Reliance on Hormone Receptor Assays of Surgical Specimens May Compromise Outcome in Patients With Breast Cancer
http://www.100md.com
《临床肿瘤学》
the Department of Surgery and Pathology, The Royal Melbourne Hospital, The University of Melbourne, Parkville, Australia
ABSTRACT
PURPOSE: To assess the concordance of breast cancer immunohistochemical receptor assays on core biopsy and surgical specimens.
PATIENTS AND METHODS: We identified 100 patients whose core biopsy and definitive surgical specimens were processed in our hospital. New sections, with core and surgical specimens on the same slides, were stained for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2) immunohistochemistry (IHC). Two pathologists assessed the sections independently. Raw scores and clinically significant groupings were compared.
RESULTS: Concordance for ER, PR, overall hormone receptor (HR), and HER-2 status was seen in 86%, 83%, 90%, and 80% of patients, respectively. The core was positive, while the surgical specimen was negative in 13%, 11%, 9%, and 1% of patients, respectively. Using a log-linear model, differences in ER, PR, and HER-2 staining were all in the direction of stronger staining in the cores, and were statistically significant. Nine percent (95% CI, 4.2% to 16.4%) of women in this group would have had endocrine therapy inappropriately withheld if management decisions were based on surgical specimen results alone.
CONCLUSION: ER and PR assays on core biopsy specimens are more reliable than assays on surgical specimens. Receptor IHC should be performed on core biopsy specimens to avoid patients with HR positive cancers not receiving appropriate hormonal therapy and being overtreated with systemic chemotherapy. Biopsy should be considered in patients with "receptor negative" cancer and recurrent disease.
INTRODUCTION
The status of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2) is critical in the management of patients with invasive breast cancer.1 Endocrine manipulation with tamoxifen, aromatase inhibitors, and ovarian ablation is a central part of adjuvant therapy for early breast cancer or therapy for advanced disease in patients with ER- and/or PR-positive cancers.2,3 Trastuzumab (Herceptin) is an important recent treatment option for advanced disease in patients with HER-2 overexpressing tumors, and is under investigation to determine if it has a role in adjuvant therapy.4 In patients whose tumors show no expression of ER and PR, any form of endocrine therapy is considered ineffective,1 and trastuzumab is not considered in those whose tumors exhibit less than 2+ staining of HER-2.4
Immunohistochemistry (IHC) assays are the basis for determining the ER, PR, and HER-2 status of breast cancers. IHC has replaced ligand-binding assays for ER and PR assessment in most pathology laboratories, after comparative studies demonstrated their superiority in terms of simplicity and reproducibility.5 For HER-2, 3+ staining on IHC correlates with gene amplification and is thus accepted as an indication for trastuzumab treatment. 2+ staining with IHC is followed by fluorescent in situ hybridization (FISH) before trastuzumab treatment.4. Thus, it is critical that IHC to measure ER, PR, and HER-2 status is accurate. The possibility of false-negative assays must be minimized to reduce the chance that potentially life-saving or life-extending relatively nontoxic therapy will be inappropriately withheld.
Some reports have demonstrated that IHC on percutaneous biopsy specimens correlates with those on surgical specimens and can be used for treatment decisions.6-9 We commenced performing IHC on core biopsy specimens to help with decisions regarding induction treatment, and noticed occasional disparity between assay results on core biopsy and surgical specimens. We hypothesized that the fixation process is more complete with core biopsies compared with surgical specimens. This study was designed to address the hypothesis that relying on IHC on surgical specimens would result in false negatives in a significant minority of cases.
PATIENTS AND METHODS
Pathology department records were used to identify 100 consecutive patients treated for invasive breast cancer who had both a core biopsy and surgical specimen reported by the pathology department of the Royal Melbourne Hospital. Patients with multiple tumors were included only if the surgical specimen corresponding to the core biopsy could be identified. The study was approved by the ethics committee of Melbourne Health.
Core biopsies were taken under ultrasound guidance. Multiple passes of a 14-g Manan Pro-Mag 2.2 automatic biopsy system (Manan Medical Products Inc, Northbrook, IL) were made. With larger tumors, various areas of the tumor were sampled. All specimens had been fixed in 10% neutral buffered formalin before processing. Core biopsies were fixed in formalin for a minimum of 3 hours before processing, while specimens received after 2 PM were usually left in formalin overnight. Small surgical specimens were allowed to fix for a minimum of 4 hours, while larger specimens and all mastectomies were fixed for at least 8 hours. Mastectomies were sliced approximately every 5 mm, and paper towel was inserted between the slices to aid fixation. Small excisional specimens were not incised before fixation, while larger specimens were, after inking of the margins. Any specimen from which tissue was taken for tissue banking was incised before being placed in formalin.
Specimens were processed into paraffin blocks using one of two automated processing machines. Breast cores were processed on the Tissue Tek VIP processor (Sakura, Tokyo, Japan). Surgical specimens and mastectomies were processed on the Medite TPC-15 tissue processor (Medite, Burgdorf, Germany). The Medite uses chloroform as its clearing agent as it is gentler on delicate fatty specimens, as opposed to xylene used in the VIP processor. Both processes are accepted as appropriate for processing breast cancer specimens.
Paraffin blocks were retrieved, and sections of the core biopsy, surgical specimen, and a known positive control block for each of the three antibodies were cut at 3 μm and all placed on the same SuperFrost Plus (Menzel-Gl?ser, Braunschweig, Germany) slide (ie, three slides per case were generated, with tumor, core biopsy, and positive control all stained on the same slide for each antibody). Slides were dewaxed and carefully brought to water. Heat-induced antigen retrieval was performed by placing the slides in a coplin jar with 85 mL of citrate buffer (pH 6.0) and heating in a microwave for 10 minutes once boiling point was reached. The slides were then allowed to cool in the coplin jar for 30 minutes.
The biotin-streptavidin immunoperoxidase staining technique with a DAKO autostainer (DAKO Corp, Carpinteria, CA) was used. After dewaxing and antigen retrieval, endogenous peroxidase activity was blocked by placing the slides in distilled water and 30% hydrogen peroxidase for 5 minutes, and then rinsing well in phosphate-buffered saline (PBS). Next, the primary antibody was added. The primary antibodies used were: ER clone 6F11 (Novocastra NCL-6F11, Newcastle, UK) dilution factor 50; PR clone PgR316 (Novocastra NCL-PgR-316) dilution factor 400; and c-erbB-2 (DAKO Polyclonal A0485; DAKO Corp) dilution factor 500.
Incubation continued at room temperature for 30 minutes, followed by washing in PBS. Biotinylated antimouse and antirabbit Ig (DAKO code K0675; DAKO Corp) was applied, and slides incubated at room temperature for 10 minutes, then washed with PBS. Streptavidin-peroxidase (DAKO code K0675; DAKO Corp) was applied, incubated at room temperature for 10 minutes, then washed again with PBS. The peroxidase activity was demonstrated by applying 2 to 3 drops of liquid 3,3-Diaminobenzidine (DAKO K3466; DAKO Corp), and the peroxidase activity developed for 8 minutes and then washed in tap water. The slides were counterstained with Lillee-Mayer hematoxylin, blued in Scott's Tap Water Substitute, dehydrated in ethanol, cleared in xylene, and mounted in DPX mounting medium (Surgipath Medical Laboratories, Richmond, IL).
The slides were labeled with the patient details at the time of paraffin sectioning. Subsequent to staining, each of the three slides for each case was given an allocated case number (1 to 100) and was identified only by this number. The slides were covered so that the core biopsy material was visible but the surgical tissue was obscured. The core biopsy of each case was then scored. Once all the cores had been scored, the core biopsy material was obscured leaving only the surgical material available for scoring. Thus, the individual core and surgical specimens were scored without direct visualization or knowledge of the other staining results. The results were entered on separate Excel (Microsoft, Redmond, WA) spreadsheets.
Two pathologists examined each slide independently. For ER and PR, the proportion of cells staining was quantified, and the intensity of staining was assessed on a 0 to 3+ rating. HER-2 staining was rated on a 0 to 3+ scale. The mean of the two assessments was used for the subsequent analysis. If the assessments of the pathologists differed by more than 20% or by 1 intensity score, or if there was a clinically significant difference in the assessment of the two pathologists, both pathologists reassessed the slides, and a consensus was reached.
For ER and PR, the results were grouped according to the thresholds of 0, less than 10%, and 10%, which are commonly used to designate receptors as negative, equivocal, and positive, respectively.10 For a more informative analysis of the staining differences for ER and PR, we calculated the Allred score.11 This is a scoring system that combines the proportion of cells staining with the intensity of staining to give a single overall rating. An assessment of overall hormone receptor status was also made: if either ER or PR had 10% of cells staining, the tumor was considered receptor positive. For HER-2, the raw scores were tabulated. They were also grouped according to clinical relevance as HER-2 negative (0 and 1+), equivocal (2+), and positive (3+).
To assess whether there was a systematic tendency for core biopsies to show a positive result more often than the corresponding surgical tissue sample, a log-linear model was used to test the symmetry of the comparisons. This model assumes the null hypothesis that where there is disagreement between the two samples, it is equally likely to be in either direction. The alternative hypothesis is that core biopsies are more likely to be positive. This is quantified by estimating the log-odds of the disagreement being in that direction: the null hypothesis corresponds to a parameter value of zero, and a log-odds greater than 0 implies that the core biopsy is more likely to have more staining than the surgical specimen.12 The test was applied separately to the 3 x 3 tables of results for the ER, PR, ER+PR, and HER-2, the 4 x 4 table of results for the raw HER-2 scores and the 9 x 9 tables of Allred ER and PR scores.
For all cases assessed as ER or PR negative on resected specimen, the proportion of cells staining and the intensity of staining was also assessed in nonmalignant breast epithelium present on each slide. This was done to assess the value of such staining as an internal control.
We retrieved clinical records of all patients for whom there was a clinically significant difference to assess the clinical advice that had been given to each patient.
RESULTS
One hundred patients who had both core biopsy and surgical specimens reported at the Royal Melbourne Hospital pathology department between April 1999 and March 2002 were identified. No patient in this series had systemic therapy administered between the core biopsy and definitive surgery.
Demographic features of the patients, and clinical, surgical, and pathologic features of the cancers are presented in Table 1.
The ER status of the surgical specimens showed that 78% were positive and 22% were negative; while for the cores 90% were positive, 1% were equivocal, and 9% were negative (Table 2). For PR of the surgical specimen, 64% were PR positive, 3% equivocal, and 33% negative; while for the cores, the corresponding results were 68%, 7%, and 25% (Table 2). If "hormone receptor–positive" is defined as either ER or PR positive, 82% were positive and 18% were negative on surgical specimens, and 90% of cores were positive, 1% equivocal, and 9% negative (Table 2).
There was concordance between the ER assessment of core and surgical specimen in 86% at the level of ER negative, equivocal, or positive. In 13%, the core was ER positive, whereas the surgical specimen was ER negative, and in 1%, the core was equivocal while the surgical specimen was ER positive (Table 2).
For PR, there was concordance in 83%. In 11%, the core showed some PR staining while the surgical specimen showed no staining. In 6%, there was more staining for PR in the surgical specimen than in the core—in two cases, the core was PR negative while the surgical specimen was equivocal; in one case, the core was PR negative, and the surgical specimen was PR positive; and in three cases, the core was equivocal, and the surgical specimen was PR positive (Table 2).
The Allred scores give an assessment of the overall extent of staining. Clinically, the most important cut point in the Allred score is greater than 2, as these are the cases that are generally considered hormone responsive. Tables 3 and 4 show how the discordance in terms of "positive" and "negative" using this cut point was essentially identical to the discordance using the 10% cutoff presented in Table 2. Analysis of the overall Allred scores shows that most of the discordances were quite substantial, with many case of strong staining in most cells in the core biopsy, with no staining at all in the surgical specimen.
For overall receptor status, there was concordance in 90%. In 9% (95% CI, 4.2% to 16.4%), the core was hormone receptor positive, and the surgical specimen was hormone receptor negative, while in 1%, the core was equivocal, and the surgical specimen was hormone receptor positive (Table 2). Assuming that core biopsies are the gold standard (see Discussion), the false-negative rates for ER, PR, and overall hormone receptor on surgical specimens in this series of 100 patients were 14% (95% CI, 7.9% to 23.4%), 15% (95% CI, 7.6% to 24.7%), and 10% (95% CI, 4.7% to 18.1%), respectively.
Regarding HER-2, 81% of surgical specimens showed 0 or 1+ staining; 12%, 2+; and 7%, 3+. For cores, 65%, 27%, and 8% showed 0 or 1+, 2+, and 3+ staining (Table 5). Concordance for clinically significant staining of HER-2 was seen in 80%. In 17%, the core was equivocal (2+), and the surgical specimen was negative (usually 1+). In 2%, the reverse was true, with a negative core and equivocal surgical specimen, while in the last case the core was positive (3+), and the surgical specimen was negative (Table 5). When the raw scores were compared, there was concordance in 42%. In 54%, the HER-2 staining was more intense in the core than in the surgical specimen, and in 4%, the reverse was the case. The most common finding (36% of cases) was 1+ staining in the core and no staining in the surgical specimen (Table 6).
When the log-linear model was applied to the results, all estimates were in the same direction, with more greater or more intense staining in the core biopsies, and all were significant or highly significant, excepting the reduced PR data. However, there was a highly significant effect in the Allred PR data (Table 7).
Assessment of staining of normal epithelium was difficult, as staining was patchy and variable throughout the specimens. There was some staining of ER and/or PR in normal epithelium in most of the specimens assessed. There were six cases with no staining at all. Four of these were false-negatives and two were true-negatives. There was definite staining in normal epithelium in another four of the false-negatives, and equivocal staining in the final one. There was at least some staining in all but two of the true-negatives.
Review of the nine cases with receptor-negative surgical specimens but receptor-positive cores showed that five had been treated as receptor positive on the basis of results of IHC on the diagnostic core biopsy. Two cases received no adjuvant systemic therapy at all. One was a 14-mm, grade 2 cancer in a 62-year-old patient, while the other was 10-mm, grade 3 cancer in a 72-year-old patient. Both would have been prescribed tamoxifen for 5 years had they been considered receptor positive. The other two cases were given adjuvant chemotherapy. One was a 64-year-old patient with a 12-mm, grade 2 cancer and a micrometastasis in the sentinel node, and the other was a 58-year-old patient with a 12-mm, grade 2 node-negative cancer. Had they been considered receptor positive, each would have been given tamoxifen, and may not have been given chemotherapy.
DISCUSSION
Endocrine therapy and chemotherapy are systemic treatment options for patients with breast cancer. Endocrine therapy is recommended in most patients with hormone receptor–positive cancer because of its efficacy and favorable side effect profile. Chemotherapy is considered in most hormone receptor–negative tumors and for higher risk hormone receptor–positive ones. The risk-benefit balance is generally more complex with chemotherapy, with well-known short-term morbidity including alopecia, nausea, and neutropaenia, as well as long-term risks of cardiac damage and leukemia.1 Accurate and reliable assessment of the ER, PR, and HER-2 status of invasive breast cancers is therefore crucial. While it is accepted that hormone receptor–negative patients should not be considered for endocrine therapy, many studies show that a proportion of these patients do respond to such treatment.13
Many studies have demonstrated that the accuracy and reliability of IHC-based assays on formalin-fixed and paraffin-embedded tissue is similar to the ligand-binding assays of the past.5 Thus, IHC remains excellent in truly experienced hands. There is, however, interlaboratory variation in IHC assessment. A review of reports of IHC assays of ER and PR showed a wide variation in the rates of positivity. For ER, this variation was 25% to 84%, and for PR, it was 24% to 66%.11 Quality-assurance studies of interlaboratory variance in IHC demonstration of ERs have suggested that the assessment is frequently inaccurate. In studies in which paraffin blocks of cancers known to express ER were distributed to various laboratories for staining and reading, false-negative rates of at least 24% were found.14,15 Our study addresses an important derivative of the general issue of quality assurance of receptor assays: if the fixation of surgical specimens results in loss of the antigen, then even a laboratory practicing excellent staining and reading practices will have false-negative results.
The negative predictive values of the assays, defined as the percentage of receptor negative patients showing no response to endocrine therapy, were generally in the order of 80% to 92%.11 While it is possible there are molecular mechanisms whereby truly receptor negative patients respond to endocrine therapy, the results of this study raise the possibility that some of the responding negatives were in fact false negatives.
There is extensive literature from the 1980s comparing the levels of receptors ER and PR by ligand-binding assay on incisional biopsy with that on mastectomy. Across a number of series, the ER and PR level was consistently lower in the mastectomy than the open surgical biopsy groups, and many false-negatives were reported in the assays on mastectomy specimens.16-19 The recommendations were to provide fresh frozen biopsy material for optimal results. Such attention to specimen processing has generally not been given emphasis in the era of IHC.
A number of small studies have addressed concordance between core biopsy and surgical specimens in breast cancer. Few have analyzed all three clinically relevant markers or the clinical and pathologic features of the cases. Core biopsies have generally been found to reflect the results of the surgical specimen, although staining in the surgical specimen was often lower. A small number of false-negative ER and PR assays on the surgical specimens were reported in many of these studies.7-9,20,21 These authors generally assumed the receptor status of the surgical specimen was the gold standard, and interpreted their results by suggesting that core biopsy specimens can be reliably used for ER assessment but are less reliable for PR assessment. Douglas-Jones et al compared ER assays on 51 pairs of core and surgical specimens and showed 35% lower staining in excised specimens compared with the core biopsies arrayed on the same slide. They also noted that the periphery of the tumors generally stained more intensely than the center. This variation in staining was not seen in core biopsies suggesting that the finding in the surgical specimens was an artifact.6
Examination of staining for ER and PR in the normal epithelial cells on the same slide as the cancer is important in every case in which the tumor cells are negative, since these may serve as an internal positive control. In cases where there is no staining, the possibility of a false-negative is present. However, it may well be a true-negative, and if there is no core biopsy specimen to assess, the clinician will not be able to distinguish between these possibilities. More importantly, and unfortunately, in four of our false-negatives, there was some staining in normal epithelium somewhere on the slide. We hypothesize that this relates to proper fixation of the peripheral normal epithelium with inadequate fixation of the central tumor epithelium.
Possible reasons for different results from core and surgical specimens include sampling error, with the core biopsy not reflecting the status of the entire tumor, and fixation artifact, with the delay in exposure of the center of a surgical specimen to formalin resulting in a false-negative result. Formalin fixation induces cross-links between proteins and nucleic acids that are necessary before IHC analysis. Problems of delayed fixation, underfixation, and overfixation with formalin before IHC have been reported.22 Delayed fixation can allow proteolytic degradation of the antigen—this degradation has been demonstrated in assays for both ER and PR.23 Underfixation means that the periphery of a specimen may be properly fixed while the center is not. This process would lead to more intense staining either at the periphery or the center of the lesion, depending on the antigen and the antibody used.22 Prolonged fixation may lead to weak or absent staining, depending on the individual epitope.23 A recent report by de Marzo et al demonstrated that inadequate formalin fixation may reduce the reliability of IHC staining. In this study, prostate surgical specimens were either processed immediately or fixed for 1 to 8 days in 10% formalin before processing. IHC for p27Kip1 was then performed in a standard manner. They found that the immediate processing was inferior to fixation for 1 or more days, and hypothesized that immediate processing resulted in cross-linking inadequate for IHC. They tested the hypothesis that needle biopsies would result in more reliable IHC and confirmed this to be true.24
Tumor heterogeneity and sampling errors in core biopsy could explain some variations in results but not those cases with strong staining in the core but none in the surgical specimen. The small number of cases where the surgical specimen showed more staining than the core biopsy may be due to sampling.
A third possible explanation for the difference is that the discordantly positive core biopsy assays were false-positives. Our series contained a high number of cases (90%) in which the core biopsy was ER positive, raising this as a possibility. If this were the case, however, one would expect that staining would not be limited to the malignant breast epithelium, and that the staining would be in an unusual pattern. Therefore, we discounted this as a possibility. As the series contained many small and low-grade lesions, the 90% positive rate is not unreasonable.
It is widely accepted that metastases from ER- or PR-negative primary tumors will also be ER or PR negative and not responsive to endocrine manipulation. This assumption may not be valid. A report of 401 patients with metastatic breast cancer who had ER assays both at primary surgery and relapse showed that 76 of 261 patients with ER-positive primaries had ER-negative recurrences, and 46 of 140 ER-negative primaries had ER-positive recurrences.25 Other studies have shown an 8% to 18% rate of conversion of receptor-negative primaries to receptor-positive recurrences.26,27 It is likely that many of the cases of "conversion" from negative to positive were in fact receptor positive from the outset. Our data reinforces the conclusion of some of these studies that routinely confirming the receptor status would be wise, not only because tumors may possibly "redifferentiate," but more importantly, because the original result may have been a false-negative.
Previous studies have reported concordance between core and surgical specimen for HER-2 of approximately 90%.21,28,29 In this study, the major variation in HER-2 staining was 2+ staining of the core but only 1+ staining of the surgical specimen. 2+ staining is followed by FISH assessment of the tumor for HER-2 amplification if treatment with trastuzumab is being considered. Without doing FISH on all of our cases, it is not possible to determine whether core biopsy or surgical specimen HER-2 IHC assessment is more accurate. We do not, therefore, rely on core biopsy IHC for HER-2 when making treatment decisions.
Our results have obvious clinical ramifications. In 9% of cases, relying on the hormone receptor assessment of the surgical specimen may have resulted in inappropriate clinical management. False-negative receptor status risks the denial of endocrine therapy in patients who may benefit, recommendation for chemotherapy where current management guidelines would suggest it is not necessary, or recommendations for chemotherapy regimens not considered optimal.
These results may not to be applicable for all institutions. If surgical specimens are processed in such a way that tumor cells are optimally fixed before processing, the assessment of ER/PR/HER-2 is probably accurate. However, we expect that many institutions will find results similar to ours. Thus, it is unwise to rely on a report of ER or PR negativity from the surgical specimen if the processes of the reporting institution or laboratory are not known.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Richard de Boer, Peter Gibbs, and Julie Miller for their critical review of the manuscript.
NOTES
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Goldhirsch A, Wood WC, Gelber RD, et al: Meeting highlights: Updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol 21:3357-3365, 2003
Coleman RE: Current and future status of adjuvant therapy for breast cancer. Cancer 97:880-886, 2003 (suppl 3)
Lake DE, Hudis C: Aromatase inhibitors in breast cancer: An update. Cancer Control 9:490-498, 2002
Vogel CL, Cobleigh MA, Tripathy D, et al: Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer. J Clin Oncol 20:719-726, 2002
Allred DC, Bustamante MA, Daniel CO, et al: Immunocytochemical analysis of estrogen receptors in human breast carcinomas. Arch Surg 125:107-113, 1990
Douglas-Jones AG, Collett N, Morgan JM, et al: Comparison of core oestrogen (ER) assay with excised tumour: Intratumoral distribution of ER in breast carcinoma. J Clin Pathol 54:951-955, 2001
Zidan A, Christie Brown JS, Peston D, et al: Oestrogen and progesterone receptor assessment in core biopsy specimens of breast carcinoma. J Clin Pathol 50:27-29, 1997
Railo M, Nordling S, Krogerus L, et al: Preoperative assessment of proliferation activity and hormonal receptor status in carcinoma of the breast: A comparison of needle aspiration and needle-core biopsies to the surgical specimen. Diagn Cytopathol 15:205-210, 1996
Jacobs TW, Sizipikou KP, Prioleau JE, et al: Do prognostic marker studies on core needle biopsy specimens of breast carcinoma accurately reflect the marker status of the tumor? Mod Pathol 11:259-264, 1998
Harvey JM, Clark GM, Osborne CK, et al: Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17:1474-1481, 1999
Allred DC, Harvey JM, Berardo M, et al: Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 11:155-168, 1998
Agresti A: 1990 Categorical Data Analysis. New York, NY, Wiley, section 10.4.3
Barnes DM, Millis RR, Beex LVAM, et al: Increased use of immunohistochemistry for oestrogen receptor measurement in mammary carcinoma: The need for quality assurance. Eur J Cancer 34:1677-1682, 1998
Rhodes A, Jasani B, Barnes DM, et al: Reliability of immunohistochemical demonstration of oestrogen receptors in routine practice: Interlaboratory variance in the sensitivity of detection and evaluation of scoring systems. J Clin Pahtol 53:125-130, 2000
Rudiger T, Hofler H, Kreipe HH, et al: Quality assurance in immunohistochemistry: Results of an interlaboratory trial involving 172 pathologists. Am J Surg Pathol 26:873-882, 2002
Newsome JF, Avis FP, Hammond JE, et al: Sampling procedures in estrogen receptor determinations. Ann Surg 193:549-554, 1981
Hasson J, Luhan PA, Kohl MW: Comparison of estrogen receptor levels in breast cancer samples from mastectomy and frozen section specimens. Cancer 47:138-139, 1981
Teicher I, Tinker MA, Auguste LJ, et al: Effect of operative devascularization on estrogen and progesterone receptor levels in breast cancer specimens. Surgery 98:784-791, 1985
Young SC, Burkett RJ, Stewart C: Discrepancy in ER levels of breast carcinoma in biopsy vs mastectomy specimens. J Surg Oncol 29:54-56, 1985
Fentiman IS, Millis RR, Chaudary MA, et al: Effect of the method of biopsy on the prognosis of and reliability of receptor assays in patients with operable breast cancer. Br J Surg 73:610-612, 1986
Connor CS, Tawfik OW, Joyce AJ, et al: A comparison of prognostic tumor markers obtained on image-guided breast biopsies and final surgical specimens. Am J Surg 184:322-324, 2002
Werner M, Chott A, Fabiano A, et al: Effect of formalin tissue fixation and processing on immunohistochemistry. Am J Surg Pathol 24:1016-1019, 2000
von Wasielewski R, Mengel M, Nolte M, et al: Influence of fixation, antibody clones and signal amplification on steroid receptor analysis. Breast J 4:33-40, 1998
De Marzo AM, Fedor HH, Wesely WR, et al: Inadequate formalin fixation decreases reliability of p27Kip1 immunohistochemical staining: Probing optimal fixation time using high-density tissue microarrays. Hum Pathol 33:756-760, 2002
Spataro V, Price K, Goldhirsch A, et al: Sequential estrogen receptor determinations from primary breast cancer and at relapse: Prognostic and therapeutic relevance—The International Breast Cancer Study Group. Ann Oncol 3:733-740, 1992
Li BD, Byskosh A, Molteni A, et al: Estrogen and progesterone receptor concordance between primary and recurrent breast cancer. J Surg Oncol 57:71-77, 1994
Raemaekers JM, Beex LV, Koenders AJ, et al: Concordance and discordance of estrogen and progesterone receptor content in sequential biopsies of patients with advanced breast cancer: Relation to survival. Eur J Clin Oncol 20:1011-1018, 1984
Mueller-Holzner E, Fink V, Frede T, et al: Immunohistochemical determination of HER2 expression in breast cancer from core biopsy specimens: A reliable predictor of HER2 status of the whole tumor. Breast Cancer Res Treat 69:13-19, 2001
Taucher S, Rudas M, Mader RM, et al: Prognostic markers in breast cancer: The reliability of HER2/neu status in core needle biopsy of 325 patients with primary breast cancer. Wein Klin Wochenschr 116:3-5, 2004(G. Bruce Mann, Vanessa D.)
ABSTRACT
PURPOSE: To assess the concordance of breast cancer immunohistochemical receptor assays on core biopsy and surgical specimens.
PATIENTS AND METHODS: We identified 100 patients whose core biopsy and definitive surgical specimens were processed in our hospital. New sections, with core and surgical specimens on the same slides, were stained for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2) immunohistochemistry (IHC). Two pathologists assessed the sections independently. Raw scores and clinically significant groupings were compared.
RESULTS: Concordance for ER, PR, overall hormone receptor (HR), and HER-2 status was seen in 86%, 83%, 90%, and 80% of patients, respectively. The core was positive, while the surgical specimen was negative in 13%, 11%, 9%, and 1% of patients, respectively. Using a log-linear model, differences in ER, PR, and HER-2 staining were all in the direction of stronger staining in the cores, and were statistically significant. Nine percent (95% CI, 4.2% to 16.4%) of women in this group would have had endocrine therapy inappropriately withheld if management decisions were based on surgical specimen results alone.
CONCLUSION: ER and PR assays on core biopsy specimens are more reliable than assays on surgical specimens. Receptor IHC should be performed on core biopsy specimens to avoid patients with HR positive cancers not receiving appropriate hormonal therapy and being overtreated with systemic chemotherapy. Biopsy should be considered in patients with "receptor negative" cancer and recurrent disease.
INTRODUCTION
The status of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2) is critical in the management of patients with invasive breast cancer.1 Endocrine manipulation with tamoxifen, aromatase inhibitors, and ovarian ablation is a central part of adjuvant therapy for early breast cancer or therapy for advanced disease in patients with ER- and/or PR-positive cancers.2,3 Trastuzumab (Herceptin) is an important recent treatment option for advanced disease in patients with HER-2 overexpressing tumors, and is under investigation to determine if it has a role in adjuvant therapy.4 In patients whose tumors show no expression of ER and PR, any form of endocrine therapy is considered ineffective,1 and trastuzumab is not considered in those whose tumors exhibit less than 2+ staining of HER-2.4
Immunohistochemistry (IHC) assays are the basis for determining the ER, PR, and HER-2 status of breast cancers. IHC has replaced ligand-binding assays for ER and PR assessment in most pathology laboratories, after comparative studies demonstrated their superiority in terms of simplicity and reproducibility.5 For HER-2, 3+ staining on IHC correlates with gene amplification and is thus accepted as an indication for trastuzumab treatment. 2+ staining with IHC is followed by fluorescent in situ hybridization (FISH) before trastuzumab treatment.4. Thus, it is critical that IHC to measure ER, PR, and HER-2 status is accurate. The possibility of false-negative assays must be minimized to reduce the chance that potentially life-saving or life-extending relatively nontoxic therapy will be inappropriately withheld.
Some reports have demonstrated that IHC on percutaneous biopsy specimens correlates with those on surgical specimens and can be used for treatment decisions.6-9 We commenced performing IHC on core biopsy specimens to help with decisions regarding induction treatment, and noticed occasional disparity between assay results on core biopsy and surgical specimens. We hypothesized that the fixation process is more complete with core biopsies compared with surgical specimens. This study was designed to address the hypothesis that relying on IHC on surgical specimens would result in false negatives in a significant minority of cases.
PATIENTS AND METHODS
Pathology department records were used to identify 100 consecutive patients treated for invasive breast cancer who had both a core biopsy and surgical specimen reported by the pathology department of the Royal Melbourne Hospital. Patients with multiple tumors were included only if the surgical specimen corresponding to the core biopsy could be identified. The study was approved by the ethics committee of Melbourne Health.
Core biopsies were taken under ultrasound guidance. Multiple passes of a 14-g Manan Pro-Mag 2.2 automatic biopsy system (Manan Medical Products Inc, Northbrook, IL) were made. With larger tumors, various areas of the tumor were sampled. All specimens had been fixed in 10% neutral buffered formalin before processing. Core biopsies were fixed in formalin for a minimum of 3 hours before processing, while specimens received after 2 PM were usually left in formalin overnight. Small surgical specimens were allowed to fix for a minimum of 4 hours, while larger specimens and all mastectomies were fixed for at least 8 hours. Mastectomies were sliced approximately every 5 mm, and paper towel was inserted between the slices to aid fixation. Small excisional specimens were not incised before fixation, while larger specimens were, after inking of the margins. Any specimen from which tissue was taken for tissue banking was incised before being placed in formalin.
Specimens were processed into paraffin blocks using one of two automated processing machines. Breast cores were processed on the Tissue Tek VIP processor (Sakura, Tokyo, Japan). Surgical specimens and mastectomies were processed on the Medite TPC-15 tissue processor (Medite, Burgdorf, Germany). The Medite uses chloroform as its clearing agent as it is gentler on delicate fatty specimens, as opposed to xylene used in the VIP processor. Both processes are accepted as appropriate for processing breast cancer specimens.
Paraffin blocks were retrieved, and sections of the core biopsy, surgical specimen, and a known positive control block for each of the three antibodies were cut at 3 μm and all placed on the same SuperFrost Plus (Menzel-Gl?ser, Braunschweig, Germany) slide (ie, three slides per case were generated, with tumor, core biopsy, and positive control all stained on the same slide for each antibody). Slides were dewaxed and carefully brought to water. Heat-induced antigen retrieval was performed by placing the slides in a coplin jar with 85 mL of citrate buffer (pH 6.0) and heating in a microwave for 10 minutes once boiling point was reached. The slides were then allowed to cool in the coplin jar for 30 minutes.
The biotin-streptavidin immunoperoxidase staining technique with a DAKO autostainer (DAKO Corp, Carpinteria, CA) was used. After dewaxing and antigen retrieval, endogenous peroxidase activity was blocked by placing the slides in distilled water and 30% hydrogen peroxidase for 5 minutes, and then rinsing well in phosphate-buffered saline (PBS). Next, the primary antibody was added. The primary antibodies used were: ER clone 6F11 (Novocastra NCL-6F11, Newcastle, UK) dilution factor 50; PR clone PgR316 (Novocastra NCL-PgR-316) dilution factor 400; and c-erbB-2 (DAKO Polyclonal A0485; DAKO Corp) dilution factor 500.
Incubation continued at room temperature for 30 minutes, followed by washing in PBS. Biotinylated antimouse and antirabbit Ig (DAKO code K0675; DAKO Corp) was applied, and slides incubated at room temperature for 10 minutes, then washed with PBS. Streptavidin-peroxidase (DAKO code K0675; DAKO Corp) was applied, incubated at room temperature for 10 minutes, then washed again with PBS. The peroxidase activity was demonstrated by applying 2 to 3 drops of liquid 3,3-Diaminobenzidine (DAKO K3466; DAKO Corp), and the peroxidase activity developed for 8 minutes and then washed in tap water. The slides were counterstained with Lillee-Mayer hematoxylin, blued in Scott's Tap Water Substitute, dehydrated in ethanol, cleared in xylene, and mounted in DPX mounting medium (Surgipath Medical Laboratories, Richmond, IL).
The slides were labeled with the patient details at the time of paraffin sectioning. Subsequent to staining, each of the three slides for each case was given an allocated case number (1 to 100) and was identified only by this number. The slides were covered so that the core biopsy material was visible but the surgical tissue was obscured. The core biopsy of each case was then scored. Once all the cores had been scored, the core biopsy material was obscured leaving only the surgical material available for scoring. Thus, the individual core and surgical specimens were scored without direct visualization or knowledge of the other staining results. The results were entered on separate Excel (Microsoft, Redmond, WA) spreadsheets.
Two pathologists examined each slide independently. For ER and PR, the proportion of cells staining was quantified, and the intensity of staining was assessed on a 0 to 3+ rating. HER-2 staining was rated on a 0 to 3+ scale. The mean of the two assessments was used for the subsequent analysis. If the assessments of the pathologists differed by more than 20% or by 1 intensity score, or if there was a clinically significant difference in the assessment of the two pathologists, both pathologists reassessed the slides, and a consensus was reached.
For ER and PR, the results were grouped according to the thresholds of 0, less than 10%, and 10%, which are commonly used to designate receptors as negative, equivocal, and positive, respectively.10 For a more informative analysis of the staining differences for ER and PR, we calculated the Allred score.11 This is a scoring system that combines the proportion of cells staining with the intensity of staining to give a single overall rating. An assessment of overall hormone receptor status was also made: if either ER or PR had 10% of cells staining, the tumor was considered receptor positive. For HER-2, the raw scores were tabulated. They were also grouped according to clinical relevance as HER-2 negative (0 and 1+), equivocal (2+), and positive (3+).
To assess whether there was a systematic tendency for core biopsies to show a positive result more often than the corresponding surgical tissue sample, a log-linear model was used to test the symmetry of the comparisons. This model assumes the null hypothesis that where there is disagreement between the two samples, it is equally likely to be in either direction. The alternative hypothesis is that core biopsies are more likely to be positive. This is quantified by estimating the log-odds of the disagreement being in that direction: the null hypothesis corresponds to a parameter value of zero, and a log-odds greater than 0 implies that the core biopsy is more likely to have more staining than the surgical specimen.12 The test was applied separately to the 3 x 3 tables of results for the ER, PR, ER+PR, and HER-2, the 4 x 4 table of results for the raw HER-2 scores and the 9 x 9 tables of Allred ER and PR scores.
For all cases assessed as ER or PR negative on resected specimen, the proportion of cells staining and the intensity of staining was also assessed in nonmalignant breast epithelium present on each slide. This was done to assess the value of such staining as an internal control.
We retrieved clinical records of all patients for whom there was a clinically significant difference to assess the clinical advice that had been given to each patient.
RESULTS
One hundred patients who had both core biopsy and surgical specimens reported at the Royal Melbourne Hospital pathology department between April 1999 and March 2002 were identified. No patient in this series had systemic therapy administered between the core biopsy and definitive surgery.
Demographic features of the patients, and clinical, surgical, and pathologic features of the cancers are presented in Table 1.
The ER status of the surgical specimens showed that 78% were positive and 22% were negative; while for the cores 90% were positive, 1% were equivocal, and 9% were negative (Table 2). For PR of the surgical specimen, 64% were PR positive, 3% equivocal, and 33% negative; while for the cores, the corresponding results were 68%, 7%, and 25% (Table 2). If "hormone receptor–positive" is defined as either ER or PR positive, 82% were positive and 18% were negative on surgical specimens, and 90% of cores were positive, 1% equivocal, and 9% negative (Table 2).
There was concordance between the ER assessment of core and surgical specimen in 86% at the level of ER negative, equivocal, or positive. In 13%, the core was ER positive, whereas the surgical specimen was ER negative, and in 1%, the core was equivocal while the surgical specimen was ER positive (Table 2).
For PR, there was concordance in 83%. In 11%, the core showed some PR staining while the surgical specimen showed no staining. In 6%, there was more staining for PR in the surgical specimen than in the core—in two cases, the core was PR negative while the surgical specimen was equivocal; in one case, the core was PR negative, and the surgical specimen was PR positive; and in three cases, the core was equivocal, and the surgical specimen was PR positive (Table 2).
The Allred scores give an assessment of the overall extent of staining. Clinically, the most important cut point in the Allred score is greater than 2, as these are the cases that are generally considered hormone responsive. Tables 3 and 4 show how the discordance in terms of "positive" and "negative" using this cut point was essentially identical to the discordance using the 10% cutoff presented in Table 2. Analysis of the overall Allred scores shows that most of the discordances were quite substantial, with many case of strong staining in most cells in the core biopsy, with no staining at all in the surgical specimen.
For overall receptor status, there was concordance in 90%. In 9% (95% CI, 4.2% to 16.4%), the core was hormone receptor positive, and the surgical specimen was hormone receptor negative, while in 1%, the core was equivocal, and the surgical specimen was hormone receptor positive (Table 2). Assuming that core biopsies are the gold standard (see Discussion), the false-negative rates for ER, PR, and overall hormone receptor on surgical specimens in this series of 100 patients were 14% (95% CI, 7.9% to 23.4%), 15% (95% CI, 7.6% to 24.7%), and 10% (95% CI, 4.7% to 18.1%), respectively.
Regarding HER-2, 81% of surgical specimens showed 0 or 1+ staining; 12%, 2+; and 7%, 3+. For cores, 65%, 27%, and 8% showed 0 or 1+, 2+, and 3+ staining (Table 5). Concordance for clinically significant staining of HER-2 was seen in 80%. In 17%, the core was equivocal (2+), and the surgical specimen was negative (usually 1+). In 2%, the reverse was true, with a negative core and equivocal surgical specimen, while in the last case the core was positive (3+), and the surgical specimen was negative (Table 5). When the raw scores were compared, there was concordance in 42%. In 54%, the HER-2 staining was more intense in the core than in the surgical specimen, and in 4%, the reverse was the case. The most common finding (36% of cases) was 1+ staining in the core and no staining in the surgical specimen (Table 6).
When the log-linear model was applied to the results, all estimates were in the same direction, with more greater or more intense staining in the core biopsies, and all were significant or highly significant, excepting the reduced PR data. However, there was a highly significant effect in the Allred PR data (Table 7).
Assessment of staining of normal epithelium was difficult, as staining was patchy and variable throughout the specimens. There was some staining of ER and/or PR in normal epithelium in most of the specimens assessed. There were six cases with no staining at all. Four of these were false-negatives and two were true-negatives. There was definite staining in normal epithelium in another four of the false-negatives, and equivocal staining in the final one. There was at least some staining in all but two of the true-negatives.
Review of the nine cases with receptor-negative surgical specimens but receptor-positive cores showed that five had been treated as receptor positive on the basis of results of IHC on the diagnostic core biopsy. Two cases received no adjuvant systemic therapy at all. One was a 14-mm, grade 2 cancer in a 62-year-old patient, while the other was 10-mm, grade 3 cancer in a 72-year-old patient. Both would have been prescribed tamoxifen for 5 years had they been considered receptor positive. The other two cases were given adjuvant chemotherapy. One was a 64-year-old patient with a 12-mm, grade 2 cancer and a micrometastasis in the sentinel node, and the other was a 58-year-old patient with a 12-mm, grade 2 node-negative cancer. Had they been considered receptor positive, each would have been given tamoxifen, and may not have been given chemotherapy.
DISCUSSION
Endocrine therapy and chemotherapy are systemic treatment options for patients with breast cancer. Endocrine therapy is recommended in most patients with hormone receptor–positive cancer because of its efficacy and favorable side effect profile. Chemotherapy is considered in most hormone receptor–negative tumors and for higher risk hormone receptor–positive ones. The risk-benefit balance is generally more complex with chemotherapy, with well-known short-term morbidity including alopecia, nausea, and neutropaenia, as well as long-term risks of cardiac damage and leukemia.1 Accurate and reliable assessment of the ER, PR, and HER-2 status of invasive breast cancers is therefore crucial. While it is accepted that hormone receptor–negative patients should not be considered for endocrine therapy, many studies show that a proportion of these patients do respond to such treatment.13
Many studies have demonstrated that the accuracy and reliability of IHC-based assays on formalin-fixed and paraffin-embedded tissue is similar to the ligand-binding assays of the past.5 Thus, IHC remains excellent in truly experienced hands. There is, however, interlaboratory variation in IHC assessment. A review of reports of IHC assays of ER and PR showed a wide variation in the rates of positivity. For ER, this variation was 25% to 84%, and for PR, it was 24% to 66%.11 Quality-assurance studies of interlaboratory variance in IHC demonstration of ERs have suggested that the assessment is frequently inaccurate. In studies in which paraffin blocks of cancers known to express ER were distributed to various laboratories for staining and reading, false-negative rates of at least 24% were found.14,15 Our study addresses an important derivative of the general issue of quality assurance of receptor assays: if the fixation of surgical specimens results in loss of the antigen, then even a laboratory practicing excellent staining and reading practices will have false-negative results.
The negative predictive values of the assays, defined as the percentage of receptor negative patients showing no response to endocrine therapy, were generally in the order of 80% to 92%.11 While it is possible there are molecular mechanisms whereby truly receptor negative patients respond to endocrine therapy, the results of this study raise the possibility that some of the responding negatives were in fact false negatives.
There is extensive literature from the 1980s comparing the levels of receptors ER and PR by ligand-binding assay on incisional biopsy with that on mastectomy. Across a number of series, the ER and PR level was consistently lower in the mastectomy than the open surgical biopsy groups, and many false-negatives were reported in the assays on mastectomy specimens.16-19 The recommendations were to provide fresh frozen biopsy material for optimal results. Such attention to specimen processing has generally not been given emphasis in the era of IHC.
A number of small studies have addressed concordance between core biopsy and surgical specimens in breast cancer. Few have analyzed all three clinically relevant markers or the clinical and pathologic features of the cases. Core biopsies have generally been found to reflect the results of the surgical specimen, although staining in the surgical specimen was often lower. A small number of false-negative ER and PR assays on the surgical specimens were reported in many of these studies.7-9,20,21 These authors generally assumed the receptor status of the surgical specimen was the gold standard, and interpreted their results by suggesting that core biopsy specimens can be reliably used for ER assessment but are less reliable for PR assessment. Douglas-Jones et al compared ER assays on 51 pairs of core and surgical specimens and showed 35% lower staining in excised specimens compared with the core biopsies arrayed on the same slide. They also noted that the periphery of the tumors generally stained more intensely than the center. This variation in staining was not seen in core biopsies suggesting that the finding in the surgical specimens was an artifact.6
Examination of staining for ER and PR in the normal epithelial cells on the same slide as the cancer is important in every case in which the tumor cells are negative, since these may serve as an internal positive control. In cases where there is no staining, the possibility of a false-negative is present. However, it may well be a true-negative, and if there is no core biopsy specimen to assess, the clinician will not be able to distinguish between these possibilities. More importantly, and unfortunately, in four of our false-negatives, there was some staining in normal epithelium somewhere on the slide. We hypothesize that this relates to proper fixation of the peripheral normal epithelium with inadequate fixation of the central tumor epithelium.
Possible reasons for different results from core and surgical specimens include sampling error, with the core biopsy not reflecting the status of the entire tumor, and fixation artifact, with the delay in exposure of the center of a surgical specimen to formalin resulting in a false-negative result. Formalin fixation induces cross-links between proteins and nucleic acids that are necessary before IHC analysis. Problems of delayed fixation, underfixation, and overfixation with formalin before IHC have been reported.22 Delayed fixation can allow proteolytic degradation of the antigen—this degradation has been demonstrated in assays for both ER and PR.23 Underfixation means that the periphery of a specimen may be properly fixed while the center is not. This process would lead to more intense staining either at the periphery or the center of the lesion, depending on the antigen and the antibody used.22 Prolonged fixation may lead to weak or absent staining, depending on the individual epitope.23 A recent report by de Marzo et al demonstrated that inadequate formalin fixation may reduce the reliability of IHC staining. In this study, prostate surgical specimens were either processed immediately or fixed for 1 to 8 days in 10% formalin before processing. IHC for p27Kip1 was then performed in a standard manner. They found that the immediate processing was inferior to fixation for 1 or more days, and hypothesized that immediate processing resulted in cross-linking inadequate for IHC. They tested the hypothesis that needle biopsies would result in more reliable IHC and confirmed this to be true.24
Tumor heterogeneity and sampling errors in core biopsy could explain some variations in results but not those cases with strong staining in the core but none in the surgical specimen. The small number of cases where the surgical specimen showed more staining than the core biopsy may be due to sampling.
A third possible explanation for the difference is that the discordantly positive core biopsy assays were false-positives. Our series contained a high number of cases (90%) in which the core biopsy was ER positive, raising this as a possibility. If this were the case, however, one would expect that staining would not be limited to the malignant breast epithelium, and that the staining would be in an unusual pattern. Therefore, we discounted this as a possibility. As the series contained many small and low-grade lesions, the 90% positive rate is not unreasonable.
It is widely accepted that metastases from ER- or PR-negative primary tumors will also be ER or PR negative and not responsive to endocrine manipulation. This assumption may not be valid. A report of 401 patients with metastatic breast cancer who had ER assays both at primary surgery and relapse showed that 76 of 261 patients with ER-positive primaries had ER-negative recurrences, and 46 of 140 ER-negative primaries had ER-positive recurrences.25 Other studies have shown an 8% to 18% rate of conversion of receptor-negative primaries to receptor-positive recurrences.26,27 It is likely that many of the cases of "conversion" from negative to positive were in fact receptor positive from the outset. Our data reinforces the conclusion of some of these studies that routinely confirming the receptor status would be wise, not only because tumors may possibly "redifferentiate," but more importantly, because the original result may have been a false-negative.
Previous studies have reported concordance between core and surgical specimen for HER-2 of approximately 90%.21,28,29 In this study, the major variation in HER-2 staining was 2+ staining of the core but only 1+ staining of the surgical specimen. 2+ staining is followed by FISH assessment of the tumor for HER-2 amplification if treatment with trastuzumab is being considered. Without doing FISH on all of our cases, it is not possible to determine whether core biopsy or surgical specimen HER-2 IHC assessment is more accurate. We do not, therefore, rely on core biopsy IHC for HER-2 when making treatment decisions.
Our results have obvious clinical ramifications. In 9% of cases, relying on the hormone receptor assessment of the surgical specimen may have resulted in inappropriate clinical management. False-negative receptor status risks the denial of endocrine therapy in patients who may benefit, recommendation for chemotherapy where current management guidelines would suggest it is not necessary, or recommendations for chemotherapy regimens not considered optimal.
These results may not to be applicable for all institutions. If surgical specimens are processed in such a way that tumor cells are optimally fixed before processing, the assessment of ER/PR/HER-2 is probably accurate. However, we expect that many institutions will find results similar to ours. Thus, it is unwise to rely on a report of ER or PR negativity from the surgical specimen if the processes of the reporting institution or laboratory are not known.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Richard de Boer, Peter Gibbs, and Julie Miller for their critical review of the manuscript.
NOTES
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Goldhirsch A, Wood WC, Gelber RD, et al: Meeting highlights: Updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol 21:3357-3365, 2003
Coleman RE: Current and future status of adjuvant therapy for breast cancer. Cancer 97:880-886, 2003 (suppl 3)
Lake DE, Hudis C: Aromatase inhibitors in breast cancer: An update. Cancer Control 9:490-498, 2002
Vogel CL, Cobleigh MA, Tripathy D, et al: Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer. J Clin Oncol 20:719-726, 2002
Allred DC, Bustamante MA, Daniel CO, et al: Immunocytochemical analysis of estrogen receptors in human breast carcinomas. Arch Surg 125:107-113, 1990
Douglas-Jones AG, Collett N, Morgan JM, et al: Comparison of core oestrogen (ER) assay with excised tumour: Intratumoral distribution of ER in breast carcinoma. J Clin Pathol 54:951-955, 2001
Zidan A, Christie Brown JS, Peston D, et al: Oestrogen and progesterone receptor assessment in core biopsy specimens of breast carcinoma. J Clin Pathol 50:27-29, 1997
Railo M, Nordling S, Krogerus L, et al: Preoperative assessment of proliferation activity and hormonal receptor status in carcinoma of the breast: A comparison of needle aspiration and needle-core biopsies to the surgical specimen. Diagn Cytopathol 15:205-210, 1996
Jacobs TW, Sizipikou KP, Prioleau JE, et al: Do prognostic marker studies on core needle biopsy specimens of breast carcinoma accurately reflect the marker status of the tumor? Mod Pathol 11:259-264, 1998
Harvey JM, Clark GM, Osborne CK, et al: Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17:1474-1481, 1999
Allred DC, Harvey JM, Berardo M, et al: Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 11:155-168, 1998
Agresti A: 1990 Categorical Data Analysis. New York, NY, Wiley, section 10.4.3
Barnes DM, Millis RR, Beex LVAM, et al: Increased use of immunohistochemistry for oestrogen receptor measurement in mammary carcinoma: The need for quality assurance. Eur J Cancer 34:1677-1682, 1998
Rhodes A, Jasani B, Barnes DM, et al: Reliability of immunohistochemical demonstration of oestrogen receptors in routine practice: Interlaboratory variance in the sensitivity of detection and evaluation of scoring systems. J Clin Pahtol 53:125-130, 2000
Rudiger T, Hofler H, Kreipe HH, et al: Quality assurance in immunohistochemistry: Results of an interlaboratory trial involving 172 pathologists. Am J Surg Pathol 26:873-882, 2002
Newsome JF, Avis FP, Hammond JE, et al: Sampling procedures in estrogen receptor determinations. Ann Surg 193:549-554, 1981
Hasson J, Luhan PA, Kohl MW: Comparison of estrogen receptor levels in breast cancer samples from mastectomy and frozen section specimens. Cancer 47:138-139, 1981
Teicher I, Tinker MA, Auguste LJ, et al: Effect of operative devascularization on estrogen and progesterone receptor levels in breast cancer specimens. Surgery 98:784-791, 1985
Young SC, Burkett RJ, Stewart C: Discrepancy in ER levels of breast carcinoma in biopsy vs mastectomy specimens. J Surg Oncol 29:54-56, 1985
Fentiman IS, Millis RR, Chaudary MA, et al: Effect of the method of biopsy on the prognosis of and reliability of receptor assays in patients with operable breast cancer. Br J Surg 73:610-612, 1986
Connor CS, Tawfik OW, Joyce AJ, et al: A comparison of prognostic tumor markers obtained on image-guided breast biopsies and final surgical specimens. Am J Surg 184:322-324, 2002
Werner M, Chott A, Fabiano A, et al: Effect of formalin tissue fixation and processing on immunohistochemistry. Am J Surg Pathol 24:1016-1019, 2000
von Wasielewski R, Mengel M, Nolte M, et al: Influence of fixation, antibody clones and signal amplification on steroid receptor analysis. Breast J 4:33-40, 1998
De Marzo AM, Fedor HH, Wesely WR, et al: Inadequate formalin fixation decreases reliability of p27Kip1 immunohistochemical staining: Probing optimal fixation time using high-density tissue microarrays. Hum Pathol 33:756-760, 2002
Spataro V, Price K, Goldhirsch A, et al: Sequential estrogen receptor determinations from primary breast cancer and at relapse: Prognostic and therapeutic relevance—The International Breast Cancer Study Group. Ann Oncol 3:733-740, 1992
Li BD, Byskosh A, Molteni A, et al: Estrogen and progesterone receptor concordance between primary and recurrent breast cancer. J Surg Oncol 57:71-77, 1994
Raemaekers JM, Beex LV, Koenders AJ, et al: Concordance and discordance of estrogen and progesterone receptor content in sequential biopsies of patients with advanced breast cancer: Relation to survival. Eur J Clin Oncol 20:1011-1018, 1984
Mueller-Holzner E, Fink V, Frede T, et al: Immunohistochemical determination of HER2 expression in breast cancer from core biopsy specimens: A reliable predictor of HER2 status of the whole tumor. Breast Cancer Res Treat 69:13-19, 2001
Taucher S, Rudas M, Mader RM, et al: Prognostic markers in breast cancer: The reliability of HER2/neu status in core needle biopsy of 325 patients with primary breast cancer. Wein Klin Wochenschr 116:3-5, 2004(G. Bruce Mann, Vanessa D.)