- © 2012 by American Society of Clinical Oncology
Low–Estrogen Receptor–Positive Breast Cancer: The Impact of Tissue Sampling, Choice of Antibody, and Molecular Subtyping
- Emad A. Rakha⇓,
- Andrew H.S. Lee,
- Joseph Roberts,
- Nadia M. Villena Salinas,
- Zsolt Hodi,
- Ian O. Ellis and
- Jorge S. Reis-Filho
- Nottingham University Hospitals National Health Service Trust; and University of Nottingham, Nottingham City Hospital, Nottingham, United Kingdom
- The Institute of Cancer Research, London, United Kingdom
- Corresponding author: Dr Emad Rakha, Department of Histopathology, University of Nottingham and Nottingham University Hospital NHS Trust, City Hospital Campus, Hucknall Rd, Nottingham, NG5 1PB, UK; e-mail: emadrakha{at}yahoo.com.
To the Editor:
Iwamoto et al1 have analyzed the clinically important subgroup of estrogen receptor (ER) –positive breast cancers in which only 1% to 9% of neoplastic cells expressed ER by immunohistochemistry. The microarray gene-expression profiles of low–ER-expressing (1% to 9%) early breast cancers were elegantly investigated. The authors concluded that approximately 25% of the 1% to 9% ER immunohistochemistry-positive breast cancers are potentially endocrine-sensitive cancers, and a second RNA-based ER assessment may help identify these ESR1 mRNA-positive immunohistochemistry-borderline breast cancers. Although we agree with the main conclusions of this study, some issues need to be considered.
Iwamoto et al1 defined ER subgroups on the basis of immunohistochemical assessment of ER on needle core biopsies (NCBs). Although a high concordance rate for the assessment of ER on NCBs and surgical excision specimens (SESs) has been established,2 it is important to note the limitations of this approach. Several studies have shown that carcinomas with 1% to 9% ER staining on NCBs frequently showed greater than 10% or 0% staining when the ER assessment was repeated on SESs.3–10 Furthermore, we have recently reviewed 3,200 cases of breast cancer consecutively reported in routine practice in the past 4 years. Thirsty-six tumors displayed ER expression in 1% to 9% of neoplastic cells on NCB. Of these, 17 cases had ER immunohistochemistry repeated on SESs (eight cases [47%] were negative [score, 0], and four cases [23.5%] showed > 10% staining on SES). It is our current practice to repeat staining on SESs in cases that show 1% to 9% ER expression.4
In Iwamoto et al,1 ER immunohistochemical assessment was based on a single antibody (ie, the monoclonal antibody 6F11). Previous studies have demonstrated that the sensitivity of detection of ER expression is related to the quality of fixation and antigen retrieval methods and the clone of antibody used.11–13 Although 6F111 is widely used and has been validated in several studies, we have experienced some difficulties with the interpretation of staining patterns, and in some instances, we observed weak granular/punctate nuclear staining on NCBs that gave false-positive results. We have identified 18 cases with this staining pattern in our routine practice and repeated the ER assessment by using different antibodies on SESs that yielded negative results (Figs 1 and 2), which suggested that the pattern observed with 6F11 is likely to constitute a false-positive staining.
Immunohistochemical staining of estrogen receptor using the 6F11 clone displaying granular/punctate brown nuclear staining in (A) the primary tumor and (B) the nodal metastatic deposit (false positive; magnification: [A] 100×, [B] 200×).
Same case as in Figure 1A was stained in the same laboratory under the same staining conditions but with a different estrogen receptor antibody (clone SP1) showing negative nuclear staining (true negative; magnification: 100×).
Importantly, of the ER immunohistochemistry-positive (> 10%) cases in the study of Iwamoto et al,1 7.6% of cases were classified as ER-negative on the basis of assessment of the level of ESR1 gene expression, and 27.9% of cases were classified as basal like or normal breast like by using a research version of PAM50.1 Furthermore, 8.7% of ER-negative cases (score = 0%) were classified as positive on the basis of ESR1 gene assessment. Hence, it remains to be determined how the results of a second RNA-based ER assessment or intrinsic subtype classification would resolve the final classification of cancers with ER immunohistochemical expression between 1% to 9% because gene-expression profiling may not have sufficient accuracy to be used as a gold standard for the definition of true ER-positive breast cancers.14
Although fine-needle aspiration biopsies have been proven to be an effective way of retrieving breast cancer cells for microarray-based studies of gene-expression profiling, the number of cells analyzed per case is relatively limited.15 If breast cancers with ER expression that ranges from 1% to 9% truly represent cases in which only a small nonmodal population of cancer cells expresses ER, the size of the cell population sampled becomes an important issue in molecular profiling studies if the ER immunohistochemistry-positive cells are restricted to a given compartment of the tumor. This is particularly relevant because of the recent observations16 that microarray profiling of distinct areas from a tumor may result in distinct gene-expression profiles. Furthermore, microarrays provide the average gene expression of a given population of cancer cells, and the results are largely representative of the modal population of cancer cells within a tumor; if a subpopulation of cancer cells account for a small minority of cells (eg, < 10%), it is unlikely that their transcriptomic features will be reliably captured by microarray profiling. Therefore, it is plausible that some low–ER-content samples were classified as basal-like or normal breast-like subtypes and of low sensitivity to endocrine therapy as a result of enrichment with ER-negative cells from a tumor classified as ER immunohistochemistry-positive on the basis of the definition of American Society of Clinical Oncology/College Of American Pathologists guidelines.17
On the basis of our evaluation, we propose that cases with 1% to 9% of ER immunohistochemistry-positive cells in NCBs should be re-evaluated in the corresponding SES, and clinical decisions should be based on the results obtained from the latter because of the reported discordant results between results obtained with NCBs and SESs.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Andrew H.S. Lee, Roche Products Research Funding: Andrew H.S. Lee, Roche Products Expert Testimony: None Other Remuneration: None
