Fluorescence in situ hybridization (FISH) is the gold standard for assessing HER-2 status for breast cancers, and paraffin-embedded tissue sections are used routinely for HER-2 FISH. Cytologic samples also are used, but to the authors' knowledge, little is known regarding the reliability of these samples. The objective of this study was to elucidate the usefulness of cytologic specimens for HER-2 FISH testing.
Histologic and cytologic specimens from 32 patients with invasive ductal carcinoma of the breast were subjected to comparative analysis of HER-2 status by FISH. FISH was performed by using a PathVysion HER-2 DNA Probe Kit according the manufacturer's instructions. The signal ratios of chromosome enumeration probe 17 (CEP17) and HER-2 were estimated and compared. In 15 cytologic specimens, the distance between signals (HER-2 and CEP17) and the nearest nuclear membrane were measured by using 3-dimensional image analysis and confocal microscopy.
Cytologic and histologic FISH results were compared. Signal ratios of HER-2/CEP17 were lower in cytologic specimens from 26 of 32 patients compared with the histologic material. Three-dimensional image analysis demonstrated that the distance between the CEP17 signal and the nuclear membrane was shorter than the distance between the HER-2 gene and the nuclear membrane.
Breast cancers with HER-2 gene amplification respond to either monotherapy or combined therapy with trastuzumab.1, 2 HER-2 testing on histologic sections has been performed widely for select patients who may benefit from the anti-HER-2 therapy, trastuzumab (Herceptin; Hoffman-La Roche, Ltd., Basel, Switzerland; Genentech, Inc., South San Francisco, Calif). For HER-2 testing, immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) have been used in most laboratories. FISH generally is considered better than IHC assays.3, 4 FISH is performed by using formalin-fixed, paraffin-embedded (FFPE) tissue blocks, and it is the gold standard method for determining HER-2 gene amplification status.5 Cytologic specimens have a potential advantage for evaluating HER-2 status, especially in patients with tumor samples that are difficult to obtain. If tumor samples are difficult to obtain from metastatic sites, such as the brain or lung, then tumor cells in pleural effusions or cerebrospinal fluid often are easily available. Only a few studies have attempted to use cytologic specimens for FISH to evaluate HER-2 gene amplification.6–13
The objective of the current study was to determine the usefulness of cytologic specimens for HER-2 FISH testing. To clarify the cause of the lower signal ratio, we also analyzed intranuclear localizations of HER-2 and chromosome enumeration probe 17 (CEP17) signals to determine the cause of lower signal ratios in cytologic specimens compared with histologic specimens.
MATERIALS AND METHODS
Thirty-two patients with breast cancer (histologically invasive ductal carcinoma) were selected from the Division of Diagnostic Pathology, Tokai University Hospital. IHC examination for HER-2 was performed by using a polyclonal antibody (HercepTest; DakoCytomation, Carpinteria, Calif), and the results were evaluated according to the recommended criteria. Of 32 tumors, 3 tumors had scores of 0, 6 tumors had scores of 1+, 18 tumors had scores 2+, and 5 tumors had scores of 3+. Freshly imprinted cytology slides from these tumors were fixed in ethanol until they were used. Routinely processed, 4-μm-thick FFPE sections and the imprint smears were evaluated by FISH on the histologic tissue sections and on the cytologic imprint smears from the same patients.
HER-2/neu Status Determination by FISH
FISH was performed by using a PathVysion HER2 DNA Probe Kit (Vysis Inc., Downers Grove, Ill) according to the manufacturer's instructions. The PathVysion HER2 DNA Probe Kit contains an HER-2 probe (Spectrum Orange) and a chromosome enumeration probe (CEP17; Spectrum Green) for the centromere region of chromosome 17. Histologic and cytologic specimens were treated according to the Paraffin Pretreatment Reagent Kit (Vysis Inc.) protocol before hybridization. Briefly, the slides were pretreated in 0.2 N hydrochloric acid for 20 minutes at room temperature, followed by pretreatment solution at 80 °C for 30 minutes, and protease solution for 10 minutes at 37 °C (pretreatment and protease solution were provided in the Paraffin Pretreatment Reagent Kit). The slides were fixed in neutral buffered formalin for 10 minutes. Then, the slides were denatured by immersing slides in denaturation solution (formamide 35 mL; 20 × saline sodium citrate buffer 5 mL; and distilled water 10 mL [pH 7.3]) for 5 minutes at 72 °C, followed by probe preparation and hybridization (12–16 hours) at 37 °C in a humidified chamber. After hybridization, the slides were immersed in posthybridization washes (2 × saline sodium citrate buffer 100 mL; NP-40 0.3 mL), then the signal was enumerated. Finally, the slides were counterstained with 4,6-diaminido-2-phenylindole dihydrochloride (DAPI) in antifade solution (DAPI I Counterstain; Vysis Inc.), and the coverslip was applied.
HER-2 and CEP17 Signal Scoring
The hybridized slides were viewed at ×1000 magnification under an Olympus BX51 fluorescence microscope (Olympus, Tokyo, Japan) using a triple band-pass filter for simultaneous detection of Spectrum Orange, Spectrum Green, and DAPI. Signals were enumerated for 60 tumor cells within an area of invasive carcinoma and were scored for both HER-2 and CEP17 numbers. FISH results were evaluated according to the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines, and HER-2 gene amplification was defined as positive with HER-2/CEP17 ratios >2.2, equivocal with ratios from 1.8 to 2.2, and negative with ratios <1.8.5
The ploidy status of chromosome 17 was determined by calculating the average number of copies of CEP17 based on a count of 60 cell nuclei. Chromosome 17 polysomy was defined as >3 CEP17 signals per nucleus from the cytologic specimen.
Measurement of Intranuclear Locations of HER-2 and CEP17 Signals
In 15 cytologic specimens, the distances between signals (HER-2 and CEP17) and the nearest nuclear membrane were measured by using a confocal microscope (FV1000; Olympus) and 3-dimensional image-analysis software Volocity (Improvision, Coventry, UK). These specimens were selected consecutively from the cases that were used to perform FISH on histologic and cytologic preparations. The distances that were measured for cancer cells included 3 fields under a ×100 objective lens for each case.
The nonparametric Mann-Whitney U test was used to evaluate statistical differences between groups.
Concordance Between Cytologic and Histologic Specimens
The HER-2/CEP17 ratios determined by FISH on histologic and cytologic specimens from 32 cases are listed in Table 1. All of the histologic and cytologic specimens were suitable for FISH analysis. Representative cases of FISH on histologic and cytologic specimens are shown in Figure 1 (highly amplified tumor) and Figure 2 (nonamplified tumor). Except for Cases 23 and 29, FISH results were in concordance between histology and cytology. The results of FISH for Case 23 with an IHC score of 2+ and Case 29 with an IHC score of 1+ were interpreted as equivocal according to the ASCO/CAP guidelines. The FISH results were equivocal on histology and negative on cytology. All 5 cases with IHC scores of 3+ and 1 case with an IHC score of 2+ demonstrated HER-2 gene amplification by FISH on both cytologic and histologic sections.
Table 1. Signal Ratio (HER-2/Chromosome Enumeration Probe 17) Examined on Histologic and Cytologic Specimens of 32 Cases
Cytologic specimen FISH HER-2/CEP17
FISH indicates fluorescence in situ hybridization; HER-2; CEP17, chromosome enumeration probe 17.
The distances between signals (HER-2 and CEP17) and the nearest nuclear membrane were measured by using a confocal microscope and 3-dimenisonal image analysis.
Comparison of CEP17 Signal and HER-2/CEP17 Between Histologic and Cytologic Specimens
We demonstrated lower signal ratios in cytologic specimens and compared them with the histologic specimens. The average signal count of chromosome 17 centromere on cytologic specimens was 2.6 ± 0.8 (mean ± standard deviation), which was significantly greater than the count on histologic specimens (2 ± 0.5; P = .0005) (Fig. 3). Average signal counts of the HER-2 gene on histologic and cytologic specimens were 4.2 ± 3.9 and 4.6 ± 3.7 (P nonsignificant), respectively.
Twelve of 32 tumors demonstrated chromosome 17 polysomy (≥3 CEP17 signals per cell on cytologic specimens). Moreover, the average signal count of centromere 17 on histologic specimens was 2.3 ± 0.5 for 12 cases with polysomy, which was significantly greater than 1.8 ± 0.5 for the other 20 cases without polysomy (P = .0194).
Twenty-six of 32 tumors demonstrated a lower signal ratio in cytologic specimens compared with histologic specimens (P = .0249) (Fig. 4). It is interesting to note that, in 26 nonamplified tumors, the signal ratio on cytologic specimens was significantly lower than on histologic specimens (P = .0029) (Fig. 5).
Distance Between CEP17 and HER-2 Signals and Nuclear Membrane by 3-Dimensional Image Analysis
To analyze the cause of the discrepancy between histologic and cytologic specimens, 3-dimensional image analysis was performed on 15 nonamplified tumors (Fig. 6). The average distance between nearest nuclear membrane and CEP17 signals in all 15 nonamplified tumors on cytologic specimens was significantly shorter than the distance between nearest nuclear membrane and HER-2 signals (P < .0001) (Fig. 7). Eight of 15 specimens demonstrated a shorter distance between the nearest nuclear membrane from CEP17 than the distance from the HER-2 signals (Fig. 8).
The current results indicated that HER-2 FISH testing in cytologic specimens potentially was more accurate than using FFPE samples. To our knowledge, the current study is the first to demonstrate that the intranuclear localization of the centeromere contributes to the lower HER-2/CEP17 ratio in cytologic specimens on 3-dimensional image analysis. Beatty et al. evaluated HER-2 status by IHC and FISH using 51fine-needle aspiration (FNA) specimens together with the corresponding histologic specimens. Those authors reported that HER-2 protein expression on cytologic preparations was insufficiently reliable for clinical use; however, HER-2 gene amplification determined by FISH demonstrated a highly consistent correlation with the HER-2 status of tissue samples.7 The majority of other studies have supported FISH analysis on cytologic specimens as a reliable procedure for HER-2/neu assessment.6, 8–12 The results of the current study, in which both tissues and cytologic imprint smears were used from identical cases, also supports the usefulness of cytologic specimens. These findings indicate that cytologic specimens are a useful and representative source of materials for the detection and quantification of HER-2 gene amplification by FISH.
We also analyzed the basic mechanisms for signal ratios of HER-2/CEP17 in cytologic specimens, because those ratios were lower than the ratios in histologic specimens in most tumors. Cytologic specimens contain whole nuclei of the cancer cells, whereas only part of the nuclei could be tested on 4-μm histologic tissue sections. It was assumed that there was no significant difference between cytologic and histologic specimens, because both centromere signals and HER-2 signals would be cut on histologic sections. However, considering the lower signal ratio on cytologic specimens, we hypothesized that these differences were caused by the differing intranuclear localization of the centromere and the HER-2 gene. Our study is the first to our knowledge documenting the finding that the distance between CEP17 signals and the nearest nuclear membrane is significantly shorter than the distance between HER-2 and its nuclear membrane. Thus, CEP17 may be lost readily by sectioning, because CEP17 is adjacent to the nuclear membrane (Fig. 9). To further compare the same group of cancer cells, additional FISH performed on histologic sections from pellets (FFPE cell blocks) remains to be investigated.
The discrepancy of CEP17 signal number between histologic and cytologic specimens raises concern about polysomy criteria. The cutoff criteria in chromosome 17 polysomy for truncated nuclei on paraffin sections reportedly differed in each study group14–20 and, thus, is controversial. According to the current results, cytologic specimens are applicable for estimating CEP17 polysomy, whereas we should be more discreet when examining histologic sections. Another issue of concern is comparing HER-2 gene amplification in FFPE histologic sections, as proposed by ASCO/CAP, with that in the cytologic sections. The population of tumor cells categorized with equivocal HER-2 gene amplification (HER-2/CEP17 ratio between 1.8 and 2.2) generally is small. In the current study, there was 1 patient with a histologic specimen in this category who had a negative cytologic specimen. It should be determined further how cytologic specimens contribute to the accurate estimation of HER-2 gene amplification for clinical use. We demonstrated that HER-2 FISH in cytologic specimens potentially is more accurate than FISH in FFPE histologic specimens.
The disadvantages of cytologic specimens for FISH are, first, distinguishing between invasive areas versus noninvasive areas, and, second, tumor cell heterogeneity. Beatty et al. pointed out that, in 1 case of invasive ductal carcinoma, overexpression of HER-2 protein was not observed in the adjacent invasive lobular carcinoma, suggesting the presence of 2 different clones. In this situation, the results of FISH analysis in FNA and histologic specimens would be discordant with gene amplification results shown only on the FNA specimens. There was no such case in our current study.
In conclusion, cytologic specimens preferentially help us obtain accurate numbers of CEP17 and the HER-2 gene. Centromere 17 may be lost readily by sectioning in histologic specimens, because CEP17 is located adjacent to the nuclear membrane. These results indicate that cytologic specimens are more applicable for HER-2 FISH analysis and make us conscious of the differences when trying to accurately estimate HER-2/CEP17 and CEP17 polysomy by using cytologic specimens.