NRAS, not only KRAS exon 2. [K2-K3]

KRAS G12C NSCLC: sotorasib and adagrasib have U.S. indications after prior systemic therapy. Testing must identify the G12C allele with an authorized test. [K6-K8]

KRAS G12C colorectal cancer: EGFR feedback limits activity of G12C inhibition alone. U.S. approvals pair adagrasib with cetuximab and sotorasib with panitumumab after specified prior chemotherapy. [K4-K5, K9-K10]

Other KRAS alleles: direct inhibitors, pan-KRAS inhibitors, degraders, vaccines, and T-cell approaches remain investigational unless a region-specific approval exists. [K1]

Monitoring role

Plasma ctDNA can detect KRAS variants when tissue is unavailable and can track clonal response or emerging resistance in advanced disease. A negative plasma result may require tissue testing because low-shedding tumors can be missed. [K6]

Postoperative ctDNA containing a tumor-specific KRAS variant can indicate molecular residual disease in research and selected clinical programs, but the testing interval and treatment response are not globally standardized. KRAS should not be used as a general population screening blood test. [K1-K3]

Common test methods

Allele-specific real-time PCR: rapid detection of defined KRAS variants, including FDA-authorized companion tests for G12C. [K4, K6]

NGS panels: assess full clinically relevant KRAS regions together with NRAS, BRAF, EGFR, and other genes; useful for uncommon alleles and co-alterations. [K2-K3]

Digital PCR: sensitive quantification of a known allele in tissue or plasma, useful for research and selected monitoring applications.

Sanger sequencing: can detect variants in high-tumor-content specimens but has lower sensitivity than many modern methods.

Preanalytic factors include tumor fraction, specimen age, fixation, decalcification, DNA quality, and plasma volume. [K2-K3]

How results may be reported

The report should provide KRAS, transcript, HGVS nucleotide and protein notation, codon/exon, variant allele fraction, specimen, assay, coverage, limit of detection, and classification. A useful result is "KRAS p.G12C detected," not simply "KRAS positive." [K2-K3]

Colorectal reports should state whether extended RAS testing was complete and whether NRAS was also assessed. If no pathogenic RAS alteration is found, "RAS wild type in tested regions" is more accurate than "KRAS negative." [K2]

Plasma reports should warn that absence of a detected variant does not assure tumor negativity and may recommend tissue testing. [K6]

General interpretation

A pathogenic activating KRAS variant indicates constitutive RAS-pathway signaling. It can explain resistance to upstream EGFR blockade in colorectal cancer and may identify direct-therapy eligibility if the allele and disease match an approval. [K2-K5]

KRAS wild type is not itself a complete treatment recommendation. In colorectal cancer, NRAS, BRAF, tumor sidedness, HER2, MSI/MMR, and other factors also matter. In lung cancer, broad testing is needed because other drivers may be actionable. [K2-K3]

Variant allele fraction is influenced by tumor purity, copy number, and clonality; it is not a direct measure of tumor burden across different specimens. [K2]

Limitations

Hotspot-only assays may miss uncommon but clinically relevant variants. NGS pipelines may vary in coverage and detection of low-frequency alleles. [K2-K3]

Plasma false negatives occur with low shedding; false interpretation can occur if a small subclone is overgeneralized to the entire tumor. Clonal hematopoiesis can occasionally complicate plasma findings and should be assessed in context. [K6]

KRAS actionability changes rapidly and remains allele specific. Evidence from G12C cannot be automatically applied to G12D or other variants. Drug access, line of therapy, and companion-test requirements differ by region. [K4-K6]

Resistance mechanisms

On-target secondary changes: substitutions in the switch-II pocket or other KRAS regions can reduce inhibitor binding; amplification of the KRAS G12C allele can increase signaling. [K11]

New RAS-pathway alterations: acquired NRAS, BRAF, MAP2K1, or other pathway changes can reactivate MAPK signaling. [K11]

Upstream feedback: EGFR and other receptor-tyrosine-kinase signaling can reactivate wild-type RAS, particularly in colorectal cancer, explaining the benefit of combining a G12C inhibitor with an anti-EGFR antibody. [K9-K11]

Parallel pathways and phenotypic change: PI3K activation, fusions, histologic transformation, and tumor-cell-state adaptation can contribute. Multiple mechanisms may coexist, so progression testing can require both tissue and plasma. [K11]

Relevant drug classes

Covalent KRAS G12C inhibitors: sotorasib and adagrasib bind the inactive GDP-bound G12C protein. [K4-K8]

KRAS G12C plus anti-EGFR antibodies: sotorasib-panitumumab and adagrasib-cetuximab are approved in specified previously treated metastatic colorectal-cancer settings. [K4-K5]

MAPK-pathway combinations: MEK, RAF, SHP2, or SOS1 inhibitors are under investigation; toxicity and adaptive feedback have limited broad use.

Non-G12C direct strategies: G12D inhibitors, pan-KRAS active-state inhibitors, degraders, vaccines, and adoptive T-cell approaches are active research areas but should be labeled investigational unless approved.

KRAS-mutated low-grade serous ovarian cancer: the U.S. FDA granted accelerated approval in 2025 to avutometinib plus defactinib after prior systemic therapy, targeting RAF/MEK-clamp and FAK pathways rather than KRAS directly. [K12]

Latest research and regulatory developments

Colorectal G12C combinations became standard regulatory options. The U.S. FDA approved sotorasib plus panitumumab in January 2025 after fluoropyrimidine, oxaliplatin, and irinotecan; adagrasib plus cetuximab received accelerated approval in June 2024 for a similar previously treated setting. [K4-K5]

Resistance-directed development is central. Molecular analyses show polyclonal resistance with on-target KRAS changes and bypass-pathway alterations. Current trials test next-generation inhibitors and vertical pathway blockade. [K11]

Non-G12C KRAS is the next frontier. G12D-selective and pan-KRAS inhibitors, RAS-ON compounds, proteolysis strategies, and immune approaches are being studied across pancreatic, colorectal, and lung cancers. The website should not imply approval until a regulator has acted.

Allele and tissue context matter. Research increasingly treats "KRAS-mutant" as a family of biologically distinct diseases, integrating allele, co-mutations, lineage, and immune microenvironment rather than using a single binary category. [K1, K7-K11]

Sources & Review

Regulatory statements describe U.S. FDA indications and should be checked against the current label for the user's country or region.