Frequency of EGFR and KRAS Mutations in Lung Adenocarcinomas in African-Americans

Structured Abstract

Introduction

The detection of mutations in the epidermal growth factor receptor (EGFR) gene, which predict sensitivity to treatment with EGFR tyrosine kinase inhibitors (TKIs), represents a major advance in the treatment of lung adenocarcinoma. KRAS mutations confer resistance to EGFR -TKIs. The prevalence of these mutations in African-American patients has not been thoroughly investigated.

Methods

We collected formalin-fixed, paraffin-embedded material from resected lung adenocarcinomas from African-American patients at three institutions for DNA extraction. The frequencies of EGFR exon 19 deletions, exon 21 L858R substitutions and KRAS mutations in tumor specimens from African-American patients were compared to data in Caucasian patients (n=476).

Results

EGFR mutations were detected in 23 of the 121 specimens from African-American patients (19%, 95% CI 13–27%), while KRAS mutations were found in 21 (17%, 95% CI 12−25%). There was no significant difference between frequencies of EGFR mutations comparing African-American and Caucasian patients, 19% vs. 13% (61/476, 95% CI 10–16%) (p=0.11). KRAS mutations were more likely among Caucasians, 26% (125/476, 95% CI 23−30%) (p=0.04).

Conclusions

This is the largest study to date examining the frequency of mutations in lung adenocarcinomas in African-Americans. Although KRAS mutations were somewhat less likely, there was no difference between the frequencies of EGFR mutations in African-American patients as compared to Caucasians. These results suggest that all patients with advanced lung adenocarcinomas should undergo mutational analysis prior to initiation of therapy.

Introduction

The discovery of activating mutations in the epidermal growth factor receptor (EGFR) gene has revolutionized our understanding and treatment of lung adenocarcinoma, a disease diagnosed in 120,000 Americans each year.¹ Somatic mutations in the tyrosine kinase domain of the EGFR gene drive the development of a substantial subset of lung adenocarcinomas, and sensitize these tumors to treatment with the small molecule tyrosine kinase inhibitors (TKIs) erlotinib and gefitinib.^2–4 When patients with EGFR mutations are treated with EGFR-TKIs, nearly all respond, with improvement in progression-free survival compared to patients without EGFR mutations.⁵ As a result, there has been significant interest in molecular characterization of EGFR and in identification of patients with EGFR mutations.

Approximately one-third of lung adenocarcinomas world-wide and an estimated 50% of tumors from East Asian patients have EGFR mutations. However, only 15% of patients from an unselected North American population will harbor these mutations.⁶ The prevalence of EGFR mutations among patients of other races, such as African-Americans, is unknown, in part due to small numbers of patients included and/or reported in prior studies.

KRAS mutations are also of interest in patients with lung adenocarcinomas. Transversion mutations (substitution of a pyrimidine for a purine or purine for a pyrimidine) are more common than transition mutations (substitutions of a purine for a purine or pyrimidine for a pyrimidine). KRAS mutations and EGFR mutations are mutually exclusive, and the presence of a KRAS mutation identifies a group of patients who are unlikely to respond to an EGFR-TKI.^7–9 While EGFR mutations are found more typically in never-smokers, KRAS mutations were first identified among patients who reported heavy tobacco use. More recently, however, KRAS mutations have also been identified in 15% of never smokers, as compared with 22% of former smokers in one study from our institution.¹⁰ KRAS mutations are found in 20–30% of lung cancer patients in North American populations, and in Asian populations they are less frequently reported. The true prevalence of KRAS mutations among African-Americans is unknown.

In a prior study of 291 tumors obtained from patients from Memorial Sloan-Kettering Cancer Center (MSKCC), 6 of 14 African Americans (43%) harbored EGFR mutations in their tumors.¹¹ Based on this higher-than expected frequency, we aimed to study the rate of both EGFR and KRAS mutations in the largest cohort of African-American patients with lung adenocarcinoma to date.

Methods

Tumor specimens were obtained from all available tissue from African-American patients from two institutions (MSKCC (New York, NY) and SUNY Downstate Medical Center (Brooklyn, NY), and population-based cases from studies conducted at Wayne State University (Detroit, MI). African-American patients were selected based on self-report of race, as determined at the time of registration. The histologic diagnosis of adenocarcinoma was verified using the pathology report provided by participating institutions.

Approval from the MSKCC Institutional Review Board/Privacy Board (IRB/PB) was obtained for the acquisition and analysis of all tumor specimens. All mutational analysis was performed in the Department of Pathology at MSKCC from formalin-fixed, paraffin-embedded material. Tumor specimen only underwent microdissection before mutational analysis when necessary. EGFR exon 19 deletions and exon 21 L858R mutations were identified using previously reported mutation-specific PCR-based methods. ^12–13 KRAS codon 12 and 13 mutation identification was performed by both mass-spectrometry (Sequenom)-based genotyping and direct sequencing. All positive cases were identified by visual review, and independently confirmed by direct sequencing.

EGFR and KRAS mutation frequencies were calculated and compared with mutation frequencies of surgically resected lung adenocarcinomas collected from Caucasian patients evaluated at MSKCC between 2006 and 2008. Mutational analysis of these tumors was also performed by the Department of Pathology.

The frequency of EGFR and KRAS mutations were summarized and compared using Chi-square tests and 95% confidence intervals.

Results

One hundred and twenty-one lung adenocarcinomas from African-American patients were collected—approximately equal numbers from MSKCC (n=50) and Wayne State University (n=47), with additional samples from SUNY Downstate Medical Center (n=24). No other clinicopathologic data was collected for this analysis.

EGFR exon 19 deletions or exon 21 L858R mutations were identified in 23 (19%) of the 121 African-American specimens (95% CI 13–27%). Among the EGFR mutations, 18 had exon 19 deletions and 5 had exon 21 L858R mutations. KRAS exon 2 codon 12 and 13 mutations were identified in 21 (17%, 95% CI 12–25%) of the 121 African-American tumor specimens (Table 1).

Table 1.

Distribution and Frequency of Mutations among African-Americans

Mutations Identified	African- Americans (n=121)	95% CI
EGFR exon 19 deletion [n, (%)]	18 (15%)	10–22%
EGFR L858R exon 21 mutation [n, (%)]	5 (4%)	2–10%
KRAS mutations [n, (%)]	21 (17%)	12–25%
EGFR/KRAS wild-type [n, (%)]	66 (55%)	46–63%

In order to evaluate mutation rates by race, we compared the data collected from African-American patient samples to mutation frequencies of all lung adenocarcinomas resected from Caucasian patients between 2006 and 2008 at MSKCC, a total of 476 patients (Figure 1). The prevalence of KRAS mutations in African-American patients was less than that observed in the Caucasian group (26%, 125/476, 95% CI 23–30%) (p=0.04). However, there was no difference in the number of EGFR mutations between African-American and Caucasian patients (13%, 61/476, 95% CI 10–16%) (p=0.11).

Figure 1.

Frequencies of EGFR and KRAS mutations among African-American and Caucasian populations

Transversion mutations were demonstrated in 17 (81%, 95% CI 59–93%) of the African-American patients harboring KRAS mutant lung cancers, and only 4 (19%, 95% CI 7−41%) had transition mutations. Similarly, 87 (70%, 95% CI 62–78%) of KRAS mutations in the Caucasian cohort were transversions, while 37 (30%, 95% CI 22–38%) were transitions (p=0.44). The distribution of KRAS mutations and their frequencies in both cohorts are listed in Table 2. The most common type of KRAS mutation found in both cohorts was the glycine to cysteine transversion (G12C). The least common KRAS mutations were G13V, G12R, G12S, and G13S. Of these, only G13S was not represented in the larger Caucasian group.

Table 2.

Distribution of KRAS mutations among Caucasian and African-American patients

KRAS Mutation	KRAS Nucleotide	AA patients (n=21)	Caucasian patients (n=125)	p-value
G12C	GGT→TGT	8 (38%)	48 (38%)	p=1
G12V	GGT→GGT	7 (33%)	18 (14%)	p=0.0547
G12A	GGT→GCT	1 (5%)	13 (10%)	*
G13C	GGC→TGC	1 (5%)	7 (6%)	*
G13V	GGC→GTC	0 (0%)	1 (1%)	*
G12R	GGT→CGT	0 (0%)	1 (1%)	*
G12D	GGT→GAT	3 (14%)	32 (26%)	*
G13D	GGC→ GAC	0 (0%)	4 (3%)	*
G12S	GGT→AGT	0 (0%)	1 (1%)	*
G13S	GGC→AGC	1 (5%)	0 (0%)	*

^*numbers too small for meaningful comparison

Discussion

This study represents the largest retrospective study to date comparing EGFR and KRAS mutation rates between African American and Caucasian populations. There was no difference in the frequency of EGFR mutations between the African American and Caucasian cohorts, while KRAS mutations were less frequent among African-American patients.

A few small studies have previously evaluated the prevalence of mutations by race. In one study performed in Louisiana, the frequency of KRAS mutations was evaluated in 116 patients with lung cancer, 60 of them African-Americans.¹⁴ In contrast to our findings, African-American patients in that study were significantly more likely to harbor a KRAS mutation as compared to Caucasian patients (37% vs 20%, p=0.048). However, consistent with our data and reports by others,¹⁴ Hunt et al. also reported a “disproportionately high number of cysteine for glycine transversions.” Cysteine missense substitutions result from a G→T transversion in the first base in either codon 12 or 13. They have been attributed to the polycyclic aromatic hydrocarbons (PAH) present in tobacco smoke.¹⁵ Hunt et. al also found a higher than expected number of serine mutations among the African-American cohort as the result of a G→A transition in the first base of either codon 12 or 13.¹⁴ This mutation was demonstrated in one African-American patient in the present study, as well as one patient in our Caucasian cohort (Table 2).

Racial differences in EGFR mutation rates have also been evaluated in smaller numbers in prior studies (Table 3). EGFR mutations were reported in one of 44 African American patients (2.4%) in a study of 219 patients with tissue specimen from the University of Maryland, the Mayo Clinic in Minnesota and the University of Milan, Italy.¹⁶ In comparison, 25 of the 177 (14.1%) tumors from Caucasian patients harbored EGFR mutations. Another study from the University of Texas M. D. Anderson Cancer Center included only eight African American patients.¹⁷ While 14 EGFR mutations were confirmed among the 159 patients included, none were reported in this small subset of African American patients. Although KRAS mutations were also reported in 18 patients overall, their races were not specified.

Table 3.

Prior Studies Comparing the Frequencies of EGFR Mutations in African-American and Caucasian Patients

Study	African- American Patients	Caucasian Patients	Frequency EGFR mutations, African-American patients (n, %, 95% CI)	Frequency EGFR mutation, Caucasian patients (n, %, 95% CI)
Yang et al, 200516	41	177	1 (2.4%) 95% CI (0–12·9%)	25 (14%) 95% CI (9·4–20·1%)
Tsao et al, 200617	8	139	0 (0%) 95%CI (0–40%)	10 (7%) 95%CI (4–13%)
Leidner et al, 200918	53	102	1 (2%) 95%CI (0–11%)	15 (17%*) 95% CI (11–26%)
Reinersman et al, 2010 (this study)	121	476	23 (19%) 95% CI (13–27%)	61 (13%) 95% CI (10–16%)

^*89 of 102 Caucasians analyzable for EGFR mutations

Most recently, Leidner et al. published an analysis of 53 tumors from African-American patients obtained from the University Hospitals Case Medical Center in Cleveland, OH.¹⁸ Patients were selected based upon availability of a tissue block as well as uninvolved paired tissue (ie. from a lymph node). Results were compared to a Caucasian cohort of 121 patients from three separate Italian centers (Bologna, Milan, and Perugia) previously enrolled in an unrelated clinical trial to receive gefitinib, who had mutational analysis performed previously. In this study, African Americans were again less likely to have EGFR mutations in their tumors, as compared to Caucasian patients (2% vs. 17%, p=0.022). However, the only EGFR mutation found among the African American patients was neither of the common activating EGFR mutations (exon 19 deletion or exon 21 L858R substitution), but a missense mutation in exon 20 L768N, a mutation in EGFR not previously reported. Leidner et al. also reported no difference in KRAS mutation frequencies between the two cohorts of patients.

Twice the number of African-American patients (n=121) were included in our analysis compared with 53 tumors from African-American patients tested in the article of Leidner et al. While not as robust a sample as our Caucasian cohort, a retrospective power calculation indicates that we have adequate power to detect a difference of 9.6% or more in either direction, assuming the frequency of EGFR mutations in the Caucasian population is 13%. Therefore, if it were true that EGFR mutations exist among African-American patients at a much lower frequency, as reported in previous studies (Table 3), we would have detected that difference in our analysis.

One potential source of bias in our study is the fact that African-American specimens were collected from three different centers, while Caucasian patients were collected only from MSKCC. Like the Leidner et al. study described above, patients were included based solely on tissue availability, a common criterion for translational tissue studies. No other sample selection was used, and because all African-American patients who had available tissue were included, we expect the study population to be representative of the African-American population of each participating institution. An analysis restricted to MSKCC patients found 16% of African-Americans had EGFR mutations, and 25% had KRAS mutations, both safely within the 95% confidence intervals for the total number of African-American patients included in the results.

We did not report clinical characteristics traditionally associated with response to EGFR-TKIs in our patients. Prior to the IPASS study, we relied upon histology, ethnicity, and smoking status to predict the likelihood of response to EGFR-TKI treatment. However, we now know that it is the presence of the EGFR mutation that truly underlies sensitivity. The most meaningful demonstration of this principle from IPASS was the paltry response (1%) of Asian women never-smokers or former light smokers randomized to receive gefitinib who were found to be without EGFR mutations in their tumors.⁵ Testing tumors for the presence of EGFR and KRAS mutations prior to treatment with the EGFR TKI gefitinib is now a standard of care in Europe. This standard will increasingly be adopted on all continents, and therefore attention to clinical characteristics for the purposes of predicting response to EGFR-TKIs will continue to lessen in importance.

Our data indicate that lung cancers arise from EGFR mutations in African-Americans just as commonly as in Caucasian patients. Therefore, African-Americans patients stand to derive similar benefit from treatment with EGFR-TKIs as other western populations. We feel these results have important diagnostic and treatment implications for African-American patients as well as any non-Asian, non-Caucasian patient with lung cancer. Irrespective of race, tumors tissue obtained from all patients with advanced lung adenocarcinoma should be screened for the presence of EGFR and KRAS mutations. Future research efforts examining other molecular mechanisms, such as MET germline polymorphisms,¹⁹ BRAF mutations, HER2 mutations and ALK translocations may further elucidate the racial variations in lung cancer development.