Main

Despite improved surgical techniques and chemotherapy protocols, overall survival of patients with lung carcinomas has not changed in the past 40 years.1, 2 Therefore, development of EGFR-targeted therapies, including monoclonal antibodies (for example, cetuximab) and tyrosine kinase inhibitors small molecules (for example, gefitinib, erlotinib), provided a new hope for better survival and quality of life of lung cancer patients. Clinical trials of small-molecule EGFR tyrosine kinase inhibitors in unselected non-small cell lung carcinoma patients showed a small proportion of patients with a radiographic response and symptomatic improvement.3, 4, 5, 6, 7, 8, 9, 10 Several studies almost simultaneously showed that responders to the EGFR tyrosine kinase inhibitors have somatic mutations in the EGFR tyrosine kinase domain.11, 12, 13 The most common mutations are exon 19 deletions that eliminate a common leucine–arginine–glutamic acid–alanine motif (LREA) and exon 21 point mutations that lead to substitution of arginine for leucine at position 858. They could be detected in about 10% Western and 40% Asian patients, who are primarily women and never smokers. In contrast, KRAS mutations are found in adenocarcinomas resistant to EGFR tyrosine kinase inhibitors, usually in current or former smokers.14, 15, 16 These mutations lead to substitutions of amino acids for glycines at positions 12 and 13, and can be identified in up to 30% adenocarcinomas. EGFR and KRAS mutations are mutually exclusive. These discoveries resulted in introduction of screening for common EGFR and KRAS mutations in patients with lung adenocarcinomas in clinical practice. At present, there are no guidelines in terms of specimen selection, methods and interpretation criteria that would represent a standard of patient care.

Numerous retrospective studies used different methodological and interpretation criteria in the assessment of EGFR and KRAS genes led to conflicting results and conclusions. EGFR FISH and DNA mutation analyses are the two most extensively studied methods for the selection of candidate patients for EGFR tyrosine kinase inhibitors therapy.17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28

Even though most laboratories accepted direct DNA sequencing or other mutational methods as the most reliable assays that predict responders to EGFR tyrosine kinase inhibitors, it is still uncertain whether EGFR FISH or chromogenic in situ hybridization may provide additional clinically useful information.17, 18, 21, 25, 27, 28, 29, 30 The complicating factor is that EGFR mutations are frequently associated with increased EGFR gene copy numbers. Method for EGFR FISH could be easily standardized because of a commercially available probe. However, the interpretation criteria, which could be arbitrary and subjective, have neither been standardized nor validated yet. The group from the University of Colorado proposed a scoring system for FISH-positive samples taking into consideration classical amplification and polysomy.17, 31 Using these criteria, Cappuzzo et al17 in an initial report showed that 33% of cases interpreted as FISH positive had a higher response rate to gefitinib (36%) than the FISH-negative patients (3%) and had a longer median survival (18.7 months vs 7.0 months). Within the same cohort, 17% of cases were EGFR mutation positive, which was associated with a response rate of 53%, compared with 5% in wild-type cases. The importance of distinguishing increased copy number of chromosome 7 without EGFR gene amplification (high polysomy) and EGFR gene amplification is still uncertain, and additional studies are needed to validate the significance of those criteria.

Since the initial reports about clinicopathological characteristics of responders to EGFR tyrosine kinase inhibitors, several studies attempted to better define the morphology of adenocarcinomas occurring in patients with EGFR and KRAS mutations. Several reports indicated that EGFR mutations were preferentially observed in bronchioloalveolar or mixed types of adenocarcinomas with bronchioloalveolar carcinoma features.32, 33, 34 More recent studies suggested papillary and micropapillary differentiation to be most likely associated with EGFR mutations.35, 36 In contrast, mucinous differentiation was associated with KRAS mutations in some studies.37, 38 This variety of morphological interpretations reflects the lack of consistency in the interpretation of the World Health Organization criteria for classifications of lung tumors. Furthermore, different types of samples, including surgical resection and cytology specimens, were used for morphological analysis. It is clear that lung adenocarcinoma represents a morphologically heterogeneous group of tumors, and precise assessment of different growth patterns may be achieved only on a large number of histological sections. Therefore, to avoid further confusion, it is absolutely necessary to strictly follow the World Health Organization criteria for classification of lung tumors and to study only large number of histological sections in an individual case.

The aim of this study was to determine whether clinicopathological characteristics and morphology of lung adenocarcinomas might be used as predictors of tumor mutational status, which then may be implemented as the selection criteria for molecular profiling of lung adenocarcinomas in clinical practice.

Materials and methods

Patient Selection

A total of 345 consecutive newly diagnosed primary lung adenocarcinomas from patients who underwent surgical resection at the University of Pittsburgh Medical Center were selected for the study. Those patients who received previous neoadjuvant or adjuvant cytotoxic chemotherapy or radiation were excluded from the analysis. Clinical informations, including gender, age, tumor stage, smoking history and surgical procedure, were obtained from the review of patients' electronic medical records. Specimens included 70 wedge resections, 48 segmentectomies, 212 lobectomies, 2 bilobectomies and 13 pneumonectomies.

There were 147 males and 198 females with an age at diagnosis ranging from 33 to 90 years (median 68). In all, 201 patients had stage I, 22 stage II, and 82 had stage III disease. There were 41 never smokers and 296 smokers, including former and current smokers. Smoking history was unknown in eight patients.

All hematoxylin and eosin (H&E) histological sections of the tumors were reviewed by three pathologists (SD, SY and PO). Rare discrepancies were resolved by consensus after discussion and review of the H&E slides at the multiheaded microscope. Histological type was determined according to the 2004 World Health Organization classification criteria.39 In all, 74% of tumors were classified as a mixed subtype of lung adenocarcinomas. Percentages of various histological subtypes in mixed subtypes of adenocarcinomas were assessed and further classified as primary or secondary histological patterns. Tumor differentiation was graded as well, moderate or poor.

KRAS and EGFR Mutational Analysis

Tumor targets were manually microdissected from the 4-μm unstained histological sections. DNA was isolated from each target using the DNeasy tissue kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. For the detection of mutations, DNA was amplified with primers flanking exon 2 of the KRAS gene (forward primer 5′-GGTGAGTTTGTATTAAAAGGTACTGG-3′ and reverse primer 5′-TCCTGCACCAGTAATATGCA-3′), exon 19 of the EGFR gene (forward primer 5′-CCCAGCAATATCAGCCTTAGGTG-3′ and reverse primer 5′-CCACTAGAGCTAGAAAGGGAAAGAC-3′) and exon 21 of the EGFR gene (forward primer 5′-CCTCACAGCAGGGTCTTCTC-3′ and reverse primer 5′-CCTGGTGTCAGGAAAATGCT-3′). Then, PCR products were sequenced in both sense and antisense directions using the BigDye Terminator v3.1 cycle sequencing kit on ABI 3130 (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. The sequences were analyzed using Mutation Surveyor software (SoftGenetics, LLC., State College, PA, USA). Each case was classified as positive or negative for the KRAS and EGFR mutation based on the sequencing results.

EGFR FISH

The FISH analysis of EGFR amplification was carried out using standard method with the dual-color EGFR SpectrumOrange/CEP7 SpectrumGreen probe (Vysis, Downers Grove, IL, USA) and paraffin pretreatment reagent kit (Vysis).40 In brief, paraffin sections were deparaffinized, dehydrated in ethanol and air dried. The sections were digested with protease K (0.5 mg/ml) at 37°C for 28 min. The slides were denatured at 75°C for 5 min and dehydrated in ethanol. The probes were denatured for 5 min at 75°C before hybridization. Slides were hybridized overnight at 37°C and washed in 2XSSC/0.3% NP40 at 72°C for 2 min. The nuclei were counterstained with DAP/antifare 1 (Vysis). Each FISH assay included normal lung tissue sections as a negative control, and sections of lung non-small cell carcinoma previously identified as carrying EGFR gene amplification as a positive control. Analyses were carried out using a fluorescence microscope (Nikon Optiphot-2 and Quips Genetic Workstation) equipped with Chroma Technology 83000 filter set with single band exitors for Texas Red/Rhodamine, FITC and DAPI (UV 360 nm). The histological areas previously selected on the H&E-stained sections were identified on the FISH-treated slides. Only individual and well-delineated cells were scored. Overlapping cells were excluded from the analysis. At least 60 cells were scored for each case and control.

The interpretation criteria for EGFR FISH have neither been standardized nor validated yet, and therefore, we applied two interpretation approaches. The first approach was to determine the EGFR gene amplification defined as a ratio between EGFR gene copy numbers and chromosome 7 >2 in all the three study groups. The second approach applied a Colorado scoring system, as previously described.31 In brief, tumors with EGFR gene amplification or with at least 40% of cells showing at least four copies of the EGFR signals were classified as EGFR FISH positive. Tumors with <40% of cells showing at least four copies of the EGFR signals and no EGFR gene amplification were classified as EGFR FISH negative.

Statistical Analysis

Statistical analyses were carried out using SAS version 9.2 (SAS Institute, Cary, NC, USA). A significance level is set at 0.05 and all the P-values reported were two sided. The correlation among morphology variables was examined by Pearson's correlation. The associations of mutation and categorical explanatory variables were examined using χ2- or Fisher's exact test. Exploratory analyses based on logistic regression were conducted to identify independent predictors of mutational profile, including the models adjusted for demographic and clinical profile. The odds ratio estimates were calculated for univariate or multivariate models using logistic regression.

Results

Clinicopathological Characteristics

Clinicopathological characteristics of 345 primary lung adenocarcinomas divided into three groups by mutation type are summarized in Table 1. Mutational analysis showed 37 (11%) EGFR+ lung adenocarcinomas, 103 (30%) KRAS+ and 205 (59%) cases negative for both EGFR and KRAS mutations. Among the cases with EGFR mutations, 19 (51%) had an exon 19 in-frame deletions and 18 (49%) had the exon 21 point mutation. No tumor had both mutations. No differences were observed in the tumor stage distribution between the three groups (P=0.8891). The median tumor size was 2 cm for KRAS–/EGFR– and KRAS+ groups, and 2.5 cm for EGFR+ tumors. Median age was 68 years for KRAS–/EGFR– group, 70 years for KRAS+ group and 68 years for EGFR+ tumors.

Table 1 Clinicopathological characteristics of 345 surgically treated primary lung adenocarcinomas grouped by EGFR and KRAS mutational status
Full size table

The EGFR mutations were significantly more frequent in women (81%) (P=0.046), whereas a relatively equal gender distribution was seen in the KRAS+ group (49% women). There was a slight predominance of women in the EGFR–/KRAS– group (56%). More patients with EGFR+ tumors had a history of never smoking (39 vs 5% KRAS+ and 11% EGFR–/KRAS–) (P<0.0001). Poorly differentiated tumors were more frequent in KRAS+ and EGFR–/KRAS– groups (19 and 21%, respectively) than in the EGFR+ tumors (3%) (P=0.0983). A prominent lymphocytic host response was more common in KRAS+ (28%) and EGFR–/KRAS– (30%) tumors than in the EGFR+ group (8%) (P=0.0187). There was no difference in the presence of parenchymal scar, angiolymphatic or visceral pleural invasion between the groups (Table 1).

The significant clinical predictors for the presence of EGFR mutations included history of never smoking (OR 5.939; 95% Wald confidence limit 1.662–21.223, P=0.0149), absent to mild lymphocytic host response (OR 4.724; 95% Wald confidence limit 1.33–1.776; P=0.0163) and female gender (OR 2.571; 95% Wald confidence limit 1.015–6.511, P=0.0463). The significant clinical predictors for the presence of KRAS mutations included older age (OR 1.034; 95% Wald confidence limit 1.007–1.062, P=0.0132) and history of smoking (OR 0.617, 95% Wald confidence limit 0.357–1.066, P=0.0412).

Mutational Status and Morphology

Mixed subtype of adenocarcinoma was the most common histological type observed in all three groups (86% in EGFR+, 69% in KRAS+ and 74% in EGFR–/KRAS– group) (P=0.1153). There were 54 pure types of adenocarcinomas in the EGFR–/KRAS– group (24 solid, 20 acinar, 8 papillary and 2 mucinous). The KRAS+ group had 32 pure types of adenocarcinomas, including nine acinar, eight solid, seven mucinous, four papillary, two clear and two nonmucinous bronchioloalveolar carcinomas. Four pure acinar and one pure papillary adenocarcinomas were identified in the EGFR+ group.

The most common primary histological types in the mixed subtype of adenocarcinomas were acinar (40%), bronchioloalveolar (15%), solid (15%), papillary (12%), mucinous (10%) and micropapillary (4%). The most common secondary histological types included acinar (29%), papillary (24%), solid (17%), bronchioloalveolar (14%), mucinous (6%) and micropapillary (4%).

Table 2 summarizes the primary and secondary histological types in mixed subtypes of adenocarcinoma in all the three study groups. Acinar growth pattern was the most common primary histological type, regardless of mutational status (62.5% EGFR+, 32% KRAS+ and 32% EGFR–/KRAS–) (Figure 1). Interestingly, no solid type was observed as a primary histological subtype in the EGFR+ mixed adenocarcinomas. Acinar (34%), followed by papillary (24%) and solid (14%) histological types were the frequently observed secondary histological types in KRAS+ tumors (Table 2). Acinar (28%), papillary (22%) and bronchioloalveolar (14%) were most the common secondary histological patterns in EGFR–/KRAS– mixed adenocarcinomas. Papillary (32%), acinar (22%) and bronchioloalveolar (19%) patterns were the most common secondary histological patterns in EGFR+ mixed adenocarcinomas (Figure 2).

Table 2 Primary and secondary histological patterns in mixed subtypes of a primary lung adenocarcinoma grouped by EGFR and KRAS mutational status
Full size table
Figure 1
figure 1

Primary histological patterns in mixed subtype adenocarcinomas and mutation type. AC, acinar; BAC, bronchioloalveolar; MUC, mucinous; PAP, papillary; SOL, solid.

Full size image
Figure 2
figure 2

Secondary histological patterns in mixed subtype adenocarcinomas and mutation type. AC, acinar; BAC, bronchioloalveolar; MUC, mucinous; PAP, papillary; SOL, solid.

Full size image

The absence of solid growth pattern in lung adenocarcinomas was a significant predictor of EGFR mutations (OR 0.01; 95% Wald confidence limit <0.001–0.34, P=0.0103). Mucinous growth pattern of adenocarcinomas was detected to be a significant predictor of KRAS mutations (OR 3.938; 95% Wald confidence limit 1.574–9.852, P=0.0034).

Mutational Status and EGFR FISH Analysis

Fluorescence in situ hybridization was successful in 344 cases. Technically suboptimal hybridization occurred in one tumor positive for EGFR mutation. A possible correlation between FISH results and mutation type was analyzed.

The EGFR gene amplification was detected in only 21 cases (6%) (Figure 3). In contrast, when the Colorado scoring criteria was applied, 50 cases (14.5%) were classified as FISH positive (Figure 3). This discrepancy in interpretation is mainly a result of high polysomy, an increased copy number of chromosome 7 without the EGFR gene amplification.

Figure 3
figure 3

Summary of FISH analysis of 344 cases of primary lung adenocarcinomas and mutational profile using two interpretation criteria.

Full size image

Only 5 cases (14%) of the EGFR-mutated tumors were positive for EGFR gene amplification defined as a ratio between EGFR gene copy numbers and chromosome 7 >2 (P=0.055). All these cases were positive for the EGFR exon 19 mutation. EGFR gene amplification was identified in 6 (6%) KRAS-mutated tumors, and in 10 (5%) tumors negative for EGFR and KRAS mutations (P=0.787) (Figure 4). No significant relationship was detected between the EGFR gene amplification and mutation type.

Figure 4
figure 4

Examples of acinar type adenocarcinomas with KRAS (a and b) and EGFR mutations (d and e). Both tumors were EGFR FISH positive. (c) FISH positivity defined by EGFR gene amplification (green signal—CEP7 probe; red signal—EGFR gene probe). (f) FISH positivity defined by high polysomy in the absence of EGFR gene amplification.

Full size image

According to the Colorado scoring criteria, 11 cases (31%) of EGFR-mutated tumors (9 exon 19 and 3 exon 21 mutations) were EGFR FISH positive. In all, 11 (11%) KRAS-mutated and 28 (14%) tumors negative for KRAS and EGFR mutations were interpreted as EGFR FISH positive. Interestingly, there was a significant relationship between the presence of EGFR mutations and FISH positivity determined by the Colorado scoring system (P=0.024). In contrast, no significant relationship was detected between the Colorado scoring system and KRAS mutations (P=0.586). Next, we were interested to determine which of the proposed criteria included in the Colorado scoring system could be a significant predictor of EGFR mutation. A high polysomy (≥four EGFR gene copies in ≥40% of the cells) was detected as a significant predictor of EGFR mutations (OR 7.096; 95% Wald confidence limit 2.144–23.486, P=0.0013). None of the proposed criteria were significant predictors of KRAS mutation.

There was no correlation between morphology and EGFR FISH positivity.

Discussion

Our results of EGFR and KRAS mutation analyses in lung adenocarcinomas generally validated and extended the observations from previous studies in North American patients. We found that in our surgically treated patients with lung adenocarcinomas, 11% were harboring EGFR mutations and 30% KRAS mutations. Our study also confirmed that EGFR mutations are more prevalent in women and nonsmokers, whereas KRAS is more frequent in smokers.

Similar to Blons et al,32 we observed no differences in tumor stage between resected tumors, with majority of patients presenting as a stage I disease regardless of tumor mutational profile. Recently, Marks et al41 reported their experience with EGFR and KRAS mutations in nearly 300 surgically resected adenocarcinomas. In their experience, almost 40% of the patients with KRAS mutations presented with stage II or higher disease, whereas 88% of the patients with EGFR mutations presented with stage I disease. Similarly, Kosaka et al42 reported higher frequency of KRAS mutations in advanced-stage lung adenocarcinomas and EGFR mutations in stage I adenocarcinomas in surgically treated Japanese patients. Even though all the studies analyzed surgically treated patients, there were differences in patient selection, which may reflect observed differences in stage presentation. In our study, only surgically treated patients with primary naïve lung adenocarcinomas were selected, whereas other two studies included some patients who received previous neoadjuvant or adjuvant cytotoxic chemotherapy before surgery.

Previous reports indicated that EGFR-mutated tumors tend to be well differentiated, whereas KRAS tumors are most likely poorly differentiated.43 We observed a similar tendency in our study group, but the trend did not reach statistical significance. This may have been a result of sample size, and if larger number of cases had been analyzed, differences in tumor grade may be a statistically significant.

Our study is the first to show that the intensity of tumor-infiltrating lymphocytes is a predictor of EGFR mutations. The significance of inflammatory cells within or surrounding solid tumors including lung is controversial. Several studies have shown that increased number of tumor-infiltrating lymphocytes are associated with better prognosis.44, 45 The limitation of our study is the lack of survival analysis. The routine clinical screening for EGFR and KRAS mutations was introduced in our practice in early 2005, and therefore, a follow-up period is too short to make any reliable and conclusive survival analysis. It would be interesting to determine a prognostic relationship between intensity of tumor-infiltrating lymphocytes and tumor genotype. Recent studies suggest that the type, not the quantity of tumor-infiltrating inflammatory cells, may be a more significant prognostic determinant, although data about prognostically significant cell type are contradictory.44, 45, 46 Pelletier et al47 suggested that peritumoral B cells, not T cells, are related to a better survival. It was shown that only stromal CD4+ tumor-infiltrating lymphocytes are associated with a favorable prognosis in non-small cell lung carcinomas.48 Hiraoka et al49 found that only concurrent infiltration by CD8+ and CD4+ T lymphocytes in the tumor epithelial cells is a good prognostic indicator. It is obvious that the role of tumor-infiltrating lymphocytes is still controversial and most of the studies analyzed relatively small number of cases. A knowledge of the type of tumor-infiltrating inflammatory cells may open a new avenue for possible immunotherapeutic manipulations that may potentially enhance a clinically desirable response to targeted therapies and outcome for lung cancer patients.

Initial reports indicated that EGFR mutations are most frequently observed in bronchioloalveolar carcinomas.11, 13 However, studies that applied strict World Health Organization definition of bronchioloalveolar carcinoma failed to show this association. Subsequently, several reports showed that mixed subtype of invasive adenocarcinoma with a bronchioloalveolar component were commonly associated with EGFR mutations, although other studies did not confirm this correlation. More recent studies found a link between papillary differentiation and EGFR mutations.38 There are several reasons for these differences, such as the patient selection criteria and different criteria and interpretations by pathologists. Furthermore, in some studies, cytology specimens or small biopsy specimens only were reviewed, whereas in others either larger surgical specimens or, most frequently, a combination of both surgical and cytology specimens was included. In our study, we have tried to exclude all these potentially confounding factors. We reviewed only the surgical resection specimens. The majority of our cases measured up to 3 cm in diameter, and they were entirely submitted for histological analysis. Pulmonary adenocarcinomas are known for their morphological heterogeneity, and therefore, it is very important to review a large number of histological sections. One of the main advantages of our study was that, we compared the histology among three genotypically different groups of lung adenocarcinomas, including tumors that are negative for both EGFR and KRAS mutations. Before any conclusions about histology–genotype correlation are made and before any consideration is given to morphology as a selection criteria for molecular analysis of lung carcinomas, it is very important to analyze tumors with EGFR and KRAS wild types, as they are the most common types in the Western population. Our study focused on naïve cases of lung adenocarcinomas only, excluding a possible influence of chemotherapy/radiation on tumor morphology.

As expected, the majority of cases in all the three genotypic groups were classified as a mixed subtype of adenocarcinoma. As recently suggested, we further assigned the percentages for each observed histological subtype and divided them into primary and secondary growth patterns.38 There was a large overlap in the histological growth patterns among the three groups. Acinar growth pattern was the most common dominant pattern of mixed subtype of adenocarcinoma in all the three groups. We showed that none of the mixed type adenocarcinomas with EGFR mutations showed a dominant solid growth pattern. In addition, logistic regression analysis for the first time showed that absence of solid growth pattern is a predictor of EGFR mutations. There are many studies from Japan that found a major papillary subtype to be associated with EGFR mutations. Similary, Motoi et al35 showed the same association in the Western population. In our study, papillary type was the most common secondary subtype in EGFR-mutated group, but it was also present in other two study groups.38 Most importantly, statistical analysis failed to show any significance of this subtype. Similar to Finberg et al,37 we found a strong correlation of mucinous differentiation in mixed subtypes of adenocarcinomas and KRAS mutations.40 Marchetti et al50 showed the same correlation in mucinous types of bronchioloalveolar carcinoma. It is very interesting that mucinous differentiation is a significant predictor of KRAS mutations. However, this was not restricted to the KRAS-mutated group only. It was also identified in EGFR-mutated group, as well as in the EGFR and KRAS wild types. It would be interesting to know whether mucinous differentiation in EGFR-mutated group has any effect on the patient's response to tyrosine kinase inhibitors therapies.

Many reported studies indicated that the EGFR FISH results correlate with the EGFR gene mutations, but the correlation is not absolute. The main issue is the definition of FISH-positive and -negative results. EGFR amplification by FISH has been reported in 7–40% of non-small cell lung carcinomas, more frequently in squamous cell carcinomas. This wide range most likely reflects a variation in techniques and interpretation criteria. However, balanced trisomy and polysomy are more frequent events leading to EGFR gain. The Colorado group introduced a very complex scoring system for EGFR FISH interpretations, which correlates very well with the patient's response to tyrosine kinase inhibitors therapies and may have prognostic significance. We clearly showed that EGFR amplification alone does not predict EGFR mutations. However, EGFR FISH positivity and negativity defined by the Colorado criteria was a good predictor of EGFR mutations. As there are many criteria included in the Colorado scoring scheme, we decided to determine which criteria is the best predictor of mutations. Many of our cases that showed EGFR amplification, also had a high polysomy. However, a significant number of cases had high polysomy alone particularly in EGFR-mutated group. Logistic regression showed that high polysomy is the best predictor of EGFR mutations. It is important to mention, that this observation was not limited to the EGFR-mutated group only. It was also observed in the other two studied groups indicating that FISH analysis cannot replace mutational analysis. Our study indicates that is important to distinguish polysomy and gene amplification for EGFR copy number assessment in lung adenocarcinomas. Furthermore, this indicates that EGFR FISH is a preferable method over chromogenic in situ hybridization, which cannot readily distinguish chromosome 7 polysomy from EGFR gene amplification. On the other hand, many studies indicate that mutational analysis is the best predictor of patient's response to EGFR tyrosine kinase inhibitors. In that case, EGFR FISH cannot be used as a replacement for mutational analysis, as EGFR FISH positivity is observed across all the three study groups regardless of their genotype.

In summary, detection of lung carcinoma genotype is important for treatment decision. There is variability in the extent of testing of lung carcinomas. Currently, some institutions are testing every lung adenocarcinoma, some are testing none and some are testing only lung carcinomas at the oncologist request. There is a need to develop a universal algorithmic approach based on clinical and histological parameters. A comprehensive statistical analysis in our study showed clinical and pathological predictors of lung adenocarcinoma genotype. Owing to the large morphological and clinical overlap between tumors with and without EGFR and KRAS mutations, none of the criteria can be used as the selection criteria for targeted molecular testing in an individual case. In the future, personalized medicine will become the standard of care for adenocarcinomas, and the possibility exists for all lung carcinomas to undergo molecular testing. Until such time and in the current economical environment, we need to develop an algorithmic method that would best reduce the cost of molecular testing. As KRAS mutations are more common in the Western population, it would be the first choice for testing; if the results are negative, EGFR mutational analysis could be carried out as a second step.