Advertisement
Journal of Clinical Oncology  
Search for:
Limit by:
  Browse by Subject or Issue
Home Search or Browse JCO My JCO Subscriptions Customer Service Site Map

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a colleague
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Save to my personal folders
Right arrow Download to citation manager
Right arrowRights & Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pauletti, G.
Right arrow Articles by Slamon, D. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pauletti, G.
Right arrow Articles by Slamon, D. J.
Journal of Clinical Oncology, Vol 18, Issue 21 (November), 2000: 3651-3664
© 2000 American Society for Clinical Oncology

Assessment of Methods for Tissue-Based Detection of the HER-2/neu Alteration in Human Breast Cancer: A Direct Comparison of Fluorescence In Situ Hybridization and Immunohistochemistry

By Giovanni Pauletti, Suganda Dandekar, HongMei Rong, Lilllian Ramos, HongJun Peng, Ram Seshadri, Dennis J. Slamon

From the Department of Medicine, Division of Hematology-Oncology, University of California at Los Angeles, UCLA School of Medicine, Los Angeles, CA; and Flinders Medical Centre, Bedford Park, South Australia, Australia.

Address reprint requests to Dennis J. Slamon, MD, PhD, Division of Hematology and Oncology, UCLA School of Medicine, 11–934 Factor Bldg, Los Angeles, CA 90095-1678; email dslamon{at}mednet.ucla.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
PURPOSE: To compare the efficacy of fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC) in detecting the HER-2/neu alteration in human breast cancer.

PATIENTS AND METHODS: Unselected stage I, II, and III breast cancer patients (N = 900) were tested for HER-2/neu gene amplification by FISH in paraffin-embedded, formalin-fixed archival material. Of these samples, 856 were tested for HER-2/neu overexpression by non–antigen-retrieval IHC with the polyclonal antibody R60, the sensitivity and specificity of which was preliminarily compared with the United States Food and Drug Administration–approved HercepTest (Dako Corp, Carpinteria, CA). Patient survival was analyzed in relation to the presence of the HER-2/neu alteration as determined by these two methodologies.

RESULTS: A total of 189 (21%) of 900 patients were positive by FISH and 147 (17.2%) of 856 were positive by IHC. This discrepancy is consistent with expected loss of IHC sensitivity associated with tissue fixation/embedding. The HercepTest did not improve sensitivity and introduced false positives. Comparison of R60-based IHC with FISH demonstrates that patient survival is associated progressively to gene amplification level as determined by FISH, whereas for IHC an association is found only in the highest (3+) immunostaining group. Among FISH-negative tumors, 45 (6.6%) of 678 were IHC-positive, with a survival probability similar to that of FISH-negative/IHC-negative cases; FISH-positive/IHC-negative patients have a survival probability similar to that of FISH-positive/IHC-positive cases.

CONCLUSION: IHC does not consistently discriminate patients likely to have a poor prognosis, whereas FISH provides superior prognostic information in segregating high-risk from lower-risk beast cancers. HER-2/neu protein overexpression in the absence of gene amplification occurs infrequently in breast cancer, in which case, patient outcome is similar to that of patients without the alteration.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE HER-2/neu GENE is overexpressed in 25% to 30% of human breast cancers,1,2 and in approximately 90% to 95% of these cases, overexpression is a direct result of gene amplification.2-4 This alteration correlates with poor clinical prognosis, meaning that malignancies with HER-2/neu amplification/overexpression are associated with shorter disease-free survival (DFS) as well as overall survival (OS). This association was initially controversial, as studies either confirmed or disputed the correlation. Review of the controversy revealed that these discrepancies were mainly due to methodological and/or technical variables.5,6 Reagents and methods used in a number of studies proved suboptimal in identifying the alteration.6 Other studies were limited by cohort size or follow-up time (see review in Gullik et al7 ). Most studies using well-characterized reagents, reliable methodologies, adequate sample sizes, and appropriate follow-up now confirm the prognostic association between the HER-2/neu alteration and clinical outcome for both node-negative and node-positive breast cancers.8-12

The need for accurate detection of the HER-2 alteration has now become even more important, because therapeutic decisions for patients are increasingly dependent on this information. Retrospective analyses of clinical trials indicate that this alteration may be predictive of response to certain types of chemotherapy.13,14 In addition, recent work demonstrates that HER-2/neu gene product, p185HER-2/neu, represents a target for specific therapy with the monoclonal antibody trastuzumab (Herceptin; Genentech, Inc, South San Francisco, CA).15,16 Optimal use of this therapy, however, requires accurate determination of HER-2/neu status because, to date, the presence of this alteration is the sole criteria for trastuzumab treatment.

Tissue-based detection of the HER-2/neu alteration by fluorescence in situ hybridization (FISH) and/or immunohistochemistry (IHC) offers clear advantages over other approaches, such as Southern, Northern, and Western blot and polymerase chain reaction–based analyses, which suffer from dilutional artifacts resulting from mixture of normal and abnormal cell populations within the tissue. Both IHC and FISH allow for the specific detection of the alteration in individual cells while maintaining critical architectural tissue information. As a result, these techniques have emerged as the preferred methods to measure HER-2/neu alterations.

FISH already has important applications in clinical hematology as well as areas of medical genetics such as prenatal chromosomal analyses.17 Its utility in solid tumor cytogenetics, however, has only recently been explored. An advantage of FISH is that it circumvents antigenic changes that occur in formalin-fixed/paraffin-embedded tissues, a major limitation inherent to IHC. Measurement of the HER-2/neu alteration in human breast cancer by FISH has already been shown to have comparable specificity and sensitivity to IHC using frozen tissue.2,6,18 In the current study, FISH is directly compared with IHC in assessing formalin-fixed, paraffin-embedded tissue for the presence of this alteration in a large, unselected, consecutively accrued breast cancer cohort.


   PATIENTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
FISH
Specimen preparation, hybridization, and microscopy were performed as previously described4,19 with the following modifications. Deparaffinized tissue sections were treated for 4 hours in 4% paraformaldehyde in phosphate-buffered saline and then three times (2 minutes each) in 0.3 mol/L glycine phosphate-buffered saline solution. Chaotropic salt treatment was 1 N sodium thiocyanate for 12 minutes at 80°C. The SpectrumOrange-labeled HER-2/neu probe and the SpectrumGreen chromosome 17 centromeric alpha-satellite (chr.17cen.) probe were cohybridized on the tissue specimen. FISH probes were obtained from Vysis, Inc, Downers Grove, IL.

IHC
IHC was performed with a rabbit polyclonal antiserum, R60, which targets the intracellular domain of p185HER-2/neu.1,2 No antigen retrieval procedure was used with this antibody. The peroxidase-antiperoxidase technique was used and immunoprecipitate formation was identified by light microscopy after using the chromogen, 3,3'-diaminobenzidinetetrahydrochloride. Each specimen was scored semiquantitatively as to the intensity of membrane immunostaining on a four-point scale, with 0 indicating absence of staining, 1+ indicating the lowest level of detectable staining and/or nonhomogeneous weak staining, 2+ indicating moderate homogeneous membrane staining, and 3+ indicating intense homogeneous membrane staining. We do not report percentage of stained tumor cells because relatively homogeneous immunostaining was observed in all positive (>= 2+) tumor specimens using this polyclonal antibody. Controls included paraffin-embedded human breast cancer cell lines with known levels of HER-2/neu expression and two human tumor specimens, one of which is known to be negative and the other 3+ positive. Scoring was done in a blinded fashion by a single observer independent of FISH results. In a simultaneously conducted pilot study, the commercially available, United States Food and Drug Administration–approved kit (HercepTest, Dako Corp, Carpinteria, CA) was also tested in an independent laboratory using the manufacturer’s detection procedures and scoring system. HercepTest analysis was performed blinded to FISH and IHC results with the R60 Ab method.

FISH Scoring Criteria
Scoring of HER-2/neu and chr.17cen. signals was restricted to cancer cells. Signal enumeration was performed with single-band pass filter combinations. FISH scores were expressed both as HER-2/neu signals per cell (continuous variable) and HER-2/neu signals per chr.17cen. signal (two categories: <= two, > two HER-2/neu per chr.17cen.). Slides were first scanned at x630 or x1,000 with the appropriate filter combination for the HER-2/neu probe, sorted into three groups, and examined as follows. The first group consisted of samples in which obvious HER-2/neu gene amplification was present, ie, consistent presence of signal cluster(s) or more than 10 to 15 scattered signals/cell. In this group, 10 to 20 nuclei were selected for signal enumeration from those that displayed the highest number of HER-2/neu signals. Precise signal enumeration of high-level amplification was frequently not possible because of coalescing fluorescence of signal cluster(s). The final FISH score (signals/cell) assigned to each tumor was based on the average number of signals in examined nuclei. On the basis of cytogenetic evidence indicating occasional loss of physical linkage between amplified HER-2/neu genes (localized to homogeneously staining regions) and chromosome 17,3,20 the HER-2/neu per chr.17cen. ratio was not evaluated. When categorizing these data for amplification corrected for chr.17 polysomy, these tumors were included in the more than two HER-2/neu per chr.17cen. group. The second group consisted of samples in which no cells with more than four HER-2/neu signals were detectable after initial scanning. In this group, more than 200 nuclei or on occasion the entire tumor cell population was examined to determine the FISH score. If any cell(s) containing more than four HER-2/neu signals were found, the specimen was scored similar to the third group of specimens (see below). Although chr.17cen. copy number was also evaluated, HER-2/neu per chr.17cen. ratio was not calculated, and all these tumors were placed into the <= two HER-2/neu per chr.17cen. group. The third group consisted of samples (n = 304) in which initial screening indicated that the tumor cell population contained cells with >= five HER-2/neu signals with no obvious gene amplification patterns (< 10 to 15 scattered signals/cell) or cells with HER-2/neu signal clusters with fewer than 10 to 12 signals/cell. In this group, for 247 (81%) of 304 patients, 100 nuclei (inclusive of nuclei with at least one HER-2/neu signal) were scored for both HER-2/neu and centromere 17 signals at x1,000. In the other 57 (19%) of 304 cases, fewer nuclei (average {cong} 50 nuclei/slide) were scored because of either the presence, often concurrent, of broken and/or overlapping nuclei, the loss of nuclear DNA, the sporadic decondensation of centromere signals, or a limited number of tumor cells. HER-2/neu doublets were counted as separate signals. Depending of the mean ratio of the two markers’ signals, a tumor was placed in either the <= two or more than two HER-2/neu per chr.17cen. category.

Clinical Features of the Patient Population
Study patients were from South Australia and diagnosed with invasive breast cancer between 1987 and 1991. The study included all patients with stage I through III breast carcinomas. The median follow-up period for the cohort at the time of analysis was 5.7 years. Patients were unselected with the exception of the exclusion of patients with stage IV disease and those for whom axillary clearance was not performed. Patients underwent either total (75%) or partial (25%) mastectomy, with axillary lymph node clearance (> eight nodes). Patients who underwent total mastectomy did not receive adjuvant postoperative radiotherapy, as opposed to 67% of the partial mastectomy patients. Adjuvant chemotherapy with six cycles of the combination of cyclophosphamide, methotrexate, and fluorouracil or tamoxifen was administered to axillary node-positive, premenopausal (15%) or postmenopausal patients (24%), respectively. Node-negative patients did not receive any adjuvant therapy. Slot-blot analysis for HER-2/neu copy number was previously performed in this cohort,9 as well as hormone receptor biochemical assays21 and assessment of tumor cell proliferation by Ki67 antigen determination.22 Grade was determined according to the Bloom-Richardson method. Tumor size corresponds to the maximum tumor diameter.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
A cohort of 923 breast cancer patient samples consecutively collected was analyzed simultaneously for HER-2/neu gene amplification using FISH and for HER-2/neu protein overexpression using IHC. FISH and IHC assays were conducted blinded from each other as well as to clinical information. No patients were excluded on the basis of age, cause of death, or the length of available follow-up. The clinicopathologic features of this cohort were consistent with those of other large breast cancer studies (Table 1), indicating that it is representative of the general breast cancer population.23 This was confirmed by evaluating the prognostic value for DFS as well as OS (DFS/OS) of several established factors, including presence of involved axillary lymph nodes (log-rank test, P = .0001/.0001), primary tumor size (P = .0001/.0001), histologic grade (P = .0005/.0002), S phase (P = .0001/.0001), and estrogen receptor (P = .0027/.0003) as well as progesterone receptor (P = .0260/.0008) status; all of which were predictive of DFS/OS survival. As expected, most cases (84.6%) were classified as invasive ductal carcinomas, however, given the inclusive nature of the study design, lobular carcinomas (6.4%) and special histologies (9.0%) were also evaluated. A total of 23 cases were disqualified from the study either on the basis of absence of tumor cells in the tissue section or an unsatisfactory FISH analysis. Criteria for successful FISH analysis included identification of single-copy HER-2/neu gene status in normal cells within the specimen, uniform hybridization throughout the section, and limited DNA loss.4,19 Therefore, a total of 900 specimens were successfully analyzed by FISH in this study.


View this table:
[in this window]
[in a new window]
  		Table 1. Cohort Characteristics
 
Definition and Frequency of HER-2/neu Gene Amplification as Determined by FISH
The FISH data were assessed for HER-2/neu amplification using two definitions. The first defined gene copy number in terms of total HER-2/neu signals per cell, whereas the second defined HER-2/neu amplification relative to chr.17cen. signals. The latter approach involves cohybridization with a chr.17cen. alpha-satellite–specific probe and HER-2/neu probe (Fig 1). For both approaches, cutoffs for the variables were defined by objective rather than by subjective criteria, ie, biologically relevant thresholds rather than selection of predetermined cutoff points based on predictive value.



View larger version (109K):
[in this window]
[in a new window]
  		Fig 1. Demonstration of HER-2/neu gene amplification in an infiltrating breast adenocarcinoma using FISH with HER-2/neu (orange) and chr.17cen. (green) specific probes. Magnification is x630.

 
The maximum number of FISH signals that can be detected for a gene in normal somatic cells is four. Cells with three to four signals are considered to be in either the S or G2 phase of the cell cycle. This biologic threshold determined the cutoff of four HER-2/neu signals/cell used to dichotomize the copy number data (maximum number of signals per cell present in a specimen) when not treated as a continuous variable. This cutoff only distinguishes those cases with a disomic HER-2/neu status from those containing a higher copy number (ie, > four signals/cell). Discrimination of true HER-2/neu amplification requires correction for chromosome 17 polysomy. In this study, a total of 487 (54%) of 900 samples have signal distributions of up to four HER-2/neu signals/cell, with most cells containing zero, one, or two signals. These cancers have equivalent chr.17cen. signal distributions; only two (0.04%) of 487 tumors contain cells with more than four chr.17cen. signals, indicating that chromosome 17 polysomy is rare in the absence of HER-2/neu signal increase and validating chr.17cen. use as a control for detecting increased HER-2/neu signals due to polysomy. In the remaining 413 (46%) of 900 cancers, HER-2/neu signal distribution range exceeded four, with approximately 60% of these cases containing a maximum of five to 10 signals/cell. The distinction between increased HER-2/neu signal caused by true gene amplification as opposed to chromosomal polysomy in these specimens is critical. When HER-2/neu gene amplification was defined in terms of HER-2/neu signals per chr.17cen., tumors with a ratio of more than two HER-2/neu signals per chr.17cen. signal were classified as amplified. A total of 189 (21%) of 900 specimens in this cohort demonstrated gene amplification using these criteria.

Definition and Frequency of HER-2/neu Overexpression as Determined by IHC
Two IHC reagents and methods were considered with the intent of selecting an IHC test for use in the FISH versus IHC comparison: these two IHC tests were one with the R60 antibody1,2 and, the other, the commercial detection kit, HerceptTest. We evaluated the performance of this new kit relative to that of the R60 antibody6 in two selected paraffin-embedded specimen cohorts (HER-2/neu copy number previously defined by FISH). One group consisted of HER-2/neu single-copy specimens (n = 43), whereas the other consisted of HER-2/neu–amplified specimens (n = 42). Analysis with the HercepTest was conducted in a blinded fashion at a facility outside our laboratory to insure that analysis with this kit was performed correctly and consistent with the manufacturer’s recommended procedures.

Specimens in the amplified group were chosen to represent the entire spectrum of HER-2/neu amplification. Five cases with an HER-2/neu signals per chr.17cen. ratio of more than 1.5 but less than two were included; a ratio of two was the cutoff used in this study to define gene amplification corrected for polysomy (Table 2). HER-2/neu mRNA and protein expression levels were available for 36 (86%) of 42 amplified cases, making them fully molecularly characterized by Northern and Western blot analyses as well as by the IHC analysis of frozen material from the same specimens.2 This group is representative of the most fully characterized cohort of human breast cancer specimens analyzed for HER-2/neu status in the current literature.2 The data show that although both IHC methods performed satisfactorily at higher amplification levels, both suffer a loss of sensitivity at low amplification levels (Table 2). A total of 28 (75.7%) of 37 cases with HER-2/neu signals per chr.17cen. ratio of more than two were positive (IHC >= 2+) with both R60 and the HercepTest (Table 2). Among the cases positive by R60 but negative by the HercepTest, two had significantly high copy number increase (> 25 to 30 HER-2/neu genes/cell). High levels of HER-2/neu expression confirmed by both Northern and Western blot analyses indicated that the HercepTest had truly misclassified these two specimens and that antigen retrieval did not overcome the decrease in sensitivity caused by fixation problem in these cases.


View this table:
[in this window]
[in a new window]
  		Table 2. Concordance Between HercepTest and R60 IHC in Tumors With HER-2/neu Amplification as Determined by FISH
 
The single gene copy cases (n = 43) for this comparative IHC analysis were selected based on the presence of <= four HER-2/neu signals/cell, which indicates single-copy status for both signals/cell and signals/chr.17cen. FISH data. Chromosome 17 polysomic samples were excluded from this analysis to avoid any potential issues involving the definition of gene amplification and the relationship between gene overexpression and chromosome 17 ploidy status. All 43 of these cases were negative by R60; however, 11 (25.6%) of 43 were positive (>= 2+) using the antigen-retrieval–based HercepTest. In addition, we tested 10 cases with no amplification, which were positive by IHC with the R60 antibody, and found six of 10 to be positive when tested with HercepTest. Because the results of this pilot study failed to indicate any improvement in sensitivity using the HercepTest, and especially given the excess positive cases in the single-copy group indicating a new problem of false positivity, we conducted the FISH versus IHC comparison study using the R60 antibody. Availability of tissue sections limited the IHC analysis to 856 (95%) of 900 cases in the cohort. The number of patients in each R60 immunostaining group were as follows: 0, 603 patients; 1+, 106 patients; 2+, 55 patients; and 3+, 92 patients. IHC categories of 0 or 1+ are considered consistent with normal levels of HER-2/neu expression, whereas the 2+ or 3+ categories correspond with protein overexpression.2 Using these criteria, a total of 17.2% of the patients in this cohort were shown to have HER-2/neu overexpression by IHC. This is consistent with previously published data for this IHC approach.2,6

Patient Survival as a Function of HER-2/neu Gene Copy Number Determined by FISH Versus HER-2/neu Protein Expression Level Determined by IHC
Clinical data were prospectively collected for this cohort. Survival probability of patients stratified for HER-2/neu amplification corrected for polysomy demonstrates that patients with more than two HER-2/neu genes per chromosome 17 have significantly decreased survival (Fig 2A). Survival as a function of HER-2/neu copy number uncorrected for chromosome 17 polysomy was also evaluated and was again found to decrease with increasing gene copy number (Fig 2B). This association was further confirmed when FISH signals per cell were evaluated as a continuous variable by Cox regression analysis (relative risk [RR] = 1.02; beta = 0.0180 ± 0.005; P > .0009).



View larger version (24K):
[in this window]
[in a new window]
  		Fig 2. Patient survival probability analysis (Kaplan-Meier): (A) <= 2 and > 2 HER-2/neu signals per chr.17cen. signal; (B) increasing levels of HER-2/neu signals/cell (patients stratified by maximum signals/cell within each specimen, 4 arbitrary strata); (C) increasing p185HER-2/neu expression levels; (D) node-negative patients (n = 589); Cox proportional hazards analysis (strata by HER-2/neu copy number and tumor size as indicated). Inset, Kaplan-Meier analysis (strata as in A). (N) = number of patients.

 
Although the number of HER-2/neu genes per cell is known to be correlated with the amount of HER-2/neu protein on the cell surface,24,25 the IHC data in the current study demonstrate a decreased survival probability in the 3+ immunostaining group only (Fig 2C). These data fail to show the same graded relationship between the level of HER-2/neu alteration and prognosis seen with FISH, indicating a discrepancy in the ability of the two techniques to predict patient survival. We found no major changes for FISH or IHC in the results nor in the conclusions described above when survival analysis was limited to patients with stage I and II disease (data not shown). To better understand the source of this discrepancy, a quantitative representation of the relationship between amplification (HER-2/neu signals per cell) as determined by FISH and overexpression as determined by IHC was generated for the 856 tumors for which both FISH and IHC data existed (Fig 3B and 3C). FISH results were grouped into five categories (determined according to maximum signals per cell within each specimen), which were <= four, five to eight, nine to 16, 17 to 32, and more than 32 HER-2/neu signals/cell. Data categorization was not determined by mean signals per cell because the means were not available for all specimens as a result of our scoring criteria. The relative frequency of the four IHC categories was then calculated for each FISH group (Fig 3B and 3C). A limited number (25 of 470) of IHC-positive cases (IHC >= 2+) were found in tumors with a disomic gene copy number (<= four HER-2/neu signals/cell). The disomic HER-2/neu overexpressing cases represented 5.3% of this group, or 2.9% of the total analyzed patients. As expected, FISH categories corresponding to progressively higher levels of gene copy number contain higher percentages of IHC-positive patients. Although the largest increase in IHC positivity (28.6%) occurs within the nine to 16 signals/cell category, it is not until the 17 to 32 category that IHC-positive patients become the majority of cases (65.6%). Even in the more than 32 signals/cell category, IHC-positive cases do not exceed 80% of the samples. The gap between sensitivity of IHC in frozen versus paraffin-embedded tissue in the current study is greater at lower levels of HER-2/neu amplification. This is again consistent with published data demonstrating that HER-2/neu antibody sensitivity is dependent on gene copy number.6 Failure of IHC to consistently detect overexpression is further documented by plotting both the FISH and IHC data against HER-2/neu amplification as determined by an independent technique (ie, slot-blot analysis) in these same samples (Fig 3A). As expected, the level of HER-2/neu protein expression generally is correlated with the level of HER-2/neu gene amplification; however, it is noteworthy that even with this less sensitive technique,3 IHC in fixed/embedded material still yields false-negatives among cases known to be amplified. The discordance between FISH and IHC in defining high-risk breast cancer group(s) was again confirmed when IHC was compared with HER-2/neu amplification corrected for chromosome 17 polysomy (in 856 patients with both FISH and IHC data). In this comparison, 102 (57.3%) of 178 tumors containing amplification (> two HER-2/neu per chr.17cen.) were >= 2+ by IHC. Conversely, IHC-positive cases were rare in tumors with <= two HER-2/neu per chr.17cen., with a frequency of 45 (6.6%)of 678 tumors.



View larger version (44K):
[in this window]
[in a new window]
  		Fig 3. (A) Relationship between HER-2/neu expression level as determined by IHC and HER-2/neu amplification as determined by FISH (HER-2/neu signals/cell) or by slot-blot analysis. {Delta} = no IHC available. (B) Relative frequency of each immunostaining category within 5 FISH categories (determined according to maximum signals/cell within each specimen): <= 4, 5 to 8, 9 to 16, 17 to 32, and > 32 HER-2/neu signals/cell. FISH category size proportional to the number of patients. Arbitrary category cutoffs follow geometrical progression from the maximum biologically normal copy number (4 signals/cell). (C) Numerical values for (B): squares contain number of patients with a given IHC score per FISH category, frequency IHC (%) per FISH category (superscript), and per the whole cohort (subscript).

 
Direct Comparison of IHC and FISH in Determining Survival Probability Based on Detection of HER-2/neu Alteration
A single cutoff, two-group categorization was used to stratify patients either by FISH, for which a ratio of <= two HER-2/neu per chr.17cen. was classified as negative for amplification and more than two was considered positive, or by IHC, for which the conventional criteria was used, with 0 and1+ classified as negative for overexpression and 2+ and 3+ categorized as positive. Estimated survival probability for each strata was plotted (Fig 4A and 4D). Again, this comparison was restricted to the patients for whom both FISH and IHC data were available (n = 856). Subsequently, FISH-based patient stratification was applied to the IHC-categorized groups (Fig 4E and 4F), and conversely, IHC stratification was applied to FISH-categorized groups (Fig 4B and 4C) to determine whether either technique, FISH or IHC, adds prognostic strength to the other. This analysis showed that FISH adds statistically significant prognostic information (Fig 4F) to the group defined as positive by IHC status, whereas IHC adds only marginal prognostic value to the FISH-categorized groups (Fig 4B and 4C). This was further demonstrated by comparing survival of patients who were negative by FISH but positive by IHC with the survival of patients who were positive by FISH but negative by IHC (Fig 5A). IHC-positive/FISH-negative patients who have the single-copy, HER-2/neu–overexpressing cancers (Fig 5A, upper curve), demonstrate a survival probability similar to that of the low-risk (IHC-negative/FISH-negative) patients (Fig 5B, upper curve). Conversely, the IHC-negative/FISH-positive patients (Fig 4A, lower curve) have a survival probability similar to that of the poor-prognosis IHC-positive/FISH-positive patient group (Fig 5B, lower curve). The results of analysis of DFS by log-rank test for each of the patient groups shown in Fig 4 were as follows: A, P = .0090; B, P = .5611; C, P = .1394; D, P = .0475; E, P = .4705; and F, P = .0465, resulting in the same conclusions stated above (graphs not shown).



View larger version (36K):
[in this window]
[in a new window]
  		Fig 4. Comparison between FISH and IHC in predicting OS. Patients stratified by (A) HER-2/neu amplification relative to chromosome 17 copy number and (D) the conventional IHC criteria (0,1+ v 2+,3+). Estimated survival probabilities of FISH or IHC groups by Kaplan-Meier method, compared by log-rank test. Data available for 856 patients (number of patients per group in parentheses). (B) and (C) IHC stratification applied to the FISH-based patient groups and (E) and (F) vice versa, ie, FISH-based stratification to the IHC-categorized patient groups.

 


View larger version (25K):
[in this window]
[in a new window]
  		Fig 5. Survival of patients stratified by either concordance or discordance between amplification (FISH) and overexpression (IHC). (A) and (B) HER-2/neu signals/chr.17cen. as in Fig 4, or (C) and (D) absolute HER-2/neu signals/cell as in Fig 6. (N) = number of patients.

 


View larger version (36K):
[in this window]
[in a new window]
  		Fig 6. Comparison between FISH and IHC in predicting OS. Patients stratified by either (A) absolute HER-2/neu number (<= 12 v > 12 HER-2/neu signals/cell) or (D) the IHC criteria (0,1+ v 2+,3+). (N) = number of patients. Survival probabilities of FISH or IHC groups are estimated by Kaplan-Meier method, compared by log-rank test. (B) and (C) IHC stratification applied to the FISH-based patient groups, and (E) and (F) vice versa, FISH-based stratification to the IHC-categorized patient groups.

 
Finally, the analysis was repeated with FISH scores expressed as an absolute HER-2/neu copy number per cell, but this time with a FISH cutoff arbitrarily set so as to maximize the difference in the percentage of patients overexpressing HER-2/neu in the two resulting FISH groups. In one FISH group (n = 730), tumors have HER-2/neu signal distributions up to a maximum of 12 signals/cell, which maximized the number of IHC-negative patients and minimized the number of IHC-positive patients. In the other FISH group (n = 126), specimens have signal distributions exceeding 12 signals/cell and the number of IHC-positive patients is maximized and the number of IHC-negative patients is minimized. In this analysis, the FISH cutoff was defined in a dependent manner that was arbitrarily determined using IHC expression data. Among all possible dichotomization strategies for FISH data (HER-2/neu signals per cell), this approach represents the one most concordant with the IHC-based patient dichotomization (ie, 0 and 1+ v 2+ and 3+). In the tumors with an upper limit of 12 signals, the mean number of HER-2/neu signals per cell was 5.54, whereas for those that had an upper limit of 13, the mean was 6.63. Survival for these FISH and IHC groups was compared as in the previous analyses (Figs 5C and 5D and 6) and yielded the same conclusion, ie, FISH was superior to IHC in predicting clinical outcome despite intentional minimizing of the differences between risk groups as defined by each method. The log-rank test results for the DFS analysis of the patient groups shown in Fig 6 were as follows: A, P = .0034; B, P = .6831; C, P = .1926; D, P = .0410; E, P = .6010; F, P = .0245 (graphs not shown).

Multivariate Analysis
Proportional hazards regression analysis26 was conducted to compare FISH-determined HER-2/neu amplification as well as IHC-determined HER-2/neu protein levels with the most commonly used clinical prognostic markers, including grade, histopathologic type, absence of estrogen and progesterone receptors, presence of nodal involvement, and S phase. The association of these prognostic factors with the HER-2/neu alteration is well documented,27-29 and the results from the current study cohort confirm these associations, further validating this study group as a representative cohort (Table 3). Both FISH (treated as continuous and dichotomized variables, ie, HER-2/neu signals per cell and <= two v > two HER-2/neu per chr.17cen. or <= four v > four HER-2/neu signals/cell) and IHC were found to be predictors of OS independent of other variables (Table 4). However, only HER-2/neu amplification corrected for chr.17cen. and IHC were found to be independent predictors of DFS (Table 4). These results were also confirmed when grade and S phase were added to the model in the 525 patients for whom these data were also available (data not shown). This is important because both grade and S phase were strongly associated with HER-2/neu amplification/overexpression (Table 3) and were both significant univariate predictors of survival in this cohort. In the node-negative patient group, cases disomic for HER-2/neu (<= four HER-2/neu signals/cell) had a better survival probability than those with more than four signals/cell, which demonstrates a statistically significant decrease of OS (RR = 1.54) independent of tumor size (Fig 2D) and ER status (RR = 1.5; beta = 0.449 ± 0.200; P = .0254; not shown in Fig). Node-negative patients with more than two HER-2/neu per chr.17cen. also had a worse survival compared with those with <= two by univariate analysis; however, this difference did not reach statistical significance (Fig 2D, inset).


View this table:
[in this window]
[in a new window]
  		Table 3. Association of Various Clinical Variables With HER-2/neu Gene Amplification or Overexpression as Determined by FISH or IHC, Respectively
 

View this table:
[in this window]
[in a new window]
  		Table 4. Multivariate Analysis
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Accurate detection of the HER-2/neu gene alteration in human breast cancer specimens has become increasingly important in determining patient prognosis as well as response to standard chemotherapeutic agents. Moreover, it is currently the sole criteria for the selection of patients for HER-2/neu–targeted therapy with the recombinant humanized anti-p185HER-2/neu antibody trastuzumab (Herceptin). To date, a number of techniques have been used to detect the presence of this alteration. These approaches have included analyses of tissue sections, tissue lysates, and serum-based assays.12 A major advantage of in situ–based technologies, such as IHC and FISH, is the maintenance of the original tissue architecture, which permits analysis in its histologic context. This aspect, along with the general availability of reagents and low cost, accounts for the popularity of IHC. IHC, however, has significant shortcomings, the most important of which is the loss of sensitivity secondary to antigenic alterations caused by standard fixation procedures.2,18 This loss, coupled with the documented wide variability in sensitivity of available anti-HER-2/neu antibodies (6% to 80%), results in major problems for IHC-based testing.6 A second and equally important issue with this technique is the lack of consistent interpretation of results,30 even with inclusion of standardized reference samples. Although quantitative imaging approaches are possible,10 they can be time-consuming and require special imaging equipment. In addition, the use of tissue culture cells as internal standards suffers from the fact that fixation and embedding of cell pellets does not replicate this process in actual human tissue specimens.

Interphase cytogenetics using FISH in paraffin-embedded tissue sections has been shown to be a valid alternative to IHC.4,11,12 These studies also demonstrate that the sensitivity and specificity of FISH is superior to IHC in formalin-fixed, paraffin-embedded tissue sections in terms of correlation with HER-2/neu overexpression as confirmed by Northern and Western blot analyses as well as by IHC in frozen tissue sections. In the current study, we evaluated FISH and IHC concurrently in a large archival breast cancer cohort. A direct cross-over comparison between the two techniques in terms of their relative ability to define high-risk patients was performed. Clinical data on the patients in this study were prospectively collected. Care was taken to avoid introduction of any possible ascertainment bias by analyzing all specimens accrued to the cohort irrespective of cause of death or follow-up time. Classical prognostic factors showed a significant correlation with prognosis, indicating that this cohort was representative of the general breast cancer patient population.

For this study, FISH-detected increases of HER-2/neu copy number were evaluated both as absolute number of HER-2/neu signals per cell and as HER-2/neu signals relative to chr.17cen. signals. A total of 413 (46%) of 900 tumors contained cells with an abnormal number of HER-2/neu genes (> four HER-2/neu signals/cell), whereas 189 (21%) of 900 contained an average of more than two HER-2/neu per chr.17cen. The latter approach was termed the chromosome 17 polysomy-corrected amplification score. In comparison, 147 (17.6%) of 856 patients in this same cohort displayed positive (2+, 3+) immunostaining. FISH results were compared directly with IHC in predicting both DFS and OS. Both assays predicted patient survival in multivariate analysis, confirming that both HER-2/neu amplification and overexpression are correlated with survival independently of other variables. However, in univariate analysis, although there was a direct correlation between survival and gene copy number as determined by FISH, patients stratified by IHC demonstrated this relationship only in the highest (3+) immunostaining group. Therefore, the ability to define high-risk groups differed significantly between the two techniques. Further analysis demonstrated that neither the 2+ nor 3+ IHC groups completely overlapped with tumors containing the highest copy number (Fig 3B). It is noteworthy that in a subgroup analysis of the phase III clinical trials that led to approval of trastuzumab,15 patients with 2+ IHC score did not seem to benefit significantly from trastuzumab therapy.31

In this study, the cutoff of more than two HER-2/neu per chr.17cen. defined polysomy-corrected amplification. This cutoff ratio is uniformly associated with HER-2/neu protein overexpression as determined by IHC in three separate studies using three different antibodies.3,4,31,32 It is critical to note that IHC in each of these published studies was performed using frozen tissue specimens rather than formalin-fixed, paraffin-embedded material. Because it is unlikely that the majority of testing in the general population will be performed on frozen material, the HER-2/neu antigenic changes inherent to formalin-fixed, paraffin-embedded tissue will likely remain problematic. This problem is further confirmed in the current study, which was conducted in formalin-fixed, paraffin-embedded material and for which the results demonstrate that 76 (42.7%) of 178 cases with more than two HER-2/neu per chr.17cen. were IHC-negative. It is important to notice, however, that although this cutoff is likely to be the most biologically valid way of defining HER-2/neu amplification status, this does not imply that ratio values of more than two HER-2/neu per chr.17cen. correlate precisely with p185HER-2/neu expression levels; within the amplified group (> two HER-2/neu per chr.17cen.), absolute copy number may correlate best with expression levels.25 In this study, survival probabilities for ratio values of more than two HER-2/neu per chr.17cen. were not evaluated. With respect to those tumors defined as nonamplified by the cutoff (<= two HER-2/neu per chr.17cen.), a small number, 45 (6.6%) of 678, have HER-2/neu overexpression in the absence of gene amplification. Using frozen material, the HER-2/neu single copy overexpression incidence previously reported was one (2.9%) of 34 cases,3 four (4%) of 101 cases,4 and seven (8.3%) of 83 cases.32 Overall, the incidence of single gene copy overexpression in this cohort is consistent with the published data and confirms that HER-2/neu overexpression in the absence of gene amplification is an infrequent event in human breast cancers.

FISH data expressed in terms of absolute HER-2/neu copy number per cell were also compared with that of IHC groups (0 and 1+ v 2+ and 3+), with dichotomization of the FISH variable into two groups: <= 12 versus more than 12 HER-2/neu signals/cell. This is not proposed as a potential FISH cutoff for actual HER-2/neu testing, but rather was used for the purpose of further comparing FISH with IHC with the intent of minimizing discordance between risk groups as defined by the two techniques. Despite this, the FISH group with more than 12 HER-2/neu signals/cell still contained 39 (31%) of 126 IHC-negative patients. The transition between these two FISH groups occurs at a mean HER-2/neu copy number of approximately six signals/cell. Above this mean level of absolute HER-2/neu gene amplification, IHC in frozen material consistently detects overexpression of the HER-2/neu gene,3 again underscoring the sensitivity problem of IHC in detecting the HER-2/neu alteration in paraffin-embedded archival material, as confirmed by the current study. No attempt was made in this study to define a cutoff for the HER-2/neu signals per cell data, which would account for HER-2/neu increase resulting from chromosome 17 polysomy.

Application of FISH-based patient stratification to IHC-classified samples and IHC stratification to FISH-classified samples demonstrates that FISH adds significant prognostic information to patients categorized by IHC status, whereas IHC adds only marginal information to FISH-classified groups. One of the more important findings in the current study is that analysis of survival results for women with HER-2/neu overexpression in the absence of gene amplification demonstrates that their probability of survival is statistically indistinguishable from patients who are negative by both techniques. This shows a significant discriminatory effect of FISH in the important category of patients treated with trastuzumab (ie, IHC >= 2+) and could have substantial implications in predicting their response to this new therapy. Conversely, patients who are positive by FISH but negative by IHC have a worse survival than those who have HER-2/neu overexpression in the absence of gene amplification. These data support the concept that failure of IHC is the cause of the IHC-negative/FISH-positive cases. Again, these findings may have therapeutic implications with respect to trastuzumab use.

Beyond technical considerations, these data may also have a biologic basis. Transcriptional activation of the HER-2/neu gene in the absence of amplification may result in insufficient levels of p185HER-2/neu to trigger the pathogenetic response that results in decreased patient survival. In a separate study, Northern and Western blot analyses showed that seven of eight breast cancer patients with single gene copy–overexpressing tumors had levels of HER-2/neu mRNA and p185HER-2/neu similar to those of single copy–nonoverexpressing tumors.10 In addition, quantitative HER-2/neu protein determination at the cell level demonstrates that single copy–overexpressing tumors consistently have less p185HER-2/neu per cell than tumors containing even the lowest levels of gene amplification.10 Another explanation for the superior prognostic value of FISH could be that overexpression of a gene or genes other then HER-2/neu but located in the same amplicon may contribute to the HER-2/neu–positive phenotype. Among these genes are PPARBP (TRAP220), which encodes an estrogen receptor–alpha coactivator,33 and GRB7. Grb7 has been implicated in cell migration and angiogenesis via the FAK34 and Tek35 pathways, respectively. In addition, Grb7 overexpression has been reported to contribute to metastatic progression of human esophageal carcinoma.36 Interestingly, Grb7 is also involved in the p185HER-2/neu signal transduction pathway.37 Coamplification of HER-2/neu and GRB7 with resulting overexpression of these interacting gene products may contribute to enhanced signaling. The functional status of the overexpressed p185HER-2/neu may also be of relevance. In a retrospective analysis of a primary breast cancer cohort, only a fraction of the p185HER-2/neu immunopositive cases were also positive by analysis with an antibody that recognizes only the tyrosine phosphorylated form of p185HER-2/neu.38 It is possible that amplification correlates better with the fraction of activated HER-2/neu receptors. Finally, proteolytic cleavage of the extracellular domain of p185HER-2/neu generates a soluble ectodomain and a transmembrane 95-kd protein that still retains kinase activity. In a study of 161 patients, p95HER-2/neu was detected in 65.6% of patients with high-level HER-2/neu overexpression and was found at a level that ranged from 10% to 100% of that of p185HER-2/neu as judged by Western blot analysis.39 Antibodies recognizing an HER-2/neu epitope on the extracellular domain of the protein will not detect p95HER-2/neu, which may still participate in downstream signaling. Whether any or all of these mechanisms are important in HER-2/neu detection technology remains to be determined.

Using the biologically significant cutoff point of two HER-2/neu genes per chr.17cen., node-negative patients with HER-2/neu amplification had a worse survival compared with that of patients without amplification (Fig 2D, inset). Although the difference was not statistically significant, it is important to point out that the survival curves of these two patient groups show a clear trend toward poorer survival with amplification. Because of the limited number of events (24 patients with amplification; 77 patients without amplification) and incidence of gene amplification (19.3%) among node-negative patients in this cohort, it is likely that to show a clear-cut association between amplification and survival in this patient group, a longer follow-up period is required. Furthermore, it is clear that within the present study’s follow-up time, the subset of node-negative patients disomic for HER-2/neu have the best prognosis independent of tumor size and estrogen receptor status. This finding is a prerequisite to demonstrating a statistically significant improved prognosis for node-negative patients with <= two HER-2/neu genes per chr.17cen. relative to those with more than 2 on extended follow-up.

The current study shows that irrespective of the manner by which FISH data were analyzed, FISH provides better prognostic information and segregates high-risk breast cancer patients more accurately than IHC. Such survival analyses provide direct evidence that HER-2/neu copy number detection by FISH is a superior predictor of patient outcome when compared with HER-2/neu overexpression detected by IHC. These data, combined with the results from previous studies comparing FISH with IHC in frozen tissue sections, raise a cautionary note on the reliability of IHC in accurately detecting patients eligible for HER-2/neu–targeted therapy such as trastuzumab when using formalin-fixed/paraffin-embedded material. This includes those IHC methods using HER-2/neu antigen retrieval to improve sensitivity, such as the United States Food and Drug Administration–approved HercepTest, which was evaluated as part of this study. The sensitivity of the HercepTest was not improved when compared with the sensitivity of antibodies previously reported to be best for detecting this alteration in fixed/embedded tissue.6 Additionally, the HercepTest seems to result in a significant incidence of false-positive cases, 11 (25.6%)of 43, which was apparent with the analysis performed in an independent laboratory facility. This same phenomenon has been reported by several other investigators who confirm both false-negative (21.7% of FISH-positive)40 and particularly high positivity rates (54% to 60%)41-43 using the HercepTest. The current data demonstrate the superiority of FISHover IHC in detecting the HER-2/neu alteration in formalin-fixed, paraffin-embedded human breast cancer specimens. Automated FISH analysis, as well as automated signal enumeration technologies, may quickly move FISH from research laboratories to clinical laboratories on a wide scale.


   ACKNOWLEDGMENTS
 
Supported by The Revlon/University of California at Los Angeles Women’s Cancer Research Program.

We thank Drs L.Q. Tang and H.J. Wang for statistical analysis, Drs M. Pegram, S. Seelig, and G. Koneckny for critically reading the manuscript, and J. Mitchell for manuscript preparation.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
1. Slamon DJ, Clark GM, Wong SG, et al: Human breast cancer: Correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235: 177-182, 1987[Abstract/Free Full Text]

2. Slamon DJ, Godolphin W, Jones LA, et al: Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 244: 707-712, 1989[Abstract/Free Full Text]

3. Kallioniemi O-P, Kallioniemi A, Kurisu W, et al: ERBB2 amplification in breast cancer analysed by fluorescence in situ hybridization. Proc Natl Acad Sci U S A 89: 5321-5325, 1992[Abstract/Free Full Text]

4. Pauletti G, Godolphin W, Press MF, et al: Detection and quantitation of HER-2/neu gene amplification in human breast cancer archival material using fluorescence in situ hybridization. Oncogene 13: 63-72, 1996[Medline]

5. Slamon DJ, Press MF, Godolphin W, et al: The HER-2/neu proto-oncogene in human breast cancer. Cancer Cells 7: 371-380, 1989

6. Press MF, Hung G, Godolphin W, et al: Sensitivity of HER-2/neu antibodies in archival tissue samples: Potential source of error in the immunohistochemical studies of oncogene expression. Cancer Res 54: 2771-2777, 1994[Abstract/Free Full Text]

7. Gullik WJ, Love SB, Wright C, et al: c-erbB-2 protein overexpression in breast cancer is risk factor in patients with involved and uninvolved lymph nodes. Br J Cancer 63: 434-438, 1991[Medline]

8. Paterson MC, Dietrich KD, Danyluk J, et al: Correlation between c-erbB-2 amplification and risk of recurrent disease in node-negative breast cancer. Cancer Res 51: 556-567, 1991[Abstract/Free Full Text]

9. Seshadri R, Figaira FA, Horsfall DJ, et al: Clinical significance of HER-2/neu oncogene amplification in primary breast cancer. J Clin Oncol 11: 1936-1942, 1993[Abstract/Free Full Text]

10. Press MF, Pike MC, Chazin VR, et al: HER-2/neu expression in node-negative breast cancer: Direct tissue quantitation by computerized image analysis and association of overexpression with increased risk of recurrent disease. Cancer Res 53: 4960-4970, 1993[Abstract/Free Full Text]

11. Press MF, Bernstein L, Thomas PA, et al: HER-2/neu gene amplification characterized by fluorescence in situ hybridization: Poor prognosis in node-negative breast carcinomas. J Clin Oncol 15: 2894-2904, 1997[Abstract]

12. Ross JS, Fletcher JA: The HER-2/neu oncogene: Prognostic factor, predictive factor and target for therapy. Semin Cancer Biol 9: 125-138, 1999[Medline]

13. Muss HB, Thor AD, Berry DA, et al: c-erbB-2 expression and response to adjuvant therapy in women with node- positive early breast cancer [published erratum appears in N Engl J Med 331: 211, 1994]. N Engl J Med 330: 1260-1266, 1994[Abstract/Free Full Text]

14. Paik S, Bryant J, Park C, et al: erbB-2 and response to doxorubicin in patients with axillary lymph node-positive, hormone receptor-negative breast cancer [see comments]. J Natl Cancer Inst 90: 1361-1370, 1998[Abstract/Free Full Text]

15. Slamon DJ, Leyland-Jones B, Shack S, et al: Addition of Herceptin (humanized anti-HER2 antibody) to first line chemotherapy for HER2 overexpressing metastatic breast cancer (HER2+/MBC) markedly increases anticancer activity: A randomized multinational controlled phase III trial. Proc Am Soc Clin Oncol 17: 98a, 1998 (abstr 377)

16. Cobleigh MA, Vogel CL, Tripathy D, et al: Multinational study of the efficacy and safety of humanized anti-HER2 monoclonal antibody in women who have HER2-overexpressing metastatic breast cancer that has progressed after chemotherapy for metastatic disease. J Clin Oncol 17: 2639-2648, 1999[Abstract/Free Full Text]

17. Gray JW, Pinkel D: Molecular cytogenetics in human cancer diagnosis. Cancer 69: 1536-1542, 1992[Medline]

18. Penault-Llorca F, Adelaide J, Houvenaeghel G, et al: Optimization of immunohistochemical detection of erbB2 in human breast cancer: impact of fixation. J Pathol 173: 65-75, 1994[Medline]

19. Pauletti G, Slamon DJ: Interphase cytogenetics of the HER-2/neu oncogene in breast cancer, in Andreeff M, Pinkel D (eds): Introduction to Fluorescence In Situ Hybridization: Principles and Clinical Application. New York, NY, Wiley-Liss, 1999, pp 153-171

20. Bernardino J, Apiou F, Gerbault-Seureau M, et al: Characterization of recurrent homogeneously staining regions in 72 breast carcinomas. Genes Chromosomes Cancer 23: 100-108, 1998[Medline]

21. Horsfall DJ, Tilley WD, Orell SR, et al: Relationship between ploidy and steroid hormone receptors in primary invasive breast cancer. Br J Cancer 53: 23-28, 1986[Medline]

22. Seshadri R, Leong AS, McCaul K, et al: Relationship between p53 gene abnormalities and other tumour characteristics in breast-cancer prognosis. Int J Cancer 69: 135-141, 1996[Medline]

23. Carter CL, Allen C, Henson DE: Relation of tumor size, lymph node status, and survival in 24,740 breast cancer cases. Cancer 63: 181-187, 1989[Medline]

24. Lewis GD, Figari I, Fendly B, et al: Differential responses of human tumor cell lines to anti-p185HER2 monoclonal antibodies. Cancer Immunol Immunother 37: 255-263, 1993[Medline]

25. Szollosi J, Balazs M, Feuerstein BG, et al: erbB-2 (HER-2/neu) gene copy number, p185HER-2 overexpression and intratumor heterogeneity in human breast cancer. Cancer Res 55: 5400-5407, 1995[Abstract/Free Full Text]

26. Cox DR: Regression models and life tables. J R Stat Soc 34: 187-220, 1972

27. Berger MS, Locher GW, Saurer S, et al: Correlation of c-erbB-2 gene amplification and protein expression in human breast carcinoma with nodal status and nuclear grading. Cancer Res 48: 1238-1243, 1988[Abstract/Free Full Text]

28. Heintz NH, Leslie KO, Rogers LA, et al: Amplification of the c-erb B-2 oncogene and prognosis of breast adenocarcinoma. Arch Pathol Lab Med 114: 160-163, 1990[Medline]

29. Borg A, Baldetorp B, Ferno M, et al: ERBB2 amplification in breast cancer with a high rate of proliferation. Oncogene 6: 137-143, 1991[Medline]

30. Allred DC, Harvey JM, Berardo M, et al: Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 11: 155-168, 1998[Medline]

31. United States Food and Drug Administration: Department of health and human services: Oncology drugs advisory committee— 58th Meeting, September 2, 1998, P.M. Session. Http://www.fda.gov/ohrms/dockets/ac/98/transcpt/3443t2.rtf

32. Persons DL, Borelli KA, Hsu PH: Quantitation of HER-2/neu and c-myc gene amplification in breast carcinoma using fluorescence in situ hybridization. Mod Pathol 10: 720-727, 1997[Medline]

33. Zhu Y, Qi C, Jain S, et al: Amplification and overexpression of peroxisome proliferator-activated receptor binding protein (PBP/PPARBP) gene in breast cancer. Proc Natl Acad Sci U S A 96: 10848-10853, 1999[Abstract/Free Full Text]

34. Han DC, Guan JL: Association of focal adhesion kinase with Grb7 and its role in cell migration. J Biol Chem 274: 24425-24430, 1999[Abstract/Free Full Text]

35. Jones N, Master Z, Jones J, et al: Identification of Tek/Tie2 binding partners: Binding to a multifunctional docking site mediates cell survival and migration. J Biol Chem 274: 30896-30905, 1999[Abstract/Free Full Text]

36. Tanaka S, Sugimachi K, Kawaguchi H, et al: Grb7 signal transduction protein mediates metastatic progression of esophageal carcinoma. J Cell Physiol 183: 411-415, 2000[Medline]

37. Stein D, Wu J, Fuqua SA, et al: The SH2 domain protein GRB-7 is co-amplified, overexpressed and in a tight complex with HER2 in breast cancer. EMBO J 13: 1331-1340, 1994[Medline]

38. DiGiovanna MP, Carter D, Flynn SD, et al: Functional assay for HER-2/neu demonstrates active signalling in a minority of HER-2/neu-overexpressing invasive human breast tumours. Br J Cancer 74: 802-806, 1996[Medline]

39. Christianson TA, Doherty JK, Lin YJ, et al: NH2-terminally truncated HER-2/neu protein: Relationship with shedding of the extracellular domain and with prognostic factors in breast cancer. Cancer Res 58: 5123-5129, 1998[Abstract/Free Full Text]

40. Jacobs TW, Grown AM, Yaziji H, et al: Comparison of fluorescence in situ hybridization and immunohistochemistry for the evaluation of HER-2/neu in breast cancer. J Clin Oncol 17: 1974-1982, 1999[Abstract/Free Full Text]

41. Roche PC, Ingle JN: Increased HER2 with U.S. Food and Drug Administration-approved antibody. J Clin Oncol 17: 434, 1999 (letter)[Free Full Text]

42. Maia DM: Immunohistochemistry assays for HER2 overexpression. J Clin Oncol 17: 434, 1999 (letter)

43. Jacobs TW, Grown AM, Yaziji H, et al: Specificity of HerceptTest in determining HER-2/neu status of breast cancer using the United States Food and Drug Administration-approved scoring system. J Clin Oncol 17: 1983-1987, 1999[Abstract/Free Full Text]

Submitted December 29, 1999; accepted July 14, 2000.




This article has been cited by other articles:


Home page
 J. Mol. Diagn.Home page
A. Carbone, G. Botti, A. Gloghini, G. Simone, M. Truini, M. P. Curcio, P. Gasparini, A. Mangia, T. Perin, S. Salvi, et al.
Delineation of HER2 Gene Status in Breast Carcinoma by Silver in Situ Hybridization is Reproducible among Laboratories and Pathologists
J. Mol. Diagn., November 1, 2008; 10(6): 527 - 536.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
S. Makkat, R. Luypaert, T. Stadnik, C. Bourgain, S. Sourbron, M. Dujardin, J. De Greve, and J. De Mey
Deconvolution-based Dynamic Contrast-enhanced MR Imaging of Breast Tumors: Correlation of Tumor Blood Flow with Human Epidermal Growth Factor Receptor 2 Status and Clinicopathologic Findings--Preliminary Results
Radiology, November 1, 2008; 249(2): 471 - 482.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Pathol.Home page
S Parry, K Savage, C Marchio, and J S Reis-Filho
Nestin is expressed in basal-like and triple negative breast cancers
J. Clin. Pathol., September 1, 2008; 61(9): 1045 - 1050.
[Abstract] [Full Text] [PDF]

Home page