Characterization of distinct types of KRAS mutation and its impact on first-line platinum-based chemotherapy in Chinese patients with advanced non-small cell lung cancer.

We performed this retrospective study to investigate whether the KRAS mutation status and its subtypes could predict the effect of first-line platinum-based chemotherapy in Chinese patients with non-small cell lung cancer (NSCLC). Patients received who had KRAS mutations were enrolled. Correlations between KRAS mutations, specific mutant subtypes and responses to chemotherapy were analyzed using Kaplan-Meier and Cox proportional hazard methods. A total of 2,183 cases who received KRAS mutation detection were included. A total of 218 of these cases were indicated to have KRAS mutations. KRAS mutations were identified more commonly in males compared with females (P=0.035). The most common subtypes were G12C, G12D and G12V. Among 73 KRAS mutant patients and 100 EGFR/ALK/KRAS wild-type patients with advanced NSCLC, KRAS-mutant NSCLC patients had a significantly shorter progression-free survival (P=0.007) compared with NSCLC patients with KRAS wild-type. In addition, there was a shorter but marginally statistically significant progression-free survival (PFS) in KRAS mutant patients with adenocarcinoma compared with those with non-adenocarcinoma (P=0.051). In the KRAS mutant group, patients with the KRAS G12V mutation had the poorest PFS compared with non-G12V mutant cases (P=0.045). In conclusion, KRAS mutation was a negative predictive factor of PFS in Chinese patients with advanced NSCLC who received first platinum-based chemotherapy. Patients with KRAS G12V mutations exhibited the poorest PFS compared with those with other KRAS mutant types.

Introduction

Lung cancer remains one of the most common malignancies and leading causes of cancer-related mortality both in China and worldwide (1,2). Approximately 80–85% of lung cancers are non-small cell lung cancer (3). In recent years, huge progress had been made in the treatment of non-small cell lung cancer (NSCLC) patients harboring EGFR mutation and ALK rearrangement (4–7). However, effective therapy specifically targeting KRAS mutation, which accounts for 25–50% of NSCLC patients in white populations and 5–10% in Asian populations, has not been developed yet (8–11).

KRAS is a member of the Ras gene family, which encodes small G proteins with intrinsic GTPase activity, contributing to activation of downstream effectors involved in multiple pathways including apoptosis, proliferation and differentiation (8,12,13). Point mutations occurred in tumors result in the loss of intrinsic GTPase activity and consequently in the deregulation of cell proliferation signals (13,14). KRAS mutation occurs mainly in codon 12, 13 or 61. Most common types of KRAS mutation are G12C, G12V, and G12D (8,9). In addition, in vitro data reported by Garassino et al suggested that NSCLC cell lines harboring a G12C, G12V or G12D KRAS mutation had differential sensitivity to chemotherapeutic agents (15). It appears that various types of KRAS mutations differ in clinical characters and drug response (16,17).

As early as 1990, KRAS mutation was already described as a negative prognostic marker for both overall survival (OS) and disease-free survival in lung cancer (18). Not until the last decades, more and more attention has been paid to the clinical significance of KRAS mutation in NSCLC. When it comes to the first-line platinum-based chemotherapy for advanced NSCLC patients, some researchers tend to believe there is no difference between KRAS mutant and wild-type patients regarding therapeutic response and prognosis (14,19,20). However, there were several studies indicated that KRAS mutation was a predictive factor of worse progression-free survival (PFS) or OS in advanced NSCLC patients treated with platinum-based chemotherapy (21–24). Considering the discrepant role of KRAS and its subtypes on effect of chemotherapy, the aim of this study was to investigate the predictive significance of KRAS mutation and its subtypes on clinical response and PFS in advanced NSCLC patients treated with first-line platinum-based chemotherapy.

Materials and methods

Study design

In this retrospective study, patients received KRAS mutation detection between August 2014 and June 2016 at Shanghai Pulmonary Hospital affiliated to Tongji University School of Medicine were included. We retrospectively reviewed patients' medical records to evaluate clinicopathological features and treatment regimens. All eligible patients' clinical data including age, sex, smoking status, histological type, TNM stage, Eastern Cooperative Oncology Group (ECOG) performance status (PS), treatment regimens, response to treatment, date of first diagnosis, date of starting chemotherapy, and date of disease progression or date of last contact. Pathological diagnosis was made by pathologists. Staging was carried out according to the staging system of the 2009 International Association for the Study of Lung Cancer (version 7) (25). Nonsmokers were defined as patients with the smoking dose of <100 cigarettes in their lifetime. Clinical response was evaluated by at least two clinicians according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (26). Inclusion criteria were: Adult (age≥18 years old) patients; recurrent or IIIB/IV NSCLC patients; patients received first-line platinum-based chemotherapy. Exclusion criteria were: Unknown mutational status; detected EGFR mutation or ALK rearrangement; no complete documentation; no response evaluation; adjuvant chemotherapy or radiochemotherapy. The study was approved by the Ethics Committees of Shanghai Pulmonary Hospital. Informed consent was obtained from all individual participants included in the present study. This study was conducted according to the Declaration of Helsinki, 7th version.

KRAS mutation analysis

Total DNA was extracted from tissue samples using AmoyDx DNA kit (Amoy Diagnostics Co., Ltd., Xiamen, China). The quality and quantity of extracted DNA were measured by NanoDrop 2000 Spectrophotometer (Thermo Scientific, Wilmington, DE, USA). KRAS mutation was identified by an AmoyDx® Human KRAS gene 7 Mutations Fluorescence Polymerase Chain Reaction (PCR) Diagnostic kit (Amoy Diagnostics Co., Ltd.). The real-time PCR conditions were as previously described (27–29). β-actin was used as an internal reference gene to ensure the quality of the extracted DNA and KRAS mutant DNA was used as positive control.

Statistical analysis

The relation between categorical parameters was tested using Pearson's χ2 test (Fishers exact test was used when n≤5). Kaplan-Meier curve was used to estimate the distribution of survival and log-rank test was used to analyze differences between groups. PFS was defined as the first day of treatment until either tumor progression or death. We used cox proportional hazards models for univariate and multivariate analysis to estimate clinicopathological features, KRAS mutation types and treatment regimens for their associations with PFS. Independent variables with P<0.10 in the univariate analysis were enrolled in multivariate analysis. P-values <0.05 were defined statistically significant. Confidence intervals were calculated at a 95% CI. Statistical tests were carried out using SPSS 20.0 software (IBM Corporation, Armonk, NY, USA).

Results

Patient characteristics

In total, 2,183 patients received KRAS mutation detection at Shanghai Pulmonary Hospital between August 2014 and June 2016 were enrolled into this study and 218 (10.0%) cases harbored KRAS mutation. Distribution of different types of KRAS mutation found within 218 patients are listed in Fig. 1. Three most common KRAS mutations were G12C (32.1%), G12D (23.4%) and G12V (21.1%). Other codon 12 mutations including G12A (12.8%), G12S (4.1%) and G12R (1.4%) were found in 20% of the patients. 3 patients had codon 13 G13D mutation. Four types of double mutations were found in 8 patients: G12C + G12R (4 patients), G12C + G12V (2 patients), G12D + G12V (1 patient) and G12A + G12V (1 patient). Based on our inclusion and exclusion criteria, we further analyzed 100 EGFR/ALK/KRAS wild-type and 70 KRAS mutant patients. The median age of whole study group was 61 years old (range 28–78). In total, 84.1% of patients were stage IV disease at diagnosis, and 77.6% of patients displayed histology of adenocarcinoma. The patient characteristics were listed in Table I. The patient basic characteristics were well-matched between KRAS mutant and wild-type groups except for sex (P=0.035). As for the treatment regimens, 74.1% of all patients received first-line chemotherapy with carboplatin-based chemotherapy, with a higher percentage of wild-type KRAS patients (78.0%) receiving carboplatin-based doublet comparing with mutant KRAS patients (68.6%). Numerically more KRAS mutant patients received a cisplatin-based chemotherapy when compared with KRAS wild-type patients (28.6% vs. 21.0%, respectively). However, there seems to be more patients in the KRAS wild-type group received platinum/pemetrexed treatments (68.0% in KRAS wild-type group vs. 57.1% in KRAS mutant group). Whereas patients with wild-type KRAS were as likely as patients with mutant KRAS to receive platinum/gemcitabine chemotherapies. Of note, 6 patients within the KRAS mutant group received platinum/docetaxel whereas only 1 patient within the KRAS wild-type group received platinum/docetaxel treatments.

Figure 1.

Distribution of KRAS mutation in whole population.

Table I.

Patient characteristics.

	KRAS mutant (n=70)	KRAS wild-type (n=100)	P-value
Mean age at diagnosis, mean ± SD	61±7.34	60±9.31	0.334
Sex, n (%)
Male	60 (85.7)	72 (72.0)	0.035
Female	10 (14.3)	28 (28.0)
Smoking history, n (%)			0.302
Smoker	42 (60.0)	52 (52.0)
Non-smoker	28 (40.0)	48 (48.0)
Histology, n (%)			0.826
Adenocarcinoma	55 (78.6)	77 (77.0)
Squamous	6 (8.6)	12 (12.0)
Other	0 (0.0)	1 (1.0)
NSCLC-NOS	9 (12.9)	10 (10.0)
Stage, n (%)
IIIB	6 (8.6)	12 (12.0)	0.475
IV	64 (91.4)	88 (88.0)
Platinum, n (%)
Cisplatin	20 (28.6)	21 (21.0)	0.287
Carboplatin	48 (68.6)	78 (78.0)
Other	2 (2.9)	1 (1.0)
Platinum doublets, n (%)
Platinum/pemetrexed	40 (57.1)	68 (68.0)	0.029
Platinum/gemcitabine	23 (32.9)	31 (31.0)
Platinum/docetaxel	6 (8.6)	1 (1.0)
Other	1 (1.4)	0 (0.0)

P-value based on Kruskal-Wallis test, otherwise P-value based on χ2 test or Fisher's exact test. NSCLC-NOS, non-small cell lung cancer-not otherwise specified; ECOG, Eastern Cooperative Oncology Group; PS, performance status.

Effect of KRAS mutation on response rate and PFS

None of the patients reached complete response. Partial response was similar between two groups (21.4% in KRAS mutant patients vs. 19.0% in KRAS wild-type patients). Comparatively, stable disease was observed more in wild-type KRAS patients than in mutant KRAS patients (67.0% vs. 44.3%, respectively). However, numerically more disease progressed in patients with mutant KRAS than wild-type KRAS (34.3% vs. 14.0%). There were no statistically significant differences in the objective response rate (ORR). In contrast, disease control rate (DCR) of KRAS wild-type patients to platinum-based chemotherapy was obviously higher than KRAS mutant patients (86.0% vs. 65.7%, P=0.002; Table II). In Table II, we also listed clinical outcomes of three most common KRAS mutation subtypes and other rare mutations. Among them, although G12V has the lowest DCR for 55.6%, response to platinum-based chemotherapy had no statistically significant differences between mutation subtypes.

Table II.

Response to first line chemotherapy in KRAS mutant vs. KRAS wild-type NSCLC patients.

		KRAS mutant (n=70)
	KRAS wild-type (n=100)	Total (n=70)	G12C (n=23)	G12V (n=18)	G12D (n=9)	Rare (n=20)	P-value[a]	P-value[b]
Response
CR	–	–	–	–	–	–
PR	19	15	6	4	1	4
SD	67	31	12	6	5	8
PD	14	24	5	8	3	8
ORR	19.0%	21.4%	26.1%	22.2%	11.1%	20.0%	0.893	0.697
DCR	86.0%	65.7%	78.3%	55.6%	66.7%	60.0%	0.442	0.002

P-value was calculated among KRAS mutation subtypes.

P-value was calculated between KRAS wild-type and mutated patients. CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; PFS, progression-free survival; ORR, objective response rate; DCR, disease control rate.

A total of 140 (82.4%) patients had progressed disease during the study period, with a median PFS for all subjects of 5.9 months (95% CI, 4.9–6.9 months). In all included patients with metastatic NSCLC at diagnosis, PFS was shorter in the KRAS mutant group vs. wild-type group (4.2 vs. 6.3 months; P=0.007; Fig. 2A). In addition, there was a shorter but only marginally statistically significant PFS in KRAS mutant patients with adenocarcinoma histology patients (4.3 months vs. 6.7 months; P=0.051; Fig. 2B). It suggested that the presence of KRAS mutation may be associated with a worse response to first-line platinum-based chemotherapy in advanced NSCLC patients. Next, we compared PFS of wild-type KRAS patients with three most common KRAS subtypes G12V, G12C, G12D and other rare mutations. When comparing patients with G12V mutant vs. wild-type, there was a statistically significant shorter PFS (2.9 months and 6.4 months, respectively; P=0.001). While other KRAS subtypes had no differences in PFS compared with wild-type KRAS (Fig. 2C). Patients with KRAS G12V mutation had inferior PFS compared with patients with non-G12V mutation (median PFS, 2.9 vs. 4.7 months; P=0.045; Fig. 3B). When comparing patients with G12C vs. non-G12C mutation and patients with G12D vs. non-G12D mutation, there was no differences in PFS, 4.4 months (95% CI, 3.3–5.5) vs. 4.2 months (95% CI, 2.3–6.1; P=0.202; Fig. 3A) and 7.0 months (95% CI, 1.1–12.8) vs. 4.3 months (95% CI, 3.8–4.8; P=0.519; Fig. 3C). It suggested that response to chemotherapy is not the same among KRAS mutation subtypes and patients with KRAS G12V mutation showed the poorest PFS than those with other KRAS mutant types.

Figure 2.

PFS of whole patients' cohort. Subgroup analysis of progression-free survival in KRAS mutant vs. wt patients with metastatic NSCLC at diagnosis (A). Subgroup analysis of progression-free survival in KRAS mutant vs. wild-type patients with adenocarcinoma histology (B). PFS in KRAS wild-type patients vs. three most common KRAS mutation subtypes and other rare mutations (C). PFS, progression-free survival; NSCLC, non-small cell lung cancer; wt, wild-type, mt, mutant.

Figure 3.

PFS of KRAS mutant patients' cohort. Three most common KRAS mutation subtypes G12C vs. wild-type (A), KRAS mutation subtypes G12V vs. wild-type (B) KRAS mutation subtypes G12D vs. wild-type (C). PFS, Progression-free survival.

Univariate and multivariate analysis

In univariate analysis, sex, smoking history and KRAS G12V mutation were significantly associated with PFS. Women had decreased risk of progressed disease when compared with men (HR, 0.616; 95% CI, 0.405–0.937; P=0.024). Smoking history also affected PFS (never smokers vs. current/former smokers; HR,0.665; 95% CI, 0.472–0.937; P=0.020). KRAS G12V was associated with shorter PFS (HR, 2.342; 95% CI, 1.378–3.981; P=0.002). In multivariate analysis, only KRAS G12V mutation was associated with shorter PFS (HR, 2.116; 95% CI, 1.211–3.696; P=0.008; Table III).

Table III.

Prognostic evaluation of clinical and histopathological characteristics in whole group and in KRAS mutant subgroup- progression free survival.

Variable	Univariate analyses 1[a] HR (95% CI) P-value	Multivariate analyses 1[a] HR (95% CI) P-value	Univariate analyses 1[b] HR (95% CI) P-value	Multivariate analyses 1[b] HR (95% CI) P-value
Age, <61 vs. ≥61 years old	1.170 (0.837–1.634)		1.622 (0.952–2.746)	–
	0.359		0.075	0.121
Sex, female vs. male	0.616 (0.405–0.937)		0.844 (0.399–1.784)
	0.024		0.657
Stage, IIIB/recurrent vs. IV	0.811 (0.457–1.438)		0.781 (0.259–1.990)
	0.473		0.525
Smoking, never	0.665 (0.472–0.937)	–	0.799 (0.462–1.379)
vs. current/former	0.020	0.126	0.420
Pathology, SQC	1.301 (0.775–2.183)		0.779 (0.306–1.981)
vs. ADC	0.319		0.599
KRAS, mutant vs. wt	1.324 (0.942–1.861)		–
	0.106
G12C vs. wt	1.107 (0.654–1.873)		–
	0.705
G12V vs. wt	2.342 (1.378–3.981)	2.116 (1.211–3.696)	–
	0.002	0.008
G12D vs. wt	1.031 (0.474–2.239)		–
	0.939
G12C vs. others	–		0.697 (0.395–1.231)
			0.214
G12V vs. others	–		1.762 (0.992–3.129)	1.831 (1.025–3.270)
			0.053	0.041
G12D vs. others	–		0.774 (0.350–1.714)
			0.528
Chemotherapy, cisplatin	1.296 (0.883–1.902)		1.720 (0.982–3.013)	–
vs. carboplatin	0.186		0.058	0.158
Chemotherapy, gemcitabine	1.390 (0.971–1.991)	–	1.527 (0.852–2.736)
vs. pemetrexed	0.072	0.335	0.155

Independent variables with P<0.10 in the univariate analyses were included in the model

Univariate and multivariate analysis for PFS in all patients after first line platinum-based chemotherapy

Univariate and multivariate analysis for PFS in KRAS mutated patients after first line platinum-based chemotherapy. PFS, progression-free survival HR, hazard ratio; CI, confidence interval; Cox's model, multivariate analyses with forward elimination; wt, wild-type.

In KRAS mutant group, univariate analysis showed that smoking history did not have impact on outcome for PFS (HR, 0.799; 95% CI, 0.462–1.379; P=0.420). And there was marginally statistic difference in outcome of G12V mutant patients vs. other mutant KRAS patients in univariate analysis (HR, 1.762; 95% CI, 0.992–3.129; P=0.053). In multivariate analysis based on age, G12V mutation status and cisplatin- or carboplatin-based chemotherapy, results showed that G12V mutant patients did have a shorter PFS than other KRAS mutant types (HR, 1.831; 95% CI, 1.025–3.270; P=0.041; Table III).

Discussion

Our treatment of NSCLC has been dramatically improved with the introduction of molecular markers. Targeted therapies, including tyrosine kinase inhibitors (TKIs), for EGFR mutation and ALK rearrangement improved PFS in patients bearing the relevant mutations (4,7,30). However, effective therapy specifically targeting KRAS mutation has not been developed yet. For patients with KRAS mutation, platinum-based chemotherapy remains their first choice. Nevertheless, the predictive value of KRAS mutation in NSCLC for chemotherapy also remains unclear.

In the last decades, although a large number of studies had been conducted focusing on KRAS mutation, the prognostic and predictive value of KRAS in lung cancer is still a highly debated issue. Considering the enormous discrepancy of studies in terms of races, tumor stage, histological types and various treatments, it is difficult to draw a definite conclusion. Therefore, we analyzed a well-defined Chinese patient cohort with advanced NSCLC received first-line platinum-based chemotherapy in our study. KRAS mutation rate in all tested population was 10.0%, which is in accordance with other studies of Asian NSCLC study cohort (10,11,29,31,32). Furthermore, we found a ratio of the major subtypes, G12C (32.1%), G12V (23.4%), G12D (21.1%), which is almost identical with the previous reports (31–35). We also identified four kinds of co-mutations in our study group: Four patients with G12C/G12R, two patients with G12C/G12V, one patient with G12D/G12V and one patient with G12A/G12V. And no significant differences in PFS between KRAS co-mutant and other KRAS mutant or wild-type patients were found (data not shown).

Prior findings indicated patients with KRAS mutation were preferably to be smokers and have histology of adenocarcinoma comparing with patients of wild-type KRAS (36,37). However, in the current study, we noted that there were no differences in smoking history and pathological types between two groups of patients. Nevertheless, we observed KRAS mutation was not exclusively found in patients with adenocarcinoma. Hence testing all patients with NSCLC for KRAS mutation is necessary. Although KRAS mutant patients and KRAS wild-type patients shared similar smoking habits, smokers had increased risk of shorter PFS compared with non-smoker in our univariate analysis of whole study group. There seemed to be more males in the KRAS-mutant group comparing with the group of patients with wild-type KRAS. But the significance of this finding was complicated to explain regarding clinical outcome. Although male sex was dramatically associated with worse outcomes in our univariate analysis, survival was similar in whole study group between KRAS mutant and wild-type groups despite the KRAS cohort had a higher percentage of males. The majority of patients in the study group received a cisplatin or carboplatin plus pemetrexed or gemcitabine chemotherapy. The different choice of chemotherapy regimens did not affect the PFS both in whole group and in KRAS mutant cohort in univariate and multivariate analysis.

There were many articles reporting inconsistent results in regards to the impact of KRAS mutation on survival of advanced NSCLC patients who received platinum-based chemotherapy. For example, a retrospective analysis performed by Mellema et al showed no significant differences in clinical response to chemotherapy or OS when compared patients with KRAS mutation with patients without KRAS mutation (19). Conversely, Metro et al demonstrated that patients with KRAS mutation had lower response rates, and shorter PFS compared with EGFR wild-type/KRAS wild-type patients (23). Besides, Hames et al reported that the presence of KRAS mutation in advanced NSCLC patients displayed a worse prognosis of platinum-based chemotherapy compared with those absence of detectable driver mutations (21). In the current analysis, our results suggested that KRAS mutant patients did have lower DCR compared with KRAS wild-type patients, but not ORR. In addition, KRAS mutant patients demonstrated a decrease PFS comparing with wild-type patients, which was in accordance with prior report (21) and we found more convincing results in patients with metastatic NSCLC at diagnosis, PFS was significantly shorter in the KRAS mutant group vs. wild-type group (4.2 vs. 6.3 months; P=0.007). In addition, there was a shorter but only marginally statistically significant PFS in KRAS mutant patients with adenocarcinoma histology patients (4.3 months vs. 6.7 months; P=0.051). Based on the above results, we made the conclusion that KRAS mutation was a negative predictive factor of PFS in Chinese patients with advanced NSCLC received first platinum-based chemotherapy. Admittedly, this study was conducted at a single institution and had limited patient samples. We considered that, to make our findings more convincing, sharing of more data from multicenter studies, especially those covering various populations should be encouraged. We will also stay focuced on this issue and further exploration of the prognostic value of KRAS and its underlying mechanism is needed. Although recent research in colorectal cancer reported that G12V mutation demonstrated poor response to therapy and survival (38), the relevance of specific mutation subtypes in KRAS and clinical outcome remains controversial in NSCLC (16,39–41). In recent studies of advanced NSCLC, effects of KRAS G12V mutation regrading as either response to chemotherapy or OS were not obvious (40). However, in our study, patients with G12V mutant not only responded poorer to platinum-based chemotherapy, although not statisticly significant, but also had a significantly shorter PFS than those with other KRAS mutations. Our finding was in accordance with results carried out by Ihle et al (16). Downstream signaling of RAS differed in mutation subtypes. KRAS G12C/G12V preferably activated RalA/B signaling while KRAS G12D activated Akt pathway and the former demonstrated decreased survival (42). Taking all our presented results together, there is reason to believe that, in NSCLC, patients with different KRAS mutant subtypes may lead to distinct response to first-line platinum-based chemotherapy. Furthermore, subtype-specific mutation analysis is necessary in clinical practice, which may help to identify the most effective treatment regimens for each individual patient. Despite some of our results were consistent with previous publication, our study was conducted among Chinese population. Considering the differences in gene background between Caucasian and East Asian people (43,44), whether previous observation is also true among East Asian population remains uncertain. The conclusions we made in the study will provide clinicians with more comprehensive evidence when making clinical decisions for NSCLC patients with KRAS mutation.

There are several limitations in the present study that should be acknowledged. First of all, selection bias was inevitable due to the nature of retrospective studies. Second this study design was at a single institution. Taking the high cost of molecular detection into consideration, not all patients in our hospital received KRAS mutation test, therefore patients included in our study may not be representative of a general population. Sufficiency of cancer samples was also one of the limitations in this study. However, according to previous reports, in white populations KRAS accounts for 25–50% of NSCLC patients but KRAS mutations are only found in 5–10% of NSCLC patients in Asian populations (8–11). When we reviewed relative studies conducted among Caucasian populations, we found our patient number was very similar to other studies. In a retrospective analysis performed by Hames et al and colleagues, they compared 70 patients with pan-mutation negative and 80 patients with KRAS-mutant advanced NSCLC patients (21). On the other hand, considering the lower incidence of KRAS mutation among Asian people, we only focused on whether KRAS mutation was a negative predictive factor of PFS in Chinese patients with advanced NSCLC received first platinum-based chemotherapy. Further studies should be done aiming at the prognostic value of KRAS mutation on chemotherapy and also comparing responses with different cytotoxic chemotherapy regimens in patients with advanced NSCLC based on KRAS mutation and subtypes. Thus, considering the above limitations, multi-centered, international cooperative and larger number of NSCLC patients should be analyzed to valid our present findings.

The current study suggested that the presence of KRAS mutation was associated with a worse response in advanced NSCLC patients received first-line platinum-based chemotherapy. Responses to cytotoxic chemotherapy are not same among KRAS mutation subtypes. As the currently available literatures are still conflicting on the predictive value of KRAS mutation and its subtypes in advanced NSCLC, future studies should be done aiming at comparing responses with different cytotoxic chemotherapy regimens in patients with advanced NSCLC based on KRAS mutation and subtypes.