Quantitative Expression of Estrogen Receptor on Relapse Biopsy for ER-positive Breast Cancer: Prognostic Impact

Discussion

In our work, ER expression levels of the relapse biopsy (continuous and categorical variable) and not of the primary tumor was a strong independent prognostic factor for patients with ER-positive primary BC and ER-positive paired recurrence. The fact that for patients with ER-positive BC, ER expression levels may have a prognostic/predictive impact in the early and advanced setting has been established (10-13).

In a recently published work, 698 patients with ER-positive primary BC were evaluated. The percentage of tumor cell nuclei positively stained for ER was significantly and independently correlated to BC-specific mortality risk (ptrend=0.006) as a continuous variable (12). Dowsett et al. evaluated quantitative expression of ER as prognostic/predictive value in 950 patients with ER-positive BC enrolled into the monotherapy arms (adjuvant anastrozole or tamoxifen) of the ATAC trial. ER expression was calculated as the H-score, taking into considerations both the percentage of stained cells and the staining intensity. Low ER H-scores were associated with a shorter time to relapse (p=0.78 and p=0.0009 for tamoxifen and anastrozole-treated patients, respectively) (11).

Evidence has also been produced as regards the metastatic setting. As an example, the percentage of ER-positive tumor cell nuclei was associated with the cumulative probability of survival in a cohort of 205 patients with ER-positive metastatic BC treated with tamoxifen within the SWOG protocol 8228 (10). More recently, in the context of the phase III EGF30008 trial comparing lapatinib alone vs. lapatinib plus letrozole for advanced BC, Finn et al. investigated whether ER expression levels predicted response to the combination treatment in a total of 821 patients with ER-positive/HER2-negative BC. Only those patients presenting a low ER expression level (first H-score quartile) derived benefit from the addition of lapatinib to letrozole (13).

However, all the studies conducted in the advanced setting evaluated ER expression mainly of the primary tumor. Nowadays, in the era of the rapidly increasing knowledge in cancer biology, this approach may be considered as inadequate. Indeed, many contributions support the concept of BC heterogeneity and the fact that molecular features of BC may change during tumor progression (5-7). How these changes affect patient outcome and management is starting to be evaluated and represents a new field for cancer research. This is the main reason why we decided to evaluate to what extent ER expression level of the relapse biopsy may influence the prognosis of patients with ER-positive primary BC and ER-positive relapse. To the best of our knowledge, this is the first time that the degree of ER expression on the metastasis sample is described as a strong independent prognostic factor.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A: Kaplan–Meier overall survival curves for patients with the same estrogen receptor (ER) expression level for both primary and recurrent tumor samples (ER-concordant, i.e. primary and relapse both ER-high or both ER-low) vs. patients with a change in ER expression level from primary to relapse (ER-discordant, i.e. primary ER-high and relapse ER-low or vice versa). B: Kaplan–Meier overall survival curves for ER-concordant patients, patients whose status changed from ER-low to ER-high, and those whose changed from ER-high to ER-low. HR: Hazard ratio; CI: Confidence interval.

Moreover, patients with the worst prognosis were those whose ER-level changed from high (>50% of positive tumor cells) for the primary tumor to low (≤50% of positive tumor cells) at relapse compared to patients who changed in the opposite direction and patients who maintained the same ER level for both samples.

One possible criticism of our study may be that technical errors might have accounted for discrepancies in ER level between the primary and relapsed tumors. However, this risk is reduced by the rigid methodology that was applied in our work. Indeed, all specimens from both primaries and recurrences were evaluated at the same laboratory, by the same pathologist and adopting the same procedures and techniques. Nevertheless, some potential bias deserves to be mentioned. This is a retrospective study with the potential bias of enriching for cases with unusual clinical course. Moreover, relapse biopsy was not taken at the same time point for all the patients during their disease history. However, the study population was mostly represented by patients with ER-positive advanced BC whose disease progressed after a previous antiestrogen treatment (either as adjuvant treatment or as therapy for advanced disease) before relapse biopsy was performed.

In conclusion, besides BC relapse confirmation and receptor status evaluation, we provide evidence to support another rationale for performing histological evaluation of recurrent BC. We demonstrate that even with the same positive ER status on matched primary and relapsed tumor samples, the biopsy of recurrence provides relevant additional prognostic information. Although it is unclear whether a change from ER-high to ER-low at relapse reflects a real switch to a more aggressive phenotype or is the consequence of prior treatments, the particularly poor outcome of this group of patients suggests that they might deserve more aggressive treatments.

Our results strengthen the concept of BC heterogeneity, widening the spectrum of prognostic phenotype changes that may be observed during clinical progression. In the future, the application of next-generation sequencing techniques to paired BC primaries and recurrences might provide further evidence to this concept (14) and, hopefully, will allow prognostic and predictive parameters to be defined at the genomic level that might help in treatment decision.