This article is mentioned in:

Introduction

Lung cancer is the most common cause of cancer-associated mortality worldwide, accounting for 1.8 million newly diagnosed cases and 1.6 million incidences of cancer-associated mortality in 2012 (1). In recent years, epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs), including gefitinib, erlotinib or afatinib, have been widely used in the treatment of non-small cell lung cancer (NSCLC) patients, resulting in markedly prolonged progression-free and overall survival times compared with those of patients undergoing standard chemotherapies (2–7). However, the therapeutic effect of EGFR-TKIs is associated with the mutation spectrum and status of EGFR (8). Deletions in exon 19 and L858R in exon 21 are known to sensitize patients to EGFR-TKI therapy (9,10), with these mutations covering ~90% of oncogenic EGFR mutations (11). By contrast, the T790M substitution in exon 20 is the most common secondary mutation among patients who acquire resistance to EGFR-TKIs (12). Osimertinib, a third-generation TKI that specifically targets the T790M mutation, was introduced to in 2015 and has exhibited clinical efficacy in NSCLC patients with the T790M mutation (13). Thus, in order to select NSCLC patients suitable for EGFR-TKI therapy, it is necessary to perform EGFR molecular testing with cancer specimens from these patients (14).

In clinical practice, EGFR mutation status is commonly analyzed by polymerase chain reaction (PCR)-based assays using biopsy specimens from advanced NSCLC patients (14); however, cytological specimens are often the only specimens available. Although tissue specimens obtained by transbronchial or transcutaneous biopsies are preferable (14), it is often difficult to obtain sufficient cancer tissue to perform morphological and molecular analyses. Cytological specimens may, however, be useful diagnostic tools for these analyses and the procedures used to obtain these specimens are less invasive than those used to obtain biopsy specimens (15).

Previous studies have documented the use of cytological specimens, including pleural effusion (PLE), for EGFR mutation testing (16–19). However, there are limited studies examining the utility of bronchial lavage fluid (BLF), which is obtained following transbronchial lung biopsy and usually contains fewer cancer cells than PLE (20–22). In addition, data on the reported performance of companion diagnostics, including the therascreen EGFR RGQ PCR kit (the therascreen EGFR assay), primarily rely on formalin-fixed paraffin-embedded (FFPE) tissue blocks from resected or biopsy specimens and as such, there are limited data on fresh cytological specimens (23).

The purpose of the present study was to compare the efficiency of EGFR mutation detection between fresh cytological samples (cell pellets and cell-free supernatants) and FFPE cell blocks prepared from 1% EGFR mutation-positive lung cancer cell line mixtures using the therascreen EGFR assay. Furthermore, the utility of fresh BLF specimens from patients with NSCLC was also validated against matched FFPE tissue specimens in EGFR mutation detection using the therascreen EGFR assay.

Materials and methods

Lung cancer cell line samples

A total of three types of cytological samples were prepared: Fresh cell pellets, fresh supernatants, and FFPE cell blocks from three series of lung cancer cell line samples, including 1% EGFR mutant cells [sample (S) 1, 2, and 3]. The following human lung cancer cell lines were used: PC9 [EGFR E746_A750del (c.2235_2249del)] and A549 (EGFR wild-type). The PC9 cell line was obtained directly from the Riken BioResource Centre (Tsukuba, Japan). The A549 cell line was obtained directly from the Japanese Cancer Research Bank (Tokyo, Japan). PC9 and A549 were mixed at a ratio of 1:99, so that the percentage of EGFR-mutant cells in the mixture was 1%. The cell line mixture was divided into three samples, S1, S2, and S3, and each sample was washed with saline, centrifuged at 500 × g for 5 min at room temperature, then divided into cell pellet samples (S1p, S2p, and S3p) and cell-free supernatant samples (S1s, S2s, and S3s). From the cell pellets, FFPE cell block samples (S1b, S2b, and S3b) were produced through fixation with 20% neutral-buffered formalin for 24 h at room temperature using a sodium alginate cell block method (Fig. 1).

Figure 1. Lung cancer PC9 [EGFR E746_A750del (c.2235_2249del); red ovals] and A549 (EGFR wild-type; blue ovals) cell lines were mixed at a ratio of 1:99, meaning the percentage of EGFR mutant cells in the mixture was 1%. The 1% EGFR mutation-positive mixture was divided into samples S1, S2 and S3, which were washed with saline, centrifuged and divided into S1p, S2p, and S3p, and S1s, S2s, and S3s. From the cell pellets, S1b, S2b, and S3b were prepared. EGFR, epithelial growth factor receptor; S1, sample 1; S1p, S1 cell pellet; S1s, S1 cell-free supernatant; S1b, S1b formalin-fixed paraffin embedded cell block.

BLF and matched FFPE tissue specimens from patients with NSCLC

Fresh BLF specimens were collected from 219 patients who were clinically diagnosed with lung cancer between January 2014 and December 2014, and the specimens provided were used for cytological diagnosis at the Department of Laboratory Medicine at Shinshu University Hospital (Matsumoto, Japan). Following cytological evaluation of specimens from 219 patients, specimens that contained normal or benign cells (123 patients), malignant lymphoma (1 patient), renal cell metastatic carcinoma (1 patient), small cell carcinoma (8 patients), and squamous cell carcinoma (9 patients) were all excluded. The remaining 77 patients (age range, 41–85; median age, 69; 54 males and 23 females) whose specimens were cytologically diagnosed as ‘primary lung adenocarcinoma’, ‘NSCLC-not otherwise specified (NSCLC-NOS)’, ‘suspicious for malignancy’ (i.e., suspicious for adenocarcinoma or NSCLC-NOS), or ‘atypical cells’ meaning indefinite for neoplasia were enrolled in this study. These patients are the same as those used in a previous study by the same authors (22).

EGFR mutation status was analyzed using fresh cell pellets from BLF specimens. Among the 77 patients, 49 patients had EGFR mutation assay results from FFPE tissue specimens obtained by surgical resection or biopsy. The assay results from the BLF specimens were compared with those from the matching FFPE tissue specimens. The present study was reviewed and approved by the Medical Ethical Committee of the Shinshu University School of Medicine. All patients provided written informed consent for inclusion in the present study.

DNA extraction

DNA was extracted from fresh samples (cell pellets and cell-free supernatants) using the QIAamp DNA Blood Mini kit (Qiagen, Inc., Valencia, CA, USA) and the QIAamp DNA FFPE Tissue kit (Qiagen, Inc.) for FFPE specimens, according to the manufacturer's protocol. DNA concentration was quantified by spectrophotometry using a NanoDrop ND100 instrument (Thermo Fisher Scientific, Inc., Waltham, MA, USA). For comparison of the three types of cytological samples using cell lines, the therascreen EGFR assays were performed following adjustment of the DNA concentrations of cell pellets, cell-free supernatants, and FFPE cell blocks to 4.5 ng/µl. For the validation study using BLF specimens, the DNA concentrations were adjusted to <10 ng/µl.

EGFR mutation analysis

For EGFR mutation detection, the Rotor-Gene Q 5plex HRM instrument was used with the therascreen EGFR RGQ PCR kit (therascreen EGFR assay; Qiagen, Inc.). This assay is approved for use in the United States, Europe, Japan, and China, and the kit is based on the amplification-refractory mutation system (ARMS) and Scorpion PCR technology, which enable the sensitive and selective site-specific detection of 29 types of somatic mutations in EGFR (24). The reaction conditions were as follows: 95°C for 15 min for 1 cycle; 95°C for 30 sec and 60°C for 60 sec for 40 cycles. The analysis was performed using the Rotor-Gene Q series software, version 2.0.2 (Qiagen, Inc.). For comparison between the three types of cytological samples (cell pellets, cell-free supernatants and FFPE cell blocks) using cell lines, the cycle quantification (Cq) value of the mutant allele in each sample was compared using the therascreen EGFR ‘Deletions’ assay, which detects EGFR E746_A750del. For the validation study comparing BLF specimens to FFPE tissue specimens, the manufacturer-supplied cut-off delta Cq (ΔCq) values were used to determine the result whether positive or negative for the mutation in each EGFR mutation reaction.

Statistical analysis

For comparison of the cell line samples, the Quade test (Statcel version 4; OMS Publishing, Tokorozawa, Japan) was used to compare the differences in Cq values between the cell pellets, cell-free supernatants, and FFPE cell blocks. The Quade test indicates if there is a significant difference in the Cq values among the three sample types: Cell pellets, cell-free supernatants and FFPE cell blocks. The Cq values in three series (S1, S2 and S3) were compared. P<0.05 was considered to indicate a statistically significant difference.

Results

Comparison of Cq values between samples

The Cq values of the cytological samples in the three series, as determined by the therascreen EGFR assay, are presented in Table I. For all cell line samples excluding the FFPE cell block of S3, the EGFR mutation (E746_A750del) was detected by therascreen EGFR assay within 40 cycles. The average Cq values for mutation detection for the three sample types in the three series (S1, S2, and S3) were as follows: 29.58 for fresh cell pellets, 34.15 for fresh cell-free supernatants and 37.12 for FFPE cell blocks. The Quade test revealed that the Cq values in each series were significantly lower in cell pellets compared with that in the other sample types, followed by cell-free supernatants and FFPE cellblocks (P<0.05). The representative amplification curve and Cq value for each sample in the S2 series are presented in Fig. 2.

Figure 2. Amplification curves and Cq values of S2p, S2s and S2b, including the manufacturer's positive control and a negative control consisting of nuclease-free water. The lowest Cq value was observed for the cell pellet, followed by the cell-free supernatant and then the FFPE cellblock. Cq, threshold cycle; FFPE, formalin-fixed paraffin-embedded; S2p, sample 2 cell pellet; S2s, S2 cell-free supernatant; S2b, S2 FFPE cellblock.

Table I. Cq values of cell line samples in therascreen EGFR assay.

Table I.

Cq values of cell line samples in therascreen EGFR assay.

Cq value	Cell pellets	Cell-free supernatants	FFPE cell blocks	P-value
S1	24.93	34.43	35.17	<0.05
S2	31.52	33.80	37.15
S3	32.28	34.21	39.05
Average	29.58	34.15	37.12

[i] Cq, threshold cycle; S1, sample 1; FFPE, formalin-fixed paraffin-embedded.

Assay results of 77 fresh BLF specimens

EGFR mutation status was successfully analyzed using cell pellets from all 77 BLF specimens. The therascreen EGFR assay detected the exon 21 L858R point mutation in 14 patients (18.2%), exon 19 deletions in 10 patients (12.3%), and an exon 20 insertion in 1 patient (1.3%).

For the 49 patients who had EGFR mutation assay results for matching FFPE tissue specimens, all EGFR assay results were completely concordant between BLF cell pellets and FFPE tissue specimens. The assay results for the BLF cell pellets and matching FFPE tissue specimens in these 49 patients are listed in Table II.

Table II. EGFR assay results of BLF cell pellets and matching FFPE tissues.

Table II.

EGFR assay results of BLF cell pellets and matching FFPE tissues.

EGFR mutation status	BLF cell pellets, n (%)	Matching FFPE tissues, n (%)
Positive	17 (34.7%)	17 (34.7%)
Exon 19 deletions	5	5
Exon 21 L858R	11	11
Exon 20 insertions	1	1
Negative	32 (65.3%)	32 (65.3%)
Total	49	49

[i] EGFR, epithelial growth factor receptor; BLF, bronchial lavage fluid; FFPE, formalin-fixed paraffin-embedded.

The association between cytological diagnosis and EGFR mutation status of BLF specimens is demonstrated in Table III. EGFR mutations were detected in specimens that were cytologically diagnosed as ‘primary lung adenocarcinoma’ or ‘NSCLC-NOS’ from 7 patients, ‘suspicious for malignancy’ from 6 patients, and ‘atypical cells’ from 4 patients.

Table III. EGFR mutation status and cytological diagnosis of 49 BLF specimens.

Table III.

EGFR mutation status and cytological diagnosis of 49 BLF specimens.

Cytological diagnosis	EGFR-mutation positive	EGFR-mutation negative	Total
ADC or NSCLC-NOS	7	27	34
Suspicious for malignancy	6	3	9
Atypical cells	4	2	6
Total	17	32	49

[i] EGFR, epithelial growth factor receptor; BLF, bronchial lavage fluid, ADC, adenocarcinoma; NSCLC-NOS, non-small cell lung cancer, not otherwise specified.

Discussion

Although the use of FFPE cell block specimens should be prioritized for EGFR mutation testing, various cytological specimens, including liquid-based cytology, smear, fresh cell pellet and supernatant specimens are also suitable (14–16,18,25,26). However, the guidelines of the College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology, currently recommend the creation of FFPE cell blocks from cytological specimens in order to perform morphological and immunohistochemical examination for pathological diagnosis, and to stock them for additional molecular diagnostic studies (14). A novel EGFR mutation assay (the EGFR d-PCR assay) using fresh liquid cytological specimens (cell pellets) was previously validated, revealing that these specimens may contribute to the rapid and simple point-of-care test (22). In the present study, the detection efficiency for fresh cytological samples (not only cell pellets but also cell-free supernatants) was demonstrated, as was the relative EGFR mutation detection accuracy for fresh cell pellets from BLF specimens versus FFPE tissue specimens.

When comparing cell pellets, cell-free supernatants and FFPE cell blocks from cancer cell lines in the present study, cell pellets had significantly lower Cq values compared with FFPE cell blocks. This is because FFPE samples suffer from DNA fragmentation and the formation of cross-links during the fixation process (27); therefore, fresh cell pellets are considered to contain DNA that is more efficacious for mutation detection compared with FFPE cell blocks. In addition, DNA derived from slices of FFPE cell blocks usually contain DNA from only a small number of cancer cells, whereas fresh cytological specimens contain DNA from a number of cancer cells (28). Hence, FFPE cell block samples are assumed to exhibit a higher risk of false-negative outcomes compared with fresh cell pellets from cytological specimens (29,30).

In the present study, cell-free supernatant samples, which are considered to contain less DNA than fresh cell pellets, exhibited significantly better amplification efficiency compared with FFPE cell blocks. The result from the present study indicated that the fixation process has a strong negative impact on the detection efficiency of EGFR mutations and is indicative of the use of fresh cell-free supernatants of cytological specimens for EGFR mutation detection assays. Previous studies have reported the feasibility of detecting EGFR mutations with fresh cell-free supernatants of PLEs by direct sequencing (31), mutant-enriched PCR (19) and Scorpion ARMS (17). Zhang et al (19) revealed a good concordance between EGFR mutation assay results from cell pellets and those from matching cell-free fluids following mutant-enriched PCR. By using highly sensitive methods for EGFR mutation detection, performing EGFR assays with cell-free supernatants of BLF specimens is possible and may be useful, as the cell pellets, which enable morphological and multiple molecular analyses are preserved.

In the validation of the EGFR mutational status of BLF specimens using cell pellets from 77 patients with NSCLC, all EGFR mutational statuses were successfully analyzed and there were no specimens that failed to amplify. The assay results for all 49 BLF cell pellets with matching FFPE tissue specimens were completely concordant between the two specimen types. Goto et al (16) reported an excellent interassay concordance between 29 bronchofiberscopic brushing cytological specimens and matching FFPE specimens using Scorpion ARMS, PCR-Invader, PNA-LNA PCR clamp, and Cycleave PCR, with concordance rates of 93.1–96.6% (κ-coefficients, 0.86–0.93). Khode et al (32) compared 37 paired cytological smears with matched FFPE surgical tissue specimens from the same anatomical sites using pyrosequencing and RT-PCR platforms, and additionally reported a concordance rate of 97% between the two sample types. The present study also demonstrated the accuracy of the EGFR assay using fresh cytological specimens for EGFR mutation detection, with high concordance following the therascreen EGFR assay results of FFPE tissue specimens.

Among the 77 fresh BLF specimens, EGFR mutations were even detected in specimens containing only a few cancer cells, which were cytologically diagnosed as ‘atypical cells’ (i.e., indefinite for neoplasia). As was demonstrated by comparing different cytological samples from cell lines, use of the EGFR mutation assay with fresh cell pellets resulted in higher sensitivity of mutation detection compared with the use of FFPE cell blocks, even with a small amount of mutant DNA derived from a few cancer cells being detected by the assay. Furthermore, the use of BLF specimens may also markedly reduce the time and effort required for DNA extraction compared with the use of FFPE specimens, as BLF specimens do not require deparaffinization or reversal of formaldehyde-induced nucleic acid modification. Thus, the EGFR mutation assay with fresh BLF specimens may represent a screening method for diagnosing whether atypical cells are cancerous or not, particularly when morphological diagnosis is difficult owing to the presence of insufficient material.

Despite the fact that fresh BLF specimens contain DNA that is more efficacious for an EGFR mutation detection assay, they also usually contain more non-cancerous cells, including inflammatory cells and benign epithelial cells, compared with tissue specimens. The amount of DNA from non-cancerous cells affects the quality of the mutation-specific PCR assay, which may lead to false-negative results. Furthermore, for the current major companion diagnostics for EGFR-TKIs, including the therascreen EGFR assay, the cut-off values have been determined for DNA samples extracted from FFPE tissue or cell block specimens, as fresh cytological specimens are not generally used for testing. Therefore, the results of testing fresh cytological specimens should be interpreted carefully, and priority should be given to FFPE tissue specimens or cell blocks if they contain sufficient cancer cells.

To conclude, EGFR mutation detection assays using fresh BLF specimens offer a sensitive, accurate, simple and time-saving method for detection of EGFR mutations, even when sufficient cancer tissues or cytological specimens are not obtained. This method enables the full use of specimens from NSCLC patients for multiple molecular analyses.

Acknowledgements

The present study was partly supported by the Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research, commonly termed KAKENHI (grant no. 25460434).

Glossary

Abbreviations

Abbreviations:

References