HER2/neu status in breast cancer is determined by immunohistochemical analysis and/or fluorescence in situ hybridization (FISH). Previous studies have found widely varying sensitivities and specificities for anti-HER2/neu antibodies, including recently developed rabbit monoclonal antibodies. The current prospective study compared rabbit monoclonal antibody SP3 and rabbit polyclonal antibody A0485 immunostaining on routinely processed consecutive cases of breast carcinoma.
Of 1,610 cases tested, 261 (16.2%) equivocal (2+) cases were evaluated by FISH. Of 253 cases equivocal with A0485 results, 125 (49.4%) were negative with SP3. In 22 (8.7%) of 253 cases equivocal with A0485, there was amplification by FISH, and 3 of these cases were SP3– (0/1+). Of the 20 (14.8%) of 135 SP3-equivocal cases amplified by FISH, 1 case was A0485–.
The reported false-negative rate with A0485 is 2.8%, and the American Society of Clinical Oncology/College of American Pathologists guidelines recommend a rate of less than 5%. Compared with A0485, the false-negative rate with SP3 is only 0.3% (3/1,156) higher, but it shows about a 50% reduction in equivocal scores, reducing the need for reflex FISH testing.
Immunohistochemical evaluation of HER2/neu receptor protein overexpression in breast carcinomas became routine by the late 1990s, following the demonstration that it predicted response to adjuvant HER2/neu-targeted therapy with trastuzumab.1 The early validation of immunohistochemical techniques sought to establish thresholds for staining that correlated with clear response to trastuzumab therapy, and staining patterns were graded by completeness of membrane staining, intensity of staining, and percentage of tumor cells staining. In these early studies, the presence of more than 10% of tumor cells showing strong, complete membranous (“chicken-wire”) staining in immunohistochemical studies was considered positive; this was revised to more than 30% following panel review by American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP).2 Such cases show a high correlation with gene amplification by fluorescence in situ hybridization (FISH).3,4 Cases with less than 10% complete membranous staining and weak staining (1+) show no response to trastuzumab.5 For cases with intermediate levels of staining (2+), correlation with the presence of HER2/neu gene amplification is recommended to determine potential benefit from targeted therapy.2 Detailed ASCO/CAP guidelines for HER2/neu testing in breast cancer were published in 2007.2 The 2007 ASCO/CAP guidelines also provided minimum acceptable false-negative rates (<5%) and false-positive rates for immunohistochemical analysis (∼10%) and recommended additional FISH testing of all equivocal cases (2+).
The 2 antibodies most frequently used in the validation studies were A0485 (DAKO, Carpinteria, CA, in the HercepTest), a rabbit polyclonal antihuman antibody directed at an epitope on the cytoplasmic portion of the receptor, and the CB11 mouse monoclonal antibody (MAb) (Novocastra, Newcastle upon Tyne, England), which also detects an epitope on the internal portion of the receptor. The largest published series correlating immunohistochemical results with FISH results are by Yaziji et al3 on more than 2,900 carcinomas with reported sensitivity and specificity of A0485 of 92.6% and 98.8%, respectively.
Polyclonal anti-HER2/neu antibodies often show weak background staining of the normal breast epithelium (should always be <1+).3 In the last few years, rabbit MAbs have become increasingly popular because they rarely show background staining and typically have higher affinities for the desired epitope, which translates into more intense positive staining and lower background staining.6,7 The rabbit MAb SP3 is one such antibody, although its affinity for HER2/neu has not been reported. Recent studies report conflicting data on the sensitivity and specificity of SP3 when compared directly with A0485, the HercepTest, and CB11 Table 1.8–12 In these studies, the sensitivity of SP3 for detecting HER2/neu amplification in breast carcinomas ranged from 52% to 100% and the specificity ranged from 88% to 99%. Of the 4 published studies examining SP3 expression, 2 used tissue microarrays, 1 used cytology cell blocks, and 1 used whole tissue sections.8–11
Given the relatively high false-positive rate for A0485, we hypothesized that SP3 might decrease the numbers of cases requiring reflex FISH testing. The aim of this prospective study was to compare MAb SP3 with A0485 in consecutive cases of breast carcinoma analyzed for routine clinical care during a 24-month period at Brigham and Women’s Hospital, Boston, MA, between May 2008 and May 2010. For all cases with equivocal (2+) immunohistochemical results with SP3 or A0485, FISH was performed.
Materials and Methods
All surgical specimens from Brigham and Women’s Hospital were received fresh and then routinely formalin fixed and paraffin embedded. Specimen fixation time was at least 6 hours for all cases. For consultation cases, whether the specimens were frozen before formalin fixation and paraffin embedding was recorded.
We immunostained 4-μm-thick sections according to manufacturer recommendations. In detail, slides were baked at 37°C overnight, then deparaffinized and rehydrated (100% xylene 4 times for 3 minutes each, 100% ethanol 4 times for 3 minutes each, and running water for 5 minutes). Endogenous peroxidase activity was blocked with 3% hydrogen peroxide in methanol for 10 minutes and washed under running water for 5 minutes. Heat-induced epitope retrieval was performed in 10 mmol/L citrate buffer (pH 6.0) with a pressure cooker (Biocare Medical, Concord, CA) at 122°C to between 14 and 17 psi with the cycle lasting, on average, 45 minutes and the cool-down period approximately 20 minutes. Immunohistochemical analysis was performed on an automated instrument (DAKO Autostainer Plus).
A range of titers was tested for both antibodies, and titers were calibrated using positive control staining of HER2/neu-amplified tumors and negative staining of normal breast epithelium. Primary antibodies SP3 (dilution 1:100; Labvision, Fremont, CA) and A0485 (dilution 1:500-1:750 according to control staining; DAKO) were incubated for 40 minutes at room temperature, followed by detection with the EnVision Plus System (DAKO) for 30 minutes in a humid chamber at room temperature. Sections were developed using 3,3′-diaminobenzidine (Sigma, St Louis, MO) as the substrate and counterstained with Mayer hematoxylin. External positive control tests were also run. The studies comparing antibodies were performed on the same day for a given case.
HER2/neu was scored by 2 of us (E.M. and J.E.B.) according to the ASCO/CAP guidelines.2 In brief, scoring was as follows: negative (0), no immunoreactivity or immunoreactivity in fewer than 10% of tumor cells; negative (1+), faint, weak immunoreactivity in more than 10% of tumor cells but only a portion of the membrane was positive; equivocal (2+), weak to moderate complete membrane immunoreactivity in more than 10% of tumor cells; and positive (3+), moderate to strong complete membrane immunoreactivity in more than 30% of tumor cells. All immunohistochemical scores were read blinded to the FISH results. Discrepancies between the 2 reviewers in scoring were resolved at a multiheaded microscope.
FISH was performed by Clarient (Aliso Viejo, CA) using the Pathvysion HER-2 DNA probe kit (Abbott Laboratories, Abbott Park, IL). FISH was scored according to ASCO/CAP guidelines.2 FISH results are reported as the average HER2/neu count to average CEP (chromosome enumeration probe) 17 count and use the total HER2/neu count and total CEP17 count to calculate the final ratio. In brief, the average number of HER2 signals per nucleus and the average number of CEP17 chromosome probes per nucleus were counted in 20 nonoverlapping cells, with counts made in at least 2 different areas. A HER2/CEP17 ratio of more than 2.2 was reported as amplified and a ratio of less than 1.8 as not amplified. An additional 20 cells were counted if the ratio was in the equivocal range (1.8–2.2). The same block of tissue used for immunohistochemical studies was used for subsequent FISH studies.
Immunohistochemical analysis for HER2/neu was performed on 1,610 consecutive cases of breast carcinoma as part of routine clinical care: about 85% were primary invasive carcinomas (of which about 64% were core biopsy and 36% were excision specimens), and approximately 15% were metastases (of which about 42% were core biopsy, 34% were excision, and 24% were fine-needle aspiration cell block specimens). Of the 1,610 cases using A0485, 1,156 (71.8%) were negative (0/1+), 253 (15.7%) were equivocal (2+), and 201 (12.5%) were positive (3+) for HER2/neu. The 261 cases with equivocal (2+) staining were sent for FISH, and 253 (96.9%) were equivocal by A0485 and 135 (51.7%) were equivocal by SP3. Table 2 shows a summary of the HER2/neu immunohistochemical results using SP3 and A0485 correlated with FISH results.
Of the 253 equivocal (2+) cases with A0485, 22 (8.7%) had a FISH ratio of 2.0 or more (18 cases) or a HER2/neu copy number of more than 6 (4 cases) and were considered eligible for HER2/neu-targeted therapy. By using the ASCO/CAP guidelines for evaluation of HER2/neu gene amplification, 13 (5.1%) showed amplification by the FISH HER2/CEP17 ratio (>2.2), 7 (2.8%) were equivocal (ratio, 1.8–2.2), and 233 (92.1%) were not amplified (ratio, <1.8).
Of 135 SP3-equivocal cases, 20 (14.8%) had a FISH ratio of 2.0 or more (17 cases) or a HER2/neu copy number more than 6 (3 cases). By using the ASCO/CAP guidelines, 12 (8.9%) showed amplification by the FISH ratio, 8 (5.9%) were equivocal, and 115 (85.2%) were not amplified.
Of the 253 A0485-equivocal cases, 125 (49.4%) were negative (0/1+) with SP3. The 3 A0485-equivocal cases that were SP3– had a FISH ratio of 2.0 or more (2 cases with ratios of 4.2/1.9 = 2.23 and 4.5/1.8 = 2.56) or a HER2/neu copy number of more than 6 (1 case with a HER2 copy number 6.8 but a ratio of less than 1.8) Image 1D, Image 1E, and Image 1F. By using the ASCO/CAP guidelines, 1 additional case was FISH-equivocal (ratio, 4.3/2.2 = 1.97). In addition to the A0485-equivocal cases outlined, 1 case among the positive (3+) A0485 cases was identified that was SP3–. FISH studies confirmed amplification (ratio, 6.4/1.8 = 3.53) Image 1G, Image 1H, and Image 1I.
Of the 135 SP3-equivocal cases, 8 (5.9%) were A0485–. One SP3-equivocal, A0485– case had a FISH ratio of 2.0 or more (ratio, 12/1.8 = 6.8) Image 1A, Image 1B, and Image 1C. No SP3-equivocal, A0485– case had a HER2/neu copy number of more than 6 without showing amplification by the ratio as well. By the ASCO/CAP guidelines, 2 additional SP3-equivocal, A0485– cases were FISH-equivocal (ratio 3.9/2.2 = 1.8 and 4.9/2.6 = 1.8). No cases were identified that were SP3+ (3+) and A0485–.
In this prospective comparison of immunohistochemical studies with SP3 and A0485, 125 (49.4%) A0485-equivocal (2+) cases were SP3– (0/1+). In 8.7% (22/253) of our A0485-equivocal cases, there was a FISH ratio of 2.0 or more (18 cases) or a HER2/neu copy number of more than 6 but a HER2/CEP17 ratio of less than 1.8 (4 cases). In 14.8% (20/135) of the SP3-equivocal cases, there was a FISH ratio of 2.0 or more (17 cases) or a HER2/neu copy number of more than 6 and a ratio of less than 1.8 (3 cases). Overall, SP3 failed to detect 4 amplified cases that A0485 detected as 2+ (3 cases) or 3+ (1 case), and, conversely, SP3 detected 1 amplified case as 2+ that was A0485–.
The ASCO/CAP guidelines have a category for an equivocal FISH ratio (1.8–2.2) reflecting the error in FISH analysis; however, in clinical practice, patients with a ratio between 2.0 and 2.2 often receive HER2-targeted therapy and can be enrolled in several but not all current clinical trials for HER2-targeted therapy. Patients with a HER2 copy number of more than 6.0 but without amplification by the ratio (1.6% of A0485-equivocal cases in our study) are eligible for HER2-targeted therapy by virtue of their HER2 copy number when single-probe FISH or chromogenic in situ hybridization is used, although they may not receive it when dual-probe FISH or chromogenic in situ hybridization corrects the HER2 copy number for the centromere CEP17 copy number intended to correct for chromosome 17 polysomy. Polysomy of chromosome 17 seems not to confer the same biologic sensitivity to HER2-targeted therapy as true HER2 gene amplification; thus, polysomic cases are not reported as HER2-amplified.13 The presence of more than 3 copies of CEP17 has been used as a surrogate marker of polysomy, and the ASCO/CAP currently recommends the correction for chromosome 17 polysomy by reporting the ratio of HER2 to CEP17.2,14
Most cases of polysomy result in 4 to 6 copies of HER2 and are estimated to be present in at least 8% of breast cancers overall and are enriched in the HER2-equivocal immunohistochemical group.15 In our study, about 16% of the A0485-equivocal cases and 17% of the SP3-equivocal cases had 4 to 6 copies of HER2 but were not amplified when corrected for CEP17. A recent study reported that an increased CEP17 copy number is due to gains or amplification of the CEP17 region in the majority of cases and in only 6% of cases tested (1/18) reflects true chromosome 17 polysomy.16 By correcting for the CEP17 copy number in the absence of true polysomy, a small number of truly HER2-amplified tumors will be reported as not amplified based on the ratio (seen in ∼6% of cases in the study reported).16 There is currently no simple way to identify this small subgroup of patients who might benefit from targeted therapy but whose tumors are currently classified as not amplified, but patients with HER2 copy numbers of 6 or more can be considered for HER2-targeted therapy when there is evidence of membrane protein overexpression.2 Unfortunately, we do not know if the “SP3-equivocal, FISH not-amplified” cases are biologically different from the “A0485-equivocal, SP3– FISH not-amplified” cases or whether higher-affinity SP3 antibody binding with a cleaner stain is reflected.
Compared with the largest previous study (Yaziji et al3) using A0485 with a very similar staining protocol, we report a lower rate of amplified cases (8.7% compared with 17% of 2+ cases amplified in the 2004 study by Yaziji et al3). However, the total proportions of negative, equivocal, and positive results at our institute for all invasive carcinomas tested are within the range recommended by ASCO/CAP based on the data generated by Yaziji et al.3 The recommended range is 70% to 75% negative, 15% equivocal, and 10% to 15% HER2/neu+, and our proportions for all carcinomas tested by A0485 in the study period were 71.8% negative, 15.7% equivocal, and 12.5% positive.3 When using SP3, these proportions become 79.2% negative, 8.4% equivocal, and 12.4% positive.
Higher rates of amplification in SP3 2+ cases have been reported in other studies (Table 1); there are several possible explanations for these different results, such as the method of testing (tissue microarray as opposed to whole sections), the threshold for considering a case amplified (ratio of 2.0 vs 2.2 vs 50% of cells with >5 HER2/neu copies), and different antibody staining techniques, including dilution and antigen retrieval. In addition, although guidelines are clear for the definition of 1+, 2+, and 3+ in reporting HER2 immunohistochemical results, there is likely some interobserver variability when placing borderline cases in each category, which could contribute to the apparent differences in performance of A0485 and SP3 across studies.
In our study, there was a trend toward greater specificity and less sensitivity for HER2 amplification with SP3 compared with A0485, but true sensitivity and specificity can only be calculated with the false-negative rate and false-positive rate (this would have required sending all negative immunohistochemical cases and positive immunohistochemical cases for FISH). The reported false-negative rate for A0485 by Yaziji et al3 was 2.8%, and a false-negative rate of less than 5% is recommended by ASCO/CAP guidelines, although a number closer to 0 is clearly desirable.
We do not know the false-negative rate in the 1,148 cases that were negative for both A0485 and SP3 in our study of 1,610 cases tested, but we found that when using SP3 compared with A0485, only 0.3% more women would receive a false-negative result (3 missed cases of 1,156 A0485– cases). Neither antibody is known to correlate better with response to current HER2-targeted therapies, although A0485 has the advantage of having been used in many of the early studies. The SP3 antibody seems to be a reasonable alternative to A0485 because it adds only 0.3% to the false-negative rate of A0485 and significantly reduces, by about 50%, the number of equivocal cases reflexively sent for FISH testing. In health care settings where resources are limited, the use of SP3 instead of A0485 may provide significant cost savings with a small but acceptable decrease in sensitivity compared with A0485.
Upon completion of this activity you will be able to:
define the breast cancer proportions of HER2/neu negative, equivocal, and positive results by immunohistochemistry recommended by American Society of Clinical Oncology/College of American Pathologists guidelines.
estimate the proportion of HER2/neu equivocal cases that are amplified by fluorescence in situ hybridization testing.
discuss the possible reasons why reported HER2/neu immunohistochemical results for a tumor sample might vary between laboratories and thus across studies.
list the possible reasons why a rabbit monoclonal antibody such as SP3 may reduce by 50% the number of equivocal results for HER2/neu immunohistochemistry compared with rabbit polyclonal A0485 antibody.
The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.
The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.
Questions appear on p 974. Exam is located at www.ascp.org/ajcpcme.