Changes in Estrogen Receptor, Progesterone Receptor and Her-2/neu Status with Time: Discordance Rates Between Primary and Metastatic Breast Cancer

Discussion

In this retrospective analysis, significant discordances were found for the hormone receptor status between primary and metastatic breast pathology samples. A change in ER status over time occurred in 17.7% of patients (switching occurred both from ER-positive to -negative and vice versa) and for PR it occurred in 37.3% of patients (all of these tumors lost PR). This meant that a total of 45.1% of patients had some sort of hormone receptor change (i.e. a change in either ER or PR). These results are not dissimilar to what has been previously reported in the literature. In an analysis of 232 patients, Hull et al. found that 19% of patients in their series lost ER and 13% gained ER over time (1). Lower et al. found a discordance rate for ER of 30 % in a chart review of 200 patients with 19.5% of tumours losing ER and 10.5% gaining ER (2). For PR, this group found a discordance rate of 39.3% . Mobbs et al. performed a retrospective pathology specimen review of 129 cases and found discordance rates of 24 and 30% for ER and PR respectively (3). Gross et al., in their series of 161 cases, found that 44% of patients lost PR, however 8% of patients gained PR (5). And finally a meta-analysis of 8 observational studies was performed by Franco et al. totaling 658 paired ER samples and 418 paired PR samples (4). They found a discordance rate of 29 and 27% for ER and PR respectively.

There are less data available on changes in Her-2/neu status with time, however all the series published to date suggest that Her-2/neu status is more stable (6-9). The highest rate of Her-2/neu discordance was found by Zidan et al. in a series of 58 patients whose tumours had been tested by either FISH or IHC (6). They found a discordance rate of 14%, with 12% of patients losing Her-2/neu positivity and 2% gaining it over time. Other published series have found lower rates of Her-2/neu discordance including one series which found that none of the 56 cases analysed had changed status (7). In the presented series of 100 patients with paired primary and metastatic pathology samples, who had had ‘some’ receptor analysis performed on both specimens, a significant discordance rate for Her-2/neu was not found, with only one patient losing Her-2/neu positivity. However, because Her-2/neu testing had not been performed in both specimens from most of these patients, few paired Her-2/neu samples were available to analyse for discordance.

There are three possible reasons for the lack of Her-2/neu testing on the patients included in this series. Firstly, there may have been an assumption by clinicians and pathologists that Her-2/neu status in the metastasis is the same as in the primary, therefore the test was not requested. Secondly, over a third of metastatic samples were collected before 2001, when Her-2/neu testing was not routinely performed. Thirdly, the most common site of metastasis biopsy was bone and there may be technical issues related to Her-2/neu testing in decalcified bone making it more difficult to perform.

Incidentally it was noted that there was a large proportion of patients with T1 and N0 tumours in this series of patients who had all relapsed with metastatic disease. One possible reason for this is that these patients would have been thought to have been at a lower risk of relapse and therefore when they presented with clinical or radiological evidence of metastatic disease, clinicians may have been more eager to obtain a biopsy to confirm or refute the diagnosis of metastatic breast cancer. This may have meant that this series of patients was skewed to be different from the typical population of patients with metastatic breast cancer with respect to their primary tumour characteristics. This is one limitation of this study. There are several other important limitations of this study. Firstly, a central pathology review was not performed. This was a chart review with the receptor status being taken typically from the pathology report and some patients had their primary tumour tested at a different pathology laboratory (commonly another Toronto community hospital) to that where their metastatic sample was analysed (Princess Margaret or Mt. Sinai Hospitals). This analysis is therefore subject to interlaboratory variation. Secondly, for both hormone receptor and Her-2/neu receptor analysis there were often different assay methods used on the primary and metastatic specimens and this could have confounded the results. EIA was the main method utilized for hormone receptor analysis on the primary specimen whereas IHC was almost invariably used when the metastasis was analysed. This represents a change in practice over recent years. FISH (as the definitive test for Her-2/neu) was used proportionally more often in the analysis of the metastasis compared to the primary and this is likely because it is thought to be a more reliable method when a core biopsy is done. Thirdly, a label of either positive or negative was ascribed for each receptor as interpreted from the pathology report. It is possible that a tumour which changes its PR status from being weakly positive to negative may have less clinical meaning than a change of greater magnitude and the analysis did not capture this information. The definition of hormone receptor “positive” varied depending on which laboratory and by what method the analysis was done. As was already mentioned, there was a considerable amount of missing data, especially for Her-2/neu status, and this negatively impacted on the power of this study.

Despite the limitations of this study, this review does add weight to the literature that significant hormone receptor discordance exists for both ER and PR. This has a number of potential clinical implications for the management of patients. Firstly, a proportion of patients with metastatic disease may be sub-optimally treated in the absence of a biopsy. This may be especially so if a patient's receptor status has become positive over time and, if this is not known, they may be deprived of potentially life-prolonging targeted treatment such as endocrine therapy. Secondly, patients may be inappropriately enrolled in clinical trials of systemic therapy. Many such trials are powered to detect small differences in efficacy between therapies. It is therefore possible that because of this discordance phenomenon, unknown imbalances in receptor status between the arms of trials (as eligibility is often based on the status of the primary tumour) may influence the results.

One strong theme that is perpetuated in both this article and elsewhere is the high rate of “loss of PR” over time. Out of all three receptor discordance phenomena discussed here, loss of PR is the one that has generated the most biological research. PR is an ER-regulated gene which is expressed in two isoforms, PR-α and -β. PR mediates the effects of progesterone on the development of mammary glands in healthy individuals and is implicated in the development of breast cancer. There is an increased breast cancer incidence in patients taking combined estrogen and progesterone hormone replacement therapy compared to estrogen therapy alone (13). ER-positive/PR-negative metastatic tumours tend to have a more aggressive course and are associated with a reduced overall survival compared to those retaining PR (5, 14). Therefore it is not surprising that there are more PR-negative tumours amongst the metastatic samples compared to the primaries. There is an increased association between Her-2/neu overexpression and ER-positive/PR-negative tumours (25%) compared to ER-positive/PR-positive (10%) tumours (15). There is also an association between the ER-positive/PR-negative phenotype and high tumour grade (14). Furthermore ER-positive/PR-negative tumours have higher levels of EGFR expression compared to ER-positive/PR-positive tumours (16). These associations with the ER-positive/PR-negative phenotype implicate high growth factor activity with a decrease or loss of PR. Therefore one of the predominating theories explaining loss of PR is growth factor-mediated downregulation of PR, independent of ER (17).

There are several other postulated mechanisms for loss of PR. One is genetic loss of the PR gene locus or loss of heterozygosity. Another is hypermethylation of the PR promoter, which is found in 21-40% of ER-positive/PR-negative tumours compared with none of the tumours which exhibit the ER-positive/PR-positive phenotype (18). One other theory attempting to explain the ER-positive/PR-negative phenotype is that in these tumours ER is non-functional and therefore unable to stimulate PR production (19). However, some ER-positive/PR-negative tumours may simply result from low levels of circulating endogenous estrogens in post-menopausal patients that are insufficient to induce PR expression even though the ER pathway is intact (20).

In conclusion, this study has demonstrated significant discordance for hormone receptor status between primary and metastatic specimens and this may have implications for the systemic treatment of patients' metastatic disease. In particular, loss of PR was found in 37.3% of paired samples, and this may reflect (or even cause) a shift to a more aggressive tumour biology. Her-2/neu status seems more stable than either ER or PR although this analysis was not sufficiently powered, because of missing data, to directly assess this. This was not a prospective study and therefore this analysis carries the burden of all the pitfalls inherent to a retrospective review. The Authors' group has recently completed what is understood to be the first prospective study assessing the rate and practical implications of hormone and Her-2/neu receptor discordance (i.e. asking the question “Does a confirmatory metastasis biopsy alter clinical management?”). The preliminary results available on twenty-nine individuals have recently been presented and demonstrate a discordance rate for hormone receptor status of 40% and for Her-2/neu of 8% (two patients becoming positive for Her-2/neu in the metastatic sample) (21). The results of a confirmatory biopsy altered management in 20% of patients. These preliminary results further support the phenomenon of receptor discordance and reinforce the importance of obtaining a confirmatory biopsy when patients present with suspicion of metastatic disease.