Context.—In 2007 the American Society of Clinical Oncology/College of American Pathologists made new recommendations for HER2 testing and redefined HER2 positivity.
Objective.—To analyze results from simultaneous HER2 testing with immunohistochemistry and fluorescence in situ hybridization (FISH) in 2590 invasive breast carcinomas between 2002 and 2010, using 2 scoring systems.
Design.—Cases from between 2002 and 2006 were scored by using original US Food and Drug Administration criteria (N = 1138) and those from between 2007 and 2010 were evaluated according to American Society of Clinical Oncology/College of American Pathologists criteria (N = 1452). Concordance between testing methods and clinicopathologic associations were determined.
Results.—Overall concordance between immunohistochemistry/FISH in the 9-year period was 96.2% (κ = 0.82), and positive concordance was lower. After 2007, the proportion of HER2/neu-positive and HER2/neu-negative cases was not significantly changed when using immunohistochemistry (10.5% versus 8.9%, P = .22 and 69.4% versus 63%, P = .13, respectively), but the number of equivocal cases was higher (19.9% versus 28%, P < .001). While the proportion of negative cases by FISH remained unchanged after 2007 (86.5% versus 88.2%, P = .76), the number of positive cases was lower (13.4% versus 9.2%, P < .001). In addition, 38 cases (2.6%) were FISH equivocal, 16 of which were also equivocal by immunohistochemistry. Overall, immunohistochemistry/FISH concordance was 95.9% between 2002 and 2006 (κ = 0.82) and 96.4% after 2007 (κ = 0.82). However, an approximately 13% lower positive assay concordance was noted in the last period.
Conclusions.—Application of American Society of Clinical Oncology/College of American Pathologists recommendations is associated with comparable overall immunohistochemistry/FISH concordance, reduced positive concordance, and increased equivocal results.
Overexpression of the epidermal growth factor receptor 2 (HER2/neu) is observed in 10% to 20% of human breast carcinomas, is associated with poor prognosis, and occurs mostly through HER2/neu oncogene amplification.1–4 HER2/neu overexpression predicts the tumor sensitivity to targeted therapy with the anti-HER2/neu humanized monoclonal antibody trastuzumab both in the adjuvant and metastatic settings.5,6 Therefore, accurate assessment of HER2 status is critical to recommend targeted therapy only to patients who are most likely to benefit and to reduce potential side effects. Current available methods to assess the HER2 status in tumor cells include chromogen-based immunohistochemistry (IHC), Western blot, enzyme-linked immunosorbent assay, reverse transcription polymerase chain reaction, the ratio of HER2/neu/CEN17 DNA probe–based in situ hybridization (ISH) using diverse visualization reagents (eg, fluorophores [FISH], chromogens, or silver preparations), and real-time polymerase chain reaction. However, the only approved and recommended protocols for clinical use include IHC and ISH–based methods.3,7 The original US Food and Drug Administration (FDA)–approved interpretation criteria for HER2 testing were used for patient enrollment into prospective randomized adjuvant trials of trastuzumab and considered positive a score of IHC 3+, defined as complete intense membrane staining of more than 10% of tumor cells, and a HER2/neu gene to chromosome 17 centromeric region (HER2/neu/CEN17) ratio of at least 2.0 by FISH.7,8 To improve the accuracy and concordance between IHC and FISH and to help standardize the assay, the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) panel made new recommendations in 2007 for HER2 testing including modifications in specimen handling, laboratory quality control/validation, and HER2 IHC and FISH interpretation criteria.3 The proposed criteria redefined the IHC 3+ score by increasing the threshold to more than 30% of membrane-positive invasive tumor cells and the HER2/neu/CEN17 FISH ratio above 2.2. In addition, the ASCO/CAP criteria introduced a “new” FISH equivocal category including those cases with HER2/neu/CEN17 scores between 1.8 and 2.2, thus raising the possibility of patients ending up without enough information for treatment recommendation (eg, “double” IHC/FISH equivocal cases). The impact of the changed interpretation criteria has been a matter of recent interest and controversy.7–9 At Yale School of Medicine, New Haven, Connecticut, a decision was made in 2002 to do both IHC and FISH on every case. Recently, the data have been retrieved into a secondary database to allow retrospective assessment of the assays in a routine clinical basis at a tertiary care center. Here we report the results obtained from simultaneous HER2 testing using IHC and FISH in a retrospective cohort including 9 consecutive years from a single academic institution, and we assess the impact of the change in the HER2 status interpretation criteria.
MATERIALS AND METHODS
Case Selection and Data Storage
We included 2590 patients with breast cancer, comprising all breast cancer patient specimens seen at the Yale Surgical Pathology suite and assessed for HER2 by both FISH and IHC between 2002 and 2010. All tests were performed in the routine clinical setting and recorded as part of the medical record. After collection, information was filtered, anonymized, and introduced into a protected database for further storage and processing by the investigators. Data collection was done under the approval of Yale Human Investigation Committee Protocol No. 9500008219.
Immunohistochemistry
Immunohistochemistry analysis was performed by staining of 4- to 5-μm sections by using the HercepTest kit (Dako Corp, Carpinteria, California) according to manufacturer's instructions and as previously described.4,8–10 Appropriate positive and negative controls were used for each run. Specimens were fixed in 10% buffered formalin for 2 to 24 hours. HER2 was scored on the scale of 0 to 3 (0, 1+, 2+ and 3+). Between 2002 and 2006 cases were scored by using FDA criteria: 3+, strong, complete membrane staining in more than 10% of the malignant cells; 2+, weak to moderate complete membrane staining in more than 10% of the malignant cells; 0 or 1+, no or fewer than 10% of cells staining. Cases between 2007 and 2010 were evaluated according to the ASCO/CAP guideline recommendations. In this system, a positive 3+ HER2/neu reflects a uniform, intense membrane staining of more than 30% of invasive tumor cells.
Fluorescence In Situ Hybridization
Fluorescence in situ hybridization was performed according to manufacturer's instructions by using freshly cut 5-μm paraffin sections and the Vysis LSI HER-2/neu kit (Vysis PathVysion, Abbott Laboratories, Abbott Park, Illinois) as described.4,8,10 Centromeric probes for chromosomes 17 (CEN17) were visualized by using Spectrum Green and locus-specific HER2/neu gene (HER2) probes were detected by using Spectrum Orange fluorescence. All specimens were read in our molecular pathology laboratory. Results were scored as the ratio of HER2/neu to CEN17 signal numbers. Before 2007 specimens were considered amplified if the ratio of HER2/neu to CEN17 signals was higher than 2.0 according to the FDA score. After 2007 gene amplification was evaluated according to the ASCO/CAP guideline recommendations by using the ratio of greater than 2.2. Cases with HER2/neu/CEN17 ratios between 1.8 and 2.2 were considered as equivocal FISH score.
Statistical Analysis
Some data sets are presented as means ± standard error (standard error of the mean). Means for each group were compared by using 1-way analysis of variance, followed by Bonferroni multiple comparison post test. Differences were considered significant at P < .05. Statistics were performed by using the Graph Pad Prism v4 (2003; GraphPad Software, Inc, La Jolla, California) software. Concordance for positive and negative results was calculated according to the previously described formula.10 The kappa agreement coefficients (κ) and 2-tailed χ2 test were calculated by using the online GraphPad software QuickCalcs tool (http://graphpad.com/quickcalcs/kappa1.cfm; accessed May to December 2012).
RESULTS
Of 2590 cases evaluated in the 9-year period, 1708 (65.9%) were reported to be HER2/neu negative by IHC (865 cases were scored as 0, 843 as 1+), 633 (24.4%) as equivocal (score 2+), and 249 (9.6%) as positive (score 3+), whereas FISH results in the same group included 2266 (87.5%) negative cases, 38 (1.5%) equivocal cases, and 286 (11%) breast carcinomas with HER2/neu amplification (Table 1). Of cases recorded as equivocal through IHC, 554 (87.5%) showed a negative FISH result, 16 cases (2.5%) had an equivocal HER2/neu/CEN17 score, and 63 (10%) were positive. In addition, 73 cases (2.8%) presented with discordant IHC/FISH results, including 47 with positive IHC and negative FISH results, and 26 cases with negative IHC and a positive FISH result. The overall concordance between nonequivocal IHC and FISH was 96.2% (κ = 0.82). However, concordance for positive results was lower than for negative scores (79.1% and 97.5%, respectively; Table 1).
HER2 Status With Immunohistochemistry and Fluorescence In Situ Hybridization (FISH) in 2590 Casesa

As expected, the average FISH HER2/neu to CEN17 ratios were comparably low and close to a value of 1 in the groups with negative HER2/neu IHC (score 0, 1.11 ± 0.26, and 1+, 1.16 ± 0.55, P = .90, Figure). Cases showing IHC 2+ showed higher FISH scores than cases with negative IHC (1.47 ± 1.18, P < .001), but they were prominently lower than in the 3+ group (4.63 ± 2.8, P < .001, Figure). The dispersion of FISH scores was higher in the IHC 3+ group. The lowest recorded FISH ratio was 0.3 and the highest was 16.2. The correlation coefficient between the FISH scores and the IHC levels was 0.83, suggesting a close relationship between both variables in the studied samples.
The fluorescence in situ hybridization (FISH) ratios are higher in cases with increased membrane HER2/neu immunoreactivity. This figure shows the distribution of FISH ratio scores in 2590 breast biopsy specimens according to their immunohistochemistry (IHC) HER2/neu score. FISH scores were expressed as absolute numbers representing the ratio of HER2/neu to CEN17 signals, counted in at least 20 tumor nuclei. The IHC HER2/neu levels were expressed as 0, 1+, 2+, and 3+ according to the relative amount of tumor cells showing complete membrane HER2/neu staining. The number of cases is indicated within each column and the dashed black line indicates the average value in each category. ***P < .001. Abbreviation: ns, not significant.
The fluorescence in situ hybridization (FISH) ratios are higher in cases with increased membrane HER2/neu immunoreactivity. This figure shows the distribution of FISH ratio scores in 2590 breast biopsy specimens according to their immunohistochemistry (IHC) HER2/neu score. FISH scores were expressed as absolute numbers representing the ratio of HER2/neu to CEN17 signals, counted in at least 20 tumor nuclei. The IHC HER2/neu levels were expressed as 0, 1+, 2+, and 3+ according to the relative amount of tumor cells showing complete membrane HER2/neu staining. The number of cases is indicated within each column and the dashed black line indicates the average value in each category. ***P < .001. Abbreviation: ns, not significant.
Most cases with discordant IHC/FISH results (94.5%) corresponded to ductal carcinomas, most of them of moderate and high histologic grade (Tables 2 and 3). Of all discordant cases, 8 (11%) showed lobular features, 5 (6.8%) showed apocrine differentiation, 20 (27%) were without detectable hormone receptor (eg, estrogen receptor/progesterone receptor) expression, and 27 cases (37%) showed altered CEN17 signal counts, 21 of which showed increased signals. In addition, 2 cases with demonstrated BRCA mutations and 3 carcinomas with metaplastic components were found. The proportion of cases with altered CEN17 signal counts was approximately 2-fold higher in the group with positive IHC and negative FISH results (21 of 47 cases, 44%, Table 3) than in cases with negative IHC and positive FISH results (6 of 26 cases, 23%, Table 2). Of note, the former group included only samples with increased CEN17 signal counts, and the latter included 2 cases with a single CEN17 signal.
Clinicopathologic and Molecular Characteristics of Breast Carcinomas With Negative HER2/neu Immunohistochemistry (IHC) and Positive Fluorescence In Situ Hybridization (FISH) Results

Clinicopathologic and Molecular Characteristics of Breast Carcinomas With Positive HER2/neu Immunohistochemistry (IHC) and Negative Fluorescence In Situ Hybridization (FISH) Results

In the 2002–2006 period, using the previous FDA 10% criteria for HER2/neu IHC positivity and dichotomous FISH category, 791 of a total 1138 cases (69.5%) were negative by IHC; 227 (19.9%) equivocal; and 120 (10.5%) positive. Fluorescence in situ hybridization results in the same group included 985 negative cases (86.6%) and 153 positive cases (13.4%) (Table 4). The overall concordance between nonequivocal IHC and FISH in this group was 95.9% (κ = 0.82). Assay concordance for negative results was 97.5% and for positive scores, 85.8% (Table 4).
HER2 Status With Immunohistochemistry and Fluorescence In Situ Hybridization (FISH) Between 2002 and 2006a

In the 2007–2010 period, using both the 3-tiered FISH classification and the 30% threshold for HER2/neu IHC positivity, a comparable proportion of cases tested with IHC were scored as HER2 negative (917 of 1452, 63.1%, P = .13) and positive (129 of 1452, 8.9%; P = .22). However, a higher proportion of cases showed equivocal IHC result (406 of 1452, 28%; P < .001, Table 5).
HER2 Status With Immunohistochemistry and Fluorescence In Situ Hybridization (FISH) Between 2007 and 2010a

Fluorescence in situ hybridization results in the same group showed a comparable proportion of negative cases (1281 of 1452, 88.2%; P = .76), a reduction in positive cases (133 of 1452, 9.2%; P = .002), and 38 (2.6%) equivocal cases. The overall IHC/FISH concordance between 2007 and 2010 was 96.4% (κ = 0.82) and the negative assay concordance remained at 97.5%. However, concordance for positive results was 72.9% in this period (Table 5). Of discordant cases, those with negative IHC and positive FISH results were considerably fewer after 2007 (20 of 1138 versus 6 of 1452), and cases with positive IHC and negative FISH results were increased (17 of 1138 versus 30 of 1452). Only 16 cases showed equivocal results both by IHC and FISH and their main clinicopathologic and molecular features are depicted in Table 6. As shown, patients with double IHC/FISH equivocal results averaged 59 years (range, 41–95 years), included ductal carcinomas of predominant moderate and high histologic grade, and had an average estrogen receptor and progesterone receptor expression of 75% and 37%, respectively. Notably, 8 cases (50%) showed altered CEN17 counts, revealing structural chromosome 17 alterations by FISH examination, 7 of which were with increased signals.
COMMENT
The overall HER2 positivity using IHC and FISH in the studied cohort was approximately 10% and IHC/FISH concordance was 96.2%, findings that are in agreement with previous reports.3,10–16 However, concordance for positive results was considerably lower than for negative scores as shown in other studies.10–16 The latter might be associated with diverse factors preferentially affecting positive cases or differentially affecting the accuracy of IHC and FISH to detect HER2 positivity, including inappropriate sample quality (eg, time of fixation and to fixation), assay-specific limitations (eg, limited assay sensitivity/specificity), interpretative considerations (eg, overestimation/underestimation of signal location/intensity), and additional biological factors (eg, cases with complex genetic background or with true protein overexpression without gene amplification).
Although prolonged fixation is known to reduce the immunoreactivity of many antibodies in paraffin sections, relatively long fixation periods (>96 hours) do not significantly affect HER2/neu IHC results.17,18 A similar lack of deleterious effect of extended fixation time (up to 7 days) has been reported for dual-probe FISH when using the PathVysion kit.18,19 However, increased time to fixation to more than 3 to 4 hours (with and without sample refrigeration) considerably reduces HER2/neu IHC signal.20 In addition, Khoury et al21 used a collection of 10 cases to show that dual-probe FISH was significantly compromised after 2 hours of cold ischemic time, and this effect was more pronounced in FISH than IHC. Moreover, the HER2 signal was more sensitive to cold ischemic time than CEN17, thus being a potential factor for case underscoring using the HER2/neu/CEN17 probe ratio. In support of this notion, a recent report including 84 samples on tissue microarray format showed a significant degradation of FISH signal intensity when cold ischemic time exceeded 3 hours.22
In addition to the effect of preanalytic variables, previously described limitations in the sensitivity of the approved HercepTest antibody used in this study might also explain the lower positive concordance between IHC and FISH identified in our data set.23,24
Inaccurate assay interpretation might also affect the IHC/FISH concordance, and only a few studies have directly addressed its effect, possibly owing to the difficulties in measuring differences in the subjective manual/visual HER2/neu IHC scoring. In a retrospective collection of 697 cases between 2004 and 2007, Grimm et al10 found the major reason for lack of positive concordance to be inappropriate interpretation of HER2/neu IHC staining (11 of 13 IHC positive/FISH negative cases).
Increased CEN17 copy number has been reported to occur in a variable proportion of breast tumors, depending on definition criteria, and to be a prominent cause of potential IHC/FISH discrepancy.14,16,25,26 One study, including 687 samples between 2000 and 2002, identified CEN17 coamplification (defined as ≥3 CEN17 signals) in 47% (9 of 19) of discordant cases and only 9.1% (54 of 592) of IHC/FISH concordant cases.16 All discordant cases in this study were IHC positive (3+), lacking gene amplification, and only 4% (3 of 67) of nonamplified cases with moderate HER2/neu IHC expression (2+) showed increased CEN17 signal counts. In addition, the number of CEN17 signals was 1.6-fold higher in the former group, suggesting that CEN17 coamplification was associated with strong protein expression without HER2/neu gene amplification.16 In support of these findings, we found a relatively high proportion (37%) of cases with CEN17 signal alterations among discordant cases, most of them showing increased CEN17 counts. Moreover, the proportion of cases with increased CEN17 signals was approximately 2-fold higher in cases with positive IHC and negative FISH results than in other discordant cases. Additional studies with rigorous evaluation of CEN17 signals and structural chromosome 17 alterations will be required to clarify this issue.
The fraction of positive cases was slightly higher with FISH than IHC (11% versus 9.6%), suggesting that using only IHC might overlook some patients with HER2/neu amplification. Alternatively, cases with positive FISH and negative IHC findings might reflect true false-positive FISH results or HER2/neu amplification without protein overexpression. Previously discussed methodologic issues may also account for at least part of these discordances. In addition, the utilization of FISH reduced the fraction of equivocal cases from 24.4% with IHC to 1.5% in the same population, indicating that FISH is more accurate in segregating positive from negative cases. However, we note that there are not follow-up data on this cohort, and as such, any relationship between IHC/FISH HER2 status and response to therapy is unknown.
Cases showing higher HER2/neu membranous immunoreactivity show higher FISH scores, confirming the close relationship between HER2/neu gene amplification and protein overexpression in breast carcinomas. The larger dispersion of FISH scores in the IHC-positive group is likely due to the marked variation in the number of HER2/neu copies between tumors (ranging from ∼2 to >15 copies). Despite showing some degree of membrane immunoreactivity, most IHC 2+ cases (88%) showed absence of HER2/neu amplification, as also communicated by others.3,10–16
Utilization of 2007 ASCO/CAP criteria for HER2 testing was associated with comparable overall IHC/FISH concordance, unchanged negative concordance, but also to a reduction in positive assay concordance of approximately 13%. One possible explanation for the latter finding is the more stringent criteria for positivity introduced in the ASCO/CAP guidelines. In support of this, we observed a significant (4.2%) reduction in FISH-positive cases after 2007 as previously communicated.9,14 As expected by the use of more stringent criteria for IHC positivity, a mild (1.6%, nonstatistically significant) reduction in IHC-positive cases and a significant 8.1% increase in IHC equivocal cases was also found after 2007 (Tables 4 and 5).
A recent study comparing cases before and after 2007 ASCO/CAP guidelines also found comparable overall IHC/FISH concordance, lower positive concordance, and reduction in HER2-positive cases using ASCO/CAP scoring criteria.14 However, the latter study also reported lack of significant variation in the amount of HER2/neu IHC equivocal cases (2+). This difference might be explained at least in part by the relatively lower number of cases included in the ASCO/CAP group of this study (421 versus 1016 in the FDA group), variations in the subjective estimation of the amount of HER2-positive cells among pathologists, and more importantly, to differences in the studied populations. Indeed, the authors of the study reported a prominent shift in the population with a cumulative 7% reduction of HER2-positive cases along their 8-year period.14 Such a shift was not observed in our cohort (not shown). In addition, the FISH scoring in the study from Vergara-Lluri et al14 did not consider the FISH equivocal category in the analysis (eg, no amplification was defined only as HER2/neu/CEP17 ratio < 2.2), thus eliminating the ASCO/CAP FISH equivocal status. Moreover, our data show that only 16 of a total 38 FISH equivocal cases (42%) were in the IHC equivocal group (Table 5). To our knowledge, the present report is the largest study addressing the impact of the 2007 ASCO/CAP HER2 guidelines.
It is important to acknowledge that in addition to changes in the IHC and FISH scoring criteria, modifications in specimen handling and laboratory practice proposed by 2007 ASCO/CAP recommendations might also account for some of the observed differences between the 2 studied periods. In addition, we cannot exclude the possible influence of unexplored variations in the study populations and/or in the assay readers in our findings.
Our results are also in agreement with those from Middleton et al27 who reported only a marginal (0.5%) increase in the overall IHC/FISH concordance after application of 2007 guidelines. In addition, the authors27 found a prominent reduction in cases that could not be interpreted by FISH after 2007 (from 10.8% to 3.8%), thus supporting a positive impact of ASCO/CAP guideline specimen-handling recommendations.
Finally, application of the new criteria also defined a small (16 cases) but meaningful group of patients whose findings fall into the IHC equivocal/FISH equivocal category and whose prognosis and management is currently not well defined. For instance, patients with FISH scores equal or above 2 are being treated as having positive findings. Interestingly, 50% of cases with equivocal IHC and FISH results showed altered CEN17 signal counts, suggesting chromosome 17 alterations. While this is not a large number of cases (0.6%), and may be resolved by retesting or other methods, it reveals a small subset that may represent another biological variant of HER2-driven breast cancer.
References
Author notes
The authors have no relevant financial interest in the products or companies described in this article.
Competing Interests
No potential conflicts of interest were disclosed by the authors.