OUP user menu

Performance of Chromogenic In Situ Hybridization on Testing HER2 Status in Breast Carcinomas With Chromosome 17 Polysomy and Equivocal (2+) HercepTest Results
A Study of Two Institutions Using the Conventional and New ASCO/CAP Scoring Criteria

Yun Gong MD, William Sweet MD, Yi-Jing Duh MD, Larry Greenfield MD, Emily Tarco, Smita Trivedi, W. Fraser Symmans MD, Jorma Isola MD, Nour Sneige MD
DOI: http://dx.doi.org/10.1309/AJCP4M2VUZCLDALN 228-236 First published online: 1 August 2009

Abstract

This study specifically addressed the performance of chromogenic in situ hybridization (CISH) on HER2 testing in 66 breast carcinomas with chromosome 17 polysomy and 49 carcinomas with an equivocal HercepTest (DakoCytomation, Carpinteria, CA) score by comparing CISH with corresponding FISH results at 2 test sites and evaluating intersite agreement of CISH results. For tumors with chromosome 17 polysomy, when using the manufacturers’ criteria, the concordance values between CISH and FISH at site A, site B, and intersite CISH agreement were 95.8%, 95.5%, and 93.5%, respectively; when using the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) criteria, the values were 100.0%, 100.0%, and 100.0%, respectively. For tumors with an equivocal HercepTest score, when using the manufacturers’ criteria, the concordance values between the 2 methods at site A, site B, and intersite CISH agreement were 88.2%, 95.1%, and 91.1%, respectively; when using the ASCO/CAP criteria, the values were 96.7%, 97.3%, and 97.4%, respectively. These results indicate that CISH is reliable for testing these 2 types of tumors, especially when the ASCO/CAP criteria are used.

Key Words:
  • HER2
  • Breast carcinoma
  • Fluorescence in situ hybridization
  • FISH
  • Chromogenic in situ hybridization
  • CISH
  • American Society of Clinical Oncology/College of American Pathologists
  • ASCO/CAP
  • Chromosome 17 polysomy
  • Equivocal
  • Immunohistochemistry

Identification of human epidermal growth factor receptor (HER2) status is a prerequisite for eligibility for anti-HER2 therapies such as trastuzumab (Herceptin) and lapatinib in patients with breast cancer. Positive HER2 status (ie, overexpression of the HER2 protein or amplification of the HER2 gene) is a strong predictor of response to trastuzumab.17 Trastuzumab therapy has shown an impressive improvement in patients’ clinical outcome in metastatic and adjuvant settings,13,5,7,8 as well as in the neoadjuvant setting.9,10 However, the significant benefits of trastuzumab are coupled with potential cardiotoxic effects and high cost.1,7,1113 Therefore, accurate, yet easily available, HER2 testing is essential to effectively and promptly identify patients who might benefit from anti-HER2 targeted therapy while sparing patients who would not.

Immunohistochemical analysis and dual-probe fluorescence in situ hybridization (FISH) are the 2 assays commonly used to test HER2 status in breast cancer specimens. FISH assesses HER2 gene amplification and is more specific and reliable than immunohistochemical analysis7,1416; however, FISH is relatively expensive, time-consuming, and labor-intensive compared with immunohistochemical analysis. Usually, testing algorithms for HER2 start with immunohistochemical analysis as a first-line method to detect protein overexpression, with FISH being performed for tumors with equivocal results. Chromogenic in situ hybridization (CISH) is a relatively new method and, like FISH, evaluates HER2 gene amplification. In addition, CISH is similar to immunohistochemical analysis in that it allows detection of HER2 amplification using a conventional peroxidase reaction and enumeration of gene copy number with simultaneous histologic examination by regular bright-field microscopy.17,18 Consequently, CISH has been viewed by some proponents as a potential alternative to FISH in the HER2 testing algorithm or simply as a primary testing method.19

The Zymed SPOT-Light HER2 CISH kit (Invitrogen, Camarillo, CA) has been recently approved by the US Food and Drug Administration (FDA). However, unlike FISH (PathVysion kit, Vysis, Downers Grove, IL), which uses dual probes, CISH usually defines HER2 amplification based on a mean absolute HER2 gene copy number per nucleus, without chromosome 17 correction. Thus, there is a concern that CISH might overestimate the presence of gene amplification as a result of chromosome 17 polysomy (polysomy 17).18,20 On the other hand, as a confirmatory test, the performance of CISH for cases with an equivocal immunohistochemical score, which might be associated with polysomy 17,2125 is also of interest.

The American Society of Clinical Oncology and the College of American Pathologists (ASCO/CAP) recently published guidelines to standardize specimen processing, testing, and scoring in an attempt to improve HER2 testing accuracy in routine clinical practice.26,27 A major component of the new guidelines was the change in the scoring criteria used to define positive and negative HER2 status. For example, 3 categories (ie, positive, equivocal, and negative), instead of the dichotomous categories established by manufacturers, are used to document the HER2 status determined by in situ hybridization methods. In addition, the cutoffs for determining positive and negative status also differ from conventional ones established by manufacturers Table 1. The ASCO/CAP guidelines mandate that a new assay should show 95% or more concordance with another validated test for positive and negative results before routine application. Results that fall into the equivocal category require further confirmation using another validated test.

A number of previous studies have demonstrated good overall concordance between CISH and FISH results and reproducibility, but in these studies, only the manufacturers’ scoring criteria were used.17,2835 We recently conducted a multicenter study on 226 consecutive breast carcinomas using the new ASCO/CAP scoring criteria and observed excellent concordance between CISH and FISH results at different test sites and very high (>98%) intersite reproducibility of CISH results.36 However, because generally a vast majority of tumors have clear-cut HER2 status, unselected tumor cohorts have limited biostatistical power to demonstrate the reliability of CISH in cases with potential borderline status. To the best of our knowledge, there have been no reports on the reliability of the CISH method for tumors with polysomy 17 and tumors with an equivocal (2+) immunohistochemical score. The present study specifically focused on these 2 types of tumors identified at either or both test sites and evaluated the impact of these cases on the HER2 status determined by CISH when using the ASCO/CAP scoring criteria.

Materials and Methods

Paraffin tissues from 286 invasive breast carcinomas were retrospectively obtained from 2 institutions. Of these, 226 were consecutive cases of surgically resected breast carcinomas: 110 from The University of Texas M.D. Anderson Cancer Center, Houston (site A); and 116 from the Institute of Medical Technology, University of Tampere, Tampere, Finland (site B). To obtain a sufficient number of tumors qualified for this study, a set of 60 tumors, known to show an immunohistochemical score of 2+ (monoclonal antibody, clone AB8, dilution 1:100; NeoMarkers, Fremont, CA) determined during routine patient care at site A, was supplemented.

All paraffin blocks were less than 2 years old. Unstained sections of the 286 tumors were sent simultaneously to the 2 test sites for HER2 testing. At each site, HER2 status of the 286 tumors was determined by using the polymer detection kit for CISH analysis (Zymed SPOT-Light HER2 CISH) and the PathVysion kit for FISH analysis. We performed the tests according to the manufacturers’ instructions, using procedures similar to what we reported previously, with appropriate control samples included.30,35,37 At each test site, immunohistochemical analysis was performed using the FDA-approved HercepTest (DakoCytomation, Carpinteria, CA) on the 60 supplementary tumors and on its own consecutive tumors (ie, 110 from site A and 116 from site B).

View this table:
Table 1

The results generated at each test were interpreted by pathologists at that site who were blinded to the results of HER2 status (previous status or the status obtained during the present study). The manufacturers’ scoring criteria and the ASCO/CAP recommendations were used to score CISH and FISH results (Table 1). To standardize the interpretation criteria for HER2 status as determined by the different methods, each interpreter was instructed with a protocol and a PowerPoint (Microsoft, Redmond, WA) tutorial provided by Invitrogen.

In the CISH analysis, at least 30 invasive tumor cells from each sample were scored for average HER2 gene copy number using a light microscope under a 40× dry objective. If the average copy number was 4 to 6 per nucleus (equivocal), another 30 invasive tumor cells were enumerated, and the final average copy number of the case was calculated from the total of 60 cells.

In the FISH analysis, at least 20 invasive tumor cells from each sample were scored for nuclear HER2 and chromosome 17 centromere probe (CEP17) signals using an epifluorescence microscope. If the average HER2/CEP17 ratio was 1.8 to 2.2 (equivocal), another 20 invasive tumor cells were scored, and the final ratio of the case was calculated from the total of 40 cells. Polysomy 17 was defined as the presence of an average of 3 or more CEP17 signals per tumor nucleus. The tumors showing polysomy 17, identified from the 286 cases at either or both sites, were selected for this study to test the impact on CISH results.

In HercepTest analysis, HER2 expression was interpreted based on the new ASCO/CAP recommendations and scored as 0 (no staining), 1+ (weak and incomplete membrane staining), 2+ (strong, complete membrane staining in ≤30% of tumor cells or weak/moderate heterogeneous complete membrane staining in ≥10% of tumor cells), or 3+ (strong, complete, homogeneous membrane staining in >30% of tumor cells). HER2 status was considered negative or not overexpressed if the immunohistochemical score was 0 or 1+, equivocal if the score was 2+, and positive if the score was 3+. Cases showing an equivocal (2+) immunohistochemical score with the HercepTest were selected for this study to test the impact on CISH results.

Image 1

An invasive breast carcinoma showing chromosome 17 polysomy with fluorescence in situ hybridization (×1,000).

The reliability of the CISH method in tumors showing polysomy 17 and tumors with an equivocal HercepTest score was assessed by comparing CISH results with the corresponding FISH results at site A and site B and by evaluating the reproducibility of CISH results between the 2 sites. Concordance and the Cohen κ coefficient, with the corresponding 95% confidence interval (CI), were used to evaluate intermethod and intersite agreement. The relationship between the κ value and the level of agreement was proposed by Landis and Koch38: 0.00 to 0.20, slight agreement; 0.21 to 0.40, fair agreement; 0.41 to 0.60, moderate agreement; 0.61 to 0.80, substantial agreement; and 0.81 to 1.00, almost perfect agreement.

This study was conducted with the approval of the institutional review boards at the M.D. Anderson Cancer Center and the University of Tampere.

Results

Cases With Chromosome 17 Polysomy

At site A, FISH results were available for 273 (95.5%) of the 286 tumors, and 51 (18.7%) of them showed polysomy 17. At site B, FISH results were available on 278 tumors (97.2%), and 22 (7.9%) of them showed polysomy 17 Image 1. The total number of tumors showing polysomy 17 at either or both sites was 66. Concordance between CISH and FISH at each site (51 cases at site A and 22 at site B) and intersite reproducibility of CISH on the 66 tumors is shown in Table 2.

When the manufacturers’ scoring criteria were used, the concordance between CISH and FISH was 95.8% at site A and 95.5% at site B, and the agreement of HER2 status determined by CISH between the 2 sites was 93.5%.

When the ASCO/CAP guidelines were used for scoring, the HER2 gene was found by FISH to be amplified in 12 (23.5%) of 51 tumors at site A and in 12 (54.5%) of 22 tumors at site B. When the 3-category criteria (amplified, equivocal, and nonamplified) were used for comparison, the concordance between CISH and FISH was 91.7% at site A and 95.5% at site B, and intersite reproducibility of CISH was 90.3% Image 2. Discrepancies between CISH and FISH results were found in 5 cases: 4 at site A and 1 at site B Table 3. The corresponding HercepTest score was 2+ Image 3 in 4 and 1+ in 1. Intersite disagreement of CISH results was found in 6 tumors Table 4. The corresponding HercepTest score was 0 in 1 tumor, 2+ in 3, and 3+ in 2. Notably, all discrepancies were seen between equivocal and positive status or between equivocal and negative status, not between positive and negative status.

View this table:
Table 2
Image 2

Chromogenic in situ hybridization test showing negative/nonamplified HER2 status (A) and positive/amplified HER2 status (B) in invasive breast carcinomas (×100).

Image 3

Immunohistochemical staining (HercepTest) for HER2 showing an equivocal (2+) score in an invasive breast carcinoma (×400).

View this table:
Table 3

When the 2-category criteria defined in the ASCO/CAP guidelines for amplified and nonamplified status were used for comparison (ie, excluding cases with equivocal status determined by CISH or FISH), the concordance between the 2 methods for positive and negative HER2 status was 100% at both sites, exceeding the minimal concordance required by the ASCO/CAP guidelines (95%) before a new test is applied routinely in the diagnostic setting. Intersite agreement on CISH was 100.0% (Table 2).

Cases With an Equivocal (2+) Score by the HercepTest

Of the 226 consecutively obtained tumors, the HercepTest was performed at each site on that site’s own specimens (110 at site A and 116 at site B), which identified a total of 21 tumors (9.3%) with a score of 2+. For the 60 supplementary tumors that had an immunohistochemical score of 2+, determined during routine patient care at site A using a monoclonal antibody from NeoMarkers, immunohistochemical analysis was reperformed using the HercepTest at both sites. A score of 2+ by the HercepTest was found in 15 (25.0%) of the 60 tumors at site A and 22 (36.7%) at site B. In this study, to evaluate the concordance between CISH and FISH on tumors with a 2+ HercepTest score, we used 36 tumors (21 from consecutive cases plus 15 from supplementary cases) at site A and 43 tumors (21 from consecutive cases plus 22 from supplementary cases) at site B. To evaluate the reproducibility of CISH results between the 2 test sites, tumors that demonstrated a 2+ HercepTest score at either or both sites were selected, and a total of 49 tumors were analyzed Table 5.

View this table:
Table 4

When using the manufacturers’ scoring criteria, the concordance between CISH and FISH was 88.2% at site A and 95.1% at site B, and the reproducibility of HER2 status as determined by CISH between the 2 sites was 91.1% (Table 5).

When the ASCO/CAP guidelines were used for scoring, the HER2 gene was found by FISH to be amplified in 5 (13.9%) of 36 tumors at site A and 5 (11.6%) of 43 tumors at site B. When using the 3-category scoring criteria for comparison, the concordance between CISH and FISH results was 85.3% at site A and 87.8% at site B, and intersite reproducibility of CISH was 86.7% (Table 5). Discrepancies between CISH and FISH results were found in 10 cases, 5 at each site Table 6. Of the 10 cases, 8 showed a discrepancy between equivocal status by one method and positive or negative status by another method, including 1 tumor that showed the same discrepant pattern at both sites (case 5 at each site). Discrepancy between negative and positive status was found only in 1 tumor, and the pattern of the discordance at both sites (case 1 at each site) was identical: CISH– and FISH+. Of the 10 cases, polysomy 17 (determined by FISH) was found in 2 tumors (case 4 at site A and case 2 at site B), in which CISH was equivocal and FISH was negative. Of the 6 cases with discrepant CISH results between the 2 sites, 5 showed discrepancy between equivocal status at one site and positive or negative status at the other site, and 1 was between negative and positive status Table 7.

View this table:
Table 5

When the 2-category criteria defined in the ASCO/CAP guidelines for the positive and negative status were used for comparison, the concordance between CISH and FISH was 96.7% at site A and 97.3% at site B, fulfilling the concordance mandated by the ASCO/CAP guidelines. Also, intersite reproducibility on the CISH results was 97.4% (Table 5).

Discussion

Currently, the dual-probe FISH is a validated, and well-accepted technique to evaluate HER2 status, which defines the status by determining the HER2 gene copies per chromosome 17 centromere because polysomy 17 has an effect on the HER2 copy number. Polysomy 17 is quite common in breast carcinomas, has an important role in breast cancer pathophysiology, and is an indicator of poor prognosis.39 However, there is no unified definition for polysomy 17. For example, some studies defined polysomy 17 when the CEP17 copy number per tumor nucleus was more than 2.25,4043 2.1 or more,22 or more than 1.85.44 We and many others14,21,24,45,46 defined polysomy 17 using a cutoff of 3 or more CEP17 copies per tumor nucleus. Depending on the criteria used to define polysomy 17 and selection criteria of the cases included in the series, the reported incidence of polysomy 17 varied from 5% to 50%, mostly 8% to 20%.2124,4043,4648 In our study, the frequency of such cases was 18.7% at site A and 7.9% at site B.

Previous studies have shown that CISH is a reliable method for HER2 testing, either as an alternative to FISH to confirm equivocal immunohistochemical results or as a frontline testing method in breast cancer.17,19,2833,36 However, a major concern remains regarding how numeric alteration of chromosome 17 affects HER2 status because CISH determines HER2 status based on the absolute number of HER2 gene copies rather than the chromosome 17–corrected HER2 gene copy number.14 Performance of CISH on testing these cases using the ASCO/CAP scoring criteria has not been previously documented.

Some of the previous studies demonstrated that correction for chromosome 17 is critical for the determination of true HER2 status and for avoiding overestimating cases that have increased HER2 gene copy number resulting from polysomy 17,7,14,40,42,47 whereas others believe that polysomy 17 alone has a minor impact on the HER2 gene copy number and HER2 protein overexpression.43,49 These studies scored HER2 status using manufacturers’ criteria. In the present study, we showed that although discordance between CISH and FISH was found in a minority of tumors at each test site using the manufacturers’ scoring criteria, complete concordance for amplified and nonamplified results was observed in tumors with polysomy 17 using the ASCO/CAP 2-category criteria. Moreover, CISH results from the 2 test sites were also completely reproducible (Table 2). This is likely because the cutoff value for positive status in the new ASCO/CAP guideline is more than 6, whereas the HER2 gene copy in nonamplified tumors showing polysomy 17 is usually 4 to 6.22,24,42

Polysomy 17 has been reported to be closely associated with equivocal (2+) or low levels of HER2 overexpression by immunohistochemical analysis in breast cancer.2125,39 In a study by Kostopoulou et al,25 43% of the equivocal (2+) cases were found to be polysomy 17. The equivocal immunohistochemical score was thought to be related to increased gene dosage by polysomy 17 in some cases. Therefore, in the present study, we examined tumors with an equivocal HercepTest score. Tumors with a 2+ score were found in 9.3% of 226 consecutive tumors. In an attempt to “enrich” this type of case, we additionally included 60 tumors that were shown to be equivocal on immunohistochemical scoring during routine patient care. Similar to results in tumors with polysomy 17, tumors with an equivocal HercepTest score showed higher concordance than required by the new ASCO/CAP guidelines for amplified and nonamplified results and also high reproducibility of CISH results between test sites. The concordance and reproducibility were suboptimal when the manufactures’ scoring criteria and ASCO/CAP 3-category criteria were used.

View this table:
Table 6
View this table:
Table 7

It is noteworthy that the reproducibility of equivocal immunohistochemical results was low when a different antibody was used on the same set of tumors. HercepTest, an FDA-approved method, was developed to overcome variability associated with other immunohistochemical methods and antibodies, and is supposed to be a more standardized immunohistochemical method.50 Of the 60 tumors that had a 2+ score using the NeoMarkers antibody (AB8), only 25.0% at site A and 36.7% at site B yielded a 2+ score by the HercepTest. The poor reproducibility of 2+ immunohistochemical cases is well known. For example, Gown et al51 observed similar findings when the results from 2 antibodies (A0485 rabbit polyclonal antibody and SP3 rabbit monoclonal antibody) were compared. Presumably, variation in antibody specificity, staining technique, and interobserver interpretation are contributing factors.

In keeping with the findings of others,2225,48,52 tumors with polysomy 17 or an equivocal HercepTest score in this study tended to be HER2 nonamplified by FISH. For example, of the tumors with polysomy 17, 23.5% at site A and 54.5% at site B were amplified. Likewise, of the tumors with an equivocal HercepTest score, 13.9% at site A and 11.6% at site B were amplified. The differences in HER2 status by FISH between sites may, in part, result from variation in technical factors and/or interobserver scoring variability. In a recent study by Press and coworkers,53 FISH was performed in 2 experienced sites: an academic reference/research laboratory and a large, high-volume commercial reference laboratory; however, concordance of HER2 status between the 2 laboratories was only 89%. This type of variation could also explain the difference in the incidence of polysomy 17 detected between the 2 test sites in our study (18.7% vs 7.9%).

Discordance between CISH and FISH when the ASCO/CAP 3-category criteria were used for comparison was less than 9% among tumors with polysomy 17 (Table 2) and less than 15% among tumors with an equivocal HercepTest score (Table 5). It is interesting that most of these tumors, except 1, showed “low-level” discordance (ie, between equivocal and positive or negative status, rather than between positive and negative status), rendering a high concordance when the ASCO/CAP 2-category criteria were used. For the only tumor that showed discordant HER2 status, a result of CISH– and FISH+ was observed at both sites, and the FISH results at each site were slightly above the positive cutoff (2.38 and 2.52, respectively; Table 6). Of the 10 discordant cases in tumors with an equivocal HercepTest score, polysomy 17 was found only in 2 cases, both of which were equivocal by CISH and negative by FISH (Table 6). We did not identify any tumor with polysomy 17 that was CISH+ and FISH–, indicating that polysomy 17–related change in the HER2 copy number does not seem to be a major factor responsible for the discordance. Therefore, correction for chromosome 17 seems to be rarely needed for the determination of true HER2 gene status using the ASCO/CAP guidelines.

To identify the underlying reasons responsible for the discordance, we examined the role of intratumoral heterogeneity because the difference in the foci selected for scoring in these cases may contribute to different HER2 status. However, we did not identify meaningful intratumoral heterogeneity in these discrepant cases. We believe that the presence of low-level discordance is probably attributable to the fact that the scoring criteria were inevitably arbitrary to some extent and may not have reflected the real status accurately in individual cases.44 Recently, using array comparative genomic hybridization or a FISH mapping technique, researchers in several groups demonstrated that the whole chromosome 17 polysomy is extremely uncommon in breast cancer. Gains of CEP17 copies are more often caused by aneusomy of 17q and/or amplification of CEP17. Therefore, increases in the copy number of CEP17 do not reliably represent chromosome 17 polysomy.54,55 Consequently, HER2 status determined on the basis of the HER2/CEP17 ratio may lead to false-negative HER2 status by FISH.56 This may also have some role in the discordant results between CISH and FISH in some cases in the present study.

By using tumors with polysomy 17 and with an equivocal HercepTest score identified at either or both test sites, we found high concordance between CISH and FISH results, especially when ASCO/CAP scoring criteria were used. The concordance values exceeded what is required by the ASCO/CAP guidelines for positive and negative results at both test sites. In addition, excellent intersite reproducibility of CISH results was found in these cases. These findings indicate that, when using the new ASCO/CAP scoring criteria, CISH can reliably determine positive or negative HER2 status of breast carcinomas. Apparently, if CISH is used as a primary HER2 testing method, any results falling into the equivocal range should be further confirmed by another validated method.

Acknowledgments

We thank scientific editor Sue Moreau, Department of Scientific Publications, M.D. Anderson Cancer Center, for editorial support.

Footnotes

  • Dr Gong received an honorarium as a speaker to present the data at an educational seminar organized by Biocare Medical at the 2009 United States and Canadian Academy of Pathology annual meeting; Drs Sneige and Isola received research funds from Invitrogen for this study; Drs Sweet and Greenfield are employees of Invitrogen; Dr Duh is a statistical consultant for Invitrogen; and Dr Symmans is a scientific advisor for Invitrogen.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
View Abstract