The importance of EGFR mutation testing in squamous cell carcinoma or non-small cell carcinoma favor squamous cell carcinoma diagnosed from small lung biopsies

Background Adenosquamous carcinoma (ADSC) of the lung, a rare but aggressive subtype of non-small cell lung cancer (NSCLC), is defined as a carcinoma containing components of adenocarcinoma (ADC) and squamous cell carcinoma (SqCC). Mutations of epidermal growth factor receptor (EGFR) are found at a frequency of 15 to 44% in Asian ADSC, and EGFR tyrosine kinase inhibitors (EGFR-TKIs) are a more effective treatment for EGFR-mutated ADSC compared to chemotherapy. However, ADSC in small lung biopsies could be misdiagnosed as SqCC or non-small cell carcinoma (NSCC) favor SqCC due to undersampling, which may result in neglecting of EGFR mutation testing and affecting patients’ clinical management, particularly in Asian patients that relatively have high prevalence of EGFR mutation. Methods A total of 148 small lung biopsy cases with pathological diagnosis of SqCC or NSCC favor SqCC were retrospectively enrolled. The frequency of EGFR mutations and the correlation between patients’ EGFR mutation status and clinicopathological characteristics were evaluated. Results EGFR mutations were found in 8.8% (13 /148) of all cases with 5.2% (7/135) in SqCC and 46.2% (6/13) in NSCC favor SqCC. There were 7 (53.8%) L858R mutation, 4 (30.8%) exon 19 deletions, and 2 (15.4%) cases with coexistent L858R and T790 M mutations. Multivariate analysis showed that EGFR mutations were more prevalent in never-smokers (83.3% versus 16.7%, p = 0.006) and patients diagnosed as NSCC favor SqCC (46.2% versus 5.2%, p = 0.001). Moreover, 75% (3/4) of EGFR mutation-positive cases with subsequent surgical resection or rebiopsy were further diagnosed as ADSC. Conclusions EGFR mutation testing should be performed in Asian patients with SqCC diagnosed from small lung biopsies, especially in never-smokers and patients with diagnosis of NSCC favor SqCC, which have a high probability of being ADSC.


Background
Lung cancer is the most common cause of cancerrelated death worldwide. Non-small cell lung cancer (NSCLC), accounting for approximately 80-85% of all the cases, is further divided into major histological subtypes including adenocarcinoma (ADC), squamous cell carcinoma (SqCC), large cell carcinoma (LCC), and other rare tumors including adenosquamous carcinoma (ADSC) [1]. ADSC is a rare and mixed-type NSCLC, counting for only 0.4 to 4% of all the cases. According to the World Health Organization (WHO) Classification of Lung Tumors, ADSC is defined as "a carcinoma showing components of both SqCC and ADC, with each comprising at least 10% of the tumor" [2]. ADSC has a more aggressive behavior and worse prognosis than ADC and SqCC [3,4]. The treatment of early-stage ADSC mainly includes surgery, supplemented by chemotherapy, and other comprehensive treatment, but the curative effect is not satisfied [5]. The poor prognosis of ADSC has raised the demand for more effective treatment approaches such as targeted therapy. Recently, the development of targeted therapy toward oncogenic "driver mutations" has revolutionized the clinical management of patients with NSCLC including ADSC [6,7].
Epidermal growth factor receptor (EGFR) mutation is the most well-characterized driver mutation in NSCLC, and is more prevalent in Asian population, never smokers, and patients with ADC subtype [8]. More than 60% of Asian ADC patients were found to have EGFR mutations [9]. EGFR tyrosine kinase inhibitors (EGFR-TKIs), giving higher response rates, longer progressionfree survival and better quality of life compared to conventional chemotherapy, are now recommended as the first-line treatment in patients with advanced EGFR activating mutation-positive NSCLC [10][11][12][13]. Currently, studies demonstrated that EGFR mutations occur at a frequency of 15 to 44% in Asian ADSC and are mainly in female and never smokers; patients with ADSC showed a similar benefit from EGFR-TKI treatment compared to those with ADC, revealing that EGFR-TKI may be a therapeutic option for ADSC patients [14][15][16][17][18]. In view of this, the current guidelines recommend that EGFR mutation testing should be performed in NSCLC patients with ADC or mixed-type tumors containing an adenocarcinomatous component, for example, ADSC, and is not necessary for patients with tumors lack of any adenocarcinomatous component, such as pure SqCC or pure LCC [19,20].
In clinical practice, accurate diagnosis of ADSC can only be made in resection specimens, because a minimum of 10% of each ADC or SqCC component is required. However, approximately 70% of lung cancer patients are diagnosed at advanced stages that are not amenable to surgical resection and small biopsies are the major tissue resources for pathological diagnosis and subtyping [21,22]. Diagnosis of ADSC before surgical resection is not possible, since tumors present in small biopsy specimens may come from either ADC or SqCC component only. In the majority of cases, ADSC in small biopsy specimens is misdiagnosed as SqCC or poorly differentiated SqCC referring to non-small cell carcinoma (NSCC) favor SqCC due to biased sampling of the lesion, which may cause the neglect of EGFR mutation testing in these cases. A previous report by Rekhtman et al. also suggested that EGFR or KRAS mutations do not occur in pure pulmonary SqCC, and the occasional detection of these mutations may come from the challenge of ADSC diagnosis in small biopsies [23]. Furthermore, two NSCLC cases, possibly ADSC, presented by Saini et al. were initially diagnosed as SqCC in small biopsies, positive for EGFR mutations and benefitted with oral EGFR-TKI treatment [24]. The lack of EGFR mutation testing in patients who are diagnosed as SqCC from small biopsies may influence the treatment strategies for patients, especially in Asian population which has a high frequency of EGFR mutations. For this reason, we conducted an Asian retrospective study investigating the frequency and clinicopathological association of EGFR mutations in patients diagnosed as SqCC in small lung biopsies to evaluate the necessity and clinical impact of EGFR mutation testing on possibly under-sampled ADSC.

Patients
From 2007 to 2016 at Taipei Veterans General Hospital, patients with a diagnosis of SqCC or NSCC favor SqCC derived from small lung biopsies and without previous history of any carcinoma elsewhere in the body were evaluated in this retrospective study. Each case was reviewed by two independent pathologists, and all final diagnoses were made according to the 2015 WHO classification of lung cancer in small biopsies and cytology [2,25]. H&E stained slides were firstly used for standard morphological diagnosis. Tumors with SqCC morphological features, such as keratinization, intercellular bridge, and keratin pearl formation were diagnosed as SqCC. For those cases composed of poorly differentiated NSCLC without definite SqCC morphology, immunohistochemistry (IHC) was used to assist the pathological diagnosis. Cases positive for SqCC markers (p40 or CK5/6) with at least moderate to diffuse staining pattern and strong intensity, and negative for ADC markers (TTF-1 and/or mucin) were classified as NSCC favor SqCC. Cases with final diagnosis of SqCC or NSCC favor SqCC determined from small biopsies were further analyzed for EGFR mutation status and those without additional tissue sections for molecular tests were excluded in this study. This retrospective study was approved by the Institutional Review Board of Taipei Veterans General Hospital (2017-05-008CC).

Scoring of immunohistochemistry
The tissue sections were immunostained with commercially available antibodies including thyroid transcription factor-1 (TTF-1) (clone 8G7G3/1, Dako), p40 (clone BC28, Biocare Medical) and CK5/6 (clone D5/16 B4, Dako) on a Leica BOND-MAX system (Leica Microsystems, Newcastle, UK). Tumors completely absent of TTF-1 staining were considered as TTF-1 negative. For p40 or CK5/6 staining, the moderate staining pattern was defined as 10 to 50% of tumor cells showing immunoreactivity, while the diffuse staining pattern was defined as more than 50% of tumor cells with immunoreactivity. Cases with staining intensity similar to internal controls represented by bronchial/bronchiolar basal cell layer were recognized as strong intensity.

EGFR mutation analysis
H&E stained tissue slides were reviewed by pathologists to select the tumor areas and the selected areas were microdissected manually from the corresponding ones in the consecutive tissue sections and, after deparaffinization, were subjected to genomic DNA extraction. Briefly, the microdissected tumor tissue was transferred to an eppendorf tube containing Proteinase K solution (PicoPure™ DNA Extraction Kit, Applied Biosystems™). The tube was then incubated at 56°C for 16 h, followed by an inactivation step at 95°C for 10 min. The Proteinase K digested extract was adequate for molecular tests. Since EGFR exon 19 deletions and L858R mutation are the most common types of EGFR mutations, together accounting for approximately 90% of TKI-sensitive mutations, TaqMan EGFR Exon 19 Deletions Assay, L858R Mutation Detection Assay and Reference Assay (Invitrogen Life Technologies Inc., Waltham, MA, USA) were used to examine EGFR mutations according to the manufacturer's instruction. Real-time PCR was performed using the StepOnePlus™ Real-Time PCR System (Applied Biosystems). The mutation status of a sample was determined by using the parameter described previously [26]. Samples were considered as mutation positive if a specific sigmoid curve was observed with Ct of mutation reaction ≤37 and/or ΔCt value ≤7.

Statistical analysis
The correlation between patients' clinicopathological features and EGFR mutation status was assessed using Pathological features in EGFR mutation-positive SqCC diagnosed from small biopsies As shown in Table 2, among 13 EGFR mutation-positive cases, 7 were diagnosed as SqCC based on standard H&E morphological criteria with keratinization, intercellular bridge or keratin pearl formation. These cases showed strong and diffuse nuclear staining of p40, and  Fig. 2. Moreover, 4 of 13 EGFR mutation-positive cases had additional follow-up tumor tissues due to further surgical resection or rebiopsy after initial diagnosis. Three of them had their subsequent lobectomy specimens showing ADSC and one showed exclusively SqCC morphology in the follow-up lung biopsy (Table 2).

Discussion
Recent advances in targeted therapy have dramatically changed the clinical management of lung cancer patients and the molecular testing for driver mutations is now routinely used to predict patients' treatment response. The current guideline of CAP/IASLC/AMP suggests that testing for EGFR mutations and anaplastic lymphoma kinase (ALK) translocations should be recommended for adenocarcinomas and mixed lung cancers with an adenocarcinomatous component, regardless of histologic grade, and are not recommended in lung cancers that lack any adenocarcinoma component, such as "pure" SqCC, "pure" SCLC, or large cell carcinomas lacking any IHC evidence of adenocarcinoma differentiation [20]. In the era of precision medicine, accurate pathologic classification of NSCLC is mandatory for tailoring patients to specific molecular diagnostic testing. However, most of advanced lung cancers are diagnosed via tumor biopsies, which increases the difficulty in making a correct diagnosis, particularly in tumors without clear differentiation or in mixed-type tumors, such as ADSC.
Our study focuses on investigating the impact of EGFR mutation testing on ADSC in small biopsy diagnostics, especially in Asian patients that have high incidence of EGFR mutations. The misdiagnosis of ADSC remains a problem in up to 20% of small biopsy cases, even using immunohistochemistry [23,27]. A recent study reported by Zhang et al. showed that 38.8% (26/67) of postoperatively pathologically proven ADSC patients were pre-operatively misdiagnosed as SqCC and 29.8% (20/ 67) were misdiagnosed as poorly differentiated SqCC via biopsy or cytology specimens [28]. In this study, we found EGFR mutations are detected in 8.8% of small biopsy-diagnosed SqCC or NSCC favor SqCC, and significantly associated with never smokers and a diagnosis of NSCC favor SqCC. It has been reported that the incidence of EGFR mutations in Asian SqCC is less than 3% [29]. The frequency of EGFR mutations in our cases is higher than that reported in SqCC, suggesting that some of these small biopsy cases may not be singlehistology SqCC. Indeed, three of 4 (75%) EGFR mutation-positive cases in our study are subsequently diagnosed as ADSC, suggesting that most of EGFRmutated SqCC diagnosed from small biopsies may result from undersampling of ADSC. The limitation of this study is that this is a retrospective study and patients who are diagnosed as SqCC in small biopsies would not receive EGFR-TKI treatment according to current guidelines, therefore, no treatment  showed that ADSC patients harboring EGFR-sensitizing mutations who were treated with EGFR-TKIs had a significantly better prognosis than those receiving chemotherapy or chemoradiotherapy alone [28]. However, these studies only included patients with clinical stage I to IIIA, and a prospective study toward efficacy of EGFR-TKIs in advanced stage ADSC patients needs to be further conducted in the future.
Our multivariate analyses showed that never-smokers and NSCC favor SqCC, but not female gender, are independently associated with EGFR mutations. Lung cancer in never smokers has significant gender and geographic variations, occurring more frequently in female and in Asia. One possible explanation is that the majority of female NSCLC patients in Asian populations have no smoking history. Studies showed that the proportion of never smokers in female lung cancer cases are 83% in South Asia, while only 15% in the United States [31]. Moreover, lung cancer in Asian never smokers are often diagnosed as ADC that can be classified into different molecular subtypes based on oncogenic driver mutations. On the contrary, SqCC is more significantly associated with smokers than other subtypes of NSCLC [32]. From a molecular point of view, several studies showed that EGFR mutations are more prevalent in tumors of never smokers, while TP53 and KRAS mutations are more frequently found in smokers [31,[33][34][35]. Taken together, in our study cohort, the clinicopathological characteristics of EGFR mutation-positive cases are similar to those of ADC, implying that some of these cases might be EGFR-mutated ADSC. Furthermore, our findings that EGFR mutations significantly associate with a diagnosis of NSCC favor SqCC, which is defined as poorly differentiated NSCLC without definite SqCC morphology features, also suggest these cases may possibly be ADSC.

Conclusions
EGFR-TKI is a therapeutic option of EGFR-mutated ADSC which is often misdiagnosed as SqCC in small lung biopsies, resulting in absence of EGFR mutation testing. Our findings demonstrated the importance of EGFR mutation testing in small biopsy-diagnosed lung SqCC, especially in never smokers and patients whose diagnoses need to be assisted by immunohistochemistry.