Clinicopathological characteristics of EGFR mutated adenosquamous carcinoma of the lung


Correspondence: Genichiro Ishii, MD, PhD, Pathology Division, Research Center for Innovative Oncology, National Cancer Center Hospital East, Kahiwa, 277-0882 Chiba, Japan. Email:; Atsushi Ochiai, MD, PhD, Pathology Division, Research Center for Innovative Oncology, National Cancer Center Hospital East, Kahiwa, 277-0882 Chiba, Japan. Email:


Adenosquamous carcinoma of the lung (Ad-Sq) is an uncommon subtype with poor prognosis. We analyzed the clinicopathological characteristics of Ad-Sq, focusing the correlation between Epidermal Growth Factor Receptor (EGFR) mutation and clinicopathological factors. A total of 67 cases were selected from September 1992 to May 2011. EGFR mutational analysis (n = 59) was performed by direct sequence. We also performed immunohistochemical staining for EGFR mutated cases using the two mutation-specific antibodies for deletion and L858R. Postoperative 3-year survival rate of Ad-Sq was 58.7%, statistically worse in comparison with adenocarcinoma (58.7% vs. 78.1%, P = 0.038). Twenty-four percent (14/59) were positive for EGFR mutations. Patients who had never been smokers and who were lymphatic permeation positive were seen more frequently in the mutation positive group (P = 0.035, 0.027, respectively). Moreover, the EGFR mutated group tended to have a more positive prognosis than negative. Focusing on the pathological features, the lepidic growth pattern was more frequently seen in the positive group (P = 0.018). Immunoreactivity for the DEL-specific and L858-specific antibody were observed in both adenocarcinoma and squamous cell carcinoma components. Our study demonstrated that EGFR mutated Ad-Sq had similar clinicopathological features as EGFR mutated adenocarcinoma.

Adenosquamous carcinoma of the lung (Ad-Sq), which is defined as a carcinoma containing components of both squamous cell carcinoma (Sq) and adenocarcinoma (Ad) and with each component comprising at least 10% of the tumor, is an uncommon subtype, accounting for 0.4 to 4% of all non-small cell lung cancer cases.[1]

Because of its rarity, the clinicopathological aspects of Ad-Sq have not received much attention, and the pathogenesis of Ad-Sq remains unknown. In general, the prognosis of Ad-Sq was reportedly worse than that of Ad and Sq.[2, 3]

Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs), such as gefitinib and erlotinib, are therapeutic options widely used for the treatment of Ad patients. The responders to EGFR-TKI have somatic mutations in the EGFR tyrosine kinase domain, including deletions in exon19 and a point mutation at codon 858 in exon21 (L858R). For these patients, EGFR-TKI has a high response rate of approximately 70% and results in better prognosis.[4-7]

However, the usefulness of EGFR-TKI treatment is limited to Ad, and few studies have analyzed the effectiveness of EGFR-TKI for Ad-Sq. Until recently, EGFR mutation in Ad-Sq has been mentioned in only a few reports. Sasaki et al. reported that 4 (15%) of 26 Ad-Sq patients were positive for EGFR mutations, which is a lower frequency than that for Ad.[8] On the other hand, Kang et al. reported that 11 (44%) of 25 Ad-Sq patients were positive for EGFR mutations among a Korean population.[9]

In the present study, we selected 67 Ad-Sq cases and investigated the clinicopathological characteristics of Ad-Sq. Focusing on the EGFR mutation in Ad-Sq, we analyzed 59 Ad-Sq cases with special reference to the correlation between mutation status and clinicopathological factors. Additionally, we performed immunohistochemical staining for the EGFR-mutated cases using two mutation-specific antibodies for a deletion and L858R to detect the distribution of mutated cells in each component.

Materials and Methods

Case selection

During the period from September 1992 to May2011, 3953 patients with lung cancer underwent surgical resection at the National Cancer Center Hospital East in Kashiwa, Japan. A total of 67 cases (2.6%) of Ad-Sq were selected. The clinicopathological characteristics of the Ad-Sq cases were examined.

Clinical information

All available clinical information was obtained from the clinical records. The records were reviewed for age, sex, smoking index, tumor markers, pathological stage, and duration of follow up. Two of the 67 patients underwent a wedge resection, 4 underwent a partial resection, 57 underwent a lobectomy, and 4 underwent a pneumonectomy. In addition, we collected clinicopathological data about 2628 Ad and 1258 Sq patients who underwent surgical resection, for the purpose of comparing with Ad-Sq.

Pathological information

In each case, the resected tissue had been fixed in 10% formalin or absolute methyl alcohol and embedded in paraffin. The tumors had been cut into 5–10 mm slices, and serial 4-μm sections had been stained with hematoxylin and eosin. The Alcian blue-periodic acid Schiff (AB-PAS) method was used to visualize cytoplasmic mucin production, and the Victoria van-Gieson (VVG) method was used to visualize elastic fibers. Two pathologists (T.S. and G.I.) reviewed all the slides through a multiheaded microscope. The tumors were classified according to the criteria of the current histological classification adopted by the World Health Organization.[1] The Ad component was characterized as tumor cells showing an acinar, papillary, bronchioalveolar, or solid with mucin growth pattern or that were positive for PAS Alcian Blue staining.[1, 10, 11] The Sq component was defined as tumor cells with keratinization, pearl formation, and/or intercellular bridges.[1] If each of the components with these features did not account for a minimum of 10% of the entire tumor, the case was excluded from the present study. As a result, we identified 67 (2.6%) cases of pathologically diagnosed Ad-Sq (Fig. 1).

Figure 1.

Two typical examples of adenosquamous carcinoma. (a) Low-power field showing tumor sections (H&E). Both adenocarcinoma and squamous cell carcinoma components are visible in the tumor sections. (b) Adenocarcinoma component with acinar pattern (H&E x100). (c) Squamous cell component with keratin pearl formation (H&E). (d) Low-power field showing tumor sections (H&E). (e) Adenocarcinoma component with lepidic growth pattern (H&E, x100). (f) Squamous cell component with keratin pearl formation (H&E).

EGFR mutational analysis using a molecular technique

From the FFPE tissue blocks, three slices of 10-μm-thick unstained sections were cut. Tissue areas in which the tumor cells occupied more than 70% of the area were macroscopically dissected and genomic DNA was isolated using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). The EGFR exon-19 and 21 fragments were amplified according to a previously described method, with some modifications. Briefly, each 50-μL PCR cocktail contained 50 ng of genomic DNA, 1.5 mM of magnesium chloride, 200 mM of deoxynucleotide triphosphates, 0.2 mM of PCR primers and 2.5 U of HotStarTaqw DNA polymerase (Qiagen). The PCR conditions were as follows: 1 cycle at 95°C for 15 min; 40 cycles at 95°C for 30 s, 65°C for 30 s and 72°C for 1 min; and 1 cycle at 72°C for 10 min. The primer sequences were


The PCR products were purified using ExoSAP-IT (Affymetrix, Santa Clara, CA, USA), and the amplicon size and amount were confirmed using DNA agarose gel electrophoresis. The purified PCR products were sequenced directly using the same primers as those used for the PCR. The BigDye Terminator v3.1 Cycle Sequencing Kit and 3500 Genetic Analyzer (Life Technologies, Carlsbad, CA, USA) were used according to the manufacturer's instructions. Analyses of the DNA sequences were performed using Sequencing Analysis Software v3.4.1 (Life Technologies).


After the detection of the EGFR mutational status using direct sequencing, immunohistochemical staining was performed in the EGFR-mutated cases. For the immunohistochemical staining, 5-μm-thick sections were deparaffinized. Positive control was taken along with the immunohistochemical analysis. Antigen retrieval was performed at 95°C for 45 min in Target Retrieval Solution, pH 9.0, x10 (Dako, Glostrup, Denmark). The primary antibodies that we used were a monoclonal antibody against human EGFR with the DEL (E746-A750del) mutation (1:75, clone 6B6; Cell Signaling Technology) and a monoclonal antibody against human EGFR with the L858R mutation (1:75, clone 43B2; Cell Signaling Technology). The antibodies were diluted in Signal Stain (Cell Signaling Technology), and the slides were incubated at 4°C overnight. After three washes in TBS for 5 min, the slides were incubated for 30 min at room temperature with anti-rabbit polymer-HRP secondary antibody (Dako). The immunoreactions were detected using 3,3-diaminobenzidine, followed by counterstaining with hematoxylin.

Statistical analysis

The statistical analysis was performed using SPSS 12.0 for Windows. Chi-square tests were used for categorical variables, and a P value < 0.05 was regarded as statistically significant.


Patient characteristics

The clinical and pathological characteristics of Ad-Sq are summarized in Table 1. Among the 67 Ad-Sq patients, 48 were men and 19 were women. The median age at surgery was 70 years (range, 45–84 years). None of them received EGFR-TKI therapy.

Table 1. Clinicopathological characteristics of the patients with adenosquamous carcinoma (Adsq)
 Ad-SqAdP valueSqP value
n = 67n = 2628n = 1258
Age, years70 (45–84)66 (22–92)0.00268 (31–88) 
<70 years321731 6890.261
≧70 years35897 569
Smoking History     
Current or former4315080.2661215<0.001
Never241120 43
<5.0 ng/dl231687<848<0.001
≧5.0 ng/dl449410.001410
<1.5 ng/ml402111<0.0014740.001
≧1.5 ng/ml27517784
Pathological stage     
Vascular invasion     
Lymphatic permeation     
Pleural involvement     

When the clinicopathological factors of Ad-Sq were compared with those of Ad and Sq, Ad-Sq was more likely to occur in men and older patients than Ad (P = 0.001, 0.002, respectively). On the other hand, Ad-Sq was less likely to occur in men and smokers than Sq (P < 0.001). T factor and pleural involvement were also seen in Ad-Sq more frequently than in Ad (P < 0.001). However, no statistically significant difference was seen between Ad-Sq and Sq with regard to the T factor and pleural involvement. The Ad-Sq was significantly correlated with vascular invasion (P < 0.001 vs. Ad, P = 0.004 vs. Sq) and lymphatic permeation (P < 0.001 vs. Ad, P = 0.001 vs. Sq).

Figure 2 shows the Kaplan–Meier survival curves for the three histological subtypes. For the 59 Ad-Sq patients with follow-up data, the postoperative 3-year survival rate was 58.7%. A statistically significant difference was recognized when compared with Ad (58.7% vs. 78.1%, P = 0.038), but not when compared with Sq (58.7% vs. 66.3% P = 0.103).

Figure 2.

Three-year survival curves after surgery, as calculated using the Kaplan–Meier method. The Ad-Sq group had a significantly poorer survival than the Ad group (log-rank test, P = 0.038). However, no significant difference compared with the Sq group was seen (log-rank test, P = 0.103).

EGFR mutational status of Ad-Sq

The EGFR mutational status was measured in 59 patients using direct sequencing. A total of 14 patients were positive for EGFR mutations (24%). DEL in exon19 was detected in nine cases. The details were as follows: del E746-A750 was detected in six cases, and the other three cases were del L747-T751, del E746-A753, and del S752-I759, respectively. A point mutation at codon 858 (L858R) in exon21 was detected in three cases, and L861Q, which is associated with an increased sensitivity to EGFR-TKIs, was detected in two cases.[12, 13]

Correlation between EGFR mutations and the clinicopathological factors of Ad-Sq

To reveal the clinical and pathological features of Ad-Sq with EGFR mutation, we assessed the relationship between EGFR mutations and the clinicopathological factors (Table 2). Patients who had never been smokers and with lymphatic permeation were more frequently seen in the mutation-positive group than in the mutation-negative group (P = 0.035, 0.027 respectively). The 3-year survival rate of the mutation-positive group tended to be better than that of the mutation-negative group (90.0% vs. 62.8% respectively, P = 0.06; Fig. 3).

Figure 3.

EGFR mutational status effects on survival, as calculated using the Kaplan–Meier method. The EGFR-mutation-positive group tended to have a better survival than the negative one (log-rank test, P = 0.06).

Table 2. Correlation between the clinicopathological factor and EGFR mutation status
 EGFR-mutation(+)EGFR-mutation(−)P value
n = 14n = 45
Age, years70 (45–832)70 (45–84)0.966
<70 years619 
≧70 years826 
Smoking History   
Current or former6330.035
<5.0 ng/dl5150.869
≧5.0 ng/dl930 
<1.5 ng/ml11250.123
≧1.5 ng/ml320 
Pathological stage   
Vascular invasion   
Lymphatic permeation   
Pleural involvement   

We compared the pathological findings of both groups (Table 3). No significant difference in the proportions of predominant subtypes in the Ad component was seen between the mutation-positive and negative groups. On the other hand, a lepidic growth pattern was more frequently seen in the mutation-positive group (P = 0.018) (Fig. 1d–f). An acinar pattern was the most frequent subtype of the Ad component in both the mutation-positive and negative groups. No statistically significant difference in the grade of differentiation was seen for the Sq component.

Table 3. Correlation between morphological characteristics and EGFR mutation
 EGFR mutation positiveEGFR mutation negativeP value
n = 14n = 45
Adenocarcinoma component
Predominant subtype
Lepidic growth pattern
Squamous cell carcinoma component grade of differentiation

Immunohistochemical staining

We performed immunohistochemical staining for EGFR-mutated cases. A DEL-specific antibody was used for the cases that were positive for a DEL in exon19, while an L858R-specific antibody was used for the cases that were positive for L858R. The clinicopathological and immunohistochemical results for the 14 EGFR-mutated cases are summarized in Table 4. Immunoreactivity for the DEL-specific antibody was observed in three of the six del E746-A750 cases and the one del E746-A753 case (identified using direct sequencing). Neither the L747-T751-positive nor the S752-I759-positive case exhibited immunoreactivity for the DEL-specific antibody. Although all the L858R mutated cases (100%) were positive for the L858R-specific antibody, neither of the L861Q-mutated cases exhibited a positive immunoreaction. As shown in Table 4 and Supplemental Table, a positive outcome for the DEL-specific antibody was more likely to occur in women and non-smokers. In contrast, a positive outcome for the L858R-specific antibody was more likely to occur in men and smokers. There were not any other clinicopathological differences every mutation. In all seven cases that were positive for mutation-specific immunostaining, the same immunohistochemical reactions were detected in both the Ad component and the Sq component (Fig. 4a–d).

Figure 4.

Immunohistochemical staining for EGFR-mutation-specific antibody. (a, b) Both the adenocarcinoma component with an (a) acinar pattern and (b) the squamous cell carcinoma component were positive for the L858R-specific antibody. (c,d) The adenocarcinoma component with (c) a lepidic growth pattern and (d) the squamous cell carcinoma component were positive for the DEL-specific antibody.

Table 4. EGFR mutation positive case (n = 14)
 AgeSexSmokingp-stageMutationIHC-stainIHC -stain





























In this study, we analyzed the clinicopathological characteristics of Ad-Sq, focusing on the correlation between EGFR mutations and clinicopathological factors. This is the first study to demonstrate that EGFR-mutated Ad-Sq had clinicopathological features, similar to those of EGFR-mutated adenocarcinoma, specifically to never having been a smoker, better prognosis, and pathologically lepidic growth pattern. It was suggested that Ad-Sq of the lung could be subclassified into distinct subtypes with different pathogenetic mechanisms according to the EGFR mutation status.

Until recently, a few studies have analyzed EGFR mutations in Ad-Sq.[8, 9, 14-17] However, we analyzed 67 cases of Ad-Sq, the largest number of cases among similar studies published to date. Moreover, we performed a clinicopathological analysis using not only a molecular method, but also immunohistochemistry with two mutation-specific antibodies.

In our study, the frequency of the EGFR mutation in Ad-Sq was 24% (14/59). In the previous studies, Sasaki et al. reported that 4 of 26 (15%) Ad-Sq were positive for an EGFR mutation, and Toyooka et al. reported that 3 of 11 (27%) cases were positive, similar to the results of the present study.[8, 15] On the other hand, a discrepancy was observed with regard to the frequency of EGFR mutation when compared with results obtained in neighboring countries. Kang et al. reported that 44% (11/25) of Ad-Sq were positive for an EGFR mutation in a Korean population, and Xiao et al. reported a frequency of approximately 38% (21/55) in a Chinese population.[9, 17] This discrepancy might be attributable to differences in the detection methods used for the mutational analysis; alternatively, genetic differences between Japanese populations and others may also affect the frequency of mutation.

The sensitivity of the DEL-specific antibody used in our study was inferior to that of the L858R-specific antibody, as previously reported.[18-22] Kozu et al. reported that the sensitivities of DEL-specific and L858R-specific antibodies were 42.2% and 75.6%, respectively.[18] Immunohistochemical staining revealed that the EGFR mutation occurred not only in the Ad component, but also in the Sq component. Staining intensity in both components was similar, as shown in Fig. 4 and Table 4. Several studies using microdissection analyses have demonstrated identical EGFR mutations in both components, but few immunohistochemical analyses have been performed. Recently, Miyamae et al. demonstrated that identical EGFR mutations were observed in the Ad and Sq components using a DEL-specific antibody, suggesting that Ad-Sq to be generated from monoclonal origin.[23] Rekhtman et al. described that squamous cell carcinoma with EGFR mutation had a component of adenocarcinoma using immunohistochemistry.[24] These findings may suggest that squamous cell carcinoma componemt was originated from adenocarcinoma somponent in Ad-Sq. In contrast, Kanazawa et al. described the monoclonal transition from squamous cell carcinoma to adenocarcinoma.[25] So, it is difficult at this time to conclude the origin of Ad-Sq.

Interestingly, a lepidic growth pattern was more frequent in the EGFR-mutated group than in the negative group. Blons et al. reported that EGFR mutations in Ad were more prevalent among cases with a lepidic growth pattern.[26] In this study, we detected the same correlation between the EGFR mutation status and the lepidic growth pattern in Ad-Sq as well as Ad. Sasaki et al. reported that the EGFR mutational status of Ad-Sq was significantly correlated with the smoking history and gender.[8] Kang et al. also reported that EGFR mutations in Ad-Sq were more frequent among people who had never been smokers and women.[9] The current study revealed a similar correlation with smoking, but we did not find a significant correlation with sex.

A Kaplan–Meier survival curve showed that the survival of the EGFR-mutated cases was better when compared with the negative cases, even though a statistically significant difference was absent. Although a similar tendency was observed in Ad cases,[27, 28] no previous studies have compared the survival of EGFR-positive Ad-Sq cases with those of EGFR-negative cases. In the future, large-scale, multi-institutional studies will be needed to verify the results.

In this study, we found several common features between EGFR-mutated Ad-Sq cases and those with Ad. Therefore, we speculated that a subtype with clinicopathological features resembling those of Ad existed among the Ad-Sq tumors with an EGFR mutation. Based on our analysis, Ad-Sq developing in a patient who had never been a smoker and pathologically exhibiting a lepidic growth pattern in the Ad component was likely to be positive for an EGFR mutation. EGFR-TKIs, such as gefitinib and erlotinib, may be effective therapeutic options for patients with EGFR-mutated Ad-Sq as well as those with EGFR-mutated Ad. Mitsudomi et al. have reported excellent clinical response to gefitinib in a patient with EGFR mutated Ad-Sq.[29] Shukuya et al. reported that efficacy of EGFR-TKIs for non-adenocarcinoma with EGFR mutation was different from that of pure adenocarcinoma.[30] So, we consider that further large-scale clinical trials of EGFR-TKI for adenosquamous carcinoma patients are needed, as is validation of an optimal therapeutic strategy.

As shown in Fig. 4, some cancer cells showed positive reaction for mutation specific antibodies. Although biological significance of this nuclear staining pattern is not clarified, further examination will be required.

In conclusion, we demonstrated that EGFR-mutated Ad-Sq exhibited characteristics similar to those of EGFR-mutated Ad in that they were more likely to occur in patients who had never been smokers and more likely to exhibit a lepidic growth component pathologically. Moreover, patients with EGFR-mutated Ad-Sq tend to have a better outcome, compared with the EGFR-mutation-negative group. So we found the correlation between EGFR mutation and the prognosis of adenosquamous carcinoma of lungs, similar to that of adenocarcinomas. In particular, Ad-Sq cases that carry an EGFR mutation might respond to EGFR-TKI therapy.