shnitt and jacobs ajcp 2000
TRANSCRIPT
-
8/3/2019 Shnitt and Jacobs AJCP 2000
1/5
AJCP / COMMENTARY
806 Am J Clin Pathol 2001;116:806-810 American Society of Clinical Pathologists
Current Status of HER2 Testing
Caught Between a Rock and a Hard Place
Stuart J. Schnitt, MD, and Timothy W. Jacobs, MD
The analysis of breast cancer specimens for alterations in
the HER2/neu (c-erb-B2) gene or its protein product has
become common practice in surgical pathology laboratories.
Although some clinicians use the information derived from
these assays for assessing patient prognosis and evaluating
the likelihood of response to various chemotherapeutic agents
and to tamoxifen, the major clinical role of assessment of
HER2 status is to help determine patient suitability for treat-
ment with trastuzumab (Herceptin), a monoclonal antibody
targeted to the HER2 protein.1,2
Although there are a variety of methods available to
assess HER2 status in clinical breast cancer specimens,
assessment of protein overexpression using immunohisto-
chemical analysis and evaluation of gene amplification
using fluorescence in situ hybridization (FISH) are the
methods most commonly used in clinical practice today.3,4
There are currently 4 commercially available, US Food
and Drug Administrationapproved assays for deter-
mining HER2 status in breast cancer samples, although
these have not all been approved for the same purpose
Table 1. Among these 4 assays, 2 are immunohisto-
chemical assays (HercepTest, DAKO, Carpinteria, CA; and
Pathway, Ventana Medical Systems, Tucson, AZ) and 2 are
FISH assays (PathVysion, Vysis, Downers Grove, IL; and
Inform, Ventana). All of these assays can be performed
using automated instrumentation. In addition, there are
numerous commercially available anti-HER2 primary anti-
bodies and immunohistochemical detection systems for
HER2 testing for which Food and Drug Administration
approval has not been sought or received.
While the clinical value of assessing HER2 status in
human breast cancers to help determine suitability for
trastuzumab therapy is indisputable, the manner in which
the test should be performed and the results evaluated and
reported is a matter of heated debate and controversy
involving several groups with somewhat different (and
sometimes conflicting) agendas including pathologists,
medical oncologists, basic scientists, commercial vendors,
the lay press, and patients and their families. The College
of American Pathologists recently issued recommenda-
tions regarding HER2 testing for patients with breast
cancer.5 However, these recommendations largely reflect
the current lack of consensus with regard to how best to
evaluate HER2 status.
Table 1US Food and Drug AdministrationApproved Assays for HER2
Test Type Indication
HercepTest (DAKO, Carpinteria, CA) Immunohistochemical Suitability for trastuzumab (Herceptin)Pathway (Ventana Medical Systems, Tucson, AZ) Immunohistochemical Suitability for trastuzumabPathVysion (Vysis, Downers Grove, IL) Fluorescence in situ hybridization Assessing prognosis; predicting response
to chemotherapyInform (Oncor/Ventana) Fluorescence in situ hybridization Assessing prognosis
-
8/3/2019 Shnitt and Jacobs AJCP 2000
2/5
AJCP / COMMENTARY
Am J Clin Pathol 2001;116:806-810 807 American Society of Clinical Pathologists
Immunohistochemical analysis offers advantages over
FISH in that it is relatively inexpensive and fast, uses a light
microscope, and can be performed in most pathology labora-
tories. In addition, the results have been linked to prognosis
and response to treatment.
Furthermore, it seems more logical to determine HER2
protein expression (by immunohistochemical analysis) than
the level of gene amplification (by FISH) when treatmentssuch as trastuzumab are specifically targeted toward the
HER2 protein on the cell surface. Immunohistochemical
analysis should, therefore, be the more biologically relevant
assay, particularly in cases in which there is discordance
between gene copy number and level of protein expression.
However, the numerous commercially available anti-HER2
antibodies vary widely in sensitivity and specificity,
staining results are subject to variations in tissue fixation
and processing, there is no standard scoring system, there is
no uniformly accepted threshold for positivity, and the
results are not quantitative. FISH offers several advantages
over immunohistochemical analysis in that the options for
reagents are limited (there are only 2 commercially avail-
able reagent kits available at this time), the threshold for
positivity has been standardized, the results are quantita-
tive, and there are internal positive controls. Moreover,
DNA, the target for FISH, seems less affected by variations
in tissue fixation and processing than does protein, the
target for immunohistochemical analysis. However, the
capability for performing FISH is limited to a much smaller
number of pathology laboratories than is immunohisto-
chemical analysis, and FISH is more expensive, is techni-
cally more cumbersome, requires longer procedure and
interpretation times than does immunohistochemical
analysis, and requires a fluorescence microscope for inter-
pretation. In addition, although published data have
supported a link between FISH results and prognosis, data
linking FISH assay results and response to therapy
(including response to trastuzumab) are more limited.6
In most studies, the concordance rates between
immunohistochemical and FISH results are in the 80% to
95% range, with the highest levels of concordance among
cases that are either completely negative or strongly positive
by immunohistochemical analysis.3,7 Much lower levels of
agreement between immunohistochemical analysis and FISHresults are seen for cases in which the immunohistochemical
result is less than strongly positive.8-10
So, given that immunohistochemical analysis and FISH
each have their advantages and disadvantages, what is the
practicing pathologist to do in order to best serve the clinical
needs of patients with breast cancer and the physicians caring
for them? Unfortunately, there is no definitive answer to this
question, and, in a way, pathologists are caught between the
proverbial rock and hard place. How, then do we move
forward? As we see it, there are a number of options, with
each having its pros and cons:
Abandon immunohistochemical analysis and use FISH
as the exclusive method for evaluating HER2 status of breast
cancers. A number of investigators have argued that because
of the inherent limitations of immunohistochemical analysis
for evaluating HER2 protein overexpression on fixed,
paraffin-embedded tissue sections, FISH should be consid-ered the method of choice for determining HER2 status in
clinical breast cancer samples and could be used as a surro-
gate for HER2 protein expression.11 While this argument
certainly has merit, there currently are numerous practical
barriers to the routine use of FISH for determining HER2
status in breast cancers. In particular, as noted, the procedure
is more complicated, time consuming, and expensive than
immunohistochemical analysis, requires a fluorescence
microscope, and requires considerably more pathologist
interpretation time than does immunohistochemical
analysis.3 It is possible that the development of automated
methods for performing the assay and screening the slides
may make FISH more practical for the routine determination
of HER2 status.
Use immunohistochemical analysis as a screening step
followed by FISH in selected cases. Many authors recently
have advocated the use of a combined approach for
assessing HER2 status.9,10 In this scenario, immunohisto-
chemical analysis is used as the initial screening step for all
cases, and, following evaluation of the immunohistochem-
ical stains, a subset of cases is subjected to FISH, which is
considered the gold standard. The most commonly
proposed algorithm is to screen cases with immunohisto-
chemical analysis and then perform FISH on cases in which
the immunohistochemical stains show only weak or
moderate membrane staining (2+ in the HercepTest
scoring system), since many such cases seem to represent
immunohistochemical false-positive results. While this
represents a reasonable suggestion, a number of studies
have clearly shown that some immunohistochemically nega-
tive cases show gene amplification and some strongly posi-
tive cases (3+ in the HercepTest scoring system) lack
gene amplification.11 Moreover, preliminary data suggest
that among cases that are strongly positive (3+) by
immunohistochemical analysis, only the cases that alsoshow gene amplification by FISH respond to trastuzumab
treatment.6 Thus, limiting FISH testing to cases scored as
2+ by immunohistochemical analysis will not eliminate
all false-positive and false-negative immunohistochemical
results, and precisely which subset of cases should be
subjected to FISH following immunohistochemical analysis
remains a matter of debate. Furthermore, it is not at all clear
which immunohistochemical assay should be used as the
screening test in such a scenario.
-
8/3/2019 Shnitt and Jacobs AJCP 2000
3/5
Schnitt and Jacobs / CURRENT STATUS OF HER2 TESTING
808 Am J Clin Pathol 2001;116:806-810 American Society of Clinical Pathologists
Develop methods to standardize immunohistochemical
analysis and/or render it more objective and quantitative.
One of the major problems with immunohistochemical
analysis is the lack of standardization with regard to primary
antibodies, detection systems, epitope/antigen retrieval, inter-
pretation, and reporting of results.12 The DAKO HercepTest
represented an important step toward standardizing immuno-
histochemical staining for HER2 by introducing predilutedreagents in kit form and a scoring system. However, some
studies have reported that this test has lower than desirable
specificity, whereas others have found that it has lower than
desirable sensitivity.8,10,11,13-15 Nevertheless, standardization
of immunohistochemical methods, with appropriate quality
control, remains an important goal.
Another major criticism of immunohistochemical
analysis has been that interpretation of the results is subjec-
tive and prone to interobserver variability, particularly at
intermediate levels of protein expression (ie, 2+ by the
HercepTest scoring system).13 Recently, a number of inves-
tigators have attempted to overcome this limitation of
immunohistochemical analysis by using computer-assisted
image analysis to quantitate HER2 immunohistochemical
results. For example, in the October 2001 issue of the
Journal, Wang et al16 retrospectively studied 189 breast
cancer cases using immunohistochemical analysis and
FISH and quantitated the immunohistochemical analysis
staining using the Automated Cellular Imaging System
(ACIS; ChromaVision Medical Systems, San Juan Capis-
trano, CA). The results of this study indicated that
immunohistochemical analysis quantitated with ACIS
showed a higher level of concordance with FISH than
immunohistochemical stains interpreted using visual
inspection (91% vs 86%) and that ACIS provided higher
sensitivity (but slightly lower specificity) than visually
interpreted immunohistochemical stains. Similar results
using other image analysis software have recently been
reported by others.17 However, Wang et al16 also appropri-
ately noted a number of limitations to the ACIS method,
which also apply to other computer-assisted image analysis
methods, including the cost of purchasing and maintaining
the hardware and computer software required for the quan-
titation, unresolved technical issues, and lack of data about
clinical validation of quantitative immunohistochemicalanalysis for HER2 protein determination. Perhaps most
important, unless the immunohistochemical analysis proce-
dure being subjected to image analysis has been optimized
by calibrating it against an appropriate standard,18 quantifi-
cation may simply provide an objective way to report an
incorrect result rather than a more reliable or accurate
result. Thus, while quantitation of immunohistochemical
results using computer-assisted image analysis could
provide certain important advantages over qualitatively or
semiquantitatively interpreted immunohistochemical stains,
there are a number of barriers and limitations to the wide-
spread implementation of this type of technology.
Develop methods to assess HER2 gene copy numbers
that can be evaluated using routine light microscopy. A
method that would combine the advantages of both immuno-
histochemical analysis and FISH might well be the most reli-
able and practical means to routinely assess HER2 status (ie,a DNA-based method that could be readily interpreted using
routine light microscopy). One such method is chromogenic
in situ hybridization (CISH). This approach permits the iden-
tification and quantitation of HER2 gene copies using light
microscopy since it uses a chromogenic rather than a fluores-
cent detection system. One study has indicated a high level
of concordance between CISH and FISH in enumerating
HER2 gene copy numbers.19 However, additional studies,
including those assessing the clinical value of CISH, are still
needed. Until such data are available, CISH should be
considered an investigational technique.
Develop tests that better predict the response of patients
to HER2-targeted therapy. Even if there were a method that
had 100% sensitivity, specificity, and accuracy in the deter-
mination of HER2 status of breast cancers, available data
suggest that this would still not permit the identification of
all patients who might or might not respond to HER2-
targeted antibody treatment with trastuzumab.2 Therefore,
there is great interest in developing a test that better predicts
response to such treatment. In particular, it is possible that
evaluation of the activated (phosphorylated) form of HER2,20
assessment of the level of expression of other members of
the epidermal growth factor receptor family in conjunction
with assessment of HER2, or assessment of downstream
effector molecules may provide more accurate information
than simply measuring HER2 status. However, additional
studies are needed to assess the clinical value of these alter-
native testing approaches.
Abandon tissue-based HER2 testing altogether. Some
investigators have studied the use of assays for circulating
HER2, in particular the measurement of the serum level of
the HER2 extracellular domain (ECD) in determining
HER2 status.21,22 Although the early results of such studies
showed promise, several potential problems exist in using a
serum assay for HER2 ECD. First, the assay has low sensi-tivity compared with immunohistochemical analysis or
FISH on corresponding tissue samples.22 Detection of ECD
is dependent on several factors, such as the level of antigen
production and release by the patients tumor and the half-
life of the protein in the serum. In addition, use of this assay
is subject to problems similar to those currently experienced
with immunohistochemical analysis and FISH, such as stan-
dardization of method and determination of appropriate
assay cutoff levels.
-
8/3/2019 Shnitt and Jacobs AJCP 2000
4/5
AJCP / COMMENTARY
Am J Clin Pathol 2001;116:806-810 809 American Society of Clinical Pathologists
Await results of studies directly comparing different
assays for HER2 with regard to their relationship to clin-
ical outcome. Perhaps more important than anything else at
this time, the various available assays for determining
HER2 status need to be directly and rigorously compared
within the setting of prospective clinical trials to determine
their relative merit and also to determine the appropriate
threshold for HER2 positivity. It should be noted, however,that the threshold for HER2 positivity may vary, depending
on the clinical question being asked. For example, the level
of HER2 protein expression required to declare a case
HER2-positive may be different when evaluating response
to treatment with trastuzumab than when determining sensi-
tivity to anthracyclines. Similarly, determining gene copy
number (eg, by FISH) may be more appropriate in some
clinical settings than others. Some recently published23 and
ongoing studies are attempting to address these issues.
In the meantime, in the absence of definitive answers,
it seems most prudent for pathologists to closely follow the
developments in this field and to work with their clinical
colleagues to formulate an approach to HER2 testing that is
most suitable at their institution. Ideally, before it is intro-
duced into clinical laboratory practice, any immunohisto-
chemical assay should be validated against another assay
such as FISH or a polymerase chain reactionbased assay.14
Based on extensive discussions with our clinical
colleagues, the approach we have settled on at our institu-
tion is to initially test all cases by immunohistochemical
analysis using a method we have validated against FISH.3,24
For cases that are strongly positive or clearly negative by
immunohistochemical analysis, no further testing is
performed. For cases that are equivocal or weakly positive
by immunohistochemical analysis (corresponding to 1+
or 2+ by HercepTest scoring) we perform FISH. We view
this as an interim solution until the issue of how best to
assess the HER2 status of breast cancers becomes more
clearly defined.
From the Department of Pathology, Beth Israel Deaconess
Medical Center and Harvard Medical School, Boston, MA.
Address reprint requests to Dr Schnitt: Dept of Pathology,
Beth Israel Deaconess Medical Center, 330 Brookline Ave,
Boston, MA 02215.Acknowledgments: We thank Anne Vincent-Salomon, MD,
and Jerome Couturier, MD, Insitut Curie, Paris, France, for their
helpful suggestions and comments.
References
1. Yamauchi H, Stearns V, Hayes DF. When is a tumor markerready for prime time? a case study of c-erbB-2 as a predictivefactor in breast cancer.J Clin Oncol. 2001;19:2334-2356.
2. Slamon DJ, Leyland-Jones B, Shak S, et al. Use ofchemotherapy plus a monoclonal antibody against HER2for metastatic breast cancer that overexpresses HER2.
N Engl J Med. 2001;344:783-792.
3. Jacobs TW, Gown AM, Yaziji H, et al. Comparison offluorescence in situ hybridization and immuno-histochemistry for the evaluation of HER-2/neu in breastcancer.J Clin Oncol. 1999;17:1974-1982.
4. Hanna W, Kahn HJ, Trudeau M. Evaluation of HER-2/neu(erbB-2) status in breast cancer: from bench to bedside.Mod Pathol. 1999;12:827-834.
5. Fitzgibbons PL, Page DL, Weaver D, et al. Prognosticfactors in breast cancer: College of American PathologistsConsensus Statement 1999.Arch Pathol Lab Med.2000;124:966-978.
6. Mass RD, Sanders C, Charlene K, et al. The concordancebetween the clinical trials assay (CTA) and fluorescence insitu hybridization (FISH) in the Herceptin pivotal trials[abstract]. Proc Am Soc Clin Oncol. 2000;19:75A.
7. Couturier J, Vincent-Salomon A, Nicolas A, et al. Strongcorrelation between results of fluorescent in situhybridization and immunohistochemistry for theassessment of the ERBB2 (HER-2/neu) gene status in breast
carcinoma. Mod Pathol. 2000;13:1238-1243.8. Jacobs TW, Gown AM, Yaziji H, et al. Specificity of
HercepTest in determining HER-2/neu status of breastcancers using the United States Food and DrugAdministrationapproved scoring system.J Clin Oncol.1999;17:1983-1987.
9. Ridolfi RL, Jamehdor MR, Arber JM. HER-2/neu testing inbreast carcinoma: a combined immunohistochemical andfluorescence in situ hybridization approach. Mod Pathol.2000;13:866-873.
10. Tubbs RR, Pettay JD, Roche PC, et al. Discrepancies inclinical laboratory testing of eligibility for trastuzumabtherapy: apparent immunohistochemical false-positives donot get the message.J Clin Oncol. 2001;19:2714-2721.
11. Pauletti G, Dandekar S, Rong H, et al. Assessment ofmethods for tissue-based detection of the HER-2/neualteration in human breast cancer: a direct comparisonof fluorescence in situ hybridization and immunohisto-chemistry.J Clin Oncol. 2000;18:3651-3664.
12. Seidal T, Balaton AJ, Battifora H. Interpretation andquantification of immunostains.Am J Surg Pathol.2001;25:1204-1207.
13. Hoang MP, Sahin AA, Ordonez NG, et al. HER-2/neugene amplification compared with HER-2/neu proteinoverexpression and interobserver reproducibility ininvasive breast carcinoma.Am J Clin Pathol.2000;113:852-859.
14. OMalley FP, Parkes R, Latta E, et al. Comparison ofHER2/neu status assessed by quantitative polymerase chain
reaction and immunohistochemistry.Am J Clin Pathol.2001;115:504-511.
15. Lebeau A, Deimling D, Kaltz C, et al. Her-2/neu analysis inarchival tissue samples of human breast cancer: comparisonof immunohistochemistry and fluorescence in situhybridization.J Clin Oncol. 2001;19:354-363.
16. Wang S, Saboorian MH, Frenkel EP, et al. Assessment ofHER-2/neu status in breast cancer: Automated CellularImaging System (ACIS)-assisted quantitation ofimmunohistochemical assay achieved high accuracy incomparison with fluorescence in situ hybridization assay asthe standard.Am J Clin Pathol. 2001;116:495-503.
-
8/3/2019 Shnitt and Jacobs AJCP 2000
5/5
Schnitt and Jacobs / CURRENT STATUS OF HER2 TESTING
810 Am J Clin Pathol 2001;116:806-810 American Society of Clinical Pathologists
17. Lehr HA, Jacobs TW, Yaziji H, et al. Quantitative evaluationof HER-2/neu status in breast cancer by fluorescence in situhybridization and by immunohistochemistry with imageanalysis.Am J Clin Pathol. 2001;115:814-822.
18. Riera J, Simpson JF, Tamayo R, et al. Use of cultured cells as acontrol for quantitative immunocytochemical analysis ofestrogen receptor in breast cancer: the Quicgel method.Am JClin Pathol. 1999;111:329-335.
19. Tanner M, Gancberg D, Di Leo A, et al. Chromogenic in situhybridization: a practical alternative for fluorescence in situhybridization to detect HER-2/neu oncogene amplification inarchival breast cancer samples.Am J Pathol. 2000;157:1467-1472.
20. Thor AD, Liu S, Edgerton S, et al. Activation (tyrosinephosphorylation) of ErbB-2 (HER-2/neu): a study ofincidence and correlation with outcome in breast cancer.
J Clin Oncol. 2000;18:3230-3239.
21. Leitzel K, Teramoto Y, Konrad K, et al. Elevated serum c-erbB-2 antigen levels and decreased response to hormonetherapy of breast cancer.J Clin Oncol. 1995;13:1129-1135.
22. Harris LN, Liotcheva V, Broadwater G, et al. Comparison ofmethods of measuring HER-2 in metastatic breast cancerpatients treated with high-dose chemotherapy.J Clin Oncol.2001;19:1698-1706.
23. Seidman AD, Fornier MN, Esteva FJ, et al. Weeklytrastuzumab and paclitaxel therapy for metastatic breastcancer with analysis of efficacy by HER2 immuno-phenotype and gene amplification.J Clin Oncol.2001;19:2587-2595.
24. Jacobs TW, Gown AM, Yaziji H, et al. HER-2/neu proteinexpression in breast cancer evaluated by immunohisto-chemistry: a study of interlaboratory agreement.Am J ClinPathol. 2000;113:251-258.