The role of targetable biomarker alterations in overall survival for non-small cell lung cancer (NSCLC) in a large integrated health system

Introduction

Lung cancer remains the second most common and deadliest malignancy worldwide, with non-small cell lung cancer (NSCLC) comprising approximately 85% of cases, with an estimated 340 people dying daily from it (1). Despite advances in screening and treatment, the prognosis for advanced NSCLC remains poor, particularly in its later stages (2). In recent years, the advent of next-generation sequencing (NGS) has expanded the utility of biomarkers in cancer treatment. These technologies allow for the simultaneous detection of multiple genetic alterations, facilitating more tailored therapeutic approaches. It is currently understood that biomarker status and histology each independently predict survival for a given targeted therapy (3-5). The role of biomarker testing is thus particularly valuable to identify the therapy that can optimize patient outcomes, as several studies have revealed that patients who receive guideline-concordant targeted therapy experience improved survival outcomes (6-9).

Developments in ease of multigene NGS-targeted panels have unveiled multiple types of biomarkers reflective of host environment and microbiome (10,11). While it has been demonstrated that survival in NSCLC differs by stage, a limitation in the current understanding is whether the nature of the biomarker itself dictates survivorship, given that the majority of biomarker-driven therapy is Food and Drug Administration (FDA) approved for stage IV disease and that survivorship is presented as a function of treatment (12,13). As revealed by a recent review by Cao et al., most studies that have looked at survivorship and biomarker status in patients with early-stage NSCLC have been limited to obtaining their histopathology sample from the resected tumor (5). For the few studies that have looked at the relationship of biomarkers collected from plasma and survival, sample sizes have been small, the follow-up period has been limited to a maximum of 1 year, and survival stratification by stage was not completed (14).

Given that activation of specific oncogenic pathways can broadly impact gene expression, we propose that the genetic makeup of cancer cells serves as a defining fingerprint, shaping the immune tumor microenvironment by driving distinct protein-targeted pathways. Herein, our study aims to evaluate and stratify overall survival (OS) in NSCLC at 1 and 3 years by biomarker alteration status. Due to the ADAURA trial not being published during our selected time frame, our study is novel in that survivorship for stages IB–IIIA is reflective of biology alone since patients did not receive targeted therapy regardless of biomarker status. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1043/rc).

Methods

This retrospective cohort study included adults aged 18–89 years diagnosed with NSCLC of any stage between 2013 and 2020 within Kaiser Permanente Northern California (KPNC), a large integrated healthcare delivery system. We used data from electronic health record (EHR) system databases and registries at KPNC. KPNC provides comprehensive medical services to more than 4.5 million enrolled members—nearly one-third of the population in its service area—through 21 hospitals and over 240 outpatient clinics across the San Francisco Bay Area and Central Valley, extending from Sacramento in the north to Fresno in the south. The KPNC membership is relatively stable, sociodemographically diverse, and broadly representative of Northern California residents.

The integrated care environment enables robust examination of the cancer care continuum from diagnosis to survivorship. In addition to comprehensive EHR data, KPNC maintains registries that support research and regulatory reporting. The KPNC Cancer Registry captures information on all patients diagnosed or treated with cancer in accordance with national cancer registry standards. Mortality data were obtained from the KPNC Vital Statistics Registry, which aggregates death information from California state records, the U.S. Social Security Administration, and the National Death Index. Together, these sources provide high-quality, longitudinal data on patient demographics, cancer characteristics, treatments, and survival outcomes.

Oncogenic alteration status was classified based on biomarker testing, and include genetic alterations (e.g., mutation, fusion, rearrangement, amplification) in nine oncogenic genes: epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK), ROS proto-oncogene 1 (ROS1), Kirsten rat sarcoma virus (KRAS), mesenchymal epithelial transition (MET), B-Raf proto-oncogene, serine/threonine kinase (BRAF), erb-b2 receptor tyrosine kinase 2 (ERBB2), rearranged during transfection (RET), and neurotrophic tyrosine receptor kinase (NTRK) genes. Patients were categorized based on the presence or absence of alterations, and OS was evaluated at 1 and 3 years among patients stratified by alteration status. OS was measured from diagnosis date of NSCLC to the date of death or end of follow-up.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics committee of Kaiser Permanente Northern California (No. 1888915-12) and individual consent for this retrospective analysis was waived.

Statistical analysis

Survival analysis was performed using the Kaplan-Meier method. The log-rank test was employed to assess the equality of survival distributions across different strata defined by oncogenic alteration status. Crude and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression to evaluate the association between oncogenic alteration status and OS. Covariates used in adjustment as confounding or precision variables included for age at NSCLC diagnosis (categorized as 18–54, 55–64, 65–74, 75–89 years), sex (male, female), race/ethnicity (Asian, Black, Hispanic, White, other/multiple/unknown race), smoking status (current, former, never), and cancer stage (stages I, II, III, IV, unknown). OS was measured continuously in the Kaplan-Meier and Cox regression analyses, but survival probabilities and associations were reported at the discrete 1- and 3-year landmarks. All statistical analyses were conducted using SAS software, with a significance threshold set at P<0.05.

Results

Demographics

The study included 9,834 patients diagnosed with incident NSCLC between 2013 and 2020 (Table 1). Overall, 38.8% (n=3,820) received biomarker testing. Most testing yielded valid results for the oncogenic alterations examined (77–92%). Of the 3,820 patients that received testing, 3,256 patients had valid results for at least 1 gene. The most commonly tested gene in this cohort was EGFR.

Prevalence

As shown in Table 2, KRAS mutations or amplifications were the most prevalent, found in 26.1% of 1,569 patients tested, closely followed by EGFR (exon 18–21) mutations, found in 25.5% of 3,412 patients tested. MET mutations or amplifications were detected in 3.9% of 1,514 patients tested, while ALK rearrangements were identified in 3.8% of 2,970 patients tested. Both BRAF and ERBB2 mutations or amplifications were observed in 3.1% of 1,479 patients tested for BRAF and 1,464 patients tested for ERBB2. ROS1 rearrangements were found in 2.1% of 2,511 patients tested, and RET rearrangements were present in 1.1% of 1,459 patients tested. The least common alteration was NTRK rearrangements, identified in 0.3% of 799 patients tested. These findings indicate that KRAS and EGFR alterations are the most prevalent genetic alterations in this cohort.

KRAS alteration

At 1 year, patients with KRAS alterations had a lower survival probability (50.4% vs. 60.5%), and at 3 years, this trend continued with KRAS-positive patients exhibiting poorer survival (26.5% vs. 34.3%). Median OS were also shorter for KRAS-positive patients (370 vs. 565 days) (Table 2, Figure 1). After accounting for covariates (Table 3), however, KRAS alteration status showed no clear association with OS at 1 year (HR 1.08; 95% CI: 0.91–1.28) or 3 years (HR 1.05; 95% CI: 0.91–1.21).

Figure 1 Kaplan-Meier survival curves by alteration status for (A) KRAS, (B) EGFR, (C) MET, (D) ALK, (E) ERBB2, (F) BRAF, (G) ROS1, (H) RET, (I) NTRK. Left panels display survival curves with 1 year of follow-up after NSCLC diagnosis; right panels display survival curves within 3 years of follow-up after NSCLC diagnosis. Shaded areas represent 95% confidence intervals. For the last image, the reason why the numbers don’t correlate with the markings is because there were missing values for NTRK, so the sample value is much different from the others (which all shared similar axes). NSCLC, non-small cell lung cancer.

EGFR alteration

Among patients with valid test results for EGFR the presence of mutations in exons 18, 19, 20, or 21 was associated with better survival outcomes (Table 2, Figure 1). At 1 year, patients with EGFR exon 18–21 mutations had a higher survival rate (73.3% vs. 47.3%) compared to those without those mutations). Similarly, patients with EGFR exon 18–21 mutations continued to have superior outcomes at 3 years (37.7% vs. 23.2%) and experienced longer median OS (770 vs. 328 days).

At 1 year, EGFR mutation status was associated with a 50% reduction in the hazard of death (adjusted HR 0.50; 95% CI: 0.43–0.58). At 3 years, the presence of EGFR mutations remained a strong predictor, with a 32% reduction in the hazard of death compared to non-mutated patients (adjusted HR 0.68; 95% CI: 0.61–0.76), further emphasizing the mutation’s impact on long-term survival (Table 3).

ALK alteration

The presence of ALK rearrangements influenced survival outcomes, with ALK-positive patients showing markedly better OS compared to ALK-negative patients at both 1 and 3 years (Table 2, Figure 1). At 1 year, the survival probability was higher in the ALK-positive group (80.3% vs. 53.6%). Similarly, at 3 years, ALK-positive patients maintained a higher survival probability (56.8% vs. 26.4%) and had longer median OS (1,351 vs. 417 days). At 1 year, ALK-positive patients had a lower risk of death (adjusted HR 0.53; 95% CI: 0.35–0.81), and this association persisted at 3 years (adjusted HR 0.54; 95% CI: 0.40–0.73) (Table 3).

ERBB2 alteration

The presence of ERBB2 mutation or amplification was associated with a lower survival probability at both 1 and 3 years (log-rank P=0.03) (Table 2, Figure 1). Among ERBB2-positive patients, the 1-year OS probability was 51.1% and the 3-year OS probability was 19.3%. Cox regression analysis revealed that ERBB2 alteration status was associated with3-year OS only (adjusted HR 1.56; 95% CI: 1.11–2.18) (Table 3).

MET alteration

The presence of MET mutations or amplifications was not associated with OS at 1 and 3 years (log-rank P=0.40). At 1 year, MET-positive patients had a slightly lower survival probability compared to MET-negative patients (51.4% vs. 58.8%), and this difference persisted at 3 years (29.1% vs. 32.8%), with a 3-year median OS of 413 vs. 524 days, respectively (Table 2, Figure 1). In Cox regression analyses, no association was observed between MET alteration status and OS at 1 year (adjusted HR 1.28; 95% CI: 0.93–1.76) or at 3 years (adjusted HR 1.17; 95% CI: 0.89–1.52) (Table 3).

ROS1 alteration

Patients with ROS1 rearrangements had a slightly lower 1-year survival but slightly higher 3-year survival compared to ROS1-negative patients, though neither difference was statistically significant (log-rank P=0.42). For ROS1-positive patients, the 1-year survival probability was 54.3%, while the 3-year survival probability was 36.6% (Table 2, Figure 1). Cox regression analysis showed that ROS1 rearrangement was associated with poorer 1-year OS (adjusted HR 1.59; 95% CI: 1.04–2.41), but not with 3-year OS (adjusted HR 1.19; 95% CI: 0.83–1.70) (Table 3).

BRAF alteration

The presence of BRAF mutations or amplifications did not affect OS at 1 or 3 years (log-rank P=0.38). The 1-year survival probability for BRAF mutation-positive patients was similar to mutation-negative patients (56.6% vs. 58.6%). Similarly, at 3 years, there was no difference in OS by alteration status (35.7% vs. 32.7%) (Table 2, Figure 1). Cox regression analysis confirmed the lack of an association between BRAF alteration status and OS, with adjusted HR estimates close to 1 (Table 3).

RET alteration

The presence of RET rearrangements did not influence OS at 1 year or 3 years. The survival probability for RET-positive patients was slightly higher at 1 year (60.6% vs. 59.1%) but slightly lower at 3 years (30.3% vs. 33.3%) compared to RET-negative patients, though these differences were not statistically significant (log-rank P=0.70) (Table 2, Figure 1). Cox regression analysis confirmed that RET rearrangement status was not associated with at either time point (Table 3).

NTRK alteration

At 1 year, the survival probability was 63.2% for NTRK-negative patients, but 100% for NTRK-positive patients. Similar results were observed at 3 years (Table 3). Cox regression analyses could not be performed, due to the extremely low alteration prevalence (0.3%) and lack of deaths observed among NTRK-positive patients.

Discussion

In this retrospective cohort study, we examined the OS of patients with NSCLC across various biomarker profiles, focusing on the impact of specific actionable biomarker alterations on survival outcomes. Our findings reinforce the critical role of certain targetable biomarkers, notably EGFR, in influencing survival rates. The study identified that patients with EGFR exon 18-21 mutations had significantly better survival outcomes at both 1 and 3 years compared to those without such mutations. This aligns with the known biology of EGFR-mutated NSCLC, where these tumors generally exhibit a more indolent course when treated appropriately (15). Similarly, patients with ALK rearrangements demonstrated improved survival, consistent with previous literature highlighting the efficacy of ALK inhibitors in extending survival (16). Conversely, mutations in ERBB2 and MET and ROS1 alterations were associated with poorer survival outcomes, particularly at the 3-year mark, indicating that these mutations may confer a more aggressive disease phenotype (17). The presence of other mutations, specifically those in ROS1, MET, BRAF, RET, and NTRK, did not significantly impact survival, suggesting that their prognostic value may be limited or context-dependent.

Our findings corroborate previous studies that have highlighted the prognostic importance of EGFR and ALK mutations in NSCLC (18,19). However, the observed survival disadvantage in patients with ERBB2 and MET mutations contrasts with an earlier report, which may be attributable to differences in treatment approaches, including the availability of targeted therapies during the study period (20). The lack of significant survival differences for ROS1, MET, BRAF, RET, and NTRK mutations may reflect the relatively smaller sample sizes for these subgroups or the relatively recent introduction of targeted therapies for these biomarkers.

Several potential limitations should be considered when interpreting our findings. First, the retrospective nature of the study introduces the possibility of selection bias, as patients who underwent biomarker testing may differ systematically from those who did not. Additionally, while we adjusted for key covariates such as age, gender, race/ethnicity, smoking status, and cancer stage, residual confounding by unmeasured variables cannot be excluded. Another limitation is the heterogeneity in treatment regimens, particularly in stage IV patients, where some received chemotherapy while others did not, potentially influencing survival outcomes independently of biomarker status, and reported associations do not adjust for any treatment received. While our study population is socio-demographically diverse and broadly representative of the Northern California population, the findings may not be generalizable to other regions or healthcare settings with different demographic compositions or treatment paradigms. Moreover, the study’s reliance on a single healthcare system may limit the applicability of the results to settings with different care delivery models. Additionally, biomarker testing is not ubiquitous throughout all stages of NSCLC and thus may limit generalizability as testing remains a challenge in both uninsured and insured populations (21). Lastly, although we report robust OS numbers, we were not able to ascertain exact cause of death.

The findings from this study underscore the importance of routine biomarker testing in NSCLC at all stages to guide treatment decisions and improve survival outcomes. As the landscape of targeted therapies continues to evolve, future research should focus on optimizing treatment strategies for patients with less common or less well-understood biomarkers, such as ERBB2 and MET (22). Furthermore, for those biomarkers that are better understood and characterized to have an associated worse OS, treatments should be modified to cater to their respective aggressiveness. It is known that not all patients who receive matched therapies respond, and for those that do respond, there exists a possibility of developing resistance via mutations. Clinicians and institutional guidelines must therefore curate treatment algorithms to keep up with the ever-changing genomic environment. Treatment may be optimized by employing more than one modality to target the tumor’s microenvironment. Future studies that characterize other associated biomarkers [i.e., status of deficient mismatch repair (dMMR)] as subgroups of the aggressive ones (i.e., MET) may offer understanding of other synergistic means of ameliorating the tumor’s ability to propagate and escape singular-modality treatment. Additionally, more robust investigations of KRAS, ROS1, BRAF, RET, and NTRK are needed to elucidate their potential roles in OS. Lastly, prospective studies are needed to validate these findings and explore the long-term outcomes associated with emerging biomarkers. Continued research is essential to refine our understanding of the prognostic and predictive value of various biomarkers in NSCLC and to translate these insights into improved patient outcomes.

Conclusions

In this large, real-world cohort of NSCLC patients, biomarker-defined subgroups demonstrated distinct survival outcomes, with EGFR and ALK alterations associated with improved prognosis and ERBB2 and ROS1 alterations linked to poorer survival. These findings underscore the critical role of routine molecular profiling to inform patient-specific treatment strategies. Future prospective studies are warranted to validate these associations and to integrate emerging biomarkers into evidence-based guidelines for NSCLC management.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1043/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1043/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1043/prf

Funding: This work was supported by funding from AstraZeneca through the Kaiser Permanente Division of Research grant (to L.C.S. and J.B.V.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1043/coif). J.B.V serves as an unpaid editorial board member of Journal of Thoracic Disease from September 2024 to August 2026, and reports funding support from AstraZeneca Industry Investigator Grant. AstraZeneca had no role in any aspects of the data and/or manuscript. L.C.S. was awarded research grants from the U.S. National Institutes of Health, California Tobacco-Related Disease Research Program, and AstraZeneca, outside the submitted work, all paid directly to her institution. Additionally, L.C.S. received payment for consultant work from the Troper Wojcicki Foundation and support from the American Cancer Society for travel to attend National Lung Cancer Roundtable meetings. J.M.S. reports grants from NCI/NIH/DHHS (No. 5UG1CA189821-11, paid to the institution), and serves as an unpaid Hope Foundation board member. Funding received is separate and unrelated to current work submission. The other 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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics committee of Kaiser Permanente Northern California (No. 1888915-12) and individual consent for this retrospective analysis was waived.

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. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12-49. [Crossref] [PubMed]
2. O'Haire S, Degeling K, Franchini F, et al. Comparing Survival Outcomes for Advanced Cancer Patients Who Received Complex Genomic Profiling Using a Synthetic Control Arm. Target Oncol 2022;17:539-48. [Crossref] [PubMed]
3. Li F, Zhai S, Lv Z, et al. Effect of histology on the efficacy of immune checkpoint inhibitors in advanced non-small cell lung cancer: A systematic review and meta-analysis. Front Oncol 2022;12:968517. [Crossref] [PubMed]
4. Cadranel J, Mauguen A, Faller M, et al. Impact of systematic EGFR and KRAS mutation evaluation on progression-free survival and overall survival in patients with advanced non-small-cell lung cancer treated by erlotinib in a French prospective cohort (ERMETIC project--part 2). J Thorac Oncol 2012;7:1490-502. [Crossref] [PubMed]
5. Cao W, Tang Q, Zeng J, et al. A Review of Biomarkers and Their Clinical Impact in Resected Early-Stage Non-Small-Cell Lung Cancer. Cancers (Basel) 2023;15:4561. [Crossref] [PubMed]
6. Biton J, Mansuet-Lupo A, Pécuchet N, et al. TP53, STK11, and EGFR Mutations Predict Tumor Immune Profile and the Response to Anti-PD-1 in Lung Adenocarcinoma. Clin Cancer Res 2018;24:5710-23. [Crossref] [PubMed]
7. Harada K, Ono S. Background and clinical significance of biomarker-based patient enrichment in non-small-cell lung cancer drug development. Sci Rep 2024;14:7194. [Crossref] [PubMed]
8. Li Y, Ma Y, Wu Z, et al. Tumor Mutational Burden Predicting the Efficacy of Immune Checkpoint Inhibitors in Colorectal Cancer: A Systematic Review and Meta-Analysis. Front Immunol 2021;12:751407. [Crossref] [PubMed]
9. Charo LM, Eskander RN, Sicklick J, et al. Real-World Data From a Molecular Tumor Board: Improved Outcomes in Breast and Gynecologic Cancers Patients With Precision Medicine. JCO Precis Oncol 2022;6:e2000508. [Crossref] [PubMed]
10. Zhang Q, Fu Q, Bai X, et al. Molecular Profiling-Based Precision Medicine in Cancer: A Review of Current Evidence and Challenges. Front Oncol 2020;10:532403. [Crossref] [PubMed]
11. Saarenheimo J, Andersen H, Eigeliene N, et al. Gene-Guided Treatment Decision-Making in Non-Small Cell Lung Cancer - A Systematic Review. Front Oncol 2021;11:754427. [Crossref] [PubMed]
12. Wang Z, Kim J, Zhang P, et al. Current therapy and development of therapeutic agents for lung cancer. Cell Insight 2022;1:100015. [Crossref] [PubMed]
13. Haslem DS, Van Norman SB, Fulde G, et al. A Retrospective Analysis of Precision Medicine Outcomes in Patients With Advanced Cancer Reveals Improved Progression-Free Survival Without Increased Health Care Costs. J Oncol Pract 2017;13:e108-19. [Crossref] [PubMed]
14. Kohan A, Kulanthaivelu R, Hinzpeter R, et al. Disparity and Diversity in NSCLC Imaging and Genomics: Evaluation of a Mature, Multicenter Database. Cancers (Basel) 2023;15:2096. [Crossref] [PubMed]
15. Paez JG, Jänne PA, Lee JC, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 2004;304:1497-500. [Crossref] [PubMed]
16. Shaw AT, Yeap BY, Mino-Kenudson M, et al. Clinical features and outcome of patients with non-small-cell lung cancer who harbor EML4-ALK. J Clin Oncol 2009;27:4247-53. [Crossref] [PubMed]
17. Wu R, Yuan B, Li C, et al. A narrative review of advances in treatment and survival prognosis of HER2-positive malignant lung cancers. J Thorac Dis 2021;13:3708-20. [Crossref] [PubMed]
18. Shigematsu H, Gazdar AF. Somatic mutations of epidermal growth factor receptor signaling pathway in lung cancers. Int J Cancer 2006;118:257-62. [Crossref] [PubMed]
19. Solomon BJ, Mok T, Kim DW, et al. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med 2014;371:2167-77. [Crossref] [PubMed]
20. Pillai RN, Behera M, Berry LD, et al. HER2 mutations in lung adenocarcinomas: A report from the Lung Cancer Mutation Consortium. Cancer 2017;123:4099-105. [Crossref] [PubMed]
21. Gross CP, Meyer CS, Ogale S, et al. Associations Between Medicaid Insurance, Biomarker Testing, and Outcomes in Patients With Advanced NSCLC. J Natl Compr Canc Netw 2022;20:479-487.e2. [Crossref] [PubMed]
22. Melosky B, Wheatley-Price P, Juergens RA, et al. The rapidly evolving landscape of novel targeted therapies in advanced non-small cell lung cancer. Lung Cancer 2021;160:136-51. [Crossref] [PubMed]