Department of Surgical, Medical, Molecular Pathology and Critical Area, Unit of Pathological Anatomy, University of Pisa, Pisa, Italy
Corresponding author: Gabriella Fontanini, MD, Department of Surgical, Medical, Molecular Pathology and Critical Area, University of Pisa, via Roma 57, 56126 Pisa, Italy; Fax: (011) 39 050 992481; email@example.com
Anaplastic lymphoma kinase (ALK) gene rearrangements are detected in approximately 5% of cases of non-small cell lung cancer (NSCLC). Patients who are positive for ALK rearrangements may be successfully treated with the ALK inhibitor crizotinib. Because advanced-stage lung cancers are not suitable for surgical resection, approximately 70% of patients are diagnosed via preoperative specimens. In the current study, the authors evaluated the suitability of stained cytologic direct smears and cell blocks for fluorescence in situ hybridization (FISH) to determine ALK status compared with small biopsies.
A total of 493 consecutive patients with NSCLC were analyzed by FISH for ALK gene rearrangements. The analyzed sample comprised 180 cytological samples, 94 direct smears, 86 cell blocks, and 313 preoperative small biopsies. Moreover, in the same series, 426 patients and 369 patients, respectively, were evaluated for epidermal growth factor receptor and KRAS mutations, respectively.
Of the total of 493 patients, 18 patients who were positive for a gene rearrangement (4.4%) demonstrated ALK FISH rearrangements, whereas 387 patients (95.6%) were negative. All other cases were classified as inadequate (7.7%) and insufficient (10.1%). A strong statistical association was found between the cytology and the small biopsy with respect to ALK rearrangements (P = .0048). Fifty-three patients (12.4%) demonstrated an epidermal growth factor receptor mutation, whereas 90 patients (24.4%) were found to have KRAS mutations. None of these patients presented with ALK gene rearrangements.
Lung cancer is the leading cause of cancer deaths worldwide. Non-small cell lung cancer (NSCLC) accounts for 80% of these cases.[1-3] Approximately 70% of patients with NSCLC are diagnosed at a late stage of disease and are not candidates for surgical resection of their primary tumor. In these patients, the only pathologic material available for diagnostic and therapeutic purpose may be small biopsy or cytological specimens. Due to an increased understanding of the molecular mechanisms underlying the disease and treatment, a rapid increase in the accessibility and use of molecular testing has been observed. Furthermore, great progress has been achieved in “tailored therapies” based on molecular markers.[5-7] Some genetic alterations, such as mutations of the epidermal growth factor receptor (EGFR), are well-known carcinogenic mechanisms in NSCLC.[8, 9] In 2007, anaplastic lymphoma kinase (ALK) gene rearrangements with the echinoderm microtubule-associated protein-like 4 (EML4) gene has been identified as an oncogenic event in a subset of patients with NSCLC.[10, 11] This fusion gene, involving a small paracentric inversion of chromosome 2p, is generated by the fusion of 3′-N-terminal portions of EML4 and the 5′-ALK catalytic domain. ALK constitutive activation causes apoptosis inhibition and cellular proliferation via multiple pathways.[11, 12] ALK rearrangement is present in 2% to 7% of cases of NSCLC, and is associated with distinct clinicopathological features, including onset at a young age, an absent or minimal smoking history, and adenocarcinoma histology.[13-18] These rearrangements are mutually exclusive with EGFR and KRAS mutations, and preclinical and clinical studies have shown that cancer cells harboring ALK abnormalities are highly sensitive to ALK inhibitors. Crizotinib, a small-molecule inhibitor of ALK tyrosine kinase, has recently demonstrated a significant clinical benefit in patients with EML4-ALK positive NSCLC in a clinical trial.[13, 20]
Several methods are currently available to screen for ALK gene rearrangements. These techniques include fluorescence in situ hybridization (FISH), immunohistochemistry, and the reverse transcriptase-polymerase chain reaction (RT-PCR).
FISH is currently the elective method. This assay has only been approved by the US Food and Drug Administration (FDA) on histological samples. Moreover, FISH is routinely not available in all pathology departments, and to the best of our knowledge its usefulness for detecting ALK rearrangements in cytological specimens of NSCLC has not been assessed fully.
Specimen adequacy remains the cornerstone for a substantial subset of cases in which molecular testing is requested. It has been widely demonstrated that direct smears, cell blocks, and small biopsy samples provide adequate material for EGFR and KRAS mutation testing,[7-31] although to our knowledge not enough is known about their adequacy for ALK FISH analysis. Nevertheless, when cytological samples are the only available specimens, they should be considered as a potential source of adequate material for molecular testing.
The objective of the current study was to assess the adequacy of FISH testing of preoperative NSCLC samples, comparing 180 cytological samples with 313 bioptic specimens, and also evaluating the relation between the ALK status and EGFR and KRAS mutations.
MATERIALS AND METHODS
Patients and Tumor Specimens: Cytological Samples and Small Biopsy Specimens
A total of 493 consecutive FISH analyses for EML4-ALK rearrangement detection were conducted in the study department between January 2010 and September 2013. The specimens included 180 cytological samples, 94 direct smears, 86 cell blocks, and 313 preoperative small biopsy specimens. Samples were sent to the Unit of Pathological Anatomy in the Department of Surgical, Medical, Molecular Pathology, and Critical Area from the Unit of Thoracic Surgery, from the Endoscopic Section of the Pneumology Unit, and from external laboratories of pathological anatomy in the Tuscany area.
The cytological sample types were exfoliative (bronchial brushing [BB] and pleural effusion [PE]) and aspirative (transbronchial needle aspiration and transthoracic needle aspiration). The cytological preparations were conducted according to a standard specimen processing protocol in our laboratory. All smears were fixed in Cytofix (BD Biosciences, San Jose, Calif) and stained with Papanicolaou staining. Cell blocks were available for cases with a sufficient needle rinse, which yielded a visible pellet after centrifugation. Paraffin sections from the cell blocks were stained with hematoxylin and eosin.
The small bioptic samples were bronchial, transbronchial, and transthoracic, and all of them were formalin-fixed and paraffin-embedded (FFPE). Slides measuring 5 μm in thickness were obtained from each cell block and stained with hematoxylin and eosin.
Cytological and histological diagnoses were reviewed by 2 pathologists according to the criteria of the 2004 World Health Organization classification of lung tumors and the International Association for the Study of Lung Cancer/Amercian Thoracic Society/European Respiratory Society multidisciplinary classification of lung adenocarcinoma.
For all samples, mutational analyses were required by clinicians after informed consent was provided by the patients. The current study received Institutional Review Board approval.
Samples were selected for molecular analysis according to clinician requests for therapeutic purposes (Fig. 1). For cytological and bioptic samples, the first step was the pathology review of the slides to document tumor cellular availability. Consequently, we were not able to conduct a paired molecular investigation on both cytological and bioptic samples in all the cases.
For cytological specimens, 1 cellular slide was used for EGFR/KRAS evaluation and ALK FISH, respectively.
For bioptic samples and cell blocks, 2 sections of FFPE tissue measuring 2-μm to 4-μm thick were used for evaluation of ALK genetic status by FISH. From the same paraffin blocks, 3 sections of tissue measuring 10-μm in thickness were then obtained for the evaluation of EGFR and KRAS molecular status.
FISH was performed on Papanicolaou-stained smears paraffin section measuring 2-μm to 4-μm thick using a commercially available break-apart probe specific to the ALK locus (Vysis LSI ALK Dual Color, Break Apart Rearrangement Probe kit; Abbott Molecular, Abbott Park, Ill) according to the manufacturer's instructions. This probe, which was used to detect any rearrangement involving the ALK gene, hybridizes to band 2p23 on either side of the ALK gene breakpoint.
Both samples were examined to identify tumor cell-enriched areas. These areas were marked on the underside of the slides with a diamond-tipped scribe.
Pretreatment and digestion phases were different in the smears and FFPE samples.
The Papanicolaou-stained smears were decoverslipped in xylene at room temperature and then air-dried. Air-dried smears were pretreated with the Vysis FISH Pretreatment Reagent Kit (Abbott Molecular) following the manufacturer's protocol. Specifically, the smears were immersed in 2× saline sodium citrate (SSC) solution for 2 minutes at 73°C ± 1°C and then in protease K solution for 8 to 10 minutes at 37°C. The length of this passage depends on the type of material analyzed. The smears were washed in 1× phosphate-buffered saline for 5 minutes at room temperature, fixed in 1% formaldehyde for 5 minutes at room temperature, and then washed again in 1× phosphate-buffered saline at room temperature. The smears were dehydrated by immersing them sequentially in 50% ethanol, 70% ethanol, 95% ethanol, and 99% ethanol for 2 minutes each.
Paraffin tissue sections measuring 2-μm to 4-μm thick were obtained from cell blocks and biopsy samples. Before hybridization, paraffin sections were deparaffinized in xylene 3 times for 10 minutes each time, dehydrated by two 5-minute washes in 100% ethanol and two 5-minute washes in 96% ethanol, and air-dried at room temperature. The tissue sections were then transferred to an 80°C pretreatment solution for 10 minutes, followed by 3-minute washes in purified water and treatment with a protease K solution for 10 minutes at 37°C to digest proteins. After briefly being washed in purified water, the slides were sequentially dehydrated in alcohol (70%, 85%, and 100%) and then air-dried at room temperature.
The hybridization and washing slides procedures were the same for both types of samples and were performed as follows. Hybridization of the Vysis LSI ALK Break Apart Rearrangement Probe (Abbott Molecular) was performed by applying 10 μL of the probe mixture to each slide. Next, the slides were covered with a coverslip, sealed with rubber cement, and finally placed in the ThermoBrite system (Abbott Molecular) for the hybridization program.
Tissue sections were denatured at 73°C for 5 minutes with Hybrite (Abbott Molecular) and probe hybridization was conducted overnight at 37°C.
The following day, at the end of the hybridization period, the slide-washing procedure was performed. The rubber cement and coverslip were removed, and the slides were immersed in 2× SSC/0.3% NP-40 for 3 minutes at room temperature, and then in 2× SSC/0.3% NP-40 for 5 minutes at 73°C ± 1°C and then again in 2× SSC/0.3% NP-40 for 3 minutes at room temperature. Finally, the slides were counterstained by applying 10 μL of DAPI (4″,6-diamidino-2-phenylindole) I counterstain and a coverslip to the target area of the slide.
Slide examinations were performed on the Olympus BX61 fluorescent microscope (Olympus Corporation of the Americas, Center Valley, Pa). The tumor samples were scored by 3 independent investigators in a blinded fashion.
According to the score proposed,[21, 26] the test was considered positive if ≥ 15% of the scored tumor cells had splitting of the 5′ (green) and 3′ (red) probe signals or had isolated 3′ signals, whereas overlapping of red and green signals (yellowish) indicated cells in which ALK was not rearranged (Fig. 2).
Moreover, certain cases were diagnosed as inadequate when technical problems, such as insufficient fixation or crush artifacts, did not allow a clear view of the red and green signals. Other cases were diagnosed as insufficient when, after FISH preparation, < 50 neoplastic cells were countable according to standard criteria.
Analysis of EGFR and KRAS Mutations
Genomic DNA was isolated from cytological specimens and histological tissues by a standard method. From a 10-μm thick section, paraffin was removed by xylene extraction and the sample was subsequently lysed by proteinase K. DNA extraction was then performed using the spin column procedure (QIAamp Tissue kit; Qiagen, Germantown, Md).
Mutational profiling of EGFR by direct dideoxy DNA sequencing
Mutational profiling of EGFR (exons 18-21) was performed as previously reported. Briefly, the eluted DNA was used as template in a standard 20- μL PCR reaction mixture. The PCR product sizes for EGFR exons 18, 19, 20, and 21 were 207 base pair (bp), 194 bp, 247 bp, and 235 bp, respectively. Because both primers had similar melting temperatures, the same PCR conditions could be used to simultaneously amplify the 4 exons (in separate reaction tubes). The conditions for the EGFR exons, after initial denaturation at 94°C (7 minutes), were 35 cycles of denaturation at 94°C for 60 seconds, annealing at 58°C for 60 seconds, and synthesis at 72°C for 60 seconds, followed by a final extension for 7 minutes. As a negative control, the DNA template was omitted in the reaction. The amplification products were separated on 2% agarose gels and visualized by staining with ethidium bromide. For the detection of mutations, PCR products were purified with the QIAquick PCR Purification kit (Qiagen), and sequenced using a cyclic sequencing kit (ALFexpress II; Amersham Biosciences, Piscataway, NJ), following the manufacturer's recommendations.
Mutational profiling of KRAS by pyrosequencing assay
Pyrosequencing assays were performed for sequencing analysis of KRAS (codons 12 and 13). Briefly, template DNA (6 ng of genomic DNA or external PCR products) was amplified using the HotStarTaq Plus DNA Polymerase KIT (Qiagen) and the standard protocol (0.2 mM of each primer, 160 mM of dNTPs, 2 units of enzyme) and cycling conditions. Reverse PCR primers were biotinylated for subsequent pyrosequencing analysis. Pyrosequencing reactions were conducted on the PSQ HS96A instrument using PSQ HS96A single-nucleotide polymorphism reagents and pyrosequencing single-nucleotide polymorphism analysis software (Biotage AB, Uppsala, Sweden, now named PyroMarkTMQ96MD by Qiagen). Pyrosequencing raw data signals were normalized using known wild-type signals.
All statistical analyses were conducted using JMP statistical software (version 8.0.2; SAS Institute Inc, Cary, NC). A chi-square test was used to analyze the associations between the different variables. The a priori level of significance was set at a P value < .05.
A total of 493 patients were studied, 315 of whom (63.9%) were male and 178 of whom (36.1%) female. The mean age at diagnosis was 66.6 years and the median age was 67 years. Of these 493 patients, 180 patients (36.5%) had a cytological sample; 94 (52.2%) had direct smears and 86 (47.8%) had cell blocks. Specifically, 53 patients (29.4%) had exfoliative cytological specimens, including 20 BBs (11.1%) and 33 PEs (18.3%). All BBs were direct smears, whereas all PEs were processed as cell blocks. Among the aspirative cytological samples, 101 transthoracic needle aspirations (56.1%) and 26 transbronchial needle aspirations (14.4%) were included. Fifty-three aspirative samples (41.7% [53 of 127]) were cell blocks, whereas 74 (46.9%) were direct smears.
Moreover, we evaluated 313 preoperative histological specimens (63.5%). All these specimens were small bioptic samples, specifically 37 pleural biopsies (11.8%), 76 transthoracic biopsies (24.3%), and 132 bronchial biopsies (42.2%). A total of 68 (21.7%) were biopsies from metastatic sites of primary lung cancer.
All analyzed cases were NSCLCs. FISH analysis was conducted on 370 adenocarcinomas (75.1%); 58 squamous cell carcinomas (11.8%); 51 NSCLC, not otherwise specified (10.3%); 1 adenosquamous carcinoma (0.2%); and 13 neoplasms categorized as “other” (2.6%). Three hundred eighty-one (77.3%) were specimens from primitive lung cancers, and 112 (22.7%) were samples from metastatic sites.
ALK FISH Analysis
Eighteen cases (4.4%) were positive for ALK FISH, and 387 (95.6%) were classified as inadequate (7.7%) or insufficient (10.1%).
The negative cases from the cytological samples were 68 direct smears (37.8%) and 59 cell blocks (32.8%), whereas the positive cases were 6 direct smears (3.3%) and 1 cell block (0.6%). We also found 10 direct smears (5.6%) and 7 cell blocks (3.9%) that were inadequate and 10 direct smears (5.6%) and 19 cell blocks (10.6%) that were insufficient.
By contrast, among the bioptic specimens, we found 11 positive cases (2.2%), 260 negative cases (52.7%), and 21 cases that were both inadequate and insufficient (4.3%).
Comparing the adequacy of FISH analysis between the cytological and bioptic samples, we found a statistically significant difference (P = .0048), indicating that the FISH analysis was more adequate when conducted on cytology (Table 1). Moreover, in the comparison between direct smears and cell blocks, we observed a statistical difference (P value of the failure rate difference of insufficient samples: P = .04; P value of the failure rate difference of inadequate samples: P = .6) that, although not as strong as that mentioned above, indicated a better adequacy of FISH for smear cytology than for cell blocks.
Table 1. Correlation Between the ALK FISH Analysis Results and the Types of Samples Analyzed
Ultimately, we did not find any discrepancies between different types of direct smears (exfoliative vs aspirative: P value of the failure rate difference, > .05; power, < 0.8).
ALK FISH Analysis and Clinicopathological Parameters
Patients found to have ALK gene rearrangements on FISH were mainly male (13 patients; 3.2%) and younger than the median age of 67 years (12 of 403 patients; 3%). The largest percentage of positive ALK FISH cases were adenocarcinoma (15 of 18 cases; 83.3%), but we also encountered gene rearrangements in 1 adenosquamous tumor; 1 NSCLC, not otherwise specified; and in 1 neoplasm classified as “other,” which was a poorly differentiated NSCLC with neuroendocrine characteristics (Table 2).
Table 2. Correlation Between the ALK FISH Status and the Clinicopathological Parameters
ALK FISH Positive
ALK FISH Negative
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; FISH, fluorescence in situ hybridization; NS, not significant; NSCLC, NOS: non-small cell lung cancer, not otherwise specified; SCC, squamous cell carcinoma.
On this basis, we could support the statistical correlation of ALK rearrangements with the adenocarcinoma histotype (P = .01). By contrast, we did not found any statistical correlation between the presence of ALK gene rearrangements and sex or age (P = .51 and P = .19, respectively).
ALK FISH Analysis and EGFR and KRAS Status
A total of 493 patients were evaluated for ALK FISH rearrangements. The EGFR mutational status was evaluated contemporarily with ALK rearrangements in 426 patients (86.4%), whereas KRAS mutations were analyzed in 369 patients (74.8%) (Table 3). Fifty-three patients (12.4%) were EGFR mutated, specifically, 28 mutations (52.8%) on exon 19, 19 mutations (35.8%) on exon 21, 2 mutations (3.8%) on exon 20, 2 mutations (3.8%) on exons 19 and 20, 1 mutation (1.9%) on exon 18, and 1 mutation (1.9%) on exons 20 and 21.
Table 3. Correlation Between the EGFR and KRAS Status and the Clinicopathological Parameters
Ninety patients (24.4%) had KRAS mutations, which were distributed as follows: 83 mutations (92.2%) on codon 12, 5 mutations (5.6%) on codon 13, and 2 mutations (2.2%) on codon 61.
None of the EGFR-mutated or KRAS-mutated cases were positive for ALK rearrangements.
In addition, 50 cases (11.7%) and 40 cases (10.8%), respectively, were not evaluable for EGFR and KRAS mutations because their samples were not sufficiently cellular or were inadequate for the molecular analysis (Table 4).
Table 4. Correlation Between EGFR and KRAS Analysis Results and the Types of Samples Analyzed
The therapeutic approach to lung cancer has dramatically changed in recent years due to the widespread use of molecular testing optimized for identifying molecular targets for biological drugs.
Targeted therapies are mostly applied in patients with advanced-stage NSCLC. Consequently, there is an increasing clinical need to investigate molecular aberrations in small biopsy specimens and cytological specimens of primary and metastatic disease.
Currently, FISH is considered the gold standard for screening NSCLC specimens for ALK rearrangements and crizotinib therapy. Actually, according to FDA guidelines, this analysis is applicable only on histological samples. Nevertheless, cytological samples are often the only available specimens and they should be considered as potential sources of adequate material for molecular testing.
To our knowledge, only 2 to date studies have been conducted to assess the adequacy of cytological specimens for ALK FISH analysis.[30, 31] Betz et al analyzed a small cohort of patients for ALK gene rearrangements using both direct smears and cell blocks. The ALK status was evaluated by FISH and RT-PCR. Neat et al instead focused on the adequacy of 52 endobronchial ultrasound-guided transbronchial needle aspiration cytology samples for identifying ALK gene rearrangements by FISH and immunohistochemical analysis. Both research groups, although using a small cohort of patients, concluded that cytological smears are feasible and effective platforms for the molecular diagnostic workflow of NSCLC, particularly for the ALK FISH assessment.
By contrast, the purpose of the current study was to examine the availability of FISH analysis for detecting ALK gene rearrangements in cytological and bioptic samples of a large series of NSCLCs, comparing the feasibility and adequacy and focusing on the specimens' adequacy.
On this basis, we compared the ALK FISH analysis in cytological samples in both direct smears and cell blocks and in histological specimens. We evaluated a cohort of 493 NSCLC specimens taken directly from our clinical ALK FISH workflow.
First, we evaluated the adequacy of cytological samples with respect to small biopsies for FISH analysis. Both direct smears and cell blocks were demonstrated to be advantageous (P = .0048); a considerable number of FISH analyses (4.3%) of bioptic specimens proved inadequate. Taking these results into consideration, we made a comparison between cytological specimens. It is interesting to note that we found direct smears to be more effective for ALK FISH analysis than cell blocks. In particular, we observed that the difference between direct smears and cell blocks is statistically significant in the subgroups of insufficient samples (P = .04). Conversely, the difference between direct smears and CBs in the inadequate subgroup was not found to be statistically significant. Nevertheless, the inadequate and insufficient cases among cell block samples were significantly more numerous compared with direct smears.
These are interesting findings that must be taken into strict consideration because, as mentioned above, histological samples are the only assays for ALK FISH detection that are approved both by the FDA and by the Italian Association of Medical Oncology and the Italian Society of Anatomic Pathology and Cytopathology. In fact, bioptic samples and cell blocks might yield the same technical artifacts, such as fixation defects or crush artifacts, that enable the signal evaluation. In addition, the isolation of neoplastic cells might be more difficult on cell block samples, in which a cellular and hematic background is frequently present.
Direct smears are advantageous for FISH assays because they do not experience problems resulting from formalin fixation and they allow for the evaluation of the whole nuclei of the tumor cells, thereby avoiding the nuclear truncation and probe signal loss that occurs in FISH performed on FFPE specimens.
In the current study, we also compared different types of direct smears: exfoliative and aspirative. According to these types of smears, we did not find any statistical differences between these methods. Although in this series the number of the exfoliative and aspirative groups was not large, the statistical power of this test was sufficient to speculate that the 2 methods are equally satisfactory for ALK FISH analysis. Evaluating these data in a larger cohort of samples could allow us to draw more definitive conclusions.
Finally, considering the clinicopathological parameters, we observed a strong association with the tumor histotype (P = .001). The rearranged tumors were mainly adenocarcinomas, confirming what was previously reported by Soda et al and Rodig et al.[10, 15] Squamous histotypes, instead, did not exhibit ALK rearrangements, as previously observed. The rational for ALK FISH analysis on squamous histotypes is that, according to the International Association for the Study of Lung Cancer/Amercian Thoracic Society/European Respiratory Society multidisciplinary group, lung cancer with mixed histology (eg, adenosquamous, combined adenosquamous/small cell) cannot be diagnosed on bioptic and cytological samples. These histotypes can have ALK gene rearrangements and demonstrate treatment response. Therefore, in specimens with incomplete sampling, such as biopsies and cytology, in which the possibility of an adenocarcinoma component cannot be excluded, ALK testing may be indicated and clinical criteria (lack of smoking history, young age) may be used to select patients. It is interesting to note that we observed the presence of an ALK gene rearrangement in an adenosquamous carcinoma, as previously described by Kwak et al and Paik et al.[21, 38] Moreover, an ALK gene rearrangement was observed in a bronchial biopsy with poorly differentiated NSCLC and neuroendocrine differentiation, which was not otherwise classified because of the small sample size. By contrast, we did not observe significant differences based on patient sex and age, although we observed that patients with ALK gene rearrangements tended to be younger than patients without ALK gene rearrangements.
Finally, a subset of patients underwent molecular testing for EGFR and KRAS mutations. EGFR mutations, of all exons, were found in 12.4% of patients, whereas KRAS alterations were encountered in 24.4% of cases analyzed. None of the patients with ALK gene rearrangements was found to have EGFR or KRAS mutations. These data support previous observations that ALK rearrangements are mutually exclusive with mutations in EGFR and KRAS,[10, 19, 39] although other studies have indicated that ALK, KRAS, and EGFR molecular defects can be harbored simultaneously.[13, 31, 40]
We conclude that the evaluation of ALK rearrangements by FISH is possible in cytological samples. Direct smears provide feasible and effective material for the molecular analysis of NSCLC, which may outperform the technical problems encountered with both cell blocks and small bioptic samples. Although further studies are needed to draw stronger conclusions, we propose this approach as a complementary platform to the molecular analysis of NSCLCs that provides a safeguard by eliminating the sole reliance on cell block preparations for molecular testing.