KRAS mutation subtypes refine prognostic stratification in colorectal cancer: a retrospective cohort study

Introduction

Colorectal cancer (CRC) is one of the leading causes of cancer-related mortality worldwide, ranking third in global incidence and second in China according to the Global Cancer Statistics 2022 report by the International Agency for Research on Cancer (IARC) (1). Despite advances in screening and systemic therapy, CRC remains a major cause of cancer-related mortality worldwide and imposes a substantial disease burden in China.

CRC is characterized by its heterogeneity at both cellular and molecular levels, with a plethora of genetic alterations and signaling pathways contributing to its pathogenesis (2). Among these, mutations in the Kirsten rat sarcoma viral oncogene homolog (KRAS) are particularly prevalent, occurring in approximately 40% of CRC cases (3). The KRAS gene, located on chromosome 12, plays a crucial role in cell signaling, acting as a molecular switch in the RAS/RAF/MEK/ERK pathway and the epidermal growth factor receptor (EGFR) signaling transduction process (4-6). Mutations in KRAS, especially at codons 12 and 13, have been associated with altered responses to anti-EGFR therapy and variable prognostic outcomes (7,8).

Despite the recognized significance of KRAS mutations in CRC, the prognostic implications of different mutation subtypes remain uncertain. While some studies suggest a poorer prognosis for patients with KRAS-mutant CRC, others report no significant association between KRAS mutations and patient outcomes. Moreover, the impact of specific KRAS mutation subtypes, such as those occurring in codons 12 and 13, on survival and treatment response is still under investigation (9-11). Notably, mutations in less common sites such as Q61, A146, K117, and A59 are less well-studied due to their lower prevalence (3).

Given the critical role of KRAS mutations in CRC pathogenesis and their potential influence on therapeutic decisions, this study aims to explore the clinical and pathological characteristics and the unique prognostic significance of various KRAS mutation subtypes in CRC patients, thereby providing a new theoretical basis for personalized treatment. By exploring the association between specific KRAS mutations and patient outcomes, we seek to provide valuable insights for personalized treatment strategies and improve the overall management of CRC. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0046/rc).

Methods

Study design

This retrospective cohort study included patients diagnosed with CRC who underwent surgical resection or endoscopic biopsy at the Department of Oncology Surgery, the First Affiliated Hospital of Kunming Medical University, between January 2017 and October 2023. Patients were eligible for inclusion if they underwent next-generation sequencing (NGS) for KRAS gene analysis. Of the 515 CRC patients who underwent KRAS testing, 172 were excluded due to incomplete clinical data or loss to follow-up, resulting in a final cohort of 343 patients. Data were collected from the hospital’s information management system and included demographic details, clinical and pathological information, and follow-up data. This study was conducted in accordance with the Helsinki Declaration and its subsequent amendments. This study was approved by the Ethics Committee of The First Affiliated Hospital of Kunming Medical University (No. 2024-L16). Written informed consent was obtained from all patients. A CONSORT diagram describing patient selection, exclusions, stage distribution, and the final analysis sets is provided as a PowerPoint file (Figure S1: CONSORT diagram).

Inclusion and exclusion criteria

Inclusion criteria included: (I) histopathologically confirmed primary CRC from surgical resection or endoscopic biopsy; (II) available tumor tissue for targeted NGS and interpretable RAS/RAF profiling; patients were categorized as KRAS-mutant or RAS/RAF wild-type (no KRAS/NRAS/HRAS/BRAF mutations); (III) no severe infectious diseases or systemic stress reactions at the time of initial diagnosis; (IV) complete clinical and follow-up data.

Exclusion criteria included: (I) non-primary CRC patients; (II) patients who received anti-tumor therapy before the first visit; (III) familial adenomatous polyposis and Lynch syndrome; (IV) patients with infectious diseases or systemic stress reactions at initial hospital admission.

Follow-up

Follow-up data, including clinic visits, survival status, and cause of death, were collected through the hospital’s information management system and telephone follow-ups. The survival time refers to the period from the patient’s diagnosis date to the end of the follow-up period or the date of death.

KRAS mutation analysis

RAS/BRAF mutation testing was performed on tumor tissue samples from CRC patients treated at the First Affiliated Hospital of Kunming Medical University to ensure accurate interpretation of KRAS mutation results. The detection method involved initially extracting DNA from fresh tumor tissue samples using the TIANamp Genomic DNA Kit (TIANGEN, China). Subsequently, the DNA was sheared to a length of 200 bp using Covaris LE220, and libraries were constructed using the KAPA Hyper Preparation Kit (Kapa Biosystems, USA) and target regions were captured using the HyperCap Target Enrichment Kit (Roche, Switzerland). Processed BAM files were analyzed using quality control, reference maps, duplicate screening, and rearrangement processes.

Statistical analysis

Continuous variables are presented as mean ± standard deviation or median (P25, P75) and compared using the independent-samples t-test or Mann-Whitney U test, as appropriate. Categorical variables are presented as n (%) and compared using the χ² test or Fisher’s exact test. Overall survival (OS) was estimated using the Kaplan-Meier method and compared by the log-rank test. Variables with clinical relevance and/or P<0.10 in univariable Cox analysis were entered into the multivariable Cox model. All tests were two-sided, and P<0.05 was considered statistically significant. Analyses were performed using IBM SPSS Statistics v27; figures were generated with GraphPad Prism v10.0. To address potential molecular confounding, we additionally performed a sensitivity Cox analysis incorporating major co-mutations (TP53, APC, PIK3CA, and BRAF) in the subset of patients with available co-mutation information; Because baseline prognosis differs substantially across tumor-node-metastasis (TNM) stages, survival analyses were performed stratified by stage (stage I–III vs. stage IV), and an interaction term was tested to assess potential stage-dependent effects of KRAS exon 2.

Results

Frequency of KRAS mutations and association with clinicopathological factors

The total study population consisted of 343 patients in whom adequate retrospective data could be obtained were further assessed for clinicopathological and survival analysis. Table 1 summarizes the baseline clinicopathological characteristics of these patients stratified by KRAS mutation status. KRAS mutations were identified in 67.3% (231 cases) of samples, The most common mutations occurred at codon 12 (41.7%) and codon 13 (13.7%) of exon 2, including mutations such as G12D, G12V, G12C, G12A, G12S, G12R, G13D, and G13C. G12D was the most frequent mutation at 20.7%, followed by G13D at 13.1%, and G12V at 10.2% (Table 2). These findings indicate a high prevalence of KRAS mutations, particularly in codons 12 and 13, among CRC patients. The relatively high KRAS mutation prevalence in this cohort may be related to patient selection for NGS testing and regional population characteristics.

The majority of patients were male (58.9%), and 53.4% of the patients were over 60 years old. Most tumors (75.7%) were located in the left colon, and 78.1% of tumors were moderately differentiated. According to TNM staging, 34.7% of patients were in stage I or II, 39.1% in stage III, and 26.2% in stage IV. Microsatellite stability (MSS) was observed in 87.5% of patients, and 53.1% had abnormal initial carcinoembryonic antigen (CEA) levels. The majority of patients (89.5%) underwent curative surgical resection, while 79.3% received chemotherapy. Comparison of clinical and pathological features among CRC patients with different KRAS mutation subtypes revealed notable differences. Patients with G12D mutations were more likely to have tumors located in the left colon, while those with G13D mutations had a higher prevalence of right colon tumors, suggesting that KRAS mutation subtypes may be associated with tumor location, further impacting patient prognosis. Additionally, poorly differentiated tumors were more common in patients with G12A mutations. These comparisons underscore the heterogeneity within KRAS mutation subtypes and their distinct clinical and pathological characteristics (Table 1).

Prognostic factors for OS

Univariate Cox regression analysis identified body mass index level, tumor differentiation degree, TNM stage, microsatellite status, initial CEA level, curative surgical resection, chemotherapy, and exon 2 mutation as prognostic factors for CRC patients (Table S1). In multivariable Cox regression, poor differentiation and TNM stage IV were independently associated with worse OS, whereas curative resection was associated with improved OS. KRAS exon 2 mutation showed a significant association with OS in the primary model (Table 3). These results underscore the importance of tumor differentiation, TNM stage, and KRAS exon 2 mutations in determining patient outcomes.

Stage-stratified survival analysis

Given the markedly different baseline prognosis across TNM stages, we performed stage-stratified survival analyses. In stage I–III disease, KRAS exon 2 status was not significantly associated with OS [hazard ratio (HR) 1.00, 95% confidence interval (CI): 0.49–2.05; P>0.99]. In stage IV disease, KRAS exon 2 status was also not significantly associated with OS (HR 0.94, 95% CI: 0.52–1.68; P=0.82) (Table 4). A formal interaction test (stage IV vs. I–III by KRAS exon 2) did not indicate effect modification (interaction P=0.75) (Table S2).

Survival analysis

Kaplan-Meier survival curves demonstrated that CRC patients with KRAS mutations had significantly poorer OS compared to those with wild-type KRAS (P=0.01) (Figure 1). Specifically, the median OS for KRAS mutant patients was 24.2 months, compared to 34.7 months for KRAS wild-type patients. Further analysis of specific KRAS mutation subtypes revealed that patients with G12A and K117N mutations had the worst prognosis, with median OS of 18.3 and 19.7 months, respectively. Conversely, patients with G12V and G13D mutations had relatively better survival outcomes, with median OS of 27.5 and 28.9 months, respectively (Figure 2). These findings highlight the significant impact of specific KRAS mutation subtypes on the prognosis of CRC patients. In a sensitivity analysis additionally adjusting for major co-mutations (TP53, APC, PIK3CA, and BRAF) within the complete-case subset with available co-mutation data, KRAS exon 2 status was not independently associated with OS (Table S3).

Figure 1 Kaplan-Meier survival curves for KRAS mutant and wild-type CRC patients. The Kaplan-Meier survival analysis compares overall survival between KRAS mutant and wild-type CRC patients. The log-rank test P value indicates statistical significance. CRC, colorectal cancer; KRAS, Kirsten rat sarcoma viral oncogene homolog; WT, wild type.

Figure 2 Kaplan-Meier survival analysis of overall survival among different KRAS mutation subtypes. The Kaplan-Meier survival curves depict overall survival for CRC patients with different KRAS mutation subtypes. Sub-figures (A-J) illustrate Kaplan-Meier survival comparisons between selected KRAS mutation subtypes and the wild-type reference group. CRC, colorectal cancer; KRAS, Kirsten rat sarcoma viral oncogene homolog.

Single-factor survival analysis

Single-factor Cox regression analysis of different mutation subtypes in KRAS gene exons 2, 3, and 4 indicated that several specific mutations were significantly associated with OS. Notably, G12A and K117N mutations were identified as high-risk factors. This further underscores the prognostic value of specific KRAS mutation subtypes in CRC (Table 5).

Multivariable Cox regression analysis

Multivariable Cox regression analysis confirmed that specific KRAS mutation subtypes independently influenced OS in CRC patients. The analysis revealed that G12A and K117N mutations were associated with significantly poorer survival outcomes. This highlights the critical need for subtype-specific prognosis and tailored therapeutic strategies in CRC management (Table 6).

Discussion

Our study provides a comprehensive analysis of the clinical and pathological characteristics and prognostic implications of various KRAS mutation subtypes in CRC patients. By focusing on a large cohort of 343 CRC patients, our findings offer valuable insights into the heterogeneity of KRAS mutations and their significant impact on patient outcomes.

The prevalence of KRAS mutations in our cohort was notably high at 67.3%, with G12D, G13D, and G12V being the most common subtypes. This high mutation rate, higher than the commonly reported 30–50%, may reflect regional genetic variability or specific characteristics of our patient population (3,10,12-15). This discrepancy may reflect selection bias, as patients undergoing NGS testing in a tertiary referral center often have more advanced or high-risk disease. Additionally, regional genetic differences and the high sensitivity of NGS may contribute to this finding. These factors limit the generalizability of our results and underscore the need for validation in multicenter, population-based cohorts. Notably, KRAS mutations were significantly associated with gender, MSS, and initial CEA levels. These associations suggest that KRAS mutations could be linked to specific biological pathways and tumor behaviors that differ by gender and microsatellite status, warranting further molecular investigations (16).

Our initial multivariate analysis suggested that KRAS exon 2 mutations were associated with worse OS. However, this association was attenuated in the co-mutation–adjusted sensitivity analysis, suggesting that the prognostic signal may partly reflect the broader molecular context of CRC. This aligns with previous studies that have highlighted the aggressive nature of poorly differentiated tumors and advanced-stage CRC (17). Pooling patients across TNM stages may obscure biomarker effects because stage is the primary determinant of survival. To address this concern, we emphasized stage-stratified analyses and interpret KRAS exon/subtype signals within stage-defined clinical contexts. Our stage-specific estimates—particularly in stage III—remain limited by event counts and require validation in larger cohorts with comprehensive co-mutation profiling.

However, our study adds a nuanced understanding by emphasizing the distinct prognostic roles of specific KRAS mutation subtypes. For instance, patients with G12A and K117N mutations exhibited the worst OS, underscoring the heterogeneity within KRAS mutations and the necessity for subtype-specific prognostic assessments (18). The identification of specific KRAS mutation subtypes with poorer prognoses has significant clinical implications. These findings may help refine prognostic risk stratification. However, because we did not analyze subtype-specific treatment response, the direct therapeutic implications of these observations remain uncertain. For example, patients with G12A or K117N mutations may warrant closer clinical monitoring and risk-adapted management strategies (19). This biological basis underlying the prognostic heterogeneity among KRAS subtypes remains incompletely understood. Different KRAS mutations may differ in intrinsic GTPase activity, downstream effector engagement, signaling amplitude, and interaction with the tumor microenvironment, thereby producing distinct oncogenic phenotypes. It is possible that rare variants such as G12A and K117N confer more aggressive biological behavior in certain molecular contexts; however, this hypothesis cannot be tested in the present clinical dataset. Accordingly, the adverse survival associations observed for these subtypes should be interpreted as clinical signals that warrant further mechanistic investigation in translational and preclinical studies. While G12A and K117N mutations emerged as potential high-risk subtypes in our cohort, the small number of cases limits the robustness of these findings. These results should be considered exploratory and hypothesis-generating. Future studies with larger sample sizes and functional validation are needed to confirm their prognostic role and uncover underlying biological mechanisms. Additionally, our findings suggest that KRAS mutation status should be a critical factor in stratifying patients for targeted therapies, particularly those involving EGFR inhibitors, which are known to be less effective in KRAS-mutant tumors (3,16).

While our study provides substantial insights, it also highlights several areas for future research. The biological mechanisms underlying the distinct behaviors of different KRAS mutation subtypes need to be elucidated (16). Functional studies exploring how these mutations interact with other genetic and epigenetic alterations could provide deeper understanding and reveal potential therapeutic targets (17). Moreover, expanding this research to multi-center studies with more diverse populations could validate our findings and address the potential single-center bias inherent in our study.

There are some limitations in this study that should be acknowledged. As a retrospective single-center study, our findings are subject to potential selection bias and unmeasured confounding. The external validity of our results is limited, and they should be validated in prospective, multicenter cohorts with larger and more diverse populations. Secondly, the rarity of certain KRAS mutation subtypes, such as G12A and K117N, precludes definitive prognostic conclusions and underscores the need for multi-center collaboration to pool data on uncommon mutations. Third, co-mutation information was not uniformly available for all patients and was limited to a predefined set of genes, which may reduce power and does not capture broader molecular context such as additional driver alterations or copy-number changes. Therefore, although we performed a co-mutation-adjusted sensitivity analysis, residual molecular confounding cannot be excluded and larger, uniformly profiled cohorts are warranted. Additionally, we did not perform subtype-specific analyses of systemic therapy response, including chemotherapy regimens, anti-angiogenic therapy, or anti-EGFR treatment. This limits the immediate clinical translatability of our findings for personalized treatment decision-making. Finally, after stratification by stage, the number of events in stage-defined subsets—particularly stage IV (n=90)—was limited, resulting in wider CIs and reduced power to detect moderate associations. Therefore, stage-specific estimates should be interpreted cautiously and considered hypothesis-generating pending validation in larger, uniformly profiled cohorts.

Conclusions

In conclusion, our study demonstrates that KRAS mutations in CRC are associated with distinct clinical and pathological characteristics and have significant prognostic implications, particularly for specific subtypes. These findings emphasize the importance of KRAS mutation subtyping in the personalized management and treatment of CRC patients. By integrating KRAS mutation subtype information into prognostic assessment, clinicians may improve risk stratification. Future multi-center studies will help validate these findings and explore the clinical applications of different KRAS mutation subtypes in broader patient populations.

Acknowledgments

We gratefully acknowledge patients who volunteered to participate in the study and agreed to provide research biopsies.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0046/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0046/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0046/prf

Funding: This work was supported by the 535 Talent Project of First Affiliated Hospital of Kunming Medical University (No. 2022535D07), National Natural Science Foundation of China (No. 82160533), The Yunnan Revitalization Talent Support Program (Nos. RLQB20200004 and RLMY20220013), and Education Teaching Research Project of The First Affiliated Hospital of Kunming Medical University (No. 2024-JY-19).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0046/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of The First Affiliated Hospital of Kunming Medical University (No. 2024-L16). Written informed consent was obtained from all patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Sagaert X, Vanstapel A, Verbeek S. Tumor Heterogeneity in Colorectal Cancer: What Do We Know So Far? Pathobiology 2018;85:72-84. [Crossref] [PubMed]
3. Takeda M, Yoshida S, Inoue T, et al. The Role of KRAS Mutations in Colorectal Cancer: Biological Insights, Clinical Implications, and Future Therapeutic Perspectives. Cancers (Basel) 2025;17:428. [Crossref] [PubMed]
4. Wang Y, Bui TA, Yang X, et al. Advancements in gene therapies targeting mutant KRAS in cancers. Cancer Metastasis Rev 2025;44:24. [Crossref] [PubMed]
5. Choucair K, Imtiaz H, Uddin MH, et al. Targeting KRAS mutations: orchestrating cancer evolution and therapeutic challenges. Signal Transduct Target Ther 2025;10:385. [Crossref] [PubMed]
6. González NS, Marchese PV, Baraibar I, et al. Epidermal growth factor receptor antagonists in colorectal cancer: emerging strategies for precision therapy. Expert Opin Investig Drugs 2024;33:613-25. [Crossref] [PubMed]
7. Simanshu DK, Nissley DV, McCormick F. RAS Proteins and Their Regulators in Human Disease. Cell 2017;170:17-33. [Crossref] [PubMed]
8. Domingo E, Camps C, Kaisaki PJ, et al. Mutation burden and other molecular markers of prognosis in colorectal cancer treated with curative intent: results from the QUASAR 2 clinical trial and an Australian community-based series. Lancet Gastroenterol Hepatol 2018;3:635-43. [Crossref] [PubMed]
9. Benmokhtar S, Laraqui A, El Boukhrissi F, et al. Clinical Significance of Somatic Mutations in RAS/RAF/MAPK Signaling Pathway in Moroccan and North African Colorectal Cancer Patients. Asian Pac J Cancer Prev 2022;23:3725-33. [Crossref] [PubMed]
10. Lin Z, Liu Y, Cai S, et al. Not All Kirsten Rat Sarcoma Viral Oncogene Homolog Mutations Predict Poor Survival in Patients With Unresectable Colorectal Liver Metastasis. Technol Cancer Res Treat 2021;20:15330338211039131. [Crossref] [PubMed]
11. Yuan Y, Liu Y, Wu Y, et al. Clinical characteristics and prognostic value of the KRAS mutation in Chinese colorectal cancer patients. Int J Biol Markers 2021;36:33-9. [Crossref] [PubMed]
12. Benmokhtar S, Laraqui A, Hilali F, et al. RAS/RAF/MAPK Pathway Mutations as Predictive Biomarkers in Middle Eastern Colorectal Cancer: A Systematic Review. Clin Med Insights Oncol 2024;18:11795549241255651. [Crossref] [PubMed]
13. Shi R, Cheng Y, Wang J, et al. Genetic feature diversity of KRAS-mutated colorectal cancer and the negative association of DNA mismatch repair deficiency relevant mutational signatures with prognosis. Genes Dis 2025;12:101245. [Crossref] [PubMed]
14. Ilhan N, Dane F, Goker E, et al. Regional and Gender-Based Distribution of KRAS Mutations in Metastatic Colorectal Cancer Patients in Turkey: An Observational Study. Medicina (Kaunas) 2025;61:694. [Crossref] [PubMed]
15. Boilève A, Smolenschi C, Lambert A, et al. KRAS, a New Target for Precision Medicine in Colorectal Cancer? Cancers (Basel) 2024;16:3455. [Crossref] [PubMed]
16. Ros J, Vaghi C, Baraibar I, et al. Targeting KRAS G12C Mutation in Colorectal Cancer, A Review: New Arrows in the Quiver. Int J Mol Sci 2024;25:3304. [Crossref] [PubMed]
17. Tria SM, Burge ME, Whitehall VLJ. The Therapeutic Landscape for KRAS-Mutated Colorectal Cancers. Cancers (Basel) 2023;15:2375. [Crossref] [PubMed]
18. Ottaiano A, Sabbatino F, Perri F, et al. KRAS p.G12C Mutation in Metastatic Colorectal Cancer: Prognostic Implications and Advancements in Targeted Therapies. Cancers (Basel) 2023;15:3579. [Crossref] [PubMed]
19. van 't Erve I, Wesdorp NJ, Medina JE, et al. Abstract 519: Clinical impact of KRASG12, G13, Q61, K117 and A146 mutations in patients with colorectal liver metastases. Cancer Res 2022;82:519.