Mutation in the RAS proto-oncogene, one of the most common mutations in NSCLC, is found in 15–30 % NSCLC. Most RAS mutations in NSCLC occur in KRAS. However NRAS and HRAS mutations occasionally have been documented in lung cancer [4]. In this chapter, after a description of the epidemiology of KRAS mutations, we review: (1) the prognostic impact of KRAS mutations in NSCLC, (2) the predictive value of KRAS mutation for standard treatments (chemotherapy and targeted therapy) and (3) therapeutic approaches to targeting KRAS mutations and downstream pathways.
Epidemiology of KRAS Mutation in NSCLC
KRAS mutations occur more frequently in lung adenocarcinomas (~20–30 %) and less frequently in squamous cell carcinoma (~5 %) [4]. KRAS also displays variation in frequency according to patient ethnicity: 25–50 % among patients of Caucasian decent and 5–15 % among those of Asian origin [9]. Many studies have found a significant association between KRAS mutations and tobacco consumption and KRAS mutations are more common in current or former smokers compared with life-time non-smokers [4]. Moreover, DNA damage as a result of tobacco smoke causes certain types of mutations. In both KRAS and p53, transversion (substituting a pyrimidine for a purine or purine for a pyrimidine e.g. G → T or G → C) are more frequent in current and former smokers while transition (substituting purine for purine or pyrimidine for pyrimidine e.g. G → A) occurs more often in never smokers [10, 11]. Usually, KRAS mutations in NSCLC are single amino acid substitutions in codon 12 of exon 2 and less commonly in codons 13 of exon 2 (approximately 10 %) and rarely 61 in exon 3 [4]. The tranversions GGT > TGT (leading to a glycine to cysteine substitution in codon 12, G12C) and GGT > GTT (glycine to valine substitution in codon 12, G12V), are the most frequent KRAS mutations in NSCLC (39–40 and 20–21 %, for G12C and G12V, respectively) [12, 13], followed by the transitions GGT > GAT (glycine to aspartate substitution in codon 12, G12D) and GGT > GCT (glycine to alanine in codon 12, G12A, are found in 17 % and 6–10 % of NSCLC, respectively [12, 13], and finally more rarely, ~3 %, the mutations GGC > TGC (glycine to valine substitution in codon 13, G13V) and GGT > AGT (glycine to serine substitution codon 12, G12S).
Recent broad-spectrum mutation profiling studies have shown that KRAS mutation can occur simultaneously with other mutations in NSCLC. Even though EGFR and KRAS mutations were thought to be mutually exclusive, some rare cases of concomitant EGFR and KRAS mutations have been reported [14]. In France, routine nation-wide testing of EGFR, KRAS, HER2, BRAF and PI3K mutations as well as ALK gene rearrangement is performed through 28 centers resulting in molecular data on the largest cohort available worldwide. Among the first 10,000 NSCLC patients tested a known target was identified in 47 % of samples and 56.9 % of the patients received a bio-guided treatment. KRAS mutations were detected in 27 % of the patients; they were more frequent in smokers (31.7 %) than in never smokers (9.6 %). Double mutations were found in 79 patients of which 44 had KRAS mutations associated another mutation, although only five occurred with EGFR mutation, 10 with ALK rearrangement, six with BRAF and 33 with PI3K mutations [14]. In the US, the Lung Cancer Mutation Consortium (LCMC) assessed the frequency of ten oncogenic drivers in tumors from 1,000 patients with advanced lung adenocarcinomas. Among 1,007 patients, 63 % had a known oncogenic driver and 28 % potentially could benefit from targeted therapy. In total, 25 % had KRAS mutation and 3 % had two drivers (specific drivers not reported). The patients with drivers who received targeted therapy had a better outcome than those who did not have targeting therapy [15].
Prognostic Value of KRAS Mutations
Prognostic factors are patient and tumor factors that predict patient outcome (usually survival) and are independent of treatment administered. The strongest clinical prognostic factors in NSCLC include stage, sex, age and performance status [16]. KRAS was the first oncogene reported to be a negative prognostic factor for lung adenocarcinoma 20 years ago in a small surgical series [17]. Since then, the prognostic significance of KRAS has been investigated extensively in NSCLC with inconsistent results (Table 11.1), likely due to considerable heterogeneity among studies using different laboratory techniques to identify mutations and inclusion of different patient populations among the studies.
Table 11.1
Selected studies of the prognostic role of KRAS mutations in NSCLC
Author/trial | N in trial/N with KRAS results | KRAS status (%) | PFS (HR; 95 % CI) | PFS p-value | OS (HR; 95 % CI) | OS p-value | |
|---|---|---|---|---|---|---|---|
Mutation | Wild-type | ||||||
Capelletti et al. (2010) [18] CALGB9633 | 344/258 | 71 (27 %) | 187 (73 %) | NR | NR | 1.1 | 0.747 |
Scoccianti et al. (2012) [19] EUELC | 762/249 | 46 (18 %) | 203 (81 %) | 1.30 (0.82–2.06) | 0.26 | NR | NR |
Ma et al. (2008) [20] IALT | 1,867/718 | 98 (14 %) | 620 (86 %) | NR | 0.03 | NR | 0.31 |
Schiller et al. (2001) [21] ECOG 4592 | 217/184 | 44 (24 %) | 140 (76 %) | NR | NR | NR | 0.38 |
Tsao et al. (2007) [22] JBR.10 | 482/450 | 117 (26 %) | 333 (74 %) | NR | NR | 1.23 (0.76–1.97) | 0.40 |
Pooled analyses and meta–analyses | |||||||
Mascaux et al. (2005) [23] | 5,216/3,779 | 695 (18 %) | 3,084 (82 %) | NR | NR | 1.30 (1.20–1.49) | 0.01 |
Meng et al. (2013) [24] | 6,939/6,939 | NR | NR | NR | NR | 1.45 (1.29–1.62) | NR |
Shepherd et al. (2013) [25] LACE-BIO | 1,718/1,543 | 300 (19.4 %) | 1,243 (81.6 %) | 1.05 (0.80–1.36) | 0.73 | 1.04 (0.78–1.38) | 0.79 |
Shepherd et al. (2013) [25] LACE-BIO ADC | 813/605 | 204 (33.7 %) | 401 (66.3 %) | 0.98 (0.78–1.24) | 0.87 | 1.00 (0.78–1.29) | 0.97 |
Randomized controlled trials testing adjuvant therapy have provided large prospective cohorts for assessment of the prognostic role of KRAS mutations in early stage NSCLC. The Eastern Cooperative Oncology Group (ECOG) E4592 randomized trial assessed adjuvant thoracic radiation +/− four cycles of cisplatin/etoposide chemotherapy in patients with resected stage II–IIIA NSCLC [21]. Among 184 assessable tumors, KRAS mutations were detected in 44 (24 %), in 4.8 % of tumors with squamous carcinomas histology and 33 % of these with non-squamous histology, respectively (p < 0.05). Although there was a trend for a prognostic role of KRAS mutations by multivariate analyses (p = 0.066), the median overall survival (OS) of patients with KRAS mutation was not statistically different (30 versus 42 months for patient with KRAS mutations versus wild-type, respectively, p = 0.38) [21].
In the North American intergroup trial JBR.10, including 482 patients with resected stage IB–I NSCLC, stratified by KRAS and randomized to receive four cycles of adjuvant cisplatin/vinorelbine or observation alone, mutations in RAS genes were detected in 113 (26 %) patients. In the observation arm, RAS mutation was again not a significant prognostic factor for OS (p = 0.40) [22].
In the European Early Lung Cancer (EUELC) cohort of 762 patients with resected NSCLC, the International Agency for Research on Cancer (IACR) detected KRAS mutations in 18.5 % of the available samples. Mutations were detected more frequently in adenocarcinoma (30.6 %) than in squamous cell carcinoma (4.3 %). KRAS mutations was not found to be a significant prognostic biomarker for progression-free survival (PFS) (p = 0.26) [19].
In the International Adjuvant Lung Cancer Trial (IALT), including 1,867 patients randomized to receive post-operative cisplatin-based adjuvant chemotherapy or observation. KRAS mutations were detected in only 14 % of assessable samples. A significant unfavourable effect of KRAS mutations on disease-free survival (DFS) was observed (p = 0.03), but not on OS (p = 0.31). However, in the small non-squamous/non-adenocarcinoma subset, there was a significant negative prognostic effect both for OS and DFS (p = 0.04 and p = 0.006, respectively) [20]
The Cancer and Leukemia Group B-9633 (CALGB-9633) phase III trial, including patients with stage IB non-small cell lung cancer randomized to observation or four cycles of carboplatin/paclitaxel, KRAS mutations were detected in 27 % of available samples and did not have any significant prognostic value (p = 0.747) [18].
As published individual studies showed discordant results, a large meta-analysis of 28 studies with a total of 3,620 patients was performed in 2005 to assess the prognostic significance of KRAS mutations on survival of patients with lung cancer [23]. Overall, the presence of KRAS mutations was a negative prognostic factor for OS (HR 1.30, 95 % CI 1.20–1.49, p = 0.01). In the subgroup of patients with adenocarcinoma (n = 1,436 patients), KRAS was prognostic for OS (HR 1.52, 95 % CI 1.30–1.78, p = 0.02) but not in squamous histology (n = 280 patients, HR 1.49, 95 % CI: 0.88–2.52; p = 0.48). KRAS mutation was a significant prognostic marker when polymerase chain reaction (PCR) sequencing (n = 2,631 patients, HR 1.40; 95 % CI 1.18–1.65; p = 0.03) was used as the detection method. No significant prognostic impact on survival was found when analysing early versus advanced stage disease [23].
In 2013, a larger meta-analysis was published based on 41 trials with a total of 6,939 patients [24]. This meta-analysis confirmed the overall unfavourable impact of KRAS mutations on survival of patients with NSCLC (HR 1.45, 95 % CI 1.29–1.62). Results were similar both in Asian patients (n = 1,524 patients, HR 1.97, 95 % CI 1.58–2.44) and in non-Asian patients (n = 4,856 patients, HR 1.37, 95 % CI 1.25–1.5), and the significant impact of KRAS mutation in adenocarcinomas was confirmed (n = 3,502 patients, HR 1.39, 95 % CI 1.24–1.55). The investigators did not aggregate the results for squamous cell carcinomas. A negative significant impact on survival was observed for early stage disease (stage I, n = 535 patients, HR 1.81, 95 % CI 1.36–2.39; stage I–IIIa, n = 474, HR 1.68, 95 % CI 1.11–2.55), but not for advanced stage disease (stage IIIb–IV, n = 975, HR 1.3, 95 % CI 0.99–1.71). The results were independent of the type of PCR used to detect the KRAS mutation: mutation-specific oligonucleotide probe (MSOP), denaturating gradient gel eletrophoresis (DGGE), RLFP and direct method sequencing. The combined HR for the 13 studies focusing on codon 12 also showed a negative prognostic impact on survival of this subgroup (n = 1,665 patients, HR 1.71, 95 % CI 1.44–2.04) [24].
In 2013, the Lung Adjuvant Cisplatin Evaluation-Bio (LACE-Bio) group published the results of a pooled analysis of KRAS mutations on 1,718 patients from four randomized trials comparing adjuvant chemotherapy (ACT) or observation (OBS) (ANITA, IALT, JBR.10, CALBG-9633) [25]. Analyses were performed in a blinded fashion in three laboratories by restriction fragment length polymorphism (RFLP), allelic specific oligonucleotide hybridization, or allelic refractory mutation system analysis and mass spectrometry; these methods have been reported to be more sensitive than direct sequencing. Analyses were successful in 1,543 samples (763 OBS and 780 ACT) and mutations were detected in 300 (19.4 %), with 275 in codon 12, 13 in codon 13 and 1 in codon 14, respectively. Consistently, KRAS mutations were more frequent in adenocarcinoma than in squamous cell carcinoma (34 versus 6 %, p < 0.001). KRAS mutations also were more frequent in female than in male patients (27 versus 17 %, p = 0.001), younger patients (trend p = 0.0003). In multivariate analysis, only age (p = 0.044) and histology (p < 0.001) remained significant. In the OBS arm, there was no prognostic effect of KRAS mutation status for OS (HR 1.04, 95 % CI 0.78–1.38, p = 0.79) or DFS (HR 1.05, 95 % CI 0.80–1.36, p = 0.73), with no significant heterogeneity among trials (p = 0.47). Importantly, KRAS was not prognostic in the adenocarcinoma subgroup for OS (HR 1.00, 95 % CI 0.78–1.29, p = 0.97) or PFS (HR 0.98, 95 % CI 0.78–1.24, p = 0.87). Trends for worse outcome were seen with KRAS mutations with non-adenocarcinoma tumors (squamous cell carcinoma HR 1.41, 95 % CI 0.89–2.23 and other non-adenocarcinoma HR 1.86, 95 % CI 1.22–2.82). There was no difference in prognosis for OS for codon 12 (mutation versus wild type, HR 1.04, CI 0.77–1.40) or codon 13 (mutation versus wild type HR 1.01, CI 0.47–2.17) mutations in the OBS arm [25]. There was also no significant difference in prognosis for the different codon 12 subgroups for OS (G12C or G12V versus wild type HR 1.04, 95 % CI 0.74–1.46, G12D or G12S versus wild type HR 0.95, 95 % CI 0.50–1.81 and G12A or G12R versus wild type HR 1.08, 95 % CI 0.49–2.37, interaction p value = 0.99) or for PFS (G12C or G12V versus wild type HR 1.04, 95 % CI 0.76–1.42, G12D or G12S versus wild type HR 1.03, 95 % CI 0.57–1.85 and G12A or G12R versus wild type HR 1.15, 95 % CI 0.55–2.39, interaction p value = 0.98). The same group further assessed the role of combined KRAS mutations with p53 mutations. No significant prognostic role of double KRAS and p53 mutations was found, but the group of patients with double mutations for KRAS and p53 was very small (24 patients with 9 deaths), and so the analyses was lacking statistical power.
Predictive Value of KRAS Mutation
Predictive factors are clinical, cellular, and molecular markers that predict tumor response to treatment (either in terms of tumor shrinkage or a survival benefit from treatment). In contrast to prognostic factors defining the effects of tumor characteristics on the patient, predictive factors define the effect of treatment on the tumor. Those measures are not always similar, as tumor response may not necessarily translate into greater survival benefit [26].
Predictive Role of KRAS Mutation Status for Epidermal Growth Factor Receptor Inhibitors
KRAS is a downstream effector from EGFR. When the intracellular tyrosine kinase of EGFR is activated after binding its ligand, the sequence homology 2 (SH2) binds the protein growth factor receptor-bound protein 2 and induces the recruitment of SOS1 and SOS2, RAS guanine nucleotide exchange factors and the dissociation of GDP, allowing binding of GTP to RAS and its activation [27]. Consequently, it was hypothesized that constitutive activation of the mutated protein KRAS would activate its downstream pathway and cell proliferation independently of upstream EGFR inhibition and therefore, could induce resistance to EGFR inhibitors. The value of KRAS mutation testing has been established in metastatic colorectal cancer where EGFR monoclonal antibodies have shown greater efficacy in patients with KRAS wild-type tumors [28, 29]. Multiple trials have assessed the predictive value of KRAS mutation for sensitivity to EGFR inhibitors in NSCLC patients (Table 11.2).
Table 11.2
Selected trials evaluating the role of KRAS mutation as a predictive marker for EGFR TKI therapy in NSCLC
Author/study | Study design | KRAS status (%) | PFS | OS | |||
|---|---|---|---|---|---|---|---|
N in trial/N with KRAS results | WT | Mut | WT | Mut | WT | Mut | |
Eberhard et al. (2005) [30] TRIBUTE | Paclitaxel/Carboplatin + erlotinib/placebo 1,070/264 | 209 (79.2 %) | 55 (20.8 %) | E – TTP 5.3 mo (CI 4.4–6.1 mo) | E – TTP 3.4 mo (CI 1.5–6.3 mo) | E – 12.1 mo (CI 9.2–15.6 mo) | E – 4.4 mo (CI 3.4–12.9 mo) |
P –TTP 5.4 mo (CI 4.4–6.1) | P – TTP 6 mo (CI 4.9–7.1) | P – 11.3 mo (CI 9.1-NR) | P – 13.5 mo (CI 11.1–15.9) | ||||
Interaction p NR | Interaction p NR | ||||||
Zhu et al. (2008) [31] BR.21 | Erlotinib versus placebo in advanced NSCLC 731/206 | 176 (85 %) | 30 (15 %) | NR | HR = 0.69 (CI 0.49–0.97) p = 0.03 | HR 1.67 (0.62–4.50) p = 0.31 | |
Interaction p NR | Interaction p = 0.0059 | ||||||
Brugger et al. (2011) [32] SATURN | Maintenance erlotinib versus placebo in stable and responding patients following first-line platinum-doublet chemotherapy 889/493 | 403 (82 %) | 90 (18 %) | HR = 0.70; (CI 0.59–0.87) p < 0.001 | HR = 0.77; (CI 0.50–1.19) p = .2246 | HR = 0.86 (CI 0.68–1.08) p = NR | HR = 0.79(CI 0.49–1.27) P = NR |
Interaction p = 0.886 | Interaction p = 0.891 | ||||||
Goss et al. (2013) [33] BR.19 | Gefitinib versus placebo in completely resected NSCLC 503/350 | 382 (72,6 %) | 96 (27,4 %) | HR = 1.08 (CI 0.74–1.59) p = 0.69 | HR = 1.77 (CI 1.00–3.13) p = 0.05 | HR = 1.13 (CI 0.78–1.65) p = 0.51 | HR = 1.51 (CI 0.84–2.70) p = 0.16 |
Interaction p = 0.15 | Interaction p = 0.36 | ||||||
Lee et al. (2012) [34] TOPICAL | Erlotinib versus placebo in first line for advanced NSCLC unsuitable for chemotherapy 670/390 | 317 (81 %) | 73 (19 %) | E – 2.7 mo (CI 2.2–2.9) | E – 3.5 mo (CI 1.7–4.8) | E – 3.7 mo (CI 2.8–4.2) | E – 4.2 mo (CI 1.8–6.2) |
P – 2.6 mo (CI 2.3–2.9) p = NR | P – 2.7 mo (CI 1.8–3.9) p = NR | P – 3.4 mo (CI 2.7–4.3) p = NR | P- 3.6 mo (CI 1.9–4.4) p = NR | ||||
Interaction p NR | Interaction p NR | ||||||
Johnson et al. (2013) [35] ATLAS | Bevacizumab with or without erlotinib in maintenance after first line chemotherapy with bevacizumab for advanced NSCLC 1,145/332 | 239 (72.0 %) | 93 (28 %) | HR = 0.67 (CI 0.49–0.91) p = 0.01 | HR = 0.93 (CI 0.55–1.56) p = 0.77 | NR | |
Interaction p NR | Interaction p NR | ||||||
Sequist et al. (2011) [36] ARQ-197-209 | Erlotinib plus tivantinib versus placebo in previously treated NSCLC 167/65 % | NR | 15 | HR 1.01 (CI 0.63–1.60) p = 0.977 | HR 0.18 (CI 0.05–0.70) p = 0.13 | NR | HR = 0.43 (CI 0.12–1.50) p = 0.17 |
Interaction p < 0.006 | Interaction p NR | ||||||
Scagliotti et al. (2013) [37] MARQUEE | Erlotinib plus tivantinib versus placebo in 2nd or 3rd line for advanced NSCLC of non-squamous histology 1,048/986 | 764 (72.9 %) | 284 (27.1 %) | HR = 1.01 | HR = 0.18 | HR = 0.94 (CI 0.77–1.14) | HR = 1.04 (CI 0.78–1.40) |
Interaction p NR | Interaction p NR | ||||||
Douillard et al. (2009) [38] INTEREST | Gefitinib versus docetaxel in 2nd line NSCLC 1,466/275 | 226 (82 %) | 49 (18 %) | HR = 1.23 (CI, 0.90–1.68) p = 0.20 | HR = 1.16 (CI, 0.56 v 2.41) p = 0.68 | HR = 1.03 (CI, 0.77–1.37) p = 0.86 | HR = 0.81 (CI, 0.44–1.49) p = 0.50 |
Interaction p NR | Interaction p = 0.51 | ||||||
Garassino et al. (2012) [39] TAILOR | Erlotinib versus docetaxel in 2nd line EGFR wild type NSCLC 222/219 | 167 (76.3 %) | 52 (23.7 %) | HR = 0.65 (CI 0.46–0.90) | HR = 0.84 (CI 0.47–1.52) | NR | |
Interaction p = 0.237 | Interaction p NR | ||||||
EGFR Tyrosine Kinase Inhibitor (TKIs)
The role of KRAS mutation as a predictive biomarker for outcome to EGFR inhibitors has been studied in five trials that compared erlotinib/gefitinib (alone or in combination) versus placebo (TRIBUTE, BR.21, SATURN, BR.19 and TOPICAL). The TRIBUTE phase III randomized trial assessed first-line chemotherapy plus erlotinib versus placebo in advanced NSCLC [30]. KRAS mutations were detected in 21 % of the 274 tested patients. Mutations were associated with significantly shorter TTP in patients treated by the combination of erlotinib plus chemotherapy than those treated with chemotherapy alone (3.4 versus 6 months, p = 0.03). The same negative predictive value of KRAS mutation was observed for OS (4.4 versus 13.5 months, respectively, p = 0.019) [30]. The NCIC CTG BR.21 phase III randomized placebo controlled trial assessed the role of erlotinib in advanced NSCLC after failure of standard chemotherapy [31]. KRAS mutations were assessed in 206 tumors of which 30 (15 %) were positive. Significant benefit in OS was observed with erlotinib in patients with KRAS wild type tumors (HR 0.69, 95 % CI 0.49–0.97, p = 0.03) but not in patients with KRAS mutations (HR 1.67, 95 % CI 0.62–4.50, p = 0.31) Response rates were 10 % for those with wild-type KRAS and 5 % for those with KRAS mutations. In the Cox model, however, the interaction between KRAS mutation status and treatment arm was only marginally significant (interaction p = 0.09), and was not significant in multivariable analysis (p = 0.13), despite the apparent trend in univariate analyses [31]. In the Sequential Tarceva® in unresectable NSCLC (SATURN) trial, patients with stage IIIB/IV NSCLC who had not experienced disease progression after four cycles of platinum-based therapy were randomized to receive maintenance erlotinib or placebo [32]. KRAS mutation was tested in 889 patients and detected in 18 %. Some degree of benefit from erlotinib in terms of PFS was seen in the KRAS wild type group (HR 0.70, 95 % CI 0.57–0.87, p < 0.001) with a non-significant trend in the KRAS mutant group (HR 0.77, 95 % CI 0.50–1.19). However, the interaction test between KRAS mutation status and treatment arm was not significant (interaction p = 0.95) [32]. The BR.19 placebo controlled randomized phase III trial assessed the role of postoperative adjuvant gefinib in NSCLC patients. KRAS mutation status was analysed in 350 tumors (169 and 181 in the gefitinib and placebo arms, respectively) [33]. Among the 254 patients with KRAS wild type tumors, gefitinib did not show any beneficial effect on DFS (HR 1.08, 95 % CI 0.74–1.59, p = 0.69) and OS (HR 1.13, 95 % CI 1.13 95 % CI 0.78–1.65, p = 0.51). Among the 96 patients with KRAS mutated tumors, gefitinib had a significant detrimental effect on DFS (HR 1.77 95 % CI 1.00–3.13, p = 0.05) although the same trend for OS was not significant (HR 1.51, 95 % CI 0.84–2.70, p = 0.16). The Cox regression model did not show any significant interaction between KRAS status and treatment (DFS, p = 0.15 and OS, p = 0.36) and no significant effect on outcome was found in multivariate analyses (DFS, p = 0.12 and OS, p = 0.5) [33]. Finally, the TOPICAL trial was a double blind, placebo controlled, phase III trial testing erlotinib as first line treatment in patients with NSCLC that were unsuitable for chemotherapy [34]. Among the 670 patients randomized, DNA for mutation testing was available in 390 patients; 73 (19 %) were positive for KRAS mutation. Among the patients with KRAS mutation, the median OS was 4.2 months (95 % CI 1.8–6.2) for erlotinib (n = 35) and 3.6 months (95 % CI 1.9–4.4) for placebo (n = 38) and the median PFS was 3.5 months (95 % CI 1.7–4.8) for erlotinib and 2.7 months (95 % CI 1.8–3.9) for placebo. Among patients with KRAS wild-type tumors, median OS was similar between the erlotinib (n = 210) versus placebo arm (n = 180) (3.7 months, CI 95 % 2.8–4.2, versus 3.4 months, CI 95 % 2.7–4.3) as was the median PFS (2.7 months, CI 95 % 2.2–2.9 versus 2.6, CI 95 % 2.3–2.9), no statistical comparison was reported [34].
The ATLAS and MARQUEE trials assessed erlotinib in combination with other targeted therapies. The ATLAS randomized, double-blind, placebo-controlled, phase IIIb trial compared bevacizumab therapy with or without erlotinib in patients with response or stable disease after first line platinum-based chemotherapy with bevacizumab for locally advanced, recurrent, or metastatic non-small cell lung cancer [40]. In 93 patients with KRAS mutant tumors, there was no PFS benefit with the addition of erlotinib (HR 0.93, 95 % CI 0.55–1.56; p = 0.7697), but in patients with wild-type KRAS tumors, there appeared to be benefit with the addition of erlotinib to bevacizumab (HR 0.67, CI 0.49–0.91; p = 0.01) [35]. In the ARQ-197-209 phase II study randomizing previously treated NSCLC patients to erlotinib plus/minus tivantinib, a MET inhibitor,, the small subset of patients with KRAS mutations (n = 15) appeared to benefit significantly from the combination (PFS HR 0.18, 95 % CI 0.05–0.70; p < 0.01, interaction p = 0.006) and OS HR, 0.43; 95 % CI, 0.12–1.50; interaction p = 0.17) [36]. Based on the results of this phase II trial, the MARQUEE phase III randomized trial assessed erlotinib plus tivantinib (ARQ 197) versus placebo in second or third line treatment in advanced NSCLC of non-squamous histology. KRAS mutations were tested in 522 patients from the placebo arm and 526 from the tivatinib arm with 148 (28.4 %) and 136 (25.9 %) patients having KRAS mutations, respectively. There was no benefit in OS for tivatinib versus placebo in patients with KRAS wild type tumors (HR 0.94, 95 % CI 0.77–1.14) or with KRAS mutated tumors (HR 1.04, 95 % CI 0.78–1.40) [37].
Finally, the effect of KRAS status was assessed in two trials comparing EGFR TKIs versus docetaxel. The Iressa Non-Small-Cell Lung Cancer Trial Evaluating Response Against Taxotere (INTEREST) trial compared gefitinib to docetaxel as second line treatment for patients with advanced NSCLC [38]. KRAS mutation status was determined in 275 tumors and 18 % were found to have mutations [41]. KRAS mutation was not found to be predictive of a differential survival benefit between gefitinib and docetaxel (HR 0.81, 95 % CI 0.44–1.49, p = 0.50 for KRAS mutant and HR 1.03, 95 % CI 0.77–1.37, p = 0.86 for KRAS wild type, respectively, interaction p = 0.51) [38]. The Tarceva Italian Lung Optimization Trial (TAILOR) was a phase III randomized trial comparing erlotinib versus docetaxel as second line treatment in NSCLC with wild type EGFR. Among the 219 assessed for KRAS mutations, 52 were positive. Again, KRAS mutation was not found to be predictive of a differential survival benefit between erlotinib and docetaxel (HR 0.84, 95 % CI 0.47–1.52, for KRAS mutant; HR 0.65, 95 % CI 0.46–0.90, for KRAS wild type, respectively, p interaction = 0.237) [39].
Two meta-analyses have evaluated the association between KRAS status and response to EGFR TKI therapy in NSCLC. Neither of them evaluated the predictive value of KRAS mutations on survival outcome. The first meta-analysis, in 2008, included 17 NSCLC trials with a total of 1,008 patients of whom 165 had KRAS mutations. Mutation was significantly associated with an absence of response to EGFR TKIs (positive Likelihood Ratio [LR] = 3.52; −LR = 0.84). The low pooled sensitivity (0.21, 95 % CI 0.16–0.28) suggests that resistance also occurs in a number of wild-type KRAS tumors and that additional mechanisms of resistance to EGFR-TKIs exist. However the test was highly specific (0.94, 95 % CI 0.89–0.97), suggesting that complete and partial responses to EGFR TKIs are highly unlikely in the presence of a KRAS mutation [42]. The second meta-analysis, in 2010, included 22 NSCLC studies with a total of 1,470 patients of whom 231 had KRAS mutations. The objective response rates (ORR) were higher for patient with KRAS wild-type NSCLC compared to these with KRAS mutation (26 and 3 %, respectively). The overall pooled relative risk for ORR was 0.29 (95 % CI 0.18–0.47, p < 0.01). The significant association between KRAS mutations and ORR was also found in subgroup analyses of the Asian population (RR 0.22, 95 % CI 0.07–0.63; p = 0.01), Caucasian patients (RR 0.31, 95 % CI 0.17–0.54; p < 0.01) in Caucasians, treated with erlotinib (RR 0.28, 95 % CI 0.12–0.63, p < 0.01) and gefitinib (RR 0.30, 95 % CI 0.16–0.57, p < 0.01) [43].
Some recent data indicate that the predictive value of KRAS mutation for response to EGFR-TKIs might differ based on the type of mutation. A very small study with only 14 and four patients having mutations in codons 12 and 13, respectively, evaluated KRAS mutation status and response to EGFR-TKI in EGFR wild type advanced NSCLC. Patients with codon 13 KRAS mutations had worse PFS (p = 0.04) and OS (p = 0.005) than patients with codon 12 mutations [44]. Very recently, Fiala et al. showed that the type of KRAS mutation might have different predictive impact for response to EGFR TKI treatment [45]. They tested 448 stage IIIB and IV NSCLC patients and identified 69 (15.4 %) with KRAS mutations. Notably, three had concomitant EGFR and KRAS mutations. KRAS mutations were more frequent in smokers than in non-smokers (17.9 % versus 5.8 %, p = 0.0048) and in adenocarcinomas than in squamous cell carcinomas (21 % versus 4.4 %, p = 0.0004), but were equally distributed between males and females. The most frequent mutation was G12C (52.2 %) followed by G12V (11.6 %) and G12D (7.2 %). A subgroup of 38 patients was treated with erlotinib or gefitinib. The PFS was significantly shorter in 24 patients with G12C KRAS mutations than in 14 patients with non-G12C KRAS mutations (median PFS 4.3 versus 9.0 weeks, respectively, HR 2.7, p = 0.009) and the OS was marginally but not significantly shorter (median OS 9.3 versus 12.1 weeks, HR 2.0, p = 0.068). These data warrant further validation in independent cohorts of NSCLC patients [45].
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
