Abstract
This cohort study examines the association of KRAS status with outcomes following immune checkpoint inhibition monotherapy vs chemoimmunotherapy.
For patients with advanced non–small-cell lung cancer (NSCLC) without a driver alteration and programmed cell death ligand 1 (PD-L1) expression of 50% or greater, immune checkpoint inhibition (ICI) monotherapy or in combination with chemotherapy is standard first-line therapy. When deciding between these options, clinicians consider disease burden and comorbidities; however, to our knowledge, no biomarkers have been shown to predict differential benefit or harm.
KRAS variants in NSCLC are associated with smoking history, higher PD-L1 expression, and responsiveness to ICI monotherapy.1,2,3 The KEYNOTE-042 study demonstrated an overall survival (OS) benefit for first-line pembrolizumab over chemotherapy in patients with PD-L1 expression of 1% or greater.4 In an exploratory analysis, this benefit was seen regardless of KRAS status, but was more pronounced in patients with KRAS variants (median OS [mOS], 28 vs 11 months; hazard ratio [HR], 0.42;, 95% CI, 0.22-0.81) than those without KRAS variants (mOS, 15 vs 12 months; HR, 0.86; 95% CI, 0.63-1.18).1 However, to our knowledge, no prior studies have evaluated the association of KRAS status with outcomes following ICI monotherapy vs chemoimmunotherapy in patients with PD-L1 of 50% or greater.
Methods
Using the Flatiron Health database, comprising 280 cancer clinics across the US,5 we analyzed patients with advanced nonsquamous NSCLC with PD-L1 expression of 50% or greater, known KRASv status, and no alteration in EGFR, ALK, or ROS1 who were treated with first-line ICI monotherapy or chemoimmunotherapy between January 2016 and May 2020. Institutional review board exemption was granted by the University of Pennsylvania after determination that the proposal met eligibility criteria for institutional review board review exemption. Kaplan-Meier methods compared OS (from first-line systemic therapy initiation to death from any cause) between groups stratified by treatment type and KRAS status (variant [v] or wild type [wt]). Cox proportional hazards models estimated adjusted HRs and 95% CIs for death associated with KRAS status and treatment regimen. Analyses were performed using Stata, version 15 (StataCorp). Statistical significance was set at 2-sided P < .05 for all tests.
Results
Among 1127 patients with advanced nonsquamous NSCLC with PD-L1 expression of 50% or greater, 573 (50.8%) had KRASv status and 554 (49.2%) had KRASwt status. Patients with KRASv status were more likely to be female (58.7% vs 47.1%; P = .002) and have smoking history (96.4% vs 87.7%; P < .001). Other characteristics, including age, race/ethnicity, performance status, and stage at diagnosis, were well balanced among the groups.
Among patients treated with ICI monotherapy, KRASv was associated with superior survival compared with KRASwt (mOS, 21.1 vs 13.6 months; P = .03); this association remained significant on an adjusted Cox model (HR, 0.77; 95% CI, 0.61-0.98; Table). However, among patients treated with chemoimmunotherapy, there was no significant survival difference between patients with KRASv and KRASwt status (mOS, 20.0 vs 19.3 months; P = .93; adjusted HR, 0.99; 95% CI, 0.70-1.40; Table).
Table. Hazard Ratios for Death in PD-L1–High NSCLC by KRAS Status and Treatment Regimen.
| Values | HR (95% CI) | |
|---|---|---|
| Unadjusted | Adjusteda | |
| Prognostic value of KRAS status | ||
| ICI monotherapy | ||
| KRAS variant | 0.79 (0.64-0.98) | 0.77 (0.61-0.98) |
| KRAS wild type | 1 [Reference] | 1 [Reference] |
| Chemoimmunotherapy | ||
| KRAS variant | 0.98 (0.73-1.33) | 0.99 (0.70-1.40) |
| KRAS wild type | 1 [Reference] | 1 [Reference] |
| Predictive value of KRAS status | ||
| KRAS variant | ||
| ICI monotherapy | 1.04 (0.80-1.35) | 1.03 (0.75-1.40) |
| Chemoimmunotherapy | 1 [Reference] | 1 [Reference] |
| KRAS wild type | ||
| ICI monotherapy | 1.27 (0.99-1.64) | 1.19 (0.89-1.58) |
| Chemoimmunotherapy | 1 [Reference] | 1 [Reference] |
Abbreviations: ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; ICI, immune checkpoint inhibition; NSCLC, non–small-cell lung cancer; PD-L1, programmed cell death ligand 1.
Multivariable Cox model adjusted for age, sex, race/ethnicity, ECOG performance status, disease stage, and smoking history. Missing values for ECOG performance status were imputed by multiple chained imputation.
Among patients with KRASv status, OS did not differ between those treated with ICI monotherapy and chemoimmunotherapy (mOS, 21.1 vs 20.0 months; P = .78; Figure; adjusted HR, 1.03; 95% CI, 0.75-1.40; Table). Among patients with KRASwt status, those treated with ICI monotherapy had numerically worse survival than those treated with chemoimmunotherapy, although this difference was not statistically significant (mOS, 13.6 vs 19.3 months; P = .06; Figure; adjusted HR, 1.19; 95% CI, 0.89-1.58; Table).
Figure. Overall Survival in Programmed Cell Death Ligand 1–High Non–Small-Cell Lung Cancer by KRAS Status and First-line Treatment Regimen.
ICI indicates immune checkpoint inhibition.
Discussion
Our findings suggest that among patients with PD-L1 expression of 50% or greater NSCLC treated with ICI monotherapy, KRASwt is associated with worse survival compared with KRASv. In contrast, survival did not differ appreciably between patients with KRASv and KRASwt status who were treated with chemoimmunotherapy. Whereas patients with PD-L1–high NSCLC with KRASv had favorable survival (mOS ≥20 months) with either ICI monotherapy or chemoimmunotherapy, patients with KRASwt who were treated with ICI monotherapy had numerically inferior survival compared with those treated with chemoimmunotherapy, although this difference was not statistically significant. These data suggest that chemoimmunotherapy might be favored over ICI monotherapy for patients with KRASwt with high PD-L1 expression. While the specific prevalence of KRASv in the PD-L1–high subset is not definitively known, the 50% prevalence observed in our cohort is similar to that reported by others.6 The limitations of this analysis include unknown KRASv subtype and covariant status, including TP53 and STK11, as well as residual confounding despite adjustment for multiple covariates. Further investigation is needed to optimize selection between multiple available treatment strategies for patients with PD-L1–high NSCLC.
References
- 1.Herbst RS. LBA4 Association of KRAS mutational status with response to pembrolizumab monotherapy given as first-line therapy for PD-L1-positive advanced non-squamous NSCLC in Keynote-042. Annals of Oncology. 2019;30:xi63-xi64. doi: 10.1093/annonc/mdz453.001 [DOI] [Google Scholar]
- 2.Liu C, Zheng S, Jin R, et al. The superior efficacy of anti-PD-1/PD-L1 immunotherapy in KRAS-mutant non-small cell lung cancer that correlates with an inflammatory phenotype and increased immunogenicity. Cancer Lett. 2020;470:95-105. doi: 10.1016/j.canlet.2019.10.027 [DOI] [PubMed] [Google Scholar]
- 3.Coelho MA, de Carné Trécesson S, Rana S, et al. Oncogenic RAS signaling promotes tumor immunoresistance by stabilizing PD-L1 mRNA. Immunity. 2017;47(6):1083-1099.e6. doi: 10.1016/j.immuni.2017.11.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mok TSK, Wu YL, Kudaba I, et al. ; KEYNOTE-042 Investigators . Pembrolizumab versus chemotherapy for previously untreated, PD-L1-expressing, locally advanced or metastatic non-small-cell lung cancer (KEYNOTE-042): a randomised, open-label, controlled, phase 3 trial. Lancet. 2019;393(10183):1819-1830. doi: 10.1016/S0140-6736(18)32409-7 [DOI] [PubMed] [Google Scholar]
- 5.Ma, X., et al. , Comparison of population characteristics in real-world clinical oncology databases in the US: Flatiron Health, SEER, and NPCR. medRxiv, 2020: p. 2020.03.16.20037143. doi: 10.1101/2020.03.16.20037143 [DOI]
- 6.Kartolo A, Feilotter H, Hopman W, Fung AS, Robinson A. A single institution study evaluating outcomes of PD-L1 high KRAS-mutant advanced non-small cell lung cancer (NSCLC) patients treated with first line immune checkpoint inhibitors. Cancer Treat Res Commun. 2021;27:100330. doi: 10.1016/j.ctarc.2021.100330 [DOI] [PubMed] [Google Scholar]