| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Cancer Diagnostics and Molecular Pathology |
a Pathology Department, b Clinical Pathology, c Biostatistics Unit, d Breast Disease Management Team, Regina Elena Cancer Institute, Rome, Italy
Key Words. Breast cancer • HER-2 • Chromogenic in situ hybridization • Fluorescence in situ hybridization • Liquid-based cytology
Correspondence: Amina Vocaturo, Ph.D., Regina Elena Cancer Institute, Via Elio Chianesi 53, 00144 Rome, Italy. Telephone: 390652666905; Fax: 390652666139; e-mail: vocaturo{at}ifo.it
Received May 8, 2006; accepted for publication July 20, 2006.
Access and take the CME test online and receive 1 AMA PRA Category 1 CreditTM at CME.TheOncologist.com
![]()
LEARNING OBJECTIVES
Top
Learning Objectives
Abstract
Introduction
Materials and Methods
Results
Discussion
Conclusion
Disclosure of Potential...
References
After completing this course, the reader will be able to:
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
Currently, two types of assays can be used for HER-2 evaluation: immunohistochemistry (IHC), which detects protein overexpression according to the U.S. Food and Drug Administration (FDA)-approved scoring staining system (0, 1+, 2+, 3+), and fluorescence in situ hybridization (FISH), which assesses gene amplification [8, 9]. Comparative studies of IHC and FISH have generally shown a high concordance rate between the two methods on primary or metastatic BC [8, 10]. Nevertheless, a high percentage of tumors displaying a 2+ score do not present concurrent gene amplification. In these discordant cases, HER-2/neu amplification appears to provide more prognostic information and better response to trastuzumab therapy than HER-2 protein overexpression [11].
Indeed, although FISH is a very accurate and sensitive assay and is regarded as the "gold standard" test with a high sensitivity and specificity in detecting HER-2/neu amplification [12], this methodology has some disadvantages as it requires expensive microscope equipment, morphological features are difficult to visualize, and fluorescence fades quickly, so it does not provide a permanent record.
Chromogenic in situ hybridization (CISH) is a recent methodology, which was introduced as an alternative to FISH. With this technique, HER-2/neu gene copies are detected using a permanent peroxidase reaction visualized by light microscopy, which makes it easy to observe both morphology and the amplification product [13, 14].
The present study analyzes the clinical usefulness of CISH, in our hospital-based practice, on a retrospective series of 111 histological samples of invasive BC. We compared the two in situ hybridization methods, CISH and FISH, with IHC by HercepTestTM (Dako, Milan, Italy) and CB11, which are the most frequently used reagents for detecting HER-2 overexpression. Moreover, in a prospective series of 53 BC patients, we investigated whether CISH can become a valid methodology for determining HER-2 status in fine-needle aspirates (FNAs) processed using ThinPrep® (Cytyc, Rome, Italy) technology, comparing the results obtained with the corresponding formalin-fixed, paraffin-embedded (FFPE) BC specimens.
| MATERIALS AND METHODS |
|---|
|
|
|---|
A prospective series of 53 cytological samples was collected by performing FNA from surgically removed BC. Cytological specimens were processed using Thin-Prep methodology. At least five slides were obtained from each sample: one for Papanicolaou staining to verify the presence of neoplastic cells, one for immunocytochemical (ICC) staining (HercepTest), two for CISH, and one for FISH. Moreover, in order to confirm our cytological findings, the same procedures (i.e., IHC with HercepTest and CISH) were performed on FFPE BC tissues from the same 53 patients. The FISH analysis was performed on 26 of 53 matched cytological and histological samples.
Cultured Cell Lines
The SKBR3, BT474, and ZR751 cell lines were kindly provided by the Laboratory of Immunology, Regina ElenaCancer Institute, Rome, Italy. These BC-derived cell lines are reported to contain highly amplified (Fig. 1A
), low amplified, and nonamplified HER-2/neu gene, respectively. The three cell lines were grown to confluence and cells were treated with trypsin-EDTA solution. After washing in phosphate buffer (0.1 M, pH 7.2), the cells were resuspended in CytoLyt® (Cytyc), prepared by the ThinPrep technique, and then processed for ICC, CISH, and FISH, according to their respective technical procedures. These cells were then used as positive and negative controls.
|
FNA specimens were obtained from 53 surgically removed samples of BC within 30 minutes of resection, before tissue fixation. The aspirates were obtained using a 22-gauge needle and placed in CytoLyt for the fixation and preparation of ThinPrep slides according to the manufacturers protocol. Corresponding BC tissue specimens were fixed in 10% neutral buffered formalin for 1824 hours according to routine procedures and embedded in paraffin.
IHC/ICC on Histological and Cytological Specimens
HER-2 overexpression was determined by the use of the pAb A0485 and mAb CB11. Both antibodies are directed against the intracytoplasmic domain of the HER-2 molecule. HercepTest had been previously performed according to the standard procedures described in the manufacturers guide accompanying the kit. Immunoperoxidase staining for CB11 was carried out by pretreating the deparaffinized and rehydrated sections in a thermostatic bath at 96°C for 40 minutes in a 10 mM citrate buffer, pH 6.
ThinPrep slides, after ethanol fixation and postfixation in buffered formalin, were washed in double-distilled water and processed as tissue specimens to determine HER-2 status.
CB11 immunoreactivity was revealed using the HRP-LSAB2 system (Dako), employing 3-amino-9-ethylcarbazole (AEC, Dako) as chromogenic substrate. Sections were slightly counterstained with Mayers hematoxylin and mounted in aqueous mounting medium (Glycergel®; Dako).
For both specimen types, appropriate positive tissue and cell line controls were included in each run. Negative controls consisted of substituting the HER-2 primary antibody with normal rabbit serum (negative control reagent, Dako).
Evaluation of the IHC/ICC results was done independently and blindly by two investigators (SB, MB).
HER-2 overexpression was determined as defined in the HercepTest kit guide: scores of 0 or 1+ were considered negative, a score of 2+ was considered weak positive, and a score of 3+ was considered strong positive. To qualify for 2+ scoring, complete membrane staining on more than 10% of tumor cells at a weak intensity had to be observed. In the majority of 3+ cases, at least 80% of the tumor cells demonstrated intense and homogeneous cell membrane staining.
FISH Procedure on Histological and Cytological Specimens
FISH was performed using the PathVysion® assay kit (Vysis, Inc., Downers Grove, IL), which includes two directly labeled DNA probes: a locus-specific probe for the HER-2/neu gene labeled with SpectrumOrangeTM (LSI HER-2) and an alpha satellite probe that targets the centromere region of chromosome 17 labeled with SpectrumGreenTM chromosome enumeration probe (CEP 17). The assay was performed according to the manufacturers instructions. In brief, after deparaffinization, FFPE sections were incubated in the pretreatment solution (80°C, 10 minutes) and then digested with protease (37°C, 15 minutes). FNA slides, obtained by the ThinPrep technique, were immersed in Carnoys fixative and stored at 20°C before FISH analysis. Cytological specimens were dehydrated before applying LSI HER-2/CEP 17 probe and then were hybridized overnight at 37°C. After washing, slides were counterstained with 4',6-diamidino-2-phenylindole (DAPI) and analyzed under a fluorescence microscope. An average of 200 nuclei was enumerated in the invasive component of tumor tissue and an average of 50100 tumor cells was enumerated in cytological preparations. HER-2/ neu gene copy number, chromosome 17 copy number, and the average HER-2/neu gene to chromosome 17 signal ratio were reported as FISH genetic variables. Samples with a ratio value
2.0 were considered to be amplified.
CISH Procedure on Histological and Cytological Specimens
FFPE sections were deparaffinized, dehydrated, air dried, and heated in boiling Tissue Heat Pretreatment Buffer for 15 minutes using a SPoT-Light® FFPE reagent kit (Zymed, Histoline, Milan, Italy). The enzymatic digestion was performed using SPoT-Light® FFPE digestion enzyme (Zymed) for 10 minutes at room temperature (RT).
FNA ThinPrep slides, prefixed in 96% ethanol, were postfixed in 99% ethanol, air dried, and then immersed in 2X standard saline preheated citrate solution (SSC; Zymed) at 37°C for 1 hour.
After dehydration, histological and cytological slides were air dried and the ready-to-use double-stranded DNA digoxygenin-labeled HER2 probe (Zymed) or the biotin-labeled Chromosome 17 Centromeric Probe (Zymed) were applied. The denaturation was performed by incubating the slides, covered with a CISH coverslip (Zymed), on a 96°C heating block for 5 minutes, and the hybridization was performed by placing the slides in a humidity chamber at 37°C overnight. After removing the coverslips, a stringent wash was performed in 0.5X SSC at 80°C for 5 minutes. The endogenous peroxidase activity and unspecific staining were blocked by applying 3% H2O2 and the CAS-BlockTM (Zymed), respectively. A mouse antidigoxygenin antibody was added to slides hybridized with HER-2 probe for 45 minutes at RT followed by incubation with a polymerized peroxidase-goat anti-mouse antibody (Dako) for 45 minutes at RT. On FFPE tissue slides, the colorimetric signal of the Chromosome 17 Centromeric Probe was improved by incubating the slides with a mouse antibiotin antibody (Dako) for 45 minutes at 37°C. For cytological slides, the Chromosome 17 Centromeric Probe was detected employing a horseradish peroxidase (HRP)-streptavidin for 45 minutes at RT. A 3,3'-diaminobenzidine (DAB) chromogen substrate system was used to generate a sensitive signal that could be viewed with a transmission light (brightfield) microscope after hematoxylin counterstaining.
Evaluation of CISH Results
Amplification was considered to be high when more than 10 copies, or large clusters, of the HER-2/neu gene were present in >50% of cancer cell nuclei, whereas 610 copies of the HER-2/neu gene, or a small HER-2/neu gene cluster, in the same percentage of cells were considered to be low amplification. Tumors were not considered to be amplified when 15 copies of the HER-2/neu gene were identified per nucleus. In the group of tumors with low amplification (610 copies/nucleus), an adjacent serial section was hybridized with the Chromosome 17 Centromeric Probe to detect polysomic cases (Fig. 1B
).
CISH signals were read by two investigators (MM, AV) independently of the results of the other two assays.
Statistics
Correlation among tests (IHC, ICC, CISH, and FISH) as well as interobserver agreement were estimated using the kappa test. For the IHC comparison of HER-2 overexpression, categories 0 and 1+ were grouped together as negative results and categories 2+ and 3+ were considered positive results. p < .05 was considered statistically significant.
Specificity, sensitivity, negative and positive predictive value (NPV and PPV), concordance, and the 95% confidence interval (CI) of the CISH assay were estimated considering FISH as the gold standard.
| RESULTS |
|---|
|
|
|---|
= 0.80; 95% CI, 69%92%; p < .0001) was evidenced in our retrospective series of 111 BC specimens. In particular, 52 cases were negative and 49 were positive with both reagents, whereas 10 cases were discordant. As summarized in Table 1
coefficient for the interassay agreement of 0.88 (95% CI, 78%97%; p < .0001). Sensitivity for CISH was 85% (95% CI, 73%95%), specificity was 100% (95% CI, 100%100%), PPV was 100% (95% CI, 100%100%), and NPV was 92% (95% CI, 86%98%). All six FISHCISH discordant tumors demonstrated a 3+ score when the HercepTest was used, whereas using the mAb CB11, three cases displayed a 3+ and three displayed a 2+ score (data not shown).
|
|
= 0.71; p < .0001; 95% CI, 82%96%) and 97% (
= 0.93, p < .0001; CI, 94%100%) concordance, respectively. On the other hand, the concordance between CB11 immunoreactivity and gene amplification was 90% for CISH (
= 0.74; p < .0001, 95% CI, 83%96%) and 93% for FISH (
= 0.84; p < .0001; 95% CI, 88%98%).
|
The ICC staining, performed in the 53 prospective FNAs, evidenced 37 (70%) cases that were HER-2 negative (score 0 or 1+) and 16 (30%) that were positive. The 37 HER-2-negative cytological cases were also negative in histological sections and none of them presented HER-2/ neu gene amplification. When we focused on the 16 HER-2-positive FNAs (Table 3
), we observed that four of these cases (three score 3+ and one score 2+) were negative in histological sections (score 1+). These discordant cytological cases were interpreted as being ICC false positive because they were nonamplified with the CISH and FISH tests, both in FNA and in FFPE samples. Six of the remaining 12 cases presented overlapping results using ICC, IHC, CISH, and FISH in both cytology and histology (all BC cases were not amplified, including one case scored 3+ by ICC and IHC). Another three cases, ICC and IHC positive (two score 3+, one score 2+), were amplified by CISH both in liquid-based cytology (LBC) FNAs and in histological tissues. In the latter tumors, because of the paucity of cellular material, we were unable to determine the HER-2/neu gene amplification status by FISH in FNAs and we did not perform FISH on the companion FFPE specimens. The last three BC specimens were all amplified both by CISH and FISH. In particular, using FISH, we found a HER-2/neu to CEP 17 ratios of 3, 2.5, and 4 in cytological preparations and of 5.1, 2.8, and 3.6 in the corresponding histological specimens. A lower HER-2 score was detected with ICC (score 2+) than with IHC (score 3+). Therefore, of 53 FNAs, 46 showed overlapping ICC results with those obtained on histological tissues (concordance, 87%;
= 0.69; p < .0001; 95% CI, 78%96%).
|
In order to further validate the CISH test in LBC, 26 of the 53 FNAs containing an adequate number of neoplastic cells (>50) and the companion histological sections were also tested by FISH. The two series showed completely overlapping results, with 3 of 26 cases (12%) amplified and 23 cases (88%) nonamplified with the two genetic tests (CISH and FISH), both on the cytological and the corresponding histological samples with 100% of concordance. (Data not shown.)
Interobserver Agreement
Interobserver agreement, calculated by
value, indicated an excellent concordance for both IHC (
= 0.96; 95% CI, 94.7%100%) and CISH (
= 0.94; 95% CI, 94.3%100%), p < .0001. Those cases in which there was a discrepancy between the two readers were resolved by reviewing the cases using a multiheaded microscope and reaching a consensus.
| DISCUSSION |
|---|
|
|
|---|
Although FISH is in general considered the gold standard methodology for detecting HER-2/neu gene amplification, the recent development of the CISH assay provides an attractive alternative to FISH, especially in laboratories that are familiar with peroxidase-based staining. In recent years, the advantages of CISH, which allows a simultaneous evaluation of gene copy number, tumor cells, and detailed surrounding tissue morphology on the same histological slide, have been widely reported by several authors [14, 16, 17]. In contrast, data on the comparative use of ICC and CISH on cytological specimens and the corresponding FFPE tissues are minimal [18, 19]. To address this issue, in our study, we first validated CISH on a retrospective series of FFPE BC samples, subsequently focusing on the feasibility of CISH in a prospective series of BC FNAs processed using ThinPrep technology.
In FFPE samples, complete agreement between FISH and CISH results was demonstrated for IHC scores 0, 1+, and 2+. A lower concordance for score 3+ tumors was found because six tumors were amplified by FISH and nonamplified by CISH. A careful re-examination of the six CISHFISH discordant tumors revealed that all these cases were collected from 1992 to 1998, and we can hypothesize that these FFPE specimens were ineffective as a result of preanalytic problems, such as bad tissue preservation or a poorly standardized formalin fixation method. The latter could represent a limit of this new assay when compared with FISH [12, 20]. The current CISH procedure is based on a single color detection of one probe and does not allow correction for HER-2/neu amplification caused by chromosome 17 polysomy, as in the dual-color FISH test provided by Vysis. Nevertheless, in our series, we did not find any CISH false-positive cases. This is probably because of the fact that, in tumors displaying 610 HER-2/neu gene copies/nucleus, an adjacent consecutive section was consistently hybridized with CEP 17 for comparison. Although Sartelet et al. [19] reported that polysomy is a rare event in cases with a low gene copy number, we strongly feel that this is a crucial point, because, without adjusting CISH results for chromosome 17 copy number, a higher number of discordant interpretations could result in borderline cases [14, 17, 2022].
In agreement with the majority of studies reporting a concordance between FISH and CISH of 83%100% [1214, 16, 17, 20, 23, 24], in our series of FFPE tissues, the interassay concordance between the two genetic tests was 95% (
= 0.88). This accurate histological validation of the CISH procedure provided the technical basis to test the feasibility of CISH in LBC, comparing our results on histologically matched tumors. We had two main aims. First, we assessed the potential value of CISH in cytological samples. Second, we verified whether LBC, which offers the possibility of adjunctive investigations on the same homogeneous samples, may be a suitable methodology to detect HER-2/neu gene amplification by CISH as already demonstrated by FISH [10].
Although FNA procedures are ever more frequently used instead of tissue biopsy to obtain diagnostic cellular material in primary inoperable or metastatic BC, to the best of our knowledge, only two studies have analyzed the CISH assay on conventional smears [18] or ThinPrep cytology [19] in parallel with the companion FFPE tissues. Moreover, neither of the two studies compared CISH with FISH findings.
In summary, our results demonstrate that CISH is easily applicable to LBC with a good correlation to results obtained with the corresponding histological sections, and a complete agreement with FISH. This issue is of major importance, because CISH could be particularly helpful in overcoming the problem of false-positive findings with ICC, one of the major criticisms of this methodology [25, 26]. In our series, we found four positive cases (three score 3+, one score 2+) using ICC with LBC that were negative using IHC (score 1+) on the companion FFPE tissues and nonamplified by CISH both in ThinPrep and histological samples. These findings seem to indicate that CISH may lower the ICC false positivity that is probably a result of the poorly assessed scoring system in cytological material. FNA BC cells, in fact, lack an architectural tissue structure, which makes it difficult, in most cases, to apply the FDA-approved histological scoring system to cytopathology. These interpretative difficulties may lead to a great variability in evaluating the immunocolorimetric signals, which consequently become highly subjective. In contrast, CISH allows a more objective interpretation of HER-2/neu gene status, mainly with LBC. This technical procedure, in fact, different from conventional smears, permits a monolayer cell assessment. Consequently, tumor cells appear cleaner and chromogenic dots are easier to score because of intact nuclei.
Considering the fact that in all ancillary tests the technical thoroughness is very important, in our cytological series, preanalytical procedures were followed very carefully, allowing a complete agreement of results between CISH and FISH (100% concordance).
| CONCLUSION |
|---|
|
|
|---|
| DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
|
|---|
| ACKNOWLEDGMENT |
|---|
|
|
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
S Di Palma, N Collins, M Bilous, A Sapino, M Mottolese, N Kapranos, F Schmitt, and J Isola A quality assurance exercise to evaluate the accuracy and reproducibility of chromogenic in situ hybridisation for HER2 analysis in breast cancer J. Clin. Pathol., June 1, 2008; 61(6): 757 - 760. [Abstract] [Full Text] [PDF] |
||||
![]() |
F C Schmitt, A Longatto-Filho, A Valent, and P Vielh Molecular techniques in cytopathology practice J. Clin. Pathol., March 1, 2008; 61(3): 258 - 267. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. A. V. Ricardo, F. Milanezi, S. T. Carvalho, D. R. A. Leitao, and F. C. L. Schmitt HER2 evaluation using the novel rabbit monoclonal antibody SP3 and CISH in tissue microarrays of invasive breast carcinomas J. Clin. Pathol., September 1, 2007; 60(9): 1001 - 1005. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| THE ONCOLOGIST | STEM CELLS | CME | ALPHAMED PRESS JOURNALS |