Immunohistochemical detection of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 expression in breast carcinomas

Comparison on cell block, needle-core, and tissue block preparations

Authors


Abstract

BACKGROUND:

Fine-needle aspiration (FNA) is a rapid and accurate procedure for the detection of breast carcinomas. The evaluation of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression by immunohistochemistry (IHC) is performed routinely on formalin-fixed, paraffin-embedded needle-core (NC) or excision tissue block (TB) preparations, according to the American Society of Clinical Oncology/College of American Pathologist guidelines. In this retrospective study, the authors compared expression levels of ER, PR, and HER2 in ethanol-fixed BC FNA cell block (CB) samples with expression levels in formalin-fixed NC and TB samples.

METHODS:

Forty-one breast carcinoma CB samples with concurrent or subsequent NC and TB samples were identified. Patients who had received neoadjuvant or adjuvant chemotherapy were excluded. CB samples initially were fixed in 50% ethanol (4-12 hours), and this was followed by formalin fixation (minimum, 6 hours). NC samples were placed promptly in formalin for a minimum of 6 hours. Within 4 to 8 hours, TB samples were fixed in formalin for 6 to 48 hours. Fluorescence in situ hybridization (FISH) results were also compared.

RESULTS:

IHC for ER on alcohol-fixed CB samples had good correlation with NC and TB samples. PR results on TB samples had excellent agreement with NC samples. A higher discordance rate wais observed when PR results were compared between CB samples and NC samples. HER2 detection on ethanol-fixed CB samples resulted in a higher rate of positive and equivocal staining than NC or TB samples. HER2 IHC on TB samples demonstrated better correlation with FISH results than CB or NC samples.

CONCLUSIONS:

Alcohol fixation did not affect ER results in breast carcinoma, but it may alter tumor cell PR antigenicity. The authors concluded that CB samples could be used to triage patients for tamoxifen therapy, but they are not reliable for the assessment of HER2 status; therefore, CB results should be correlated with results from NC or TB samples. Cancer (Cancer Cytopathol) 2009. © 2009 American Cancer Society.

Fine-needle aspiration (FNA) is a widely used, well accepted, and minimally invasive diagnostic method for the evaluation of superficial and deep-seated lesions. Its sensitivity, specificity, and positive predictive value reportedly are the highest for evaluating primary and metastatic carcinomas and melanomas.1 Although the diagnosis on a cytology sample is established primarily on cytomorphologic features, adjunctive tools, such as immunohistochemistry (IHC) and molecular tests, often are used for prognostic purposes. When IHC is anticipated, additional material can be collected for cell block (CB) preparation and placed into 10% neutral-buffered formalin, Bouin fixative, 50% ethanol, or alcohol-based preservatives that are used for liquid-based (LB) processing.

The prognostic and therapeutic implications of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor 2 (HER2) status in breast carcinomas (BC) are well established and been studied extensively.2-4 Up to 75% of primary BCs express ER, and about 50% of these tumors coexpress PR.4 Only a minority of BCs have no ER or PR expression (20%), and an even smaller percentage of primary BCs express PR but not ER (<10%). HER2 overexpression is present in 20% to 30% of BCs and should be evaluated by IHC and/or fluorescence in situ hybridization (FISH) according to The American Society of Clinical Oncology and College of American Pathologists (ASCO/CAP) guidelines for all primary BC.5 The routine method for detecting ER and PR expression is IHC, whereas HER2 overexpression can be detected by IHC and/or FISH. The IHC analysis of ER, PR, and HER2 expression has been standardized for formalin-fixed, paraffin-embedded tissue sections.6, 7 Therefore, accurate detection of these prognostic markers is essential, and consistent results can be achieved by compliance with the ASCO/CAP national guidelines.5 The guidelines also require validation of HER2 testing by documenting 95% concordance rates between IHC HER2 positive (3+ staining) and IHC HER2 negative (0 or 1+ staining) with FISH amplified and nonamplified cases, respectively. The detection of these markers on FNA samples has been studied extensively on direct smears, cytospin slides, LB preparations, and CB sections with variable success rates.8-11 The objective of this study was to evaluate the role of FNA in hormone receptor and HER2 status assessment of primary and metastatic BC and to determine its accuracy compared with needle-core (NC) and tissue block (TB) preparations.

MATERIALS AND METHODS

Study Design and Case Selection

Surgical pathology and cytopathology archives were searched at Emory University Hospital for patients who were diagnosed with primary or metastatic BC by FNA and who also had concurrent or subsequent NC and surgical excision (TB) samples available. Forty-one BCs, including 27 primary tumors, 12 lymph node metastases (10 axillary, 2 supraclavicular), and 2 solid organ metastases (1 liver, 1 lung) from 39 patients (38 women, 1 man) were identified retrospectively. Histologic types included 35 ductal BCs, 5 lobular BCs, and 1 mucinous BC, all of which had ER, PR, and HER2 status determined by IHC on CB, NC, and TB preparations. Samples from patients who had received neoadjuvant or adjuvant chemotherapy, paucicellular samples, and tumors in which hormone receptor status (HRS) was assessed at the referring laboratory were excluded. All FNA samples were assessed for adequacy on site, and additional passes were obtained for CB samples.

Fixation and Immunohistochemistry for Estrogen Receptor, Progesterone Receptor, and Human Epidermal Growth Factor Receptor 2

CB samples were fixed initially in 50% ethanol (4-12 hours), followed by fixation in 10% neutral-buffered formalin (minimum, 6 hours). NC preparations were placed promptly into 10% neutral-buffered formalin for a minimum of 6 hours, and TB preparations were formalin-fixed within 4 to 8 hours for 6 to 48 hours, followed by formalin fixation (minimum, 6 hours). IHC was performed, after epitope retrieval, with a polymer-based detection system (Envision+; Dako, Carpinteria, Calif) using mouse monoclonal antibodies for ER and PR (Dako), ER (1D5; 1:50 dilution), PR (PgR636; 1:400 dilution), and a Herceptin kit (HercepTest; Dako) according to the manufacturer's instructions. For ER and PR, antigen retrieval was performed as follows: sections were deparaffinized and rehydrated with deionized water. Then, they were heated in citrate buffer, pH 6.0, using an electric pressure cooker for 3 minutes at 12 to 15 pounds per square inch at approximately 120°C and cooled for 10 minutes before immunostaining. All slides were loaded onto an automated system (Autostainer; Dako) and exposed to 3% hydrogen peroxide for 5 minutes, incubated with primary antibody for 30 minutes, incubated with labeled polymer (Envision+ dual link) for 30 minutes, incubated in 3,3′-diaminobenzidine (DAB) as a chromogen for 5 minutes, and counterstained with hematoxylin for 5 minutes. These incubations were performed at room temperature. Between incubations. sections were washed with Tris-buffered saline (TBS). Coverslipping was performed using the Tissue-Tek SCA coverslipper (Sakura Finetek USA, Inc., Torrance, Calif). Positive controls of known positive tissues (endometrium and breast) and negative controls with primary antibody replaced by TBS were run with the patient/study slides. Nuclear staining in >10% of tumor cells was considered positive for ER and PR.

Antigen retrieval for HER2 using HercepTest was performed by immersing and incubating the slides in 10 mmol/L citrate buffer in a calibrated water bath (95-99°C) for 40 minutes (±1 minute). After decanting the epitope-retrieval solution, the sections were rinsed in the wash buffer and, later, soaked in the buffer for 5 to 20 minutes before staining. The slides were loaded onto the autostainer using the HercepTest program as described in the manufacturer's insert. In the autostainer, the slides were rinsed, placed in 200 μL peroxidase-blocking reagent for 5 minutes, rinsed, placed in 200 μL primary anti-HER2 protein (or negative control reagent) for 30 minutes, rinsed twice, and finally immersed in 200 μL substrate chromogen solution (DAB) for 10 minutes. The slides were counterstained with hematoxylin and finally were coverslipped. HER2 results were determined based on the maximum area of staining intensity according to the package insert and the ASCO/CAP guidelines as follows. Strong, circumferential membranous, staining in >30% of invasive carcinoma cells was scored as 3+; moderate, circumferential, membranous staining in ≥10% of invasive tumor cells or 3+ staining in ≤30% of cells was scored as 2+ staining; weak and incomplete membranous staining in invasive tumor cells was scored as 1+; and no staining was scored as 0 (Figs. 1, 2, 3, 1-3). Tumors with 0 and 1+ staining were considered negative, cases scored as 2+ were considered equivocal, and cases with 3+ staining were considered positive as evaluated at ×4 and ×10 magnification.

Figure 1.

This photomicrograph reveals immunohistochemical staining for human epidermal growth factor receptor 2 (HER2) using HercepTest in an alcohol-fixed breast carcinoma cell block section. Weak and incomplete membranous staining in tumor cells was scored as 1+: interpretation, negative for HER2 amplification (original magnification, ×60).

Figure 2.

This photomicrograph reveals immunohistochemical staining for human epidermal growth factor receptor 2 (HER2) using HercepTest in an alcohol-fixed breast carcinoma cell block section. Moderate, circumferential, membranous staining in >10% of tumor cells or strong and circumferential staining in <30% of tumor cells was scored as 2+: interpretation, equivocal for HER2 amplification (original magnification, ×60).

Figure 3.

This photomicrograph reveals strong, circumferential immunohistochemical staining for human epidermal growth factor receptor 2 (HER2) using HercepTest in an alcohol-fixed breast fine-needle aspiration cell block section. Strong, circumferential staining in >30% of tumor cells was scored as 3+: interpretation, positive for HER2 amplification (original magnification, ×60).

Fluorescence In Situ Hybridization for Human Epidermal Growth Factor Receptor 2 Amplification

The FISH PathVysion Her2 DNA probe kit (Abbott Molecular Inc., Downers Grove, Ill) was used as a reference standard for HER2. For FISH analysis, the slides were deparaffinized by immersing them in CitriSolv for 10 minutes, followed by dehydration in 100% ethanol at room temperature, and finally air dried in a slide warmer at 45°C to 50°C. Then, the slides were pretreated by immersing them in 0.2 N HCL for 20 minutes at room temperature, followed by purified water for 3 minutes, wash buffer for 3 minutes, pretreatment solution at 80°C for 30 minutes, purified water for 1 minute, and wash buffer for 5 minutes. The slides then underwent protease treatment by immersion in protease solution for 10 minutes at 37°C, followed by wash buffer for 5 minutes at room temperature, and finally air dried on a slide warmer at 45°C to 50°C. Next, the slides were subjected to denaturation by immersing them in denaturing solution at 72°C ± 1°C for 5 minutes, followed by 70% ethanol for 1 minute, 85% ethanol for 1 minute, and 100% ethanol for 1 minute, and finally air dried on a slide warmer at 45°C to 50°C. The slides underwent hybridization by applying 10 μL of probe mixture to the target area of the slide, a 22 × 22 mm glass coverslip was placed over the probe to spread evenly, and the edges of the coverslip were sealed with rubber cement. The slides then were placed in a prewarmed, humidified hybridization chamber and incubated at 37°C overnight (14-18 hours). After removing the coverslips, the slides were immersed in 2 × standard saline citrate (SSC)/0.3% NP-40 (100 mL 20 × SCC, pH 5.3; 847 mL purified water; and 3 mL NP-40; pH adjusted to 7.0-7.5 with 1 N NaOH) at 72°C ± 1°C for 2 minutes. The slides were air dried in the dark in an upright position. Then, 4′,6-diamidino-2-phenylindole (DAPI) counterstain (10 μL) was applied to the target area of the slide, and a 22 × 40 mm glass coverslip was placed over the DAPI. In total, 30 cells were counted, and the ratio of HER2 to chromosome enumeration probe 17 (CEP-17) was calculated for each sample. The HER2 gene was considered amplified if the signal ratio of HER2 to CEP-17 was >2.2 (1.8-2.2 was considered equivocal, and <1.8 was considered negative for amplification).

Statistical Analysis

ER, PR, and HER2 IHC analyses performed on CB and TB samples were compared with analyses on NC samples. FISH results, if available, were compared between the 3 methods. Statistical analysis, including positive and negative agreement and discrepancy rates, were calculated using Spearman rank-correlation analysis (ρ) and the Cohen κ test of agreement. Values of κ > 0.6 were correlated with good agreement, values between 0.4 and 0.6 were considered moderate agreement, values <0.4 corresponded to fair, and values <0.2 reflected poor agreement.

RESULTS

Comparison of Estrogen Receptor Status in Breast Carcinomas by Immunohistochemistry on Cell Block, Tissue Block, and Needle-Core Samples

Evaluation of ER expression by IHC on CB samples had 82% positive agreement and 100% negative agreement with ER results on NC samples (Table 1). There were 5 more tumors that were positive for ER on NC samples than on CB samples (Tables 2, 3). All of these samples had strong staining for ER on the excision specimen. ER status on TB samples had excellent correlation with NC results (96.4% positive agreement and 92.3% negative agreement) with a slightly better κ value than on CB samples (0.88 vs 0.74) (Table 1).

Table 1. Positive and Negative Agreement of Estrogen Receptor, Progesterone Receptor, and Human Epidermal Growth Factor 2 Receptor Expression by Immunohistochemistry on Cell Block and Tissue Block Samples Compared With Needle-Core Samples
VariablePositive Agreement, %Negative Agreement, %Discrepant, %KappaPCorrelation (Spearman ρ)
  1. ER indicates estrogen receptor; CB, cell block; NC, needle core; TB, tissue block; PR, progesterone receptor; Her2, human epidermal growth factor 2.

ER CB vs NC8210012.20.74<.00010.74 (<.0001)
ER TB vs NC96.492.34.80.88<.00010.90 (<.0001)
PR CB vs NC43.79226.80.38.00980.41 (<.009)
PR TB vs NC100887.30.85.00010.87 (<.0001)
Her2 CB vs NC87.566.614.60.45<.00010.65 (<.0001)
Her2 TB vs NC8362.57.30.44.00010.60 (<.0002)
Table 2. Estrogen Receptor Status in Breast Carcinoma Needle-Core and Fine-Needle Aspiration Cell Block Specimens by Immunohistochemistry
FNA (CB)ER Status (NC)
PositiveNegativeTotal
  1. ER indicates estrogen receptor; NC, needle core; FNA, fine-needle aspiration; CB, cell block.

Positive23023
Negative51318
Total281341
Table 3. Estrogen Receptor Status in Breast Carcinoma Needle-Core and Excision Tissue Block Specimens by Immunohistochemistry
Excision (TB)ER Status (NC)
PositiveNegativeTotal
  1. ER indicates estrogen receptor; NC, needle core; FNA, fine-needle aspiration; CB, cell block.

Positive27128
Negative11213
Total281341

There were 2 discrepant cases on TB samples. One tumor was negative for ER on TB but had ER 3+ staining on NC (Tables 2, 3). It is noteworthy that, on an FNA CB sample, this tumor had ER 2+ staining. The other discrepant tumor had ER 2+ staining on TB but was negative for ER on NC and CB (Tables 2, 3).

Comparison of Progesterone Receptor Status of Breast Carcinomas by Immunohistochemistry in Cell Block and Tissue Block Samples Versus Needle-Core Preparations

The detection of PR expression on CB samples resulted in significantly more PR-negative tumors (n = 32) compared with the results on NC samples (n = 25) (Tables 4, 5). There was poor agreement between the 2 preparation methods (CB vs NC) for the evaluation of PR status (κ, 0.38; ρ, 0.41) (Table 1). In contrast, PR expression on TB samples correlated well with NC results (100% positive agreement, 88% negative agreement) and had excellent overall agreement (κ, 0.85) (Table 1). In fact, there was a slightly higher number of PR-positive tumors on TB samples (n = 19) than on NC samples (n = 16) (Tables 4, 5). All 9 PR-negative tumors on CB samples were PR-positive on TB samples. In addition, 4 of these cases also were negative for ER on CB samples but demonstrated ER and PR expression on NC and TB samples (Tables 4, 5).

Table 4. Progesterone Receptor Status in Breast Carcinoma Needle-Core and Fine-Needle Aspiration Cell Block Specimens by Immunohistochemistry
FNA (CB)PR Status (NC)
PositiveNegativeTotal
  1. PR indicates progesterone receptor; NC, needle core; FNA, fine-needle aspiration; CB, cell block.

Positive729
Negative92332
Total162541
Table 5. Progesterone Receptor Status in Breast Carcinoma Needle-Core and Excision Tissue Block Specimens by Immunohistochemistry
Excision (TB)PR Status (NC)
PositiveNegativeTotal
  1. PR indicates progesterone receptor; NC, needle core; TB, tissue block.

Positive16319
Negative02222
Total162541

Comparison of Human Epidermal Growth Factor Receptor 2 Status in Breast Carcinomas by Immunohistochemistry and Fluorescence in Situ Hybridization on Cell Blocks and Needle-Core Samples

Good positive agreement (87.5%) and fair overall agreement (κ, 0.45) was observed between NC and CB preparations when evaluating HER2-positive cases (3+) and HER2-negative cases (Table 1). From the 14 equivocal HER2 (2+) cases on NC samples (3 of 14 were FISH amplified), only 7 also were equivocal on CB samples (Table 6). The corresponding CB samples from the remaining 7 cases had 3+ staining in 6 cases and a negative HER2 result in 1 case (Table 6). All of these equivocal CB samples failed to reveal HER2 amplification by FISH analysis (Table 6). There were no tumors in which the CB sample was negative for HER2 by IHC but the NC sample was positive.

Table 6. Results of Human Epidermal Growth Factor Receptor 2 Amplification by Fluorescence In Situ Hybridization in Cases with Discrepant Immunohistochemistry Results: Comparison of Cell Block and Tissue Block Samples With Needle-Core Samples
HER2 Status by IHC
NC Results (No.)CB Results (No.)No. of Cases/Total No.TB Results (No.)No. of Cases/Total No.
FISH AmplifiedFISH Not AmplifiedFISH AmplifiedFISH Not Amplified
  1. HER2 indicates human epidermal growth factor receptor 2; IHC, immunohistochemistry; NC, needle core; CB, cell block; FISH, fluorescence in situ hybridization; TB, tissue block; 3+, strong, circumferential, membranous staining in >30% of invasive carcinoma cells; 2+, moderate, circumferential, membranous staining in >10% of invasive tumor cells or 3+ in ≤30% of cells; Neg, negative.

3+ (8)3+ (7)  3+ (6)  
2+ (1)1/1 2+ (1)1/1 
Neg (0)  Neg (1) 1/1
2+ (14)3+ (6)3/63/63+ (1)1/1 
2+ (7) 7/72+ (7)3/74/7
Neg (1) 1/1Neg (6)1/65/6
Neg (19)3+ (2) 2/23+ (0)  
2+ (4) 4/42+ (4) 4/4
Neg (13)  Neg (15)  

Comparison of Human Epidermal Growth Factor Receptor 2 Status in Breast Carcinomas by Immunohistochemistry and Fluorescence in Situ Hybridization on Tissue Blocks and Needle-Core Samples

Similar to the CB samples, good positive agreement (83%), but lower discrepancy rate (7.3%) was observed, when TB HER2 IHC results were compared with NC results (Table 1). Overall agreement was almost identical (κ, 0.44 vs 0.45) (Table 1). The rate of equivocal HER2 (2+) results on TB specimens was identical to the rate on CB specimens; however, a greater number of equivocal tumors on NC samples were staining negative on TB samples (6 of 14 cases) than on CB samples (1 of 14 cases) (Tables 7, 8). Only 1 of 6 TB HER2-negative tumors had amplification of HER2 by FISH (Table 6). Four tumors had equivocal staining on TB samples but negative staining on NC samples. FISH performed on TB samples in all 4 tumors was negative (Table 6).

Table 7. Human Epidermal Growth Factor Receptor 2 Expression Status in Breast Carcinoma Needle-Core and Fine-Needle Aspiration Cell Block Specimens by Immunohistochemistry
FNA (CB)HER2 Status (NC)
3+2+NegativeTotal
  1. HER2 indicates human epidermal growth factor receptor 2; NC, needle core; FNA, fine-needle aspiration; CB, cell block; 3+, strong, circumferential, membranous staining in >30% of invasive carcinoma cells; 2+, moderate, circumferential, membranous staining in >10% of invasive tumor cells or 3+ in ≤30% of cells.

3+76215
2+17412
Negative011314
Total8141941
Table 8. Human Epidermal Growth Factor Receptor 2 Expression Status in Breast Carcinoma Needle-Core and Excision Tissue Block Specimens by Immunohistochemistry
Excision (TB)HER2 Status (NC)
3+2+NegativeTotal
  1. HER2 indicates human epidermal growth factor receptor 2; NC, needle core; TB, tissue block; 3+, strong, circumferential, membranous staining in >30% of invasive carcinoma cells; 2+, moderate, circumferential, membranous staining in >10% of invasive tumor cells or 3+ in ≤30% of cells.

3+6107
2+17412
Negative161522
Total8141941

The rate of equivocal cases was similar with all 3 preparations (CB, NC, TB) (Tables 7, 8). However, HER2 amplification was detected in 21.4% (3 of 14) of NC samples, in 8.3% (1 of 12) of CB samples, and in 33.3% (4 of 12) of TB samples (Table 6).

DISCUSSION

The role of FNA in breast lesions is primarily diagnostic but also can be used as a therapeutic procedure in the setting of simple cysts, abscesses, or a galactocele. It is a rapid, cost-effective, outpatient procedure. A preliminary diagnosis can be rendered on site, and the patient can be triaged during the same office visit. Along with its good accuracy and sensitivity, it is a minimally invasive method for obtaining tumor cells for additional studies, such as prognostic biomarkers. Evaluation of HRS in newly diagnosed BC is performed routinely, based on which adjuvant or neoadjuvant treatments are planned. However, it is well known that several factors, including preanalytic factors (length and time of tissue fixation), analytic factors (type of antibody, antigen retrieval), and postanalytic factors (interpretation of results), can have major implications on the accuracy of results. Interlaboratory and intralaboratory variabilities can be reduced by strict adherence to the national guidelines established by ASCO/CAP.5, 6 The timing and length of formalin fixation has been studied in both NC and TB samples and, based on those studies, a minimum 6 hours of formalin fixation for ER and PR and not more than 48 hours for HER2 are recommended for optimal staining results by ASCO/CAP.5, 6

The monoclonal antibody trastuzumab (Herceptin) appears to have promising results in patients with HER2-positive tumors.12, 13 However, because of the cost of treatment and its toxicity, patients who may benefit from trastuzumab therapy should be selected carefully, and HER2 testing should be accurate and reliable. Whether these biomarker studies should be performed on NC biopsy or on TB preparations remains controversial. Some studies have concluded that NC biopsy is the optimal specimen for assessment of HRS because of immediate, rapid formalin fixation and standardized specimen handling guidelines.6, 14, 15 Regardless of which specimen is used to assess prognostic factors in BC, several studies have produced high concordance rates for ER (95-95.8%), PR (89-90.3%), and HER2 (86.6%-96%) results on NC biopsy versus TB specimens.14-16 Moreover, histologic and immunohistochemical heterogeneity of breast tumors does not seem to play a significant role in these differences.14, 17

Because of the application of formalin-free fixatives in cytology, validation of ER, PR, and Her2 IHC results is essential before these HRS results can be used for therapeutic purposes. IHC and FISH methods, which are approved by the US Food and Drug Administration (FDA) for the assessment of BC, require tissue samples to be fixed in 10% neutral-buffered formalin.5 Therefore, if alternative fixatives are used, then these tests should be validated against the formalin-fixed samples using the same assay based on the CAP guidelines. Several study groups have compared BC FNA IHC results with formalin-fixed NC or TB results, including IHC results on CB collected in Roswell Park Memorial Institute (RPMI) media,18 FNA smears fixed in alcohol, or Carnoy solution,19 to assess their effect on ER staining in cytologic samples.

Our current study indicated good concordance rates (82%, 96.4%) for ER between CB and NC results and between NC and TB results, respectively. These findings are in agreement with concordance rates of 84% to 96% reported in the literature.8, 10, 16, 19 Our results suggest that initial ethanol fixation may have a minimal effect on tumor cell antigenicity. Therefore, FNA is a reliable method to triage patients for tamoxifen therapy.

However, a higher discordance rate (26.8%) was observed in our study when PR results were compared between CB and NC. A similar higher discrepancy rate was noted by several study groups, ranging from 23.2% to 29%8, 10 with moderate agreement between PR results on cytology samples and formalin-fixed tissues.8, 10 The exact reason for these findings is unknown. Moreover, the therapeutic implication of BC PR status is not understood well. Although currently, for ductal carcinoma in situ, therapeutic decisions are based solely on ER status, it has been suggested that PR positivity also carries prognostic implications and is associated with increased disease-free and overall survival.20 Studies have also demonstrated that loss of PR positivity in invasive carcinoma can be encountered with tumor progression, which can reflect resistance to hormone therapy as a result of alteration of ER pathways.2, 13 However, this would not explain our findings, because all samples in our study were obtained before adjuvant or neoadjuvant treatments were initiated. The significant discordance rate between CB and NC PR results is attributed to a relatively large number of false-negative cases on CB. Furthermore, 4 of 9 cases that were PR negative on the CB were also negative for ER (all 4 cases were ER-positive and PR-positive on NC and TB samples), indicating that our results could be related to alteration in tumor cell antigenicity due to fixation. This hypothesis also is supported by a significantly lower PR discordance rate (7.3%) and excellent positive agreement (100%) between the formalin-fixed samples (NC vs TB).

Analysis of HER2 overexpression in BC has become the standard of care and allows clinicians to select patients who will benefit from Herceptin therapy. The HercepTest (Dako) is an IHC test that is approved by the FDA and uses a scoring system from 0 to 3+ to grade the intensity of staining for HER2 in tumor cells. PathVysion (Vysis, Downers Grover, Ill) is an FDA-approved FISH assay for the detection of HER2 amplification in BC. Studies have produced excellent correlation between HercepTest results and FISH results in HER2-negative (0 to 1+ staining) and positive (3+ staining) cases. Tumors that have equivocal (2+) staining for HER2 have poor interobserver reproducibility7, 21, 22 and should be evaluated further by FISH for HER2 gene amplification, because 8% to 25%23 or even up to 48%7 of tumors can reveal HER2 amplification with this method. Both HercepTest and FISH are optimized and standardized for formalin-fixed, paraffin-embedded tissue. There are limited data available in the literature regarding the performance and accuracy of these 2 tests on tissues fixed by alternative fixatives. In a study by Willmore-Payne et al24 6 nonformalin-based fixatives were evaluated to assess their impact on FISH testing for HER2 amplification. The most common interferences with FISH analysis were high background fluorescence, faint or absent signal, and bright cytoplasmic autofluorescence, all of which caused significant difficulties in the interpretation of results. Sumiyoshi and his group evaluated HER2 detection by IHC and chromogenic in situ hybridization on alcohol-fixed CB samples and compared the results with results from formalin-fixed tissue sections.25 Although, those authors reported 84% sensitivity, 87.9% specificity, and 86.2% accuracy for the assessment of HER2 status on cytology specimens, the concordance rate between the 2 methods was lower when IHC HER2 equivocal (2+ staining) cases were studied. Studies evaluating HER2 status by IHC on alcohol-fixed FNA smears detected significantly stronger staining intensity in a greater percentage of tumor cells compared with formalin-fixed tumor sections.26-28 Similar findings were observed by Vocaturo et al,29 who performed HER2 IHC on Thin Prep-processed BC FNA slides. IHC HER2 results on ethanol-fixed, methanol-fixed (CytoLyt-ThinPrep, Cytyc, Boxborough, Mass), and formalin-fixed cytologic specimens were deemed insufficiently reliable for clinical use compared with FISH results on formalin-fixed, paraffin-embedded tissue sections in a study by Beatty et al.30 The highest discordance rate in HER2 status was noted in ethanol-fixed FNA samples. Our results confirm that alternative fixatives not only affect FISH results but also may interfere with HER2 IHC analysis.

Significantly more tumors were stained positive (3+) for HER2 by IHC on CB preparations than on NC or TB preparations. On the basis of our data, we conclude that IHC for HER2 on CB tends to overestimate the rate of HER2 equivocal and positive (2+, 3+) tumors, whereas TB preparations rather underestimate the frequency of HER2 positivity when NC is used as the gold standard. These conclusions are also reflected by the greater number of HER2-negative tumors by IHC on TB preparations than on NC preparations (22 vs 19). In our study, the correlation between NC and TB IHC HER2 results was not as good as reported by others, who calculated a concordance rate between 86.6% and 96%.15-17

An optimal CB sample for the evaluation of HRS in BC should be comparable to a biopsy, including adequate cellularity to reliably differentiate negative versus positive (>10% of tumor cells) nuclear staining for ER and PR and to assess the completeness of membranous staining for HER2. Some authors have suggested obtaining a “normalized IHC score” to decrease the false-positive HER2 rate on IHC.31 This method is based on the difference in HER2 staining intensity in non-neoplastic breast tissue and in tumor, which can than increase the concordance rate between IHC and FISH results. Results from a study by Gown et al32 suggest that this method may be of value in alcohol-fixed tumor samples. In addition, the presence of an internal control is helpful if there is concern about false-negative results. However, internal control (non-neoplastic) breast tissue usually is absent in cytology preparations. Overall, there are many confounding factors (bloody sample, paucicellular specimen, use of alternative fixative, sampling problems) that can alter antigen expression/detection and, thus, can influence IHC and FISH results on cytologic preparations. These limitations have to be addressed and kept in mind before ER, PR, and HER2 results from BC FNA are applied for treatment decisions.

In summary, the use of an ideal fixative for FNA material and the development of optimal processing conditions to preserve antigenicity of tumor cells are essential for the accurate determination of marker status in BC on CB preparations. Following the previously well established guidelines is crucial, and all new techniques, including alternative fixatives, should be validated before they are used for diagnostic purposes. Therefore, a comparative study using formalin-fixed CB preparations would be beneficial to evaluate whether the discordance rates noted in the current study can be eliminated and may be attributed solely to ethanol fixation. Although the concordance rate of CB ER results with NC and TB is high, there is significant discrepancy noted when PR and HER2 are evaluated on CB samples. Despite the small number of cases included in this study, our data indicate that correlation of CB results with corresponding formalin-fixed tissue samples is essential to accurately identify all patients who would benefit from hormone therapy.

Acknowledgements

We thank Gladwyn Leiman, MBBCh, and Sue Ellen Martin, MD, PhD for their suggestions.

Conflict of Interest Disclosures

The authors made no disclosures.

Ancillary