Estrogen Receptor Testing of Breast Cancer in Current Clinical Practice: What's the Question?

Assessment of estrogen receptor (ER) status is an essential component of the evaluation of breast cancers. Although ER status provides prognostic information, currently the major clinical value of determining ER status is to assess the likelihood that a patient will respond to endocrine therapy. The first ER assays used routinely in the clinical setting were radiolabeled ligand binding assays (LBAs), such as the dextran-coated charcoal method. Assays of this type had the disadvantage of requiring prospective collection of fresh tumor tissue, among other drawbacks, but resulted in quantitative determination of ER content of the tumor, expressed as femtomoles of ER protein per milligram of cytosol protein. When determined using LBA, breast cancers were found to exhibit a broad range of values for ER, and the magnitude of the benefit from endocrine therapy was shown to be related to the quantity of ER protein in the tumor. During the last two decades, mammographic screening resulted in such a dramatic decrease in the size of the average breast cancer that it became impossible to collect tissue prospectively for ER determination by LBA in many cases because of the small tumor size. The availability of antibodies that recognize ER in formalin-fixed, paraffin-embedded (FFPE) tissue permitted the development of immunohistochemical assays to detect ER retrospectively in such specimens, and immunohistochemistry (IHC) in FFPE tissue gradually replaced LBA as the most common method for determination of ER status of breast cancers. Several studies have indicated that ER status as determined by IHC is not only predictive of response to endocrine therapy, but that the ability of ER status as determined by IHC to predict such responses is superior to that of ER status as determined by LBA. Thus, the use of IHC to assess the ER status of breast cancers in paraffin sections is now a routine part of pathology practice worldwide. As a consequence of the experience with LBA, there had been (and still remains) the expectation that IHC assays for ER should result in a broad range of values among ER-positive patients, similar to that observed with LBA. However, this presupposes that there is a direct, linear relationship between the amount of ER protein present in the tumor cells and the amount of ER antigen detected by IHC. Although some studies have certainly suggested such a relationship, others have not. For example, two recent studies that together included the analysis of almost 7,000 breast cancers found that the distribution of ER values using contemporary IHC methodology was essentially bimodal, with more than 90% of tumors being either completely ER negative or unequivocally and strongly ER positive. These data clearly are at odds with the continuum of values observed when ER status was determined by LBA. However, this may simply reflect the fact that IHC is not an intrinsically quantitative method, and that there is not a direct, linear relationship between the intensity and distribution of the chromogenic reaction product as determined by IHC and the amount of ER protein in breast cancer cell nuclei when highly sensitive anti-ER antibodies and detection systems are used on adequately fixed tissue samples. In fact, the relationship between the actual quantity of ER protein in the tumor cell nuclei and the apparent amount of ER antigen demonstrated by IHC assays is highly complex and may be as much a function of preanalytic factors (such as details of tissue fixation and processing) and assay sensitivity as of the actual amount of antigen present in the tumor cells. A number of recent studies that support this contention are particularly noteworthy. Rhodes et al, in an analysis of data from 66 laboratories participating in a United Kingdom external quality assurance program, found that ER IHC staining results were highly affected by the efficiency of the antigen retrieval step, and that this was, in fact, the single most important factor contributing to interlaboratory reproducibility. Goldstein et al noted that IHC staining results for ER were highly dependent on the time of tissue fixation. Using the assay employed in their laboratory, the minimum fixation time for optimal ER IHC staining results was 6 to 8 hours, regardless of specimen type or size. Of note in that study, underfixation had more of a detrimental effect on IHC staining results than did overfixation. Vassallo et al performed IHC assays for ER on 20 invasive ductal carcinomas using two different anti-ER antibodies (1D5 and 6F11), two different antigen retrieval methods, and three different detection systems. Thus, for each patient, 12 different technical variations were studied. ER IHC results were scored semiquantitatively on a scale of 0 to 4. In five of these 20 patients, the ER scores varied from 0 to 4, and in three patients, the scores ranged from 1 to 4, depending on the assay conditions used. Moreover, there was not a single patient in which all 12 assay variations resulted in the same IHC staining score. Umemura et al studied 44 breast cancers with a biochemical assay and with two different IHC assays, one considered by the authors to be “highly sensitive” and the other “nonhighly sensitive.” The same anti-ER antibody (1D5) was used in both IHC procedures. These JOURNAL OF CLINICAL ONCOLOGY COMMENTS AND CONTROVERSIES VOLUME 24 NUMBER 12 APRIL 2