Screening for ALK abnormalities in central nervous system metastases of non‐small‐cell lung cancer

Abstract

Anaplastic lymphoma kinase (ALK) gene rearrangement was reported in 3%–7% of primary non‐small‐cell lung cancer (NSCLC) and its presence is commonly associated with adenocarcinoma (AD) type and non‐smoking history. ALK tyrosine kinase inhibitors (TKIs) such as crizotinib, alectinib and ceritinib showed efficiency in patients with primary NSCLC harboring ALK gene rearrangement. Moreover, response to ALK TKIs was observed in central nervous system (CNS) metastatic lesions of NSCLC. However, there are no reports concerning the frequency of ALK rearrangement in CNS metastases. We assessed the frequency of ALK abnormalities in 145 formalin fixed paraffin embedded (FFPE) tissue samples from CNS metastases of NSCLC using immunohistochemical (IHC) automated staining (BenchMark GX, Ventana, USA) and fluorescence in situ hybridization (FISH) technique (Abbot Molecular, USA). The studied group was heterogeneous in terms of histopathology and smoking status. ALK abnormalities were detected in 4.8% (7/145) of CNS metastases. ALK abnormalities were observed in six AD (7.5%; 6/80) and in single patients with adenosuqamous lung carcinoma. Analysis of clinical and demographic factors indicated that expression of abnormal ALK was significantly more frequently observed (P = 0.0002; χ ² = 16.783) in former‐smokers. Comparison of IHC and FISH results showed some discrepancies, which were caused by unspecific staining of macrophages and glial/nerve cells, which constitute the background of CNS tissues. Their results indicate high frequency of ALK gene rearrangement in CNS metastatic sites of NSCLC that are in line with prior studies concerning evaluation of the presence of ALK abnormalities in such patients. However, they showed that assessment of ALK by IHC and FISH methods in CNS tissues require additional standardizations.

Introduction

Central nervous system (CNS) metastases are the most frequent progression sites of non‐small‐cell lung cancer (NSCLC) 3, 26, 36. They are associated with high rate of morbidity and mortality 3, 26. Because of blood–brain barrier (BBB), which ensures restrict transit of agents into the brain parenchyma, they are considered as pharmacological sanctuary lesions that show limited sensitivity to anti‐cancer therapy 3, 26. Advanced disease stage, large primary tumor size or non‐squamous histology may increase the risk of CNS metastases 5, 26, 36. Moreover, increased number of CNS metastases is considered as a consequence of prolonged survival during treatment with tyrosine kinase inhibitors (TKI) of epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinase (ALK) 3, 5, 8, 10, 26.

The primary ALK gene aberration (development of EML4‐ALK fusion gene) was described in 3%–7% of primary NSCLC patients and it was predominantly associated with adenocarcinoma (AD) type, early age of diagnosis (median 52 years) and never/light smoking history (< 10 pack‐years) 17, 34, 35, 40. ALK TKIs, such as crizotinib, alectinib and ceritinib, showed antitumor activity in NSCLC patients harboring ALK rearrangement 8, 12, 16, 30, 31, 36. Despite treatment improvement in primary tumors, CNS metastases remain a significant challenge 10, 30, 33, 43. However, alectinib and ceritinib showed a potent activity also in CNS lesions what may extend survival of patients with advanced stages of NSCLC 8, 12, 31.

Till date there is limited data about driver mutations incidence in CNS metastases of NSCLC, which could be considered as a regiment for targeted treatment. For this reason we undertook the present retrospective study to determine the frequency of ALK abnormalities in CNS metastases of NSCLC. Additionally, this is the first report concerning IHC and FISH analysis performed in CNS specimens worldwide, which may provide some practical information about ALK assessment in CNS.

Materials and Methods

Patients

ALK abnormalities were evaluated on 3‐μm‐thick formalin‐fixed paraffin‐embedded (FFPE) tissue samples obtained from 145 CNS metastatic tumors of NSCLC. Patients qualified to the studied group had diagnosis of advanced NSCLC with spread lesions to the brain between 2004 and 2011. Only in single patients concurrent thoracic surgery was performed therefore the histopathology was reported based on histopathological diagnosis of CNS metastases. Moreover, the material from corresponding primary tumors was limited and poorly available (e.g., materials from fine‐needle aspiration biopsies). For this reason samples from matched primary tumors were not evaluated for ALK abnormalities. Patient demographic and clinical characteristics are summarized in Table 1[TQ1]. All patients were radiotherapy and chemotherapy naive and underwent routine neurosurgical procedures with a palliative manner. The median overall survival (OS) was 13.5 months (range 0.1–78.2 months—information available from 119 patients). In the following study all patients had diagnosed advanced stages of NSCLC with dissemination to CNS. Therefore, both diagnosis of primary and metastatic lesions were considered as a beginning of the disease from which OS was calculated to death. Expression of ALK abnormal protein was evaluated using automated immunohistochemistry (IHC) staining on BenchMark GX (Ventana, USA) device as a screening method. To confirm the IHC positive results, fluorescence in situ hybridization (FISH) technique was performed to visualize the presence of ALK rearrangement. The study was approved by the Ethics Committee of the Medical University of Lublin, Poland (No. KE‐0254/86/2013).

Table 1.

Characteristics of the studied group.

Gender
Male, n (%)	100 (69)
Female, n (%)	45 (31)
Age
Median age ± SD (years)	60 ± 8.8
≥60 years, n (%)	72 (49.7)
<60 years, n (%)	73 (50.3)
Histopathology
Adenocarcinoma, n (%)	82 (56.6)
Squamous‐cell carcinoma, n (%) (include 1 ADSC case)	29 (20)
Large‐cell carcinoma, n (%)	22 (15.1)
NSCLC‐NOS, n (%)	12 (8.3)
Smoking status
Current smokers, n (%)	73 (50.4)
Former smokers, n (%)	21 (14.5)
Non‐smokers, n (%)	36 (24.8)
Lack of data, n (%)	15 (10.3)
Performance status
0, n (%)	22 (15.2)
1, n (%)	76 (52.4)
2, n (%)	31 (21.4)
3, n (%)	16 (11)

IHC staining of ALK overexpression

For IHC evaluation of ALK abnormal protein expression on tumor cells, three FFPE slides: one for H&E (to determine the localization and content of tumor cells) and two for IHC staining [reactions with Positive Rabbit Monoclonal Antibody D5F3 (14 µg/mL) and Rabbit Monoclonal Negative Control (10 µg/mL)] were used. Automated staining was performed on BenchMark GX system using CE‐IVD ALK‐Ventana reagents with OptiView DAB Detection Kit (Ventana, USA). Reagents kit contain: peroxidase inhibitor (3%); HQ Universal Linker (<50 µg/mL); HRP Multimer (<40 µg/mL); OV DAB (0.2%); OV H₂0₂ (0.04%); OV Copper (5.0 g/L); OptiView Amplification Kit (AMP Amplifier (<40 µg/mL); AMP H₂0₂ (0.04%); AMP Multimer (<10 µg/mL); caunterstainer reagent (Hematoxilin ≤48%; Ventana, USA) and post counterstainer reagent (Bluing Reagent: 0.1M Li₂CO₃; 0.5M Na₂CO_3; Ventana, USA). The staining was performed in accordance to manufacturer procedure (BenchMark GX programmed for VENTANA ALK (D5F3) CDx) in following steps: deparaffinization and epitope retrieval in cell conditioner 1 (pH = 8)—90 min, incubation with: UV inhibitor (peroxidase inhibitor)—4 min; anti‐ALK antibody incubation (positive or negative reaction)—36 C—120 min; Amplifier A—8 min; Amplifier B—8 min; UV HRP Universal Multimer—8 min; DAB—8 min; UV DAB H₂O₂—8 min; OV Copper—4 min, counterstaining with Hematoxylin—4 min and post counterstaining in Bluing Reagent—4 min.

All stained slides were evaluated by a pathologist, who assessed presence or lack of DAB signal in cancer cells. In case of positive reaction, a strong granular cytoplasmic or cytoplasmic/membranous (Figure 1) signal was shown in the form of a homogeneous or heterogeneous staining of tumor tissue in at least 10% of analyzed cancer cells. Positive staining of alveolar macrophages, glial/nerve cells or lymphocytes was considered as negative. Moreover, staining of mucus background of neoplastic infiltration and areas of tumor necrosis were considered as artifacts.. Slides without a strong granular cytoplasmic staining of tumor cells were considered as negative. The analysis was performed according to IASCL and Ventana (Roche, USA) guidelines 37, 38. Slides without a strong granular cytoplasmic staining of tumor cells were considered as negative. In case of inconclusive staining with a numerous macrophages, the samples were stained immunohistochemically with antibody against CD68‐marker of macrophages (dilution 1:50) using EnVision^TM Detection System (DakoReal^TMEnVision^TM Detection System, Peroxidase/DAB+, Rabbit/Mouse, Dako, Denmark). All unspecific and discussable slides were re‐reviewed by second pathologist and based on double revision results were placed.

Figure 1.

Example of positive ALK protein staining (600×). (a) Represents ALK‐positive cancer cells with granular cytoplasmic reaction. (b) Represents ALK‐positive cancer cells with membranous reaction.

FISH assessment of ALK gene rearrangement

The Vysis ALK Break Apart FISH Probe Kit (Abbot Molecular, USA) was used to detect ALK gene rearrangement by FISH technique using fluorescence microscope (Nikon Eclipse 55i, Japan). Additionally, Paraffin‐Pretreatment IV and Post‐Hybridization Wash Buffer Kit (Abbot Molecular, USA) was also used for the pre‐staining procedure. The localization and content of tumor cells in the specimens were examined with H&E staining in serially prepared slides. The cells were considered as positive (with ALK gene rearrangement) when adjacent orange and green signals were more than two signals' diameters apart and/or one fused signal coexists with one orange signal in the analyzed cells.

Using a fluorescence microscope the fluorescent signals were counted in 50 tumor nuclei by first reader. Samples were considered as positive if >25/50 cells (>50%) showed single reds or separated signals. The samples were considered as equivocal if 5 to 25/50 cells (10‐50%) were positive. All samples with <5/50 positive cells (<10%) were considered as negative. Moreover, the cells were considered as negative (without ALK rearrangement) when two or one fusion signals were observed with one green signal without the corresponding orange signal. All equivocal samples were re‐evaluated by a second reader who counted an additional 50 cells for a total of 100 cells and the specimen was then considered as positive if ≥15% (15/100) of cells showed single reds or separated signals. An increased copy number of fused, non‐rearranged ALK signals corresponds to ALK polysomy (≥4 ALK copies in ≥10% nuclei) or ALK amplification (≥10 ALK copies in ≥10% nuclei). Interpretation of IHC and FISH results was performed according to recommendations presented by the American Food and Drug Administration (FDA) 19.

Statistical analysis

Statistical analysis was performed using Statistica version 9.0 (Statsoft, USA) and MedCalc10 (MedCalc Software, Belgium). Associations between the occurrence of ALK abnormalities and patient clinical factors were examined using the Chi square test. The Kaplan–Meier method was used to compare the probability of OS in patients with different ALK status. P values <0.05 were considered as statistically significant.

Results

IHC automated staining was performed in all 145 samples from CNS metastases of NSCLC and the ALK abnormal protein was observed in ten patients (6.9%). Then, all tumors with expression of ALK abnormal protein were evaluated using FISH method to confirm the presence of ALK gene rearrangement. Results of IHC and FISH analysis showed some discrepancies. In three samples ALK gene rearrangement was confirmed by FISH technique, while in three samples, FISH results were equivocal. In three sample, ALK gene rearrangement was not detected by FISH analysis. Moreover, one FISH analysis was determined as non‐diagnostic. Then, we re‐evaluated all 10 selected samples with IHC method to resolve any inconsistency. Three samples (one sample with equivocal FISH result, two samples with negative FISH result) showed high percentage of macrophages and necrotic background. We decided that this problem could induce false positive results of ALK abnormal protein expression in first IHC examination. In remaining samples, expression of abnormal ALK protein on tumor cells was confirmed by second IHC examination. Based on these results we finally found ALK abnormal protein expression in seven (4.8%; 7/145) samples obtained from CNS metastases of NSCLC. However, three samples that were falsely pre‐selected by IHC method as ALK‐positive tumors showed FISH result with ALK gene high polysomy, and with 8%–12% of ALK rearranged nuclei. On the other hand, one sample with FISH negative result and one FISH non‐diagnostic sample were determined as positive in IHC method. The examples of positive and negative IHC staining of ALK abnormal protein are presented in Figures 2 and 3. Figure 4 presents the example of false positive staining. Summary of IHC and FISH analysis is presented in Table 2.

Figure 2.

Example of ALK gene rearrangement analysis performed in the same part of the tissue from CNS metastases of NSCLC. Figure 2a shows H + E staining, 2b shows negative IHC staining with Rabbit Monoclonal Negative Control and 2c shows positive IHC staining with Positive Rabbit Monoclonal Antibody D5F3 against ALK abnormal protein. Figure 2d shows FISH result obtained from the same patient and confirmed presence of ALK rearranged cells. The white arrows marked single red and split signals.

Figure 3.

Example of negative analysis of abnormal ALK protein expression performed in the same part of the tissue from CNS metastases of NSCLC. Figure 3a shows H + E staining, 3b shows negative IHC staining with Rabbit Monoclonal Negative Control and 3c shows lack of IHC staining with Positive Rabbit Monoclonal Antibody D5F3 against ALK abnormal protein.

Figure 4.

Example of false positive ALK rearrangement analysis performed in the same part of the tissue from CNS metastases of NSCLC. Figure 4a shows H + E staining, 4b shows negative IHC staining with Rabbit Monoclonal Negative Control and 4c shows positive IHC staining with Positive Rabbit Monoclonal Antibody D5F3 in glial and nerve cells. 4d shows FISH result obtained from the same patient. The white arrows marked red and green fused signals, which are observed in case of absence of ALK rearranged nuclei.

Table 2.

Comparison of IHC and FISH results.

No	Histopathology	1st IHC	FISH			Final result after 2nd IHC examination
		% of cells with abnormal protein ALK expression	% of nuclei with ALK rearrangement	Types of positive signals (% out of positive nuclei)	% of nuclei with ≥4 gene copy number
1	SCC	50	8	87.5—SR; 12.5—SS	70	Negative
2	NSCLC‐NOS	10	12	83.3—SR; 16.7—SS	72	Negative
3	AD	75	8	100—SR	94	negative
4	Male‐differentiated AD	10	0	0	24	positive
5	AD	100	25	100—SS	26	positive
6	Male‐differentiated AD	99	Non‐diagnostic			positive
7	AD	75	55	100—SR	32	positive
8	AD	50	13	90—SR; 10—SS	74	positive
9	ADSC	25	82	100—SR	86	positive
10	AD	25	17	64.3—SR; 35.7—SS	54	positive

Characteristic of ALK rearranged patients

ALK abnormal protein expression was observed in 8.9% of female (4/45) and in 4% of male (4/100) patients. Six patients with expression of abnormal ALK had AD diagnosis (7.3% of all AD; 6/82), including two patients with male‐differentiated adenocarcinoma and one patients had adenosquamous lung carcinoma (ADSC) diagnosis. It is more likely that in this patient ALK rearrangement in metastatic site was dependent on AD component. However, we could not perform laser‐capture microdissection to confirm this hypothesis. The median age of patients with ALK abnormal protein expression was 55 years. Expression of abnormal ALK was significantly more frequently observed (P = 0.0002; χ ² = 16.783) in former‐smokers (23.8%; 5/21) than in current (1.4%; 1/73) and non‐smokers (2.7%; 1/36). The average pack‐years of smokers with abnormal ALK expression was 30 pack‐years. Analysis of other clinical and demographic factors indicated no significant differences between patients with and without abnormal ALK expression. In spite of differences between median overall survival (mOS) in patients with ALK abnormal expression (34 months) and in patients without this abnormality (15 months), survival analysis did not show statistical significance (P = 0.2609, χ ² = 1.264; HR = 0.6514; 95% CI: 0.33–1.2859; Figure 5). The ALK gene rearrangement was absent in tumors with present mutations in KRAS; NRAS, EGFR; DDR2; PIK3CA; HER2; MEK1; PTEN and AKT1 genes. These mutations were evaluated in our previous studies. The detailed clinical summary of patients with ALK abnormal protein expression and with ALK rearrangement is presented in Table 3.

Figure 5.

Kaplan–Meier survival curves in patients with (mOS = 34 months) or without ALK abnormalities (mOS = 15.5 months; P < 0.2609; χ ² = 1.264; HR = 0.6514; 95% CI: 0.33–1.2859).

Table 3.

Characteristic of patients with ALK abnormalities (patients 1‐3 were false positive in first IHC staining).

ID	Gender	Age (years)	OS (months)	Smoking status (pack‐years)	Histology of NSCLC
1	M	64	2.5	Former smoker 10	SCC
2	F	74	12	Non‐smoker	NSCLC ‐ NOS
3	M	60	52.7	Former smoker 20	AD
4	F	54	5.5	Current smoker 15	Male‐differentiated AD
5	F	53	78.2	Non‐smoker	AD
6	F	59	2.3	Former smoker (60)	Male‐differentiated AD
7	M	64	9.2	Former smoker 25	AD
8	F	58	‐	Former smoker 40	AD
9	M	58	34	Former smoker 40	ADSC
10	M	41	1.7	Former smoker (50)	AD

Characteristic of IHC and FISH analysis

IHC staining and FISH results differed between positive samples. Five of seven samples with ALK abnormal protein expression showed heterogeneous granular cytoplasmic reaction (with accompanying membranous reaction in three cases) and all of them had <50% of cancer cells with ALK abnormalities. Two samples with ALK abnormal protein expression showed homogenous cytoplasmic reaction with 50%–100% of cancer cells with ALK abnormalities.

According to FISH results, two samples were classified as equivocal, three samples presented at least 25% of ALK rearranged nuclei, one sample was FISH negative and one sample had non‐diagnostic results of FISH analysis. In two patients with at least 50% of cells expressing abnormal ALK protein in IHC method, FISH analysis revealed polysomy of ALK gene and ALK rearrangement in 25% and 55% of analyzed nuclei. In third patient with at least 50% of cells expressing abnormal ALK protein in IHC method, FISH result was non‐diagnostic. In fourth patient, IHC method allowed to detect ALK abnormal protein in 25% of cells and FISH method showed high‐grade polysomy and 82% of ALK rearranged nuclei. However, in the equivocal group of FISH results (10 and 17% of ALK rearranged nuclei with high‐grade polysomy of ALK gene), one sample showed ALK abnormal protein in 50% of cells and second sample demonstrated ALK abnormal protein in 25% of cells. Patients with FISH negative result showed ALK abnormal protein expression in IHC method in 10% of cells (Table 3, Figure 6).

Figure 6.

Analysis of abnormal ALK protein expression by IHC method in patients with no ALK gene rearrangement in FISH analysis (patients no 4). Figure 6a shows H + E staining, 2b shows negative IHC staining with Rabbit Monoclonal Negative Control and 2c shows positive IHC staining with Positive Rabbit Monoclonal Antibody D5F3 against ALK abnormal protein.

The specificity of the metastatic lesions in the CNS had an influence on the result interpretation. Metastatic tumors contained high number of CD68‐positive macrophages as well as nerve and glial cells, which could have made a false positive background in IHC staining (Figure 7).

Figure 7.

Example of false positive analysis of expression of abnormal ALK protein. Figures 7a and 7b show H + E staining and negative (with Rabbit Monoclonal Negative Control) IHC staining. Figure 7c shows false positive IHC staining caused by reaction between Positive Rabbit Monoclonal Antibody D5F3 and nerve cells, and strings of macrophages. Figure 7d shows positive IHC reaction with anti‐CD68 antibody in strings of macrophages.

Discussion

To date the knowledge on the effectiveness of targeted therapies in NSCLC patients with CNS metastases is relatively scarce. This may partly be because of exclusion of patients with untreated CNS metastases from a proportion of recent trials investigating new therapies in lung cancer 23. Data from small series indicate that EGFR TKIs (erlotinib, gefitinib, afatinib, ozimertinib) may be effective in a proportion of NSCLC patients with CNS metastases and EGFR mutations 3, 4, 6, 7, 41. Likewise, second‐generation of ALK inhibitors (alectinib, ceritinib) have shown relatively high activity in patients harboring ALK gene rearrangement, probably because of their effective penetration through the BBB 3, 8, 12, 31. Moreover, a promising activity of sequential or concurrent application of EGFR TKIs and WBRT has been recently postulated to increase the efficacy of CNS metastases management 4, 6, 7, 41.To date there are single data describing incidence of ALK abnormal protein expression in CNS metastases 4, 26. Based on promises from clinical observations we undertook the retrospective study in order to extend the knowledge about incidence of ALK abnormal protein expression in CNS metastases of NSCLC. The group of NSCLC patients in whom we evaluated the expression of ALK abnormal protein in brain metastases is one of the largest worldwide. In addition to our results we focused on technical aspects concerning analysis of IHC slides obtained from CNS metastases that ware not precisely show in previous study 4, 26. Therefore, we extremely noted that additional protocol standardizations are required to avoid unspecific staining of macrophages, glial/nerve cells or lymphocytes, as well as mucus background of neoplastic infiltration and areas of tumor necrosis.

The presence of ALK rearrangement is considered as the indicator of effectiveness of ALK TKIs therapy in primary NSCLC. However, these therapies also showed promising effects in corresponding CNS metastases of NSCLC 8, 12, 31, 43. We undertook this study in order to extend and problems with diagnosis of ALK abnormalities using IHC and FISH methods.

We detected ALK abnormalities in seven samples (7/145; 4.8%) obtained from CNS metastases of NSCLC. Our unpublished data showed that ALK abnormalities in primary adenocarcinoma were present in 8% of Polish patients (21/263). We also confirmed that ALK abnormalities are more common in young (<60) adenocarcinoma patients who declared never or light/former smoking status 34, 35, 40, although we found one ALK gene rearrangement in CNS metastases of ADSC lung carcinoma, in former smoking patient. Furthermore, in our study ALK abnormalities were more frequently detected in female than male (8.9% vs. 4%) patients.

IASCL indicated that pre‐selection of patients according to clinical (histopathology, wild type of EGFR gene) and demographic features (age, gender, race) may increase the probability of detection of ALK gene rearrangement 37. Our results partially remain in concordance with observations about frequency of ALK gene rearrangement in primary tumors of NSCLC (3%–7%) reported in the literature. Several studies confirmed that ALK gene rearrangement is observed frequently in young adenocarcinoma patients (especially in male patients), who are non‐smokers or former smokers 17, 34, 35, 40. However, the overall data described that 9%–11% of ADSC) tumors expressed ALK gene rearrangement 21, 29, 31, 37, 44. Martelli et al and Shaozhang et al suggested that the presence of ALK abnormalities in ADSC is more likely associated with the presence on AD component in the tumor 21, 29. It is worth to mention that ALK rearrangement was also reported in SCC primary tumors 2, 21, 29, 34, 37, 40, 44, 46. However, the overall literature data summarized that only 1.3% (18/1411) of SCC may harbor ALK gene rearrangement 37. Additionally, Shaw et al and Kwak et al mentioned that the variabilities in detectability of ALK rearrangement in SCC may arise from smoking habit and light/former smoking status may increase the frequency of ALK gene rearrangement in SCC 17, 32. Moreover, the pre‐selection of patients with NSCLC‐NOS and male‐differentiated NSCLC tumors according to smoking status may also increase the probability of ALK gene rearrangement detection in such group of tumors 13, 27. IASLC showed that ALK gene rearrangement may reach 4.5% (17/376) in NSCLC‐NOS tumors 37. Our high frequency of expression of ALK abnormal protein in low differentiated adenocarcinoma remains in concordance with the results of studies which analyzed ALK gene rearrangement in NSCLC‐NOS tumors. Martinez et al found ALK gene rearrangement in 15.4% of NSCLC‐NOS tumors (2/13), Park et al in 10.5% (8/76), Wong et al in 8.7% (2/23), An et al and Zhang et al in 8.3% (1/12) of such tumors 2, 22, 25, 40, 45.

In literature data, ALK gene rearrangement was mainly assessed using three methods: IHC, FISH and reverse transcription‐polymerase chain reaction (RT‐PCR). However, the utility of RT‐PCR is limited because of specificity of the assays which are not able to distinguish the variability in the EML4‐ALK fusion structure, what creates common false negative results 1, 9, 18. However, results of IHC and FISH showed high concordance in qualification of patients to ALK TKIs targeted therapies 1, 9, 24, 28. For this reason, in laboratory routine, IHC staining is considered as a valuable and cost‐effective screening technique of ALK abnormal protein and FISH is used as a reference technique in confirmation of equivocal IHC results (FDA recommendation) 1, 19, 20, 24, 28, 37. In our study, FISH analysis was performed in ten CNS metastatic specimens of NSCLC, which were primarily selected in IHC staining and which showed the expression of ALK abnormal protein. Using FISH technique three samples confirmed the presence of ALK rearrangement, while three were equivocal, three negative and one non‐diagnostic. Re‐assessed IHC staining in freshly prepared slides confirmed that three of two negative and one equivocal FISH tumor samples did not show expression of abnormal ALK protein.

Ali et al and Liu et al reported that discrepancies between IHC and FISH results might be caused by the heterogeneity of tumor tissues, which are full of false positive background from alveolar macrophages, nerve and ganglion cells, normal mucosa and necrotic areas 1, 28. Moreover, Ying et al suggested that type of intrachromosomal inversion of ALK gene and creation of different variants of EML4‐ALK fusion gene may change the fluorescent signal in FISH positive samples leading to equivocal or false negative results 42. On the other hand, Weickhardt et al suggested that IHC staining detects ALK expression for ALK fusion gene regardless of variant and fusion partner, which may influence protein expression and IHC results interpretation 39. It is also worth to mention that microscopic interpretation is still considered as subjective, which may show variability between readers and slides 9, 28. Moreover, the sensitivity of IHC antibodies and FISH molecular probes were validated on tissues from primary NSCLC and there is no data comparing their utility in tissues from CNS metastases.

Moreover, in our study six out of nine FISH tested samples showed (>40%) ALK gene polysomy, which was extremely high (>70%) in all three samples with positive results of first IHC examination. However, it was noted that ALK abnormal protein appears before chromosomal instability as an early sign of disease progression 1, 9, 20, 24, 38. On the other hand Katayama et al and Illie et al reported that ALK gene amplification may influence ALK expression in samples without ALK gene rearrangements, that can also explain the discordances between IHC and FISH results 14, 15. ALK gene copy numbers have recently been described in NSCLC as a frequent alternation 14, 47. Discussable is how polysomy affects ALK signaling pathways 14 and whether it could be predictive of either sensitivity or resistance to ALK TKIs targeted therapies 11, 14, 47.

To conclude, in this retrospective study we evaluated the expression of ALK abnormal protein and ALK gene rearrangement in extremely unique material which are CNS metastatic lesions of NSCLC. The frequency of ALK abnormalities in this material could be higher or comparable to frequency of ALK gene rearrangement in primary NSCLC tumors. However, the comparison of IHC and FISH results showed discrepancies that arose from unspecific background, which was made by cells with nonmalignant origin. For this reason assessment of ALK gene rearrangement in CNS tissues require additional standardizations.

Disclosure/Conflict of Interest

The authors declare no conflict of interest.