The rate of estrogen receptor‐conversion associated with tumor progression in estrogen receptor‐positive breast cancer patients following adjuvant Tamoxifen administration

Abstract Background Hormone Receptor (HR)‐discordance between primary breast cancer and metastasis is a known biological phenomenon. Discordance studies usually comprise a heterogeneous group of HR‐positive and negative patients and allow for the comparison of changes in HR‐status from the primary to the recurrent disease. However, in a clinical setting, the rate of estrogen receptor‐conversion following endocrine therapy with agents such as Tamoxifen (TAM) in estrogen receptor‐positive cancers is of primary interest as opposed to total receptor discordance. Aim To investigate the rate of estrogen receptor‐conversion associated with tumor progression in estrogen receptor‐positive breast cancer patients following adjuvant TAM administration and to compare the results with the meta‐analysis data of HR‐discordance studies. Methods and Results A retrospective double‐center review of biomarkers in 67 estrogen receptor‐positive breast cancer patients who underwent TAM treatment in the adjuvant setting. The estrogen and progesterone receptor‐status were compared at the time of diagnosis and following relapse and the Disease‐free Survival, mean duration of TAM treatment as well as the operative, radiation, and cytotoxic therapies registered before TAM treatment, were recorded. Initially, all patients were estrogen receptor‐positive. The average age at the time of diagnosis was 52.8 ± 12.4 years. After recurrence, only 47 patients (70.1%) were still estrogen receptor‐positive with a highly significant loss of estrogen receptor‐expression in 29.9% of cases. The mean duration of TAM treatment was 40.7 ± 19.9 months. 45 patients (i.e., 67.2%) progressed during the TAM treatment and the remaining 22 patients (32.8%) developed relapse after the TAM treatment had finished. Initially, there were 82.1% progesterone receptor‐positive and 17.9% progesterone receptor‐negative, but after relapse the progesterone receptor‐positive cases diminished significantly to 53.7%, showing a progesterone receptor‐loss of 28.4%. Conclusion The rate of estrogen receptor‐loss associated with tumor progression following TAM treatment is approximately 30%, which is of clinical relevance in order to evaluate further endocrine efficacy in these patients. This rate of receptor conversion is roughly 6‐13% higher compared to the recently published meta‐analysis data of discordance studies. This discrepancy could possibly be due to anti‐hormonal therapy with TAM accentuating receptor conversion.


Conclusion:
The rate of estrogen receptor-loss associated with tumor progression following TAM treatment is approximately 30%, which is of clinical relevance in order to evaluate further endocrine efficacy in these patients. This rate of receptor conversion is roughly 6-13% higher compared to the recently published meta-analysis data of discordance studies. This discrepancy could possibly be due to anti-hormonal therapy with TAM accentuating receptor conversion. Immuno-Histo-Chemistry (IHC) testing. This is currently the cut-off for HR-positive cancers in the clinical setting and is in accordance with major guidelines such as ASCO, ESMO, German AGO, and the St. Gallen International expert consensus for the use of endocrine therapy in these patients. 5,6 Formerly, the cut-off limit was set at 10% positive cells in IHC, which from a study point of view is much better for distinguishing between real positive HR-status and borderline cases. Nonetheless, the ER-expression serves as a predictive and prognostic biomarker and ER-positivity has been found to confer a survival advantage in patients treated with endocrine therapy, specifically TAM, which has not been observed in the ER-negative disease. 7 Microarray studies differentiate gene expression patterns of cells and have helped to distinguish breast cancer into various intrinsic subtypes. These different subtypes are referred to as Luminal A, Luminal B, HER2-enriched, Triple-negative/basal-like, or normal-like.
Some gene expression subtypes that are referred to as Luminal A, Luminal B, normal-like, code for HR-positive breast cancers and present a major treatment option in these patients for AHT. 8,9 There are genetic events that alter the generation of estrogen and the metabolism of TAM in breast cancer patients. A recent report indicated that the aromatase gene responsible for the generation of estradiol in postmenopausal women, called CYP19A1, may be amplified in endocrine-resistant breast cancer. 10 Another known phenomenon is the metabolism of TAM to a more active metabolite, called Endoxifen, via cytochrome P450 2D6 (CYP2D6). 11 Evidence suggests that polymorphism, existing for the CYP2D6-gene, can lead to either a reduced or increased metabolism to the active metabolite Endoxifen and hence show better or worse clinical outcomes in these cancer patients.
Although there has been enormous progress in the treatment of HR-positive breast cancer, the resistance to drug treatment remains a problem. The Early Breast Cancer Trialists Collaborative Group 12 recently reported a meta-analysis based on 20 years follow-up of 88 clinical trials that involved 62 923 women with ER-positive breast cancer treated for 5 years with AHT, mainly TAM. Breast cancer recurrences occurred at a steady rate throughout the period of 5-20 years. There are several mechanisms that HR-positive breast cancers may adapt to resist treatment. The most common mechanism for the lack of efficacy of AHT seems to be the loss of the HR-expression, dysregulation of co-regulators, and cross talk with growth factor signaling pathways. 2,13 Furthermore, the loss of the ER from the adjuvant to the metastatic setting seems to be associated with worse overall survival and therefore the ER-status seems to be a predictor of survival, whereas changes to the PR from the adjuvant to the metastatic setting were not associated with a change in survival. 14 There are several studies that have shown the ER-evolution in HR-positive breast cancer patients. The retrospective study by Broom et al 15 comprised the data of 100 patients with breast cancer, of which 73 were initially ER-positive and were consequently followed up. After progression, they observed a change in ER-status in 17.7% of cases (switching occurred both from ER + !ERand vice versa) and a reduction of 37.3% for PR (all tumors lost the PR). The problem with the study was that the authors did not elaborate precisely on the adjuvant treatment that each patient received, presumably AHT for the ER-positive patients, and since the cohort chosen was heterogenic in terms of ER-status, there were also some patients with increased ER-expression after relapse. Hull et al 16 18 encompassing 459 patients showed a conversion rate from primary ER-positive to relapse ER-negative as being at 24.6% and ERpositive staying as ER-positive at 47.1%, while finding that the conversion rate from primary ER-negative to relapse ER-positive was 7.8% and that of primary ER-negative staying ER-negative was 20.5%.
It should be noted, that none of the above studies elaborated exactly on the administration and duration of the AHT in their collectives.
In many of these studies, the common denominator was always the presence of HR-negative patients in their initial collective, some of which turned from ER-or PR-negative to positive over time. As one has to assume that the HR-negative patients did not receive AHT, a switch from negative to positive may be better explained by the tumor biology of progression, rather than the specific clinical effects that one might observe due to adjuvant therapy.
Furthermore, clinically the total rate of discordance is of limited value, since it measures changes to ER-status from ER-positive to ERnegative and vice versa as well as changes from PR-positive to PRnegative and vice versa. For the clinician, it is more pragmatic to know the rate of ER-loss that is, ER + !ER-, following adjuvant AHT with endocrine agents such as TAM, as this information is pertinent for the use of endocrine therapy in relapsed patients.
The aim of the study was to investigate the rate of ER conversion in ER-positive breast cancer patients from the adjuvant to recurrent disease, when subjected to TAM therapy, as one would find in a clinical setting. At the same time, any concomitant changes that might occur to PR-status in this patient cohort were monitored. Finally, the data were compared with the published meta-analysis of discordance studies from 2018.

| Data source
The study was a retrospective double-center study and the acquisition  The patients that were found to qualify for the cohort study were then reviewed and the required demographic information was extracted from their medical records as listed in Table 1.
All patients selected had breast cancer stage I-III with a complete TNM classification (AJCC/UICC). All patients underwent an R0 resection of primary cancer. The data does not specify the type of lymph node resection carried out. Patients that had documented interruptions in their AHT with TAM were excluded from the study.  Table 2.

| Histopathological tumor grading and HRstatus
For the evaluation of the ER-and PR-status in our collective of breast cancer patients, the tumor was classified as HR-positive, if the associated IRS-score was ≥3 and at least 10% of the tumor cells stained positive for the HR. Scores of ≤3 and HR-expression less than 10% were regarded as negative.
Because the reporting of HR-status has changed over time, the HR interpretation was considered positive or negative based on the standard criteria at the time of the evaluation (2008). Therefore, the cut-off for ER-positivity was chosen at 10%, since it adhered to the international guidelines asserting that 10% or higher staining of cancer cells can be interpreted as a real positive ER-status. The low staining ER-positive tumors with a lower ERexpression between 1%-10% were not considered as adequate to be included in the study.

| Statistical analysis
Wherever appropriate, the data in this study is quoted as the mean with its corresponding Standard Deviation. The statistical analysis was carried out using the non-parametric test of McNemar applied to paired frequencies. 22 Statistical significance is assumed for p ≤ .05. The mean Disease-free Survival (DFS) was 54.9 ± 34.6 months.

| Study characteristics
All the demographic data are summarized in Table 3.

| The duration of TAM treatment and sites of metastasis
The mean duration of TAM administration was 40.7 ± 19.9 months.   Table 6.
Out of the 17 patients whose tumors had documented grading before and after TAM, there were 10 patients whose grade of differentiation did not alter. From the remaining 7 cases, a change in the grading was noted. In 3 out 7 patients the grading increased and in 4 cases the grading decreased, that is, the differentiation became better. Out of the 3 patients whose grade of differentiation became worse, in 2 cases a change in the HR-status was also observed while in 1 case HR-status remained unaltered. Out of the 4 patients whose grading decreased, 3 had no change in the HR-status and in only 1 case had the HR-status changed.

Feature Value
Patients studied (n) 6   to Aromatase Inhibitor (AI) therapy and found to confer a shorter progression-free survival with subsequent AI therapy. 23,26 The latest and largest meta-analysis to date, published in 2018 by Schrijver et al, 27  to 30.0%) and from ER-! ER+ was 21.5% (95% Cl = 18.1% to 25.5%). They further divided studies into two groups, one group for which the threshold of ER-positivity was set at 1% in the IHC, and another group, for which the ER-positivity was set at 10% in the IHC.
Moreover, the frequency of conversion in the subgroup analysis ER + !ERwas given as 16 It must be noted that the population of patients in the metaanalysis was very heterogeneous and it incorporated the data of 39 studies, some newer and some older, and there was no exact definition of the adjuvant therapy in terms of AHT the patients received in this subgroup. Therefore, at best, only a portion of patients received AHT and hence the data cannot be directly compared to that in this study. Interestingly, the meta-analysis data showed that in the subgroup with 10% as a threshold for ER-positivity, the ER conversion rate was markedly higher than that for 1% ER-positivity threshold, and the value becomes closer to the estimates in this study.
Although there are limitations to this study due to the small number of cases and retrospective design, this study primarily used only a homogenous ER-positive cohort, fixed the variable of treatment to TAM, and used a high threshold of 10% as criteria for ERpositivity. This setting allows for differences to classic discordance studies to be accentuated. As in any experimental setting, it is fundamentally important to compare a treatment arm with a control arm to accentuate differences. Since a control arm with ER-positive patients devoid of TAM administration is unethical, the best replacement at present would be a comparison to the metaanalysis data.
We are bringing forward this data at the present time, in order to compare our data in corroboration with the recent large-scale meta-analysis. With the above-mentioned limitations, the discrepancy between ER-loss in this study and the meta-analysis when adjusted for ER-loss that is, ER + !ER-, is approximately between 6%-13%. This discrepancy may be explained by several possibilities.
Assuming a flawless meta-analysis, then the discrepancy could be either due to methodological shortcomings in this study or that TAM acts as a stress-factor and maybe a significant driving force for breast cancer cells to further lower their ER-expression in order to accommodate this pharmacological stress. Interestingly, Lindström et al 18 showed, in their retrospective analysis of 459 ERpositive patients, that AHT was associated with a significant higher proportion of patients losing ER.
This present study concludes that as a rule of thumb, TAM administration in the adjuvant setting, results in a 30% loss of Estrogenreceptor expression in ER-positive patients following tumor progression and this information is pertinent for a clinician to judge the further use of endocrine therapy in metastatic and recurrent breast cancers, where biopsy is not readily available.
These differences warrant further and larger investigations such as prospective randomized trials and molecular biology testing to elaborate as to what extent different AHT may influence ER-loss beyond tumor biology or evolution.

ACKNOWLEDGMENTS
We thank Dr. S. Yong (Department of Statistics and Biomedical Statistics, University Hospital Düsseldorf, Germany) for helpful comments.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.