Advertisement

Prognostic Impact of KRAS G12C Mutation in Patients With NSCLC: Results From the European Thoracic Oncology Platform Lungscape Project

Open ArchivePublished:February 25, 2021DOI:https://doi.org/10.1016/j.jtho.2021.02.016

      Abstract

      Introduction

      KRAS mutations, the most frequent gain-of-function alterations in NSCLC, are currently emerging as potential predictive therapeutic targets. The role of KRAS-G12C (Kr_G12C) is of special interest after the recent discovery and preclinical analyses of two different Kr_G12C covalent inhibitors (AMG-510, MRTX849).

      Methods

      KRAS mutations were evaluated in formalin-fixed, paraffin-embedded tissue sections by a microfluidic-based multiplex polymerase chain reaction platform as a component of the previously published European Thoracic Oncology Platform Lungscape 003 Multiplex Mutation study, of clinically annotated, resected, stage I to III NSCLC. In this study, -Kr_G12C mutation prevalence and its association with clinicopathologic characteristics, molecular profiles, and postoperative patient outcome (overall survival, relapse-free survival, time-to-relapse) were explored.

      Results

      KRAS gene was tested in 2055 Lungscape cases (adenocarcinomas: 1014 [49%]) with I or II or III stage respective distribution of 53% or 24% or 22% and median follow-up of 57 months. KRAS mutation prevalence in the adenocarcinoma cohort was 38.0% (95% confidence interval (CI): 35.0% to 41.0%), with Kr_G12C mutation representing 17.0% (95% CI: 14.7% to 19.4%). In the "histologic-subtype" cohort, Kr_G12C prevalence was 10.5% (95% CI: 9.2% to 11.9%). When adjusting for clinicopathologic characteristics, a significant negative prognostic effect of Kr_G12C presence versus other KRAS mutations or nonexistence of KRAS mutation was identified in the adenocarcinoma cohort alone and in the "histologic-subtype" cohort. For overall survival in adenocarcinomas, hazard ratio (HR)G12C versus other KRAS is equal to 1.39 (95% CI: 1.03 to 1.89, p = 0.031) and HRG12C versus no KRAS is equal to 1.32 (95% CI: 1.03 to 1.69, p = 0.028) (both also significant in the "histologic-subtype" cohort). For time-to-relapse, HRG12C versus other KRAS is equal to 1.41 (95% CI: 1.03 to 1.92, p = 0.030). In addition, among all patients, for relapse-free survival, HRG12C versus no KRAS is equal to 1.27 (95% CI: 1.04 to 1.54, p = 0.017).

      Conclusions

      In this large, clinically annotated stage I to III NSCLC cohort, the specific Kr_G12C mutation is significantly associated with poorer prognosis (adjusting for clinicopathologic characteristics) among adenocarcinomas and in unselected NSCLCs.

      Keywords

      Introduction

      Lung cancer represents one of the most mutated of solid tumors,
      • Lawrence M.S.
      • Stojanov P.
      • Polak P.
      • et al.
      Mutational heterogeneity in cancer and the search for new cancer-associated.
      highlighting the role and impact of tobacco carcinogenesis in tumor development in most cases. Over the years, oncogenic drivers have been identified in NSCLC, particularly in adenocarcinoma, and several targeted therapies have been developed with major benefit for patients.
      • Yuan M.
      • Huang L.L.
      • Chen J.H.
      • Wu J.
      • Xu Q.
      The emerging treatment landscape of targeted therapy in non-small-cell lung cancer.
      Molecular genotyping is now routinely used to guide clinical care of patients with locally advanced or inoperable and metastatic lung adenocarcinoma, largely owing to clinical trials revealing superior efficacy of targeted kinase inhibitors as compared with standard chemotherapy among patients with targetable oncogenic drivers.
      • Gadgeel S.
      • Peters S.
      • Mok T.
      • et al.
      Alectinib versus crizotinib in treatment-naive anaplastic lymphoma kinase-positive (ALK+) non-small-cell lung cancer: CNS efficacy results from the ALEX study.
      • Soria J.C.
      • Ohe Y.
      • Vansteenkiste J.
      • et al.
      Osimertinib in untreated EGFR-mutated advanced non-small-cell lung cancer.
      • Camidge D.R.
      • Kim H.R.
      • Ahn M.J.
      • et al.
      Brigatinib versus crizotinib in ALK-positive non-small-cell lung cancer.
      The impetus of targeted therapies is shifting toward the adjuvant setting as evidenced by the ADAURA trial.
      • Dong Z.Y.
      • Zhong W.Z.
      • Zhang X.C.
      • et al.
      Potential predictive value of TP53 and KRAS mutation status for response to PD-1 blockade immunotherapy in lung adenocarcinoma.
      Multiple adjuvant targeted therapy trials are ongoing for different oncogenic alterations,
      • Friedlaender A.
      • Addeo A.
      • Russo A.
      • Gregorc V.
      • Cortinovis D.
      • Rolfo C.D.
      Targeted therapies in early stage NSCLC: hype or hope?.
      and their results may alter the treatment paradigm in this setting.
      Previously, within the Lungscape project, we have revealed the importance of ALK gene rearrangements as independent prognostic factors
      • Blackhall F.H.
      • Peters S.
      • Bubendorf L.
      • et al.
      Prevalence and clinical outcomes for patients with ALK-positive resected stage I to III adenocarcinoma: results from the European Thoracic Oncology Platform Lungscape project.
      and that for other molecular mutations the effect is still unclear with no prognostic associations found for major known drivers.
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      Of these, KRAS mutations are the most frequent gain-of-function alterations found in patients with NSCLC, accounting for up to 25% of all oncogenic mutations. The significance of KRAS mutations has evolved from having uncertain prognostic value and from being “undruggable” to the current status of potential predictive therapeutic target with great promise.
      • Torralvo J.
      • Friedlaender A.
      • Achard V.
      • Addeo A.
      The activity of immune checkpoint inhibition in KRAS mutated non-small cell lung cancer: a single centre experience.
      ,
      • Friedlaender A.
      • Drilon A.
      • Weiss G.J.
      • Banna G.L.
      • Addeo A.
      KRAS as a druggable target in NSCLC: rising like a phoenix after decades of development failures.
      In particular KRAS-G12C (Kr_G12C) mutation emerged as a possible therapeutic target with the discovery and preclinical analyses of two different covalent inhibitors of Kr_G12C (AMG-510 and MRTX849).
      • Canon J.
      • Rex K.
      • Saiki A.Y.
      • et al.
      The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity.
      ,
      • Hallin J.
      • Engstrom L.D.
      • Hargis L.
      • et al.
      The KRAS G12C inhibitor MRTX849 provides insight toward therapeutic susceptibility of KRAS-mutant cancers in mouse models and patients.
      The European Thoracic Oncology Platform (ETOP) Lungscape database (iBiobank) is a virtual biobank of fully clinically annotated, surgically resected NSCLC. In this study, we describe, in the cohort of patients with multiplex testing results in the ETOP Lungscape database, the prevalence and association of the KRAS gene, specifically Kr_G12C, to clinicopathologic characteristics and postoperative patient outcomes.

      Materials and Methods

      The ETOP Lungscape database, iBiobank (https://etopdata.etop-eu.org), stores clinicopathologic and molecular data on more than 2700 patients with surgically resected stage I to III NSCLC with follow-up greater than 3 years. All patients had a histologic diagnosis of lung cancer and were pathologically staged at the time of resection. Central review regarding completeness of mandatory clinical parameters and the seventh TNM staging accuracy was performed. In the frame of Lungscape multiplex substudy,
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      a multigene, multiplex platform was used to generate mutation profiles at the time of resection. Testing was carried out on a high-throughput microfluidic-based polymerase chain reaction platform running an allele-specific multiplex test. The panel used included coverage of 13 genes (AKT1, BRAF, EGFR, ERBB2, FLT3, HRAS, JAK2, KIT, KRAS, MET, MYD88, NRAS, and PIK3CA) incorporating 130 hotspot mutations found in various tumor types. In this study, we specifically focus on the KRAS Kr_G12C mutation. In addition, data from four other Lungscape studies contributed to the correlation matrix such as expression of PTEN,
      • Rulle U.
      • Tsourti Z.
      • Casanova R.
      • et al.
      Computer-based intensity measurement assists pathologists in scoring phosphatase and tensin homolog immunohistochemistry—clinical associations in NSCLC patients of the European Thoracic Oncology Platform Lungscape Cohort.
      ALK rearrangement,
      • Blackhall F.H.
      • Peters S.
      • Bubendorf L.
      • et al.
      Prevalence and clinical outcomes for patients with ALK-positive resected stage I to III adenocarcinoma: results from the European Thoracic Oncology Platform Lungscape project.
      MET overexpression and amplification,
      • Bubendorf L.
      • Dafni U.
      • Schöbel M.
      • et al.
      Prevalence and clinical association of MET gene overexpression and amplification in patients with NSCLC: results from the European Thoracic Oncology Platform (ETOP) Lungscape project.
      and programmed death-ligand 1 (PD-L1).
      • Kerr K.M.
      • Thunnissen E.
      • Dafni U.
      • et al.
      A retrospective cohort study of PD-L1 prevalence, molecular associations and clinical outcomes in patients with NSCLC: results from the European Thoracic Oncology Platform (ETOP) Lungscape project.
      The Lungscape studies were conducted according to Lungscape master and substudy protocols, with adherence to country-specific ethics and regulatory requirements and REMARK recommendations.

      Statistical Considerations

      The analysis focuses on comparing cohort characteristics and outcome, between the specific Kr_G12C and other KRAS mutations and with KRAS nonmutated cases. Our primary analysis is based on patients of adenocarcinoma histology only, whereas a secondary analysis on the full cohort of patients irrespectively of their histology type (“histologic-subtype”) is also presented.
      Prevalence of KRAS mutation, specifically the Kr_G12C mutation, was derived, with corresponding 95% exact binomial confidence intervals (CIs).
      Balance of baseline characteristics (either between Kr_G12C-mutated cohort versus other KRAS mutations or all KRAS nonmutated cases) was tested by Fisher’s exact test and Mann-Whitney U tests, for categorical and continuous variables, correspondingly.
      The association of KRAS mutations (Kr_G12C and others), with previously published Lungscape molecular data, EGFR, PIK3CA, MET genes and PD-L1 expression, MET (clone SP44), ALK (clone 5A4), and PTEN immunohistochemistry (IHC) (clone SP218), was explored through Fisher’s exact test.
      Clinical outcome was evaluated as overall survival (OS), relapse-free survival (RFS), and time-to-relapse (TTR), estimated as time from surgery date to time to death from any cause, time to first relapse or death from any cause, and time to first relapse, respectively. Median follow-up time was estimated using the reverse censoring method for OS.
      • Peters S.
      • Weder W.
      • Dafni U.
      • et al.
      Lungscape: resected non-small-cell lung cancer outcome by clinical and pathological parameters.
      All time-to-event end points were graphically depicted by the Kaplan-Meier method, whereas observed differences in hazard were assessed by the log-rank test. The effect of Kr_G12C mutation on outcome was further explored through Cox proportional hazard regression models. Prognostic ability for several clinicopathologic characteristics was evaluated in the multivariate Cox models: sex, ethnicity, smoking history, age, adjuvant chemotherapy, adjuvant radiotherapy, previous history of cancer, performance status at time of surgery, stage, primary tumor localization, tumor size, histology, surgery year, technique, and anatomy. Backward elimination method with a removal criterion p value greater than or equal to 0.10 was used to identify the significant prognostic factors. The proportional hazard assumption was verified, visualizing the Schoenfeld residuals and testing the time-dependent covariates of interaction of mutation groups with survival time. Subgroup analysis by stage was also performed for exploratory, descriptive purposes. Lastly, as an exploratory, sensitivity analysis, we have further evaluated the effect of pure KRAS mutations, excluding all detected cases of co-mutations or other mutations. Survival analysis was performed for pure KRAS G12_C, pure other KRAS mutations, pure EGFR mutations, and wild-type cases without any targetable mutation (according to multiplex testing).
      In all exploratory analyses, results with two-sided p value less than or equal to 0.05 were considered significant.
      Statistical analyses were carried out in Statistical Analysis System version 9.4.

      Results

      Analysis Cohort

      A total of 2055 patients (75.9% of all 2709 Lungscape cases), derived from 16 Lungscape centers in 10 European countries (Belgium, Germany, Denmark, Ireland, Italy, The Netherlands, Poland, Spain, Switzerland, United Kingdom), had an assessable multiplex testing on KRAS gene (Supplementary Table S.B1).
      Among the “histological-subtype” cohort of 2055 patients, 1014 patients had adenocarcinoma histology constituting the cohort of primary analysis.
      Most patients with adenocarcinoma (Table 1) were of male sex (54%), with median age of 66 years, 81% former or current smokers, and 57% had 0 or 1 performance status at the time of surgery (with additional 42% missing information). The distribution by stage was 53%, 24%, and 22% for stage I, II, and III, respectively, whereas 72% of the patients had tumor size up to 4 cm. In addition, with respect to T staging, 36% were classified as T1a or T1b, 46% T2a or T2b, 15% T3, and only 4% as T4, whereas regarding lymph nodes, 68% were N0, 17% N1, 16% N2, and only 0.2% N3. For 66% of the patients, surgery had taken place after 2005, with the vast majority, 93% of them, within 2003 to 2010. Adjuvant chemotherapy was administered to 23% of the adenocarcinomas (7%, 33%, and 50% for stage I, II, and III, respectively), whereas only 5% had received adjuvant radiotherapy (1%, 4%, and 17% for stage I, II, and III, respectively), of which 33 (3% overall; 0%, 2%, and 13% for stage I, II, and III) had received adjuvant chemoradiotherapy (Supplementary Table S.A1: for adenocarcinomas, Supplementary Table S.B2: for “histologic-subtype”). Previous history of cancer was reported for 17% of the patients with adenocarcinoma.
      Table 1Baseline Characteristics; Overall and by KRAS Mutation Status (N = 1014)
      CharacteristicsOverall (N = 1014)Kr_G12C MD (n = 172)Other KRAS MD (n = 213)KRAS MND (n = 629)p Value
      Fisher’s exact test for categorical or Kruskal-Wallis test for continuous variables.
      Patient characteristics
      Age (in y)
       Median65.663.165.966.2A: 0.0092
       (Min–Max)(23.1–88.7)(40.3–86.1)(23.1–88.7)(38.5–84.7)B: 0.0059
      Sex
       Female465 (45.9)83 (48.3)107 (50.2)275 (43.7)A: 0.76
       Male549 (54.1)89 (51.7)106 (49.8)354 (56.3)B: 0.30
      Ethnicity
       Caucasian1001 (98.7)172 (100.0)212 (99.5)617 (98.1)A: >0.99
       Other13 (1.3)0 (0.0)1 (0.5)12 (1.9)B: 0.080
      Smoking
       Current313 (30.9)70 (40.7)66 (31.0)177 (28.1)A: 0.099
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Former509 (50.2)86 (50.0)117 (54.9)306 (48.7)B: <0.001
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Never148 (14.6)9 (5.2)19 (8.9)120 (19.1)
       Unknown44 (4.3)7 (4.1)11 (5.2)26 (4.1)
      PS at time of surgery
       0384 (37.9)67 (39.0)88 (41.3)229 (36.4)A: 0.89
      Excluding category “unknown,” “missing,” or “unknown or missing.”
      ,
      Categories “1,” “2,” and “3” are combined.
       1192 (18.9)32 (18.6)39 (18.3)121 (19.2)B: 0.73
      Excluding category “unknown,” “missing,” or “unknown or missing.”
      ,
      Categories “1,” “2,” and “3” are combined.
       210 (1.0)2 (1.2)3 (1.4)5 (0.8)
       32 (0.2)0 (0.0)0 (0.0)2 (0.3)
       Missing426 (42.0)71 (41.3)83 (39.0)272 (43.2)
      Adj. radiotherapy
       Yes52 (5.1)4 (2.3)11 (5.2)37 (5.9)A: 0.19
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       No822 (81.1)143 (83.1)177 (83.1)502 (79.8)B: 0.076
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Unknown or missing140 (13.8)25 (14.5)25 (11.7)90 (14.3)
      Adj. chemotherapy
       Yes230 (22.7)41 (23.8)53 (24.9)136 (21.6)A: >0.99
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       No647 (63.8)107 (62.2)136 (63.9)404 (64.2)B: 0.53
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Unknown or missing137 (13.5)24 (14.0)24 (11.3)89 (14.2)
      Previous history of cancer
       Yes173 (17.1)24 (13.9)30 (14.1)119 (18.9)A: 0.88
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       No690 (68.0)126 (73.3)145 (68.1)419 (66.6)B: 0.11
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Missing151 (14.9)22 (12.8)38 (17.8)91 (14.5)
      Tumor characteristics
      Stage
       Ia279 (27.5)45 (26.2)62 (29.1)172 (27.3)A: 0.93
      Categories combined to “Ia or Ib,” “IIa or IIb,” and “IIIa or IIIb.”
       Ib263 (25.9)49 (28.5)55 (25.8)159 (25.3)B: 0.47
      Categories combined to “Ia or Ib,” “IIa or IIb,” and “IIIa or IIIb.”
       IIa139 (13.7)23 (13.4)36 (16.9)80 (12.7)
       IIb108 (10.7)21 (12.2)21 (9.9)66 (10.5)
       IIIa207 (20.4)33 (19.2)35 (16.4)139 (22.1)
       IIIb18 (1.8)1 (0.6)4 (1.9)13 (2.1)
      TNM staging: T
       T1a188 (18.5)31 (18)39 (18.3)118 (18.8)A: 0.21
      Categories combined to “T1a or T1b,” “T2a or T2b,” and “T3 or T4.”
       T1b172 (17)25 (14.5)46 (21.6)101 (16.1)B: 0.84
      Categories combined to “T1a or T1b,” “T2a or T2b,” and “T3 or T4.”
       T2a379 (37.4)69 (40.1)77 (36.2)233 (37)
       T2b85 (8.4)13 (7.6)20 (9.4)52 (8.3)
       T3147 (14.5)31 (18)22 (10.3)94 (14.9)
       T443 (4.2)3 (1.7)9 (4.2)31 (4.9)
      TNM staging: N
       N0684 (67.5)115 (66.9)149 (70)420 (66.8)A: 0.74
      Categories N2 and N3 combined.
       N1168 (16.6)34 (19.8)36 (16.9)98 (15.6)B: 0.24
      Categories N2 and N3 combined.
       N2160 (15.8)22 (12.8)27 (12.7)111 (17.6)
       N32 (0.2)1 (0.6)1 (0.5)0 (0)
      Tumor localization
       Central tumor7 (0.7)2 (1.2)1(0.5)4 (0.6)A: 0.91
       Lower lobe L130 (12.8)29 (16.9)37 (17.4)64 (10.2)B: 0.033
       Lower lobe R131 (12.9)28 (16.3)29 (13.6)74 (11.8)
       Middle lobe R57 (5.6)13 (7.6)13 (6.1)31 (4.9)
       Overlapping41 (4.0)6 (3.5)6 (2.8)29 (4.6)
       Upper lobe L291 (28.7)43 (25.0)62 (29.1)186 (29.6)
       Upper lobe R357 (35.2)51 (29.7)65 (30.5)241 (38.3)
      Tumor size (in cm)
       ≤4733 (72.3)128 (74.4)155 (72.8)450 (71.5)A: 0.82
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       >4279 (27.5)44 (25.6)57 (26.8)178 (28.3)B: 0.50
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Missing2 (0.2)0 (0.0)1 (0.5)1 (0.2)
      Mean (95% CI)3.6 (3.5–3.7)3.7 (3.3–4.0)3.6 (3.3–3.9)3.6 (3.4–3.8)A: 0.85
      Median (Min–Max)3.0 (0.2–15.0)3.1 (0.5–15.0)3.0 (0.5–13.0)3.0 (0.2–14.0)B: 0.81
      Surg. characteristics
      Surg. anatomy
       Lobectomy807 (79.6)133 (77.3)171 (80.3)503 (80.0)A: 0.25
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Pneumonectomy75 (7.4)16 (9.3)10 (4.7)49 (7.8)B: 0.39
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Bilobectomy46 (4.5)8 (4.7)14 (6.6)24 (3.8)
       Wedge resection46 (4.5)6 (3.5)9 (4.2)31 (4.9)
       Segmentectomy23 (2.3)7 (4.1)4 (1.9)12 (1.9)
       Other or missing17 (1.7)2 (1.2)5 (2.4)10 (1.6)
      Surg. technique
       Open thoracotomy884 (87.2)159 (92.4)184 (86.4)541 (86.0)A: 0.0086
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Thoracoscopy97 (9.6)7 (4.1)25 (11.7)65 (10.3)B: 0.0099
      Excluding category “unknown,” “missing,” or “unknown or missing.”
       Missing33 (3.3)6 (3.5)4 (1.9)23 (3.7)
      Surg. year
       <2006341 (33.6)54 (31.4)71 (33.3)216 (34.3)A: 0.74
       ≥2006673 (66.4)118 (68.6)142 (66.7)413 (65.7)B: 0.52
      Note: Bold emphasis is used to indicate statistically significant comparisons. Italic emphasis is used to describe ‘tumor size (in cm)’ as a continuous characteristic.
      A, comparison of Kr_G12C versus all other KRAS mutations; Adj., adjuvant; B, comparison of Kr_G12C versus KRAS MND; CI, confidence interval; Kr_G12C, KRAS-G12C; Max, maximum; MD, mutation detected; Min, minimum; MND, mutation not detected; PS, performance status; Surg., surgery.
      a Fisher’s exact test for categorical or Kruskal-Wallis test for continuous variables.
      b Excluding category “unknown,” “missing,” or “unknown or missing.”
      c Categories “1,” “2,” and “3” are combined.
      d Categories combined to “Ia or Ib,” “IIa or IIb,” and “IIIa or IIIb.”
      e Categories combined to “T1a or T1b,” “T2a or T2b,” and “T3 or T4.”
      f Categories N2 and N3 combined.

      KRAS Prevalence

      The number of KRAS-mutated cases detected among patients with adenocarcinoma was 385 (KRAS mutation prevalence: 38.0%, 95% CI: 35.0%–41.0%), with 172 cases representing the Kr_G12C mutation type (with corresponding prevalence of 17.0%, 95% CI: 14.7%–19.4%).
      In the overall cohort of 2055 patients with NSCLC, 473 KRAS-mutated cases were identified (prevalence in the NSCLC cohort: 23.0%, 95% CI: 21.2%–24.9%), with 216 cases with Kr_G12C (prevalence 10.5%, 95% CI: 9.2%–11.9%).
      An analytical description of the KRAS mutation types observed is provided in Supplementary Table S.A2 (separately for patients with adenocarcinoma and non-adenocarcinoma).

      Association of Kr_G12C With Baseline Clinicopathologic Characteristics

      Patient, tumor, and surgery characteristics, for the three adenocarcinoma sub-cohorts considered (Kr_G12C-mutated patients, other-KRAS-mutated patients, and KRAS-nonmutated patients) are presented in Table 1.
      When exploring differences in patient baseline characteristics, Kr_G12C-mutated patients seem to be significantly younger (63.1 y median) compared with other-KRAS-mutated (median 65.9, p = 0.0092) and KRAS-nonmutated (median 66.2, p = 0.0059) patients. Smoking status is also a differentiating factor of Kr_G12C-mutated with KRAS-nonmutated (95% current or former smokers, excluding unknown, versus 80%, p < 0.001) but not different from the other-KRAS-mutated patients (91% current or former smokers).
      Among tumor characteristics, only a significant differentiation with respect to tumor localization is detected between Kr_G12C-mutated (with 33% lower lobes and 55% upper lobes) versus KRAS-nonmutated (22% lower, 68% upper lobes, p = 0.033) patients. Furthermore, the other-KRAS-mutated patients do not significantly differ from the Kr_G12C-mutated patients in any tumor characteristic.
      Analogous results for the full cohort of patients by "histologic-subtype" are provided in the Supplementary Data (Supplementary Table S.B1). As expected, adenocarcinomas are the vast majority in both Kr_G12C-mutated (80%) and other-KRAS-mutated (83%) patients with 40% in the KRAS-nonmutated (p < 0.001) patients. Furthermore, there are more female patients in the Kr_G12C-mutated (44%) and other-KRAS-mutated (47%) cohorts versus 31% in the KRAS-nonmutated (p < 0.001) cohort. Moreover, differentiation to tumor localization is not evident anymore, whereas a difference in tumor size emerges, with 69% of Kr_G12C-mutated having a tumor size up to 4 cm, compared with 61% of the KRAS-nonmutated (p = 0.017).

      Association of KRAS With Other Molecular Alterations and PD-L1 Expression

      The association of KRAS mutations (Kr_G12C or any other type) with other molecular alterations investigated in the Lungscape cohort for adenocarcinoma patients is summarized in Table 2.
      Table 2Association of KRAS Mutation With Other Molecular Alterations (N = 1014 Adenocarcinomas)
      GeneKr_G12C (n = 172)All Other KRAS MD (n = 213)KRAS MND (n = 629)p Value
      Fisher’s exact test.
      EGFR
       MD0 (0.0%)2 (1.0%)92 (15.2%)
      Statistical test not performed owing to table sparsity.
       MND164 (100.0%)203 (99.0%)512 (84.8%)
       No call8825
      PIK3CA
       MD2 (1.2%)5 (2.4%)25 (4.2%)
      Statistical test not performed owing to table sparsity.
       MND160 (98.8%)200 (97.6%)570 (95.8%)
       No call10834
      MET gene
       MD10 (5.8%)17 (8.0%)44 (7.0%)A: 0.43
       MND162 (94.2%)196 (92.0%)584 (93.0%)B: 0.73
       No call001
      ALK IHC
       ALK+5 (3.1%)8 (4.1%)34 (5.8%)A: 0.78
       ALK−158 (96.9%)186 (95.9%)553 (94.2%)B: 0.23
       Missing91942
      MET IHC
       MET+52 (31.5%)67 (33.3%)175 (29.4%)A: 0.74
       MET−113 (68.5%)134 (66.7%)421 (70.6%)B: 0.63
       Missing71233
      PTEN
       Loss64 (41.6%)72 (38.5%)223 (39.1%)A: 0.58
       No Loss90 (58.4%)115 (61.5%)347 (60.9%)B: 0.58
       Missing182659
      Note: No call or missing or not assessable missing categories are excluded from the calculation of percentages and p values.
      A, comparison of Kr_G12C versus all other KRAS mutations; B, comparison of Kr_G12C versus KRAS MND; IHC, immunohistochemistry; Kr_G12C, KRAS-G12C; MD, mutation detected; MND, mutation not detected.
      a Fisher’s exact test.
      b Statistical test not performed owing to table sparsity.
      Coexistence of KRAS and either EGFR or PIK3CA mutations was exceptionally rare. Among the two cases of KRAS and EGFR coexistence, none was Kr_G12C, whereas among the seven KRAS and PIK3CA and coexistence cases, only two were Kr_G12C mutations. Furthermore, no significant associations were detected when comparing Kr_G12C with either other-KRAS-mutated or KRAS-nonmutated patients (with respect to MET gene, ALK IHC, MET IHC or PTEN, or PD-L1 expression; Table 2 and Supplementary Table S.A3, respectively).
      Of note, when associations were examined in the “histological-subtype” cohort (Supplementary Table S.B3), MET IHC positivity was found to be significantly higher, almost double in both the Kr_G12C-mutated and the other-KRAS-mutated groups compared with the KRAS nonmutated group (33% versus 19%, p < 0.001) (overall KRAS group, p < 0.001, previously reported
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      ). PTEN loss was also marginally significantly lower in the Kr_G12C-mutated group (similar to the KRAS other mutation group) compared with the KRAS-nonmutated group (40% versus 48%, p = 0.047, Supplementary Table S.B3).

      Clinical Outcome and Mutation Status

      The clinical outcome of the patients (OS, RFS, and TTR) was evaluated at a median-follow-up of 57.1 months (interquartile range: 44.3–72.3 mo), comparable between the three sub-cohorts (p = 0.74). One-third of patients had died with disease (33.3%), whereas close to half of the patients were disease-free at their last follow-up (48.1%).
      The total number of deaths recorded in the adenocarcinoma analysis cohort was 446 (44.0%), with median OS (95% CI) of 59.4 months (44.9 mo–not estimable [NE]) for Kr_G12C-mutated patients, 81.7 months (55.3 mo–NE) for other-KRAS-mutated patients, and 72.0 months (62.2–82.4 mo) for nonmutated patients (Fig. 1A). Although nonsignificant in univariate analysis, a significant Kr_G12C effect on OS is detected when accounting for ethnicity, age, stage, performance status, and tumor size (Table 3). The hazard of death for patients harboring the Kr_G12C mutation is 39% higher compared with patients with other KRAS mutations (hazard ratio (HR)G12C versus other KRAS = 1.39, 95% CI: 1.03 to 1.89, p = 0.031) and 32% higher compared with patients with no KRAS mutation detected (HR G12C versus no KRAS = 1.32, 95% CI: 1.03 to 1.69, p = 0.028) (Table 3). Consistent OS results are found for all patients, irrespective of histology type (Supplementary Fig. S.B1a and Supplementary Table S.B4a). With 922 deaths observed in total, a significant detrimental Kr_G12C effect is reported with adjusted HR G12C versus other KRAS equals to 1.35 (95% CI: 1.02 to 1.75, p = 0.036) and HR G12C versus no KRAS equals to 1.25 (95% CI: 1.01 to 1.54, p = 0.040).
      Figure thumbnail gr1ab
      Figure 1(A) OS by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.12 (Kr_G12C versus other KRAS) and 0.24 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.031 (Kr_G12C versus other KRAS) and 0.028 (Kr_G12C versus “MND”). (B) RFS by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.23 (Kr_G12C versus other KRAS) and 0.36 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.070 (Kr_G12C versus other KRAS) and 0.066 (Kr_G12C versus “MND”). (C) TTR by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.16 (Kr_G12C versus other KRAS) and 0.33 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.030 (Kr_G12C versus other KRAS) and 0.063 (Kr_G12C versus “MND”). CI, confidence interval; Kr_G12C, KRAS-G12C; MND, mutation not detected; NE, not estimable; OS, overall survival; RFS, relapse-free survival; TTR, time-to-relapse.
      Figure thumbnail gr1c
      Figure 1(A) OS by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.12 (Kr_G12C versus other KRAS) and 0.24 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.031 (Kr_G12C versus other KRAS) and 0.028 (Kr_G12C versus “MND”). (B) RFS by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.23 (Kr_G12C versus other KRAS) and 0.36 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.070 (Kr_G12C versus other KRAS) and 0.066 (Kr_G12C versus “MND”). (C) TTR by KRAS mutation status (N = 1014 adenocarcinomas). Log-rank p values equal to 0.16 (Kr_G12C versus other KRAS) and 0.33 (Kr_G12C versus “MND”). Adjusted Cox p values equal to 0.030 (Kr_G12C versus other KRAS) and 0.063 (Kr_G12C versus “MND”). CI, confidence interval; Kr_G12C, KRAS-G12C; MND, mutation not detected; NE, not estimable; OS, overall survival; RFS, relapse-free survival; TTR, time-to-relapse.
      Table 3Multivariate Cox Proportional Hazards Model for OS in Patients With Adenocarcinoma
      No. of Pts: 1012
      Excluding two patients with missing tumor size measurement.
      No. of Deaths: 444
      HR95% CIp Value
      KRAS
      The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.39 (1.03–1.89) and Kr_G12C versus KRAS MND: 1.32 (1.03–1.69).
       Other KRAS mutations vs. Kr_G12C0.72(0.53–0.97)0.031
       MND vs. Kr_G12C0.76(0.59–0.97)0.028
      Ethnicity
       Other vs. Caucasian2.68(1.36–5.28)0.0044
      Age in y
       60–70 vs. <601.29(1.02–1.64)0.034
       >70 vs. <601.40(1.10–1.78)0.0064
      Stage
       Ib vs. Ia1.30(0.95–1.77)0.099
       IIa vs. Ia2.16(1.54–3.03)<0.001
       IIb vs. Ia2.00(1.36–2.93)<0.001
       IIIa vs. Ia4.08(3.02–5.51)<0.001
       IIIb vs. Ia5.02(2.70–9.35)<0.001
      Performance status at surgery
       1 or 2 or 3 vs. 01.22(0.93–1.61)0.15
       Missing vs. 01.43(1.16–1.77)0.0010
      Tumor size in cm
       >4 vs. ≤41.25(1.01–1.55)0.044
      CI, confidence interval; HR, hazard ratio; Kr_G12C, KRAS-G12C; MND, mutation not detected; OS, overall survival; Pt, patient.
      a Excluding two patients with missing tumor size measurement.
      b The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.39 (1.03–1.89) and Kr_G12C versus KRAS MND: 1.32 (1.03–1.69).
      Regarding RFS, 530 events were observed in the adenocarcinoma cohort (52.3% of 1014), with median RFS (95% CI) of 46.7 months (28.5–66.6 mo) for the Kr_G12C patients, 55.3 months (44.3–84.4 mo) for other KRAS mutations, and 50.3 months (43.1–63.2 mo) for KRAS nonmutated patients (Fig. 1B, log-rank p = not significant). In the backward selection of multivariate Cox model (Table 4), a negative effect of Kr_G12C mutation is apparent with p value less than 0.10 and adjusted for ethnicity, stage, surgery anatomy, performance status, and tumor size (HR G12C versus other KRAS = 1.30, 95% CI: 0.98–1.72, p = 0.070 and HR G12C versus noKRAS = 1.23, 95% CI: 0.98–1.56, p = 0.066).
      Table 4Multivariate Cox Proportional Hazards Model for RFS in Patients With Adenocarcinoma
      No. of Pts: 1012
      Excluding two patients with missing tumor size measurement.
      No. of RFS Events: 528
      HR95% CIp Value
      KRAS
      The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.30 (0.98–1.72) and Kr_G12C versus KRAS MND: 1.23 (0.98–1.56).
       Other KRAS mutations vs. Kr_G12C0.77(0.58–1.02)0.070
       MND vs. Kr_G12C0.81(0.64–1.02)0.066
      Ethnicity
       Other vs. Caucasian2.07(1.06–4.04)0.034
      Stage
       Ib vs. Ia1.16(0.88–1.54)0.29
       IIa vs. Ia2.18(1.61–2.96)<0.001
       IIb vs. Ia1.73(1.21–2.48)0.0027
       IIIa vs. Ia3.94(2.98–5.21)<0.001
       IIIb vs. Ia5.79(3.35–9.98)<0.001
      Surg. anatomy
       Pneumonectomy vs. lobectomy1.03(0.76–1.40)0.83
       Bilobectomy vs. lobectomy0.79(0.51–1.22)0.29
       Segmentectomy vs. lobectomy1.35(0.80–2.30)0.26
       Wedge resection vs. lobectomy1.84(1.25–2.71)0.0021
       Other or missing vs. lobectomy1.76(1.00–3.09)0.050
      Performance status at surg.
       1 or 2 or 3 vs. 01.24(0.97–1.59)0.085
       Missing vs. 01.33(1.09–1.62)0.0044
      Tumor size
       >4 vs. ≤41.31(1.07–1.61)0.011
      CI, confidence interval; HR, hazard ratio; Kr_G12C, KRAS-G12C; MND, mutation not detected; Pt, patient; RFS, relapse-free survival; Surg., surgery.
      a Excluding two patients with missing tumor size measurement.
      b The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.30 (0.98–1.72) and Kr_G12C versus KRAS MND: 1.23 (0.98–1.56).
      In “histologic-subtype’” patient cohort, (1061 RFS events, Supplementary Fig. S.B1b), adjusted for other characteristics, the negative effect of the Kr_G12C presence is found significant when compared with KRAS nonmutated patients (adjusted HR 17C versus no KRAS = 1.27, 95% CI: 1.04–1.54, p = 0.017, Supplementary Table S.B4b).
      With respect to TTR, reported medians for the adenocarcinoma patients are: 66.6 months, 95% CI: 39.5 months to NE for Kr_G12C mutated patients, versus 85.8 months, 95% CI: 66.2 months to NE for other-KRAS-mutations and 84.6 months, 95% CI: 59.4 months to NE for KRAS-nonmutated (Fig. 1C, log-rank p = not significant). A significant TTR difference emerges when accounting for stage, surgery anatomy and tumor size (Table 5), where the hazard of relapse is significantly higher for patients with the Kr_G12C mutation compared with those with other KRAS mutations by 41% (HR G12C versus other KRAS = 1.41, 95% CI: 1.03–1.92, p = 0.030). In addition, higher, although not reaching significance, is the hazard of relapse compared with KRAS-nonmutated patients, with HR G12C versus no KRAS equal to 1.27, 95% CI: 0.99 to 1.64, p value equal to 0.063.
      Table 5Multivariate Cox Proportional Hazards Model for TTR in Patients With Adenocarcinoma
      No. of Pts: 1012
      Excluding two patients with missing tumor size measurement.
      No. of Relapses: 430
      HR95% CIp Value
      KRAS
      The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.41 (1.03–1.92) and Kr_G12C versus KRAS MND: 1.27 (0.99–1.64).
       Other mutations vs. Kr_G12C0.71(0.52–0.97)0.030
       MND vs. Kr_G12C0.79(0.61–1.01)0.063
      Stage
       Ib vs. Ia1.07(0.77–1.49)0.68
       IIa vs. Ia2.31(1.65–3.25)<0.001
       IIb vs. Ia1.86(1.25–2.78)0.0023
       IIIa vs. Ia4.46(3.27–6.08)<0.001
       IIIb vs. Ia6.09(3.32–11.16)<0.001
      Surg. anatomy
       Pneumonectomy vs. lobectomy1.00(0.72–1.39)0.98
       Bilobectomy vs. lobectomy0.62(0.37–1.06)0.079
       Segmentectomy vs. lobectomy1.42(0.80–2.50)0.23
       Wedge resection vs. lobectomy2.03(1.32–3.10)0.0011
       Other or missing vs. lobectomy2.13(1.18–3.85)0.012
      Tumor size in cm
       >4 vs. ≤41.38(1.10–1.73)0.0048
      CI, confidence interval; HR, hazard ratio; Kr_G12C, KRAS-G12C; MND, mutation not detected; Pt, patient; Surg., surgery; TTR, time-to-relapse.
      a Excluding two patients with missing tumor size measurement.
      b The corresponding HR (95% CI) of Kr_G12C versus the other groups are the following: Kr_G12C versus all other KRAS mutations: 1.41 (1.03–1.92) and Kr_G12C versus KRAS MND: 1.27 (0.99–1.64).
      In the “histologic-subtype” cohort, strong significant differences of TTR between mutation subgroups are detected both in unadjusted analysis (log-rank) and in adjusted Cox models. Kr_G12C mutated patients (median: 70.2 mo, 95% CI: 39.5 mo–NE), have significantly lower TTR versus KRAS-nonmutated (104.0 mo, 95% CI: 98.6–NE, p = 0.039, Supplementary Fig. S.B1c). TTR differences are also significant when adjusting for: age, stage, surgery anatomy, histology and tumor size. The relapse hazard is significantly higher for patients with the Kr_G12C mutation compared with patients with other KRAS mutations or KRAS nonmutated, by 41% and 28%, respectively (HR G12C versus other KRAS = 1.41, 95% CI: 1.06–1.85, p = 0.016 and HR G12C versus no KRAS = 1.28, 95% CI: 1.03–1.59, p = 0.027, Supplementary Table S.B4c).
      Subgroup analysis by stage is provided in the Supplementary Data (Kaplan-Meier, Supplementary Figs. S.A1a–S.A3c and S.B2a–S.B4c). The purpose of this subgroup analysis is only exploratory and descriptive, because in all cases (adjusted Cox models for all three end points of OS, RFS, and TTR, in both the primary adenocarcinoma or "histologic-subtype" cohort) the interaction of the stage (I or II or III) with KRAS mutation status was not found to be significant. In general, the G12C-mutated patients do exhibit a numerically poorer performance, whereas the significance found in our overall cohort is mainly driven by the patients with stage III.
      Lastly, the exploratory sensitivity analysis, excluding all cases of co-mutations and other detected mutations, and also presenting separately EGFR mutations, is provided in the Supplementary Data (Kaplan-Meier, Supplementary Fig. S.A4a–c for the adenocarcinomas and Supplementary Fig. S.B5a–c for histologic-subtype patients). The poorer performance of G12C-mutated patients compared with other KRAS mutations (for all three end points) and wild-type cases (for RFS and TTR) is verified by the corresponding multivariate Cox models. Of note, in the OS case in particular, G12C-mutated patients perform significantly worse compared with EGFR-mutated patients.

      Discussion

      This is one of the largest cohorts of surgically resected European patients with NSCLC exploring associations of KRAS mutations with clinicopathologic features. The patient demographics are in line with what would be expected for such a cohort in terms of stage distribution and postoperative outcomes. As previously discussed,
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      the microfluidic-based polymerase chain reaction platform used for mutation testing was originally developed for testing multiple tumor types, with a corresponding range of clinically relevant genes and mutations. The platform provides an allele-specific range of mutations to be tested. Up to now, no gene alteration, except for ALK rearrangement, has proven to have a prognostic role in the Lungscape cohort.
      The mutations found and their relation to tumor histology and smoking status are in line with those of previous reports. By using limited multiplex technology focused on specific KRAS gene alterations, the assessment of variant allele frequency may differ compared with that obtained by a broader approach, such as whole-genome sequencing, because the latter offers broader coverage.
      As expected, KRAS mutation was one of the dominant findings in adenocarcinomas and its prevalence was 38%, predominantly represented by G12C change (17%; G12D the most common in never smokers
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      ). In the “histologic-subtype” cohort, apart from the predominance of adenocarcinomas, in the Kr_G12C mutation group, more female patients (44% versus 31%) and more patients with small tumors (69% versus 61%) are reported compared with KRAS-nonmutated patients.
      Furthermore, in the adenocarcinoma cohort, a difference in age was found between groups, of approximately 3 years. The median age was 63.1 in the Kr_G12C group of adenocarcinomas whereas it was 65.9 in other-KRAS-mutations group. The median age for KRAS mutations has been previously found to be associated with poor postoperative survival, but more recently, this has been challenged.
      • Yu H.A.
      • Sima C.S.
      • Shen R.
      • et al.
      Prognostic impact of KRAS mutation subtypes in 677 patients with metastatic lung adenocarcinomas.
      • Zer A.
      • Ding K.
      • Lee S.M.
      • et al.
      Pooled analysis of the prognostic and predictive value of KRAS mutation status and mutation subtype in patients with non-small cell lung cancer treated with epidermal growth factor receptor tyrosine kinase inhibitors.
      • Renaud S.
      • Seitlinger J.
      • Falcoz P.-E.
      • et al.
      Specific KRAS amino acid substitutions and EGFR mutations predict site-specific recurrence and metastasis following non-small-cell lung cancer surgery.
      In our adenocarcinoma cohort, a statistically significant Kr_G12C effect was found on OS when adjusting for ethnicity, age, stage, performance status, and tumor size. Notably, the hazard of death is higher for Kr_G12C-mutated patients compared with KRAS-nonmutated patients and other-KRAS-mutated patients by 32% and 39%, respectively (p = 0.028, 0.031). This significant effect is apparent in the “histologic-subtype” cohort too.
      The negative effect of Kr_G12C mutation is also found in RFS. In the full cohort of all patients with NSCLC (irrespective of histology), the risk of an RFS event, adjusted for clinicopathologic variables, is 27% higher for Kr_G12c patients compared with KRAS nonmutated patients (p = 0.017).
      The hazard of relapse is also higher for adenocarcinoma patients with Kr_G12C mutation compared with those with other KRAS mutations or KRAS nonmutated patients by 41% (p = 0.030) and 27% (p = 0.063), respectively. These differences are both significant when evaluated in the cohort of all patients, irrespective of histology.
      The significant (negative) prognostic effect of Kr_G12C mutation on resected stage I to III NSCLC, adjusted for clinicopathologic characteristics, and consistent for different efficacy end points evaluated in both the adenocarcinoma cohort and in the “histologic-subtype” cohort, is of particular importance taking into account the fact that in the same cohort of patients
      • Kerr K.M.
      • Dafni U.
      • Schulze K.
      • et al.
      Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
      the overall KRAS effect and the effect of KRAS codon 12 were not found to be significant. These findings have important clinical impact. First, after decades of failures, Kr_G12C mutation has emerged as a possible therapeutic target and different treatment are under investigation in clinical trials.
      • Canon J.
      • Rex K.
      • Saiki A.Y.
      • et al.
      The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity.
      ,
      • Hallin J.
      • Engstrom L.D.
      • Hargis L.
      • et al.
      The KRAS G12C inhibitor MRTX849 provides insight toward therapeutic susceptibility of KRAS-mutant cancers in mouse models and patients.
      Second, they are relevant for the possible future adoption of such drugs in the adjuvant setting, even more in a context of a disease with higher relapse rate and poorer OS.
      In the metastatic setting, unlike other patients, those with oncogene-addicted NSCLC with KRAS mutation seem to derive benefit from immune checkpoint inhibitors (ICIs), similarly to wild-type patients.
      • Torralvo J.
      • Friedlaender A.
      • Achard V.
      • Addeo A.
      The activity of immune checkpoint inhibition in KRAS mutated non-small cell lung cancer: a single centre experience.
      ,
      • Friedlaender A.
      • Drilon A.
      • Weiss G.J.
      • Banna G.L.
      • Addeo A.
      KRAS as a druggable target in NSCLC: rising like a phoenix after decades of development failures.
      In an exploratory analysis based on the KEYNOTE-042 trial of front-line pembrolizumab in advanced NSCLC, the role of KRAS mutations seems to be predictive of numerically superior objective response rate (57% versus 29%), median progression-free survival (12 mo versus 6 mo), and OS (28 mo versus 15 mo).
      • Herbst R.S.
      • Lopes G.
      • Kowalski D.M.
      • et al.
      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.
      This difference was even more evident in the Kr_G12C subset. There is indeed a potential biological rationale for a positive interaction between KRAS and Kr_G12C mutations and ICI benefit. These tumors often exhibit increased tumor mutation burden that might lead to increased ICI sensitivity.
      • Dong Z.Y.
      • Zhong W.Z.
      • Zhang X.C.
      • et al.
      Potential predictive value of TP53 and KRAS mutation status for response to PD-1 blockade immunotherapy in lung adenocarcinoma.
      Furthermore, as we know, KRAS mutations are more common in smokers, and smoking is associated with elevated somatic tumoral DNA mutations and in general higher tumor mutation burden.
      • Addeo A.
      • Banna G.L.
      • Weiss G.J.
      Tumor mutation burden—from hopes to doubts.
      ,
      • Rizvi N.A.
      • Hellmann M.D.
      • Snyder A.
      • et al.
      Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer.
      Having said that, the role of ICIs in the adjuvant setting is yet to be established. Many phase III trials exploring the possible benefit of ICI after complete lung resection are still ongoing. No data have been released yet. As soon as the readouts will become available, it will be worth exploring the impact and benefit of ICI in the subgroup of Kr_G12C-mutated patients. A strong positive signal favoring ICI in this subgroup could potentially lead to further explore the Kr_G12C mutation as a predictive biomarker for ICI in the adjuvant setting.
      A limitation of the analysis is the lack of information on the impact of other co-mutations typically available with panel next-generation sequencing, particularly tumor-suppressor genes. Indeed, most of the current targeted panels in clinical use do not detect tumor-suppressor genes. However, the occurrence of co-mutation in other driver oncogenes in KRAS-driven tumors is rare (reviewed by Skoulidis et al.
      • Skoulidis F.
      • Heymach J.V.
      Co-occurring genomic alterations in non-small-cell lung cancer biology and therapy.
      ). Concomitant alterations may affect the immunogenicity of KRAS-mutant tumors. For instance, cooccurring TP53 mutations are associated with enhanced tumor cell proliferation and inflammation, allowing for an immune-rich microenvironment. In contrast, cooccurring STK11 mutations which could be present in up to 20% of KRAS-mutant NSCLC could decrease immune surveillance by modulating the NF-kB pathway.
      • Dong Z.Y.
      • Zhong W.Z.
      • Zhang X.C.
      • et al.
      Potential predictive value of TP53 and KRAS mutation status for response to PD-1 blockade immunotherapy in lung adenocarcinoma.
      ,
      • Skoulidis F.
      • Goldberg M.E.
      • Greenawalt D.M.
      • et al.
      STK11/LKB1 mutations and PD-1 inhibitor resistance in KRAS-mutant lung adenocarcinoma.
      Further analysis might clarify what and if there is a clear role of these co-mutations.
      Our findings emphasize the potential benefits of merging large databases, to identify and describe rare, but potentially clinically useful, mutation profiles. It also highlights the possible role of Kr_G12C mutation as prognostic factor for OS and RFS. This may be of relevance as treatments directed against Kr_G12C become established in clinical practice.

      Acknowledgments

      The analysis of the prognostic impact did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The original KRAS research was supported by F. Hoffmann-La Roche Ltd. and Genentech Inc. , South San Francisco, California.

      Appendix

      Lungscape Consortium
      European Thoracic Oncology Platform Lungscape Contributors
      Lungscape 003 Multiplex Mutation Testing
      European Thoracic Oncology Platform Lungscape Consortium
      Lungscape Steering Committee: Rolf A. Stahel, Rafael Rosell, Fiona Blackhall, Urania Dafni, Keith M. Kerr, Miguel Ángel Molina, Lukas Bubendorf, Walter Weder, Erik Thunnissen, Solange Peters, Stephen Finn
      Coordinating Center, Bern, Switzerland: Roswitha Kammler, Nesa Marti, Thomas R. Geiger, Barbara Ruepp
      Statistical Center: Frontier Science Foundation-Hellas, Athens, Greece: Urania Dafni, Zoi Tsourti, Marie Kassapian, Varvara Polydoropoulou, Georgia Dimopoulou
      Study Support: Genentech Inc., South San Francisco, USA: David Shames, Katja Schulze; Genentech Central Laboratory: Teiko Sumiyoshi, An Do, Rachel Tam, Anna Cheung.
      Lungscape Collaborating Sites, listed by contributions
      Pangaea Oncology, Barcelona, Spain: Miguel A Molina, Jordi Bertrán, Rafael Rosell
      St James’s Hospital Dublin, Dublin, Ireland: Stephen Finn, Kathy Gately, Sinead Cuffe, Ronan Ryan
      University Hospital Aarhus, Aarhus, Denmark: Peter Meldgaard, Henrik Hager, Line B. Madsen
      University Hospital KU Leuven, Leuven, Belgium: Johan Vansteenkiste, Els Wauters, Stefanie Lepers, Sara Van Der Borght
      Medical University Gdansk, Gdansk, Poland: Rafal Dziadziuszko, Wojciech Biernat, Ania Wrona, Jacek Jassem
      Vall d’Hebron University Hospital, Barcelona, Spain: Enriqueta Felip, Javier Hernandez-Losa, Irene Sansano
      Center of Predictive Molecular Medicine, CeSI-MeT, Chieti, Italy: Antonio Marchetti, Alessia Di Lorito, Graziano De Luca, Sara Malatesta
      Royal Infirmary Aberdeen, Aberdeen, United Kingdom: Keith M Kerr, Marianne Nicolson, David AJ Stevenson, William Mathieson
      University Hospital Basel, Basel, Switzerland: Lukas Bubendorf, Spasenija Savic, Didier Lardinois
      Free University Medical Center (VUMC), Amsterdam, The Netherlands: Erik Thunnissen, Egbert Smit, Coralien van Setten, Joop de Langen
      Maastricht University Medical Center, Maastricht, The Netherlands: Anne-Marie Dingemans, Ernst-Jan M. Speel
      Roswell Park Cancer Institute, Buffalo, New York: Richard Cheney, Mary Beth Pine, Mary Reid, Elizabeth Taylor
      The Netherlands Cancer Institute, Amsterdam, The Netherlands: Paul Baas, Jeroen de Jong, Kim Monkhorst
      Lung Cancer Group Manchester, Manchester, United Kingdom: Fiona Blackhall, Daisuke Nonaka, Anne Marie Quinn, Lynsey Franklin
      Institute of Pathology and Thoracic Hospital at Heidelberg University, Heidelberg, Germany: Hendrik Dienemann, Thomas Muley, Arne Warth
      General University Hospital Valencia, Valencia, Spain: Carlos Camps, Miguel Martorell, Eloisa Jantus-Lewintre, Ricardo Guijarro
      University Hospital Zürich, Zurich, Switzerland: Walter Weder, Isabelle Opitz, Alex Soltermann, Alessandra Curioni

      Supplementary Data

      References

        • Lawrence M.S.
        • Stojanov P.
        • Polak P.
        • et al.
        Mutational heterogeneity in cancer and the search for new cancer-associated.
        Nature. 2013; 499: 214-218
        • Yuan M.
        • Huang L.L.
        • Chen J.H.
        • Wu J.
        • Xu Q.
        The emerging treatment landscape of targeted therapy in non-small-cell lung cancer.
        Signal Transduct Target Ther. 2019; 4: 61
        • Gadgeel S.
        • Peters S.
        • Mok T.
        • et al.
        Alectinib versus crizotinib in treatment-naive anaplastic lymphoma kinase-positive (ALK+) non-small-cell lung cancer: CNS efficacy results from the ALEX study.
        Ann Oncol. 2018; 29: 2214-2222
        • Soria J.C.
        • Ohe Y.
        • Vansteenkiste J.
        • et al.
        Osimertinib in untreated EGFR-mutated advanced non-small-cell lung cancer.
        N Engl J Med. 2017; 378: 113-125
        • Camidge D.R.
        • Kim H.R.
        • Ahn M.J.
        • et al.
        Brigatinib versus crizotinib in ALK-positive non-small-cell lung cancer.
        N Engl J Med. 2018; 379: 2027-2039
        • Dong Z.Y.
        • Zhong W.Z.
        • Zhang X.C.
        • et al.
        Potential predictive value of TP53 and KRAS mutation status for response to PD-1 blockade immunotherapy in lung adenocarcinoma.
        Clin Cancer Res. 2017; 23: 3012-3024
        • Friedlaender A.
        • Addeo A.
        • Russo A.
        • Gregorc V.
        • Cortinovis D.
        • Rolfo C.D.
        Targeted therapies in early stage NSCLC: hype or hope?.
        Int J Mol Sci. 2020; 21: 6329
        • Blackhall F.H.
        • Peters S.
        • Bubendorf L.
        • et al.
        Prevalence and clinical outcomes for patients with ALK-positive resected stage I to III adenocarcinoma: results from the European Thoracic Oncology Platform Lungscape project.
        J Clin Oncol. 2014; 32: 2780-2787
        • Kerr K.M.
        • Dafni U.
        • Schulze K.
        • et al.
        Prevalence and clinical association of gene mutations through multiplex mutation testing in patients with NSCLC: results from the ETOP Lungscape project.
        Ann Oncol. 2018; 29: 200-208
        • Torralvo J.
        • Friedlaender A.
        • Achard V.
        • Addeo A.
        The activity of immune checkpoint inhibition in KRAS mutated non-small cell lung cancer: a single centre experience.
        Cancer Genomics Proteomics. 2019; 16: 577-582
        • Friedlaender A.
        • Drilon A.
        • Weiss G.J.
        • Banna G.L.
        • Addeo A.
        KRAS as a druggable target in NSCLC: rising like a phoenix after decades of development failures.
        Cancer Treat Rev. 2020; 85: 101978
        • Canon J.
        • Rex K.
        • Saiki A.Y.
        • et al.
        The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity.
        Nature. 2019; 575: 217-223
        • Hallin J.
        • Engstrom L.D.
        • Hargis L.
        • et al.
        The KRAS G12C inhibitor MRTX849 provides insight toward therapeutic susceptibility of KRAS-mutant cancers in mouse models and patients.
        Cancer Discov. 2020; 10: 54-71
        • Rulle U.
        • Tsourti Z.
        • Casanova R.
        • et al.
        Computer-based intensity measurement assists pathologists in scoring phosphatase and tensin homolog immunohistochemistry—clinical associations in NSCLC patients of the European Thoracic Oncology Platform Lungscape Cohort.
        J Thorac Oncol. 2018; 13: 1851-1863
        • Bubendorf L.
        • Dafni U.
        • Schöbel M.
        • et al.
        Prevalence and clinical association of MET gene overexpression and amplification in patients with NSCLC: results from the European Thoracic Oncology Platform (ETOP) Lungscape project.
        Lung Cancer. 2017; 111: 143-149
        • Kerr K.M.
        • Thunnissen E.
        • Dafni U.
        • et al.
        A retrospective cohort study of PD-L1 prevalence, molecular associations and clinical outcomes in patients with NSCLC: results from the European Thoracic Oncology Platform (ETOP) Lungscape project.
        Lung Cancer. 2019; 131: 95-103
        • Peters S.
        • Weder W.
        • Dafni U.
        • et al.
        Lungscape: resected non-small-cell lung cancer outcome by clinical and pathological parameters.
        J Thorac Oncol. 2014; 9: 1675-1684
        • Yu H.A.
        • Sima C.S.
        • Shen R.
        • et al.
        Prognostic impact of KRAS mutation subtypes in 677 patients with metastatic lung adenocarcinomas.
        J Thorac Oncol. 2015; 10: 431-437
        • Zer A.
        • Ding K.
        • Lee S.M.
        • et al.
        Pooled analysis of the prognostic and predictive value of KRAS mutation status and mutation subtype in patients with non-small cell lung cancer treated with epidermal growth factor receptor tyrosine kinase inhibitors.
        J Thorac Oncol. 2016; 11: 312-323
        • Renaud S.
        • Seitlinger J.
        • Falcoz P.-E.
        • et al.
        Specific KRAS amino acid substitutions and EGFR mutations predict site-specific recurrence and metastasis following non-small-cell lung cancer surgery.
        Br J Cancer. 2016; 115: 346-353
        • Herbst R.S.
        • Lopes G.
        • Kowalski D.M.
        • et al.
        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.
        Ann Oncol. 2019; 30: XI63-XI64
        • Addeo A.
        • Banna G.L.
        • Weiss G.J.
        Tumor mutation burden—from hopes to doubts.
        JAMA Oncol. 2019; 5: 934-935
        • Rizvi N.A.
        • Hellmann M.D.
        • Snyder A.
        • et al.
        Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer.
        Science. 2015; 348: 124-128
        • Skoulidis F.
        • Heymach J.V.
        Co-occurring genomic alterations in non-small-cell lung cancer biology and therapy.
        Nat Rev Cancer. 2019; 19: 495-509
        • Skoulidis F.
        • Goldberg M.E.
        • Greenawalt D.M.
        • et al.
        STK11/LKB1 mutations and PD-1 inhibitor resistance in KRAS-mutant lung adenocarcinoma.
        Cancer Discov. 2018; 8: 822-835