Postanalytical Issues in HR Expression in Breast Cancer
Postanalytical Issues in HR Expression in Breast Cancer
The current mainstay of hormone receptor assessment in breast cancer is immunohistochemistry (IHC). Since the 1960s, when Jensen et al first characterised the ER protein and in the 1970s when McGuire and colleagues confirmed its predictive and prognostic utility in breast cancer, ER was initially quantified using ligand-binding assays (LBAs) that required fresh tumour tissue to be homogenised and centrifuged, producing a cytosol whose ER content was assessed by subjecting the cytosol to competitive binding with radioactive oestradiol-17β and non-radioactive oestrogen mimics. In this method, dextran-coated charcoal was frequently used to separate bound oestrogen; hence, this has often been regarded as synonymous with LBAs.
With availability of polyclonal and later monoclonal antibodies to ER and recognition of the many advantages the immunohistochemical platform possesses over LBAs, IHC has become a standard tool in hormone receptor assessment.
PR is considered a part of the post-ER pathway, and the presence of PR is believed to reflect functionality of ER. It has been debated, however, whether PR assessment is routinely necessary in breast cancer, with some studies suggesting that the predictive importance of PR may not be as relevant as ER. In the UK, the National Institute for Clinical Excellence commented that PR expression may have limited prognostic value and therefore need not be a mandatory diagnostic requirement for women with early breast cancer. In contrast to this view, several authors have discovered evidence of independent prognostic impact for PR, and in many laboratories, PR assessment is an integral part of the so-called 'breast panel' applied to newly diagnosed invasive breast cancers. In this discussion, both ER and PR evaluation are considered together.
ER and PR are nuclear proteins and the expression is assessed in nuclei of tumour cells. Cytoplasmic staining is discounted and may reflect issues with the analytical phase of testing. Nevertheless, extranuclear staining has been reported more frequently in PR than ER, and believed to reflect crosstalk with Akt/HER-2-signalling pathways and activation of aromatase.
The proportion of tumour cells stained, the intensity of staining and a final conclusion of a positive or negative result are the key parameters that should be provided in a pathology report on ER and PR expression in invasive breast cancer, together with antibody and fixation details. Assessment should be made on a representative part of the tumour that is well fixed, and which includes adjacent benign breast tissue.
The proportion or percentage of tumour cells that expresses hormone receptors can be visually estimated, quantified by counting positively stained cells or via image analysis. Although image analysis may improve intraobserver and interobserver variability, it is neither widely available nor validated extensively.
For positively stained nuclei, staining intensity is graded as weak (1+), moderate (2+) or strong (3+) based on the average staining intensity as compared with positive controls (figure 1). With information on the proportion and intensity of tumour cells stained, various scores can be calculated. The H-score, a composite of the different proportions of tumour cells stained with varying intensities, is used by some institutions to provide a more quantitative result. It has a range of 0–300 and is calculated as follows: (3×percentage of strong nuclear staining)+(2×percentage of moderate nuclear staining)+(1×percentage of weak nuclear staining). In one study, a positive cut-off of H-score of 50 and above is approximated to ≥10% tumour cells with at least 2+staining intensity. Tumours with H-scores between 50 and 99 were considered as receptor poor and those exceeding 100 were regarded as receptor rich.
(Enlarge Image)
Figure 1.
Tumour cells in the rectangles A–C represent examples of 1+ (weak), 2+ (moderate) and 3+ (strong) nuclear staining, respectively, for oestrogen receptor (Neomarker RM9101-S, SP1, dilution 1 : 50).
A quick score was proposed as a faster alternative to the H-score without the need for counting cells.
The Allred score is similarly based on the proportion and intensity of tumour cells that are positively stained, with values of 0–5 for various proportions (0, no staining; 1, <1%; 2, 1–10%; 3, 10–33%; 4, 33–66%; and 5, >66% tumour cells stained) and 0–3 for different staining intensities (0, no staining; 1, weak staining; 2, moderate staining; and 3, strong staining). These values are then added to give the Allred score, with scores of 3 and above representing receptor positivity and predicting response to hormonal therapy.Table 1 summarises the commonly used scoring methods.
Benign breast epithelium present in tumour sections subjected to IHC serves as a useful internal positive control, and this is why block selection for hormone receptor staining ought to include adjacent benign tissue. It is recommended that repeat testing is carried out when there is no staining of benign epithelium.
In the absence of an internal positive control, an external positive control should be run with each batch. In some laboratories, the external control tissue is placed on the same slide as the tissue to be tested, to ensure identical analytical conditions.
ER and PR status should be established on all invasive breast cancers that are newly diagnosed. For women with multiple synchronous tumours, it has been recommended that receptor status is determined in the largest lesion, which stands to reason as T-staging is also based on the largest tumour size. When there are two distinct tumours with different morphological appearances and histological grades, some authors would advocate testing for ER and PR in both lesions.
Although hormone receptor status is always accomplished and reported on the invasive component of a tumour that harbours invasive and in situ disease, there is increasing acknowledgment that it is also beneficial to determine ER and PR status for pure ductal carcinoma in situ (DCIS) as well. This trend is largely based on the findings from a subset analysis of the NSABP B-24 clinical trial that found a significant reduction in subsequent development of either ipsilateral or contralateral breast cancer in women originally diagnosed with ER-positive DCIS who were treated with tamoxifen. It is uncertain whether the same hormone receptor thresholds and cut-offs have similar biological impact for DCIS, even though it would seem reasonable to adopt analogous principles to achieve a standardised approach. Although it is considered advantageous and is recommended by some experts, routine hormone receptor assessment for DCIS is still not regarded as a mandatory requirement.
As part of the quality process of ensuring that interpretation of ER and PR results is reproducible among pathologists, regular slide circulations to assess concordance, gather data on interpretive differences or detect shifts in reporting trends are advisable.Table 2 shows quality assurance and proficiency testing results of various studies.
Postanalytical Assessment
The current mainstay of hormone receptor assessment in breast cancer is immunohistochemistry (IHC). Since the 1960s, when Jensen et al first characterised the ER protein and in the 1970s when McGuire and colleagues confirmed its predictive and prognostic utility in breast cancer, ER was initially quantified using ligand-binding assays (LBAs) that required fresh tumour tissue to be homogenised and centrifuged, producing a cytosol whose ER content was assessed by subjecting the cytosol to competitive binding with radioactive oestradiol-17β and non-radioactive oestrogen mimics. In this method, dextran-coated charcoal was frequently used to separate bound oestrogen; hence, this has often been regarded as synonymous with LBAs.
With availability of polyclonal and later monoclonal antibodies to ER and recognition of the many advantages the immunohistochemical platform possesses over LBAs, IHC has become a standard tool in hormone receptor assessment.
PR is considered a part of the post-ER pathway, and the presence of PR is believed to reflect functionality of ER. It has been debated, however, whether PR assessment is routinely necessary in breast cancer, with some studies suggesting that the predictive importance of PR may not be as relevant as ER. In the UK, the National Institute for Clinical Excellence commented that PR expression may have limited prognostic value and therefore need not be a mandatory diagnostic requirement for women with early breast cancer. In contrast to this view, several authors have discovered evidence of independent prognostic impact for PR, and in many laboratories, PR assessment is an integral part of the so-called 'breast panel' applied to newly diagnosed invasive breast cancers. In this discussion, both ER and PR evaluation are considered together.
Cellular Localisation of Staining
ER and PR are nuclear proteins and the expression is assessed in nuclei of tumour cells. Cytoplasmic staining is discounted and may reflect issues with the analytical phase of testing. Nevertheless, extranuclear staining has been reported more frequently in PR than ER, and believed to reflect crosstalk with Akt/HER-2-signalling pathways and activation of aromatase.
Assessment of Staining
The proportion of tumour cells stained, the intensity of staining and a final conclusion of a positive or negative result are the key parameters that should be provided in a pathology report on ER and PR expression in invasive breast cancer, together with antibody and fixation details. Assessment should be made on a representative part of the tumour that is well fixed, and which includes adjacent benign breast tissue.
The proportion or percentage of tumour cells that expresses hormone receptors can be visually estimated, quantified by counting positively stained cells or via image analysis. Although image analysis may improve intraobserver and interobserver variability, it is neither widely available nor validated extensively.
For positively stained nuclei, staining intensity is graded as weak (1+), moderate (2+) or strong (3+) based on the average staining intensity as compared with positive controls (figure 1). With information on the proportion and intensity of tumour cells stained, various scores can be calculated. The H-score, a composite of the different proportions of tumour cells stained with varying intensities, is used by some institutions to provide a more quantitative result. It has a range of 0–300 and is calculated as follows: (3×percentage of strong nuclear staining)+(2×percentage of moderate nuclear staining)+(1×percentage of weak nuclear staining). In one study, a positive cut-off of H-score of 50 and above is approximated to ≥10% tumour cells with at least 2+staining intensity. Tumours with H-scores between 50 and 99 were considered as receptor poor and those exceeding 100 were regarded as receptor rich.
(Enlarge Image)
Figure 1.
Tumour cells in the rectangles A–C represent examples of 1+ (weak), 2+ (moderate) and 3+ (strong) nuclear staining, respectively, for oestrogen receptor (Neomarker RM9101-S, SP1, dilution 1 : 50).
A quick score was proposed as a faster alternative to the H-score without the need for counting cells.
The Allred score is similarly based on the proportion and intensity of tumour cells that are positively stained, with values of 0–5 for various proportions (0, no staining; 1, <1%; 2, 1–10%; 3, 10–33%; 4, 33–66%; and 5, >66% tumour cells stained) and 0–3 for different staining intensities (0, no staining; 1, weak staining; 2, moderate staining; and 3, strong staining). These values are then added to give the Allred score, with scores of 3 and above representing receptor positivity and predicting response to hormonal therapy.Table 1 summarises the commonly used scoring methods.
Internal and External Positive Controls
Benign breast epithelium present in tumour sections subjected to IHC serves as a useful internal positive control, and this is why block selection for hormone receptor staining ought to include adjacent benign tissue. It is recommended that repeat testing is carried out when there is no staining of benign epithelium.
In the absence of an internal positive control, an external positive control should be run with each batch. In some laboratories, the external control tissue is placed on the same slide as the tissue to be tested, to ensure identical analytical conditions.
Component of Tumour to be Assessed
ER and PR status should be established on all invasive breast cancers that are newly diagnosed. For women with multiple synchronous tumours, it has been recommended that receptor status is determined in the largest lesion, which stands to reason as T-staging is also based on the largest tumour size. When there are two distinct tumours with different morphological appearances and histological grades, some authors would advocate testing for ER and PR in both lesions.
Although hormone receptor status is always accomplished and reported on the invasive component of a tumour that harbours invasive and in situ disease, there is increasing acknowledgment that it is also beneficial to determine ER and PR status for pure ductal carcinoma in situ (DCIS) as well. This trend is largely based on the findings from a subset analysis of the NSABP B-24 clinical trial that found a significant reduction in subsequent development of either ipsilateral or contralateral breast cancer in women originally diagnosed with ER-positive DCIS who were treated with tamoxifen. It is uncertain whether the same hormone receptor thresholds and cut-offs have similar biological impact for DCIS, even though it would seem reasonable to adopt analogous principles to achieve a standardised approach. Although it is considered advantageous and is recommended by some experts, routine hormone receptor assessment for DCIS is still not regarded as a mandatory requirement.
Reproducibility and Quality Issues
As part of the quality process of ensuring that interpretation of ER and PR results is reproducible among pathologists, regular slide circulations to assess concordance, gather data on interpretive differences or detect shifts in reporting trends are advisable.Table 2 shows quality assurance and proficiency testing results of various studies.