Characteristics and Prognosis of Estrogen Receptor Low-Positive Breast Cancer

Abstract

Purpose

The updated American Society of Clinical Oncology/College of American Pathologists guideline for estrogen receptor (ER) testing recommends that breast cancer with ER expression in 1–10% of tumor cells should be reported as ER-low positive (ER^low), although limited data are available on the overall benefits of endocrine therapy. We investigated the clinicopathological characteristics and clinical outcomes of ER^low breast cancer and to compare them with those of ER-negative (ER^neg) and ER-high (> 10% of tumor cells, ER^high) breast cancers.

Methods

Consecutive patients with invasive breast cancer who underwent curative surgery between November 2007 and December 2014 were included. Clinicopathological characteristics and disease-free survival (DFS) of ER^low tumors were compared with those of ER^neg and ER^high tumors.

Results

Of the 2,309 cases included, 46 (2%), 643 (27.8%), and 1,620 (70.2%) were ER^low, ER^neg, and ER^high, respectively. ER^low tumors were associated with no special type of histology (p = 0.011), advanced pT (p = 0.017), pN (p = 0.009) and anatomic stages (p < 0.001), high grade (p < 0.001), negative/low progesterone receptor (PR) status (p < 0.001), human epidermal growth factor receptor 2 positivity (p < 0.001), high Ki-67 (p < 0.001), and recurrence (p = 0.006) compared to ER^high tumors. DFS was significantly dependent on ER status, and ER^low tumors showed poorer DFS than ER^high tumors (p = 0.001), however, there was no significant survival difference between ER^low and ER^neg tumors. Furthermore, DFS in ER^high patients was affected by hormone therapy (p < 0.001), while it was not affected in ER^low patients.

Conclusion

Patients with ER^low breast cancer have clinicopathological characteristics that differ from those with ER^high tumors. Although this study was limited by the small sample size of the ER^low group, no benefit from hormone therapy was observed in the ER^low group compared with the ER^high group.

INTRODUCTION

Breast cancer is a heterogeneous disease in biology and behavior and is the most common malignancy among females worldwide and in Korea [1,2]. Invasive breast cancers (IBCs) can be grouped into biomarker-defined subtypes for treatment purposes based on the status of hormone receptors (HRs) (estrogen receptor [ER] and progesterone receptor [PR]) and human epidermal growth factor receptor 2 (HER2) [3]. Therefore, an assessment of HR and HER2 status is mandatory in all IBC cases.

ER is an important prognostic and predictive marker for the benefit of endocrine therapy. According to Korean Breast Cancer Society data from 2002 to 2018 [2], the incidence of HR-positive breast cancer has increased by more than 1% annually since 2002, 58.2% of breast cancers were ER-positive in 2002, and 78.9% displayed ER positivity in 2018 [2]. Threshold changes in ER-positive expression, advances in immunohistochemistry (IHC) techniques, and the development of highly sensitive antibodies probably contributed to this increase in ER-positive tumors. IHC has been the gold standard method for ER determination in surgical specimens, and a 10% cutoff was widely used to define ER positivity until the American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) released guidelines for ER and PR testing of breast cancer in 2010 [4,5]. Subsequently, ≥ 1% tumor cell immunoreactivity for ER or PR was interpreted as positive, according to the ASCO/CAP guidelines. However, a recently updated ASCO/CAP guideline recommends that cases with 1%–10% tumor cell staining should be reported as ER-low positive (ER^low), with a comment explaining that there is limited data on the overall benefit of endocrine therapy, although these patients may benefit [6]. Approximately 2%−3% of breast cancers have been reported to be ER^low, which are known to be heterogeneous in terms of their biological behavior and clinicopathological characteristics [7,8,9].

In this study, we aimed to investigate the clinicopathological characteristics and prognosis of ER^low breast cancer and to compare them with those of ER-negative (< 1% of tumor cells positive, ER^neg) and ER-high (> 10% of tumor cell positive, ER^high) breast cancers.

METHODS

This study included 2,309 consecutive cases with primary IBC who underwent curative surgery at Yeungnam University Hospital between November 2007 and December 2014. IHC staining for ER, PR, HER2, and Ki-67 was performed routinely in all primary and recurrent breast cancers. Patients with metachronous contralateral advanced stage breast cancer, microinvasive carcinoma, or those who received neoadjuvant chemotherapy were excluded. Patients’ clinicopathological information (age at diagnosis, sex, date and method of operation, extent of axillary dissection, tumor size, histological grade, histological type, lymphovascular invasion, pathological T and N stages, IHC results for ER, PR, HER2, and Ki-67, and clinical outcomes) were obtained from medical records and pathology reports. Since December 2008, SP1 monoclonal antibody has been used for ER staining. The percentages of cells positive for ER or PR estimated by visual assessment under a microscope were recorded in pathology reports. Ki-67 result was expressed as the percentage of positively staining cells among the total number of tumor cells by counting of at least 500 invasive cancer cells. In cases with equivocal HER2 IHC results, the presence or absence of gene amplification was routinely confirmed by in situ hybridization according to the ASCO/CAP guidelines [10].

The cases were classified into 3 groups: ER^neg (< 1% of tumor cells positive), ER^low (1%–10% of tumor cells positive), and ER^high (> 10% of tumor cells positive) based on IHC results for ER. Ki-67 results were classified as low (< 20%) or high (≥ 20%) [11]. Disease-free survival (DFS) was defined as the time from the date of diagnosis to the date of first recurrence (local recurrence or distant metastasis).

Statistical analysis was performed using SPSS version 25.0 for Windows (IBM, Armonk, USA). Comparisons of group clinicopathological characteristics were made using the χ², Fisher’s exact, Kruskal-Wallis, and Mann-Whitney tests. Survival analysis was performed using the Kaplan-Meier method, and group survival differences were compared using the log-rank test. Univariate and multivariate Cox regression analyses were performed to determine the prognostic impact of variables on recurrence. The significant variables identified in the univariate analysis were further analyzed by multivariate analysis using backward stepwise selection. Adjusted hazard ratios and associated 95% confidence intervals were calculated for each variable. The p-values of < 0.05 were considered statistically significant.

This study was approved by the institutional review board of Yeungnam University Medical Center (YUMC2022-01-005), which waived the requirement for informed consent.

RESULTS

Of the 2,309 patients with IBC, 2,304 were females, and 5 were males. A total of 1,267 (54.9%) patients underwent breast-conserving surgery, while 1,042 (45.1%) patients underwent mastectomy. Axillary lymph node dissection was performed in 813 (35.2%) patients and 1,487 (64.4%) patients underwent sentinel lymph node biopsy alone. Nine patients (0.4%) did not undergo a sentinel lymph node biopsy or axillary lymph node dissection. There were 46 (2%), 643 (27.8%) and 1,620 (70.2%) cases of ER^low, ER^neg and ER^high, respectively. The clinicopathological characteristics of the patients stratified by ER expression status are summarized in Table 1. Compared to ER^neg or ER^high tumors, ER^low tumors were more likely to be invasive carcinoma of no special type (p = 0.041 and p = 0.011, respectively), more frequently presented in advanced pN (p = 0.042 and p = 0.009, respectively) and anatomic (p = 0.017 and p < 0.001, respectively) stages and were more likely to be HER2-positive (p = 0.002 and p < 0.001, respectively). Furthermore, these tumors were present more frequently in advanced pT (p = 0.017) and were associated with a higher grade (p < 0.001), negative/low PR status (p < 0.001) and high Ki-67 status (p < 0.001) than ER^high tumors. Patients with ER^low tumors received chemotherapy more frequently (p < 0.001) but less frequently received hormone therapy (p < 0.001) than those with ER^high tumors. ER^low tumors were more strongly associated with lymph node metastasis (p = 0.008) and PR positivity (p < 0.001) than ER^neg tumors.

Table 1. Clinicopathological characteristics and biomarkers according to ER expression in invasive breast carcinomas.

Clinicopathologic factor		ER expression				p-value
		Total (n = 2,309)	Negative (n = 643)	Low (n = 46)	High (n = 1,620)	All	Negative vs. Low	Low vs. High
Age (yr)		50 (26–90)	52 (26–90)	49.5 (30–73)	49 (26–87)	< 0.001	0.087	0.812
Tumor size (cm)		1.8 (0.2–12)	2 (0.2–12)	2.2 (0.4–6.9)	1.7 (0.2–12)	< 0.001	0.313	0.003
Histologic type						0.001	0.041	0.011
	IC-NST	2,055 (89)	591 (91.9)	46 (100)	1,418 (87.5)
	Others	254 (11)	52 (8.1)	0 (0)	202 (12.5)
LN metastasis						0.003	0.008	0.072
	Absent	1,388 (60.3)	416 (65)	21 (45.7)	951 (58.9)
	Present	912 (39.7)	224 (35)	25 (54.3)	663 (41.1)
pT						< 0.001	0.577	0.017
	1	1,430 (61.9)	343 (53.3)	20 (43.5)	1,067 (65.9)
	2	791 (34.3)	270 (42)	24 (52.2)	497 (30.7)
	3	84 (3.6)	28 (4.4)	2 (4.3)	54 (3.3)
	4	4 (0.2)	2 (0.3)	0 (0)	2 (0.1)
pN						< 0.001	0.042	0.009
	0	1,388 (60.3)	416 (65)	21 (45.7)	951 (58.9)
	1	593 (25.8)	128 (20)	12 (26.1)	453 (28.1)
	2	175 (7.6)	40 (6.3)	6 (13)	129 (8)
	3	144 (6.3)	56 (8.8)	7 (15.2)	81 (5)
Anatomic stage						< 0.001	0.004	< 0.001
	I	1,092 (47.5)	270 (42.1)	13 (28.3)	809 (50.1)
	II	867 (37.7)	269 (42)	18 (39.1)	580 (35.9)
	III	329 (14.3)	98 (15.3)	13 (28.3)	218 (13.5)
	IV	13 (0.6)	4 (0.6)	2 (4.3)	7 (0.4)
Grade						< 0.001	0.768	< 0.001
	1	357 (15.5)	3 (0.5)	0 (0)	354 (21.9)
	2	623 (27)	42 (6.5)	4 (8.7)	577 (35.6)
	3	1,329 (57.6)	598 (93)	42 (91.3)	689 (42.5)
LVI						< 0.001	0.206	0.982
	Absent	1,268 (54.9)	396 (61.6)	24 (52.2)	848 (52.3)
	Present	1,041 (45.1)	247 (38.4)	22 (47.8)	772 (47.7)
PR						< 0.001	< 0.001	< 0.001
	Negative	856 (37.1)	620 (96.4)	36 (78.3)	200 (12.3)
	Positive	1,453 (62.9)	23 (3.6)	10 (21.7)	1,420 (87.7)
HER2						< 0.001	0.002	< 0.001
	Negative	1,845 (79.9)	411 (63.9)	19 (41.3)	1,415 (87.3)
	Positive	464 (20.1)	232 (36.1)	27 (58.7)	205 (12.7)
Ki-67						< 0.001	0.140	< 0.001
	Low (< 20%)	793 (34.3)	45 (7)	6 (13)	742 (45.8)
	High (≥ 20%)	1,516 (65.7)	598 (93)	40 (87)	878 (54.2)
Hormone therapy						< 0.001	< 0.001	< 0.001
	No	651 (28.2)	603 (93.8)	18 (39.1)	30 (1.9)
	Yes	1,658 (71.8)	40 (6.2)	28 (60.9)	1,590 (98.1)
Chemotherapy						< 0.001	0.608	< 0.001
	No	820 (35.5)	87 (13.5)	5 (10.9)	728 (44.9)
	Yes	1,489 (64.5)	556 (86.5)	41 (89.1)	892 (55.1)
Recurrence						< 0.001	0.078	0.006
	No	2,043 (88.5)	551 (85.7)	35 (76.1)	1,457 (89.9)
	Yes	266 (11.5)	92 (14.3)	11 (23.9)	163 (10.1)

Values are presented as median (range) or number (%).

ER = estrogen receptor; IC-NST = invasive carcinoma of no special type; LN = lymph node; LVI = lymphovascular invasion; PR = progesterone receptor; HER2 = human epidermal growth factor receptor 2.

Recurrence included locoregional and distant metastasis. During a median follow-up period of 99 months (range, 1–165 months), recurrence occurred in 266 (11.5%) patients. The recurrence rate for ER^low tumors was higher than that for ER^high tumors (p = 0.006). DFS was significantly influenced by ER status (p < 0.001, Figure 1A). Pairwise comparisons showed that ER^high tumors were associated with better DFS than ER^neg (p = 0.001) and ER^low (p = 0.001) tumors, but DFSs of ER^neg and ER^low tumors were not significantly different (p = 0.105). Similar results were found in patients who received hormone therapy (n = 1,658) (Figure 1B).

Figure 1. DFS according to ER status in (A) 2,309 patients with IBC and (B) 1,658 patients with IBC who received hormone therapy. DFS according to ER status in (C) 1,759 patients with IBC who underwent surgical treatment during the period when trastuzumab was available as adjuvant therapy for HER2-positive breast cancer and (D) 1,280 patients with IBC who received hormone therapy among the 1,759 patients.

DFS = disease-free survival; ER = estrogen receptor; HER2 = human epidermal growth factor receptor 2; IBC = invasive breast cancer.

Anti-HER2 therapy affects the survival outcomes of patients with HER2- positive breast cancer. Because our study population was heterogeneous regarding indications for anti-HER2 therapy, we performed a survival analysis in patients who underwent surgical treatment during the period in which trastuzumab was approved as an adjuvant treatment for HER2-positive breast cancer. Of the 1,811 patients who underwent surgical treatment after July 2009 (483 ER^neg, 38 ER^low, and 1,290 ER^high), 360 (19.9%) had HER2-positive breast cancer (176 ER^neg, 22 ER^low, and 162 ER^high). Of these, 288 patients with lymph node metastasis or tumors > 1 cm were eligible for adjuvant anti-HER2 therapy according to the Korean medical insurance guidelines, and 236 of them received trastuzumab treatment. After excluding 52 HER2-positive breast cancer patients who did not receive trastuzumab (17 ER^neg, 1 ER^low, and 34 ER^high), 1,759 of the 1,811 patients were finally included in the survival analysis. In this cohort, the ER^low group showed the worst DFS among the 3 groups (p = 0.001) and there were significant differences in the DFSs of ER^neg and ER^low tumors (p = 0.022), ER^low and ER^high tumors (p < 0.001), and ER^neg and ER^high tumors (p = 0.032) (Figure 1C). In a cohort of 1,759 patients, 1,280 received hormone therapy, of which ER^high tumors had better DFS than ER^low tumors (p = 0.002) (Figure 1D).

Some ER^low and ER^high patients did not receive hormone therapy due to the different thresholds of ER positivity (10%) applied in earlier cases and treatment refusal. When the DFSs of patients who received or did not receive hormone therapy-were compared to evaluate endocrine responsiveness, ER^high patients who received hormone therapy had significantly better DFS (p = 0.001) (Figure 2A). However, the DFS of ER^low patients who received or did not receive hormone therapy were similar (p = 0.979) (Figure 2B).

Figure 2. DFS according to hormone therapy in patients with (A) ER^high or (B) ER^low breast cancer.

DFS = disease-free survival; ER^high = estrogen receptor-high; ER^low = estrogen receptor-low positive.

In addition to ER status, histological grade 3, large tumor size (≥ 2 cm), lymph node metastasis, lymphovascular invasion, PR negative, HER2 positive, high Ki-67, chemotherapy status, and omission of hormone therapy were found to be associated with a higher risk of recurrence in univariate analyses. However, in the multivariate analysis, ER status was not an independent factor for DFS, but histological grade 3, large tumor size (≥ 2 cm), lymph node metastasis, lymphovascular invasion and PR negativity were independently significant factors for poorer DFS (Table 2).

Table 2. Univariate and multivariate analyses of clinicopathologic factors affecting disease-free survival.

Clinicopathologic factor		Univariate analysis		Multivariate analysis
		HR (95% CI)	p-value	HR (95% CI)	p-value
ER expression			< 0.001
	Positive	Reference
	Negative	1.548 (1.199–1.999)	0.001	-	-
	Low positive	2.593 (1.407–4.777)	0.002	-	-
Histologic grade
	1 & 2	Reference		Reference
	3	2.988 (2.221–4.021)	< 0.001	1.915 (1.380–2.656)	< 0.001
Tumor size
	≤ 2 cm	Reference		Reference
	> 2 cm	2.573 (2.015–3.287)	< 0.001	1.515 (1.161–1.978)	0.002
LN metastasis
	Absent	Reference		Reference
	Present	2.538 (1.981–3.250)	< 0.001	1.405 (1.051–1.879)	0.022
LVI
	Absent	Reference		Reference
	Present	3.243 (2.483–4.237)	< 0.001	2.247 (1.648–3.065)	< 0.001
PR expression
	Positive	Reference		Reference
	Negative	1.651 (1.298–2.101)	< 0.001	1.369 (1.054–1.777)	0.019
HER2 status
	Negative	Reference
	Positive	1.382 (1.046–1.826)	0.023	-	-
Ki-67
	Low (< 20%)	Reference
	High (≥ 20%)	2.377 (1.743–3.241)	< 0.001	-	-
Chemotherapy
	No	Reference
	Yes	2.360 (1.735–3.212)	< 0.001	-	-
Hormone therapy
	No	Reference
	Yes	0.640 (0.498–0.823)	< 0.001	-	-

HR = hazard ratio; CI = confidence interval; ER = estrogen receptor; LN = lymph node; LVI = lymphovascular invasion; PR = progesterone receptor; HER2 = human epidermal growth factor receptor 2.

DISCUSSION

ER status was incorporated into the 8th American Joint Committee on Cancer pathological prognostic stage (PPS) together with PR, HER2, and histological grade [12]. In PPS, IBC patients with a positive ER status (including ER^high and ER^low) were downstaged compared with the corresponding anatomic stage. Therefore, the biological behavior of ER^low cases needs to be clearly defined because downstaging ER^low cases can lead to undertreatment, even if they are biologically similar to ER^neg cases.

In the present study, ER^low cases constituted 2% (46/2,309) of all IBC cases and 2.8% (46/1,666) of ER-positive IBC cases, which is consistent with previous studies that reported ER^low cases accounted for 2%–5.1% of all IBC and 2.5%–7.4% of ER-positive IBC cases [6,7,9,13,14,15]. Some studies [9,13] have defined ER^low as 1%–9% positive staining instead of 1%–10% (the ASCO/CAP definition). Furthermore, a recent study that used tissue microarray IHC data reported a slightly higher incidence of ER^low cases (3% of all IBCs and 4.1% of ER-positive IBCs) [8].

Most ER^low cases (93.5%) in the present study and > 50% of ER^low cases in previous studies presented with low or negative PR expression [7,8,9], which may contribute to the lack of significant survival benefits of adjuvant hormone therapy for patients with ER^low IBC. Previous studies have reported HER2 positivity in 29%–60% of ER^low cases [7,8,9]. Furthermore, the current study revealed that ER^low IBCs are heterogeneous, with some cases similar to triple-negative breast cancer and others similar to HER2-positive breast cancer.

Regarding clinicopathological characteristics, ER^low cases were associated with younger age, larger tumor size, higher histological grade, higher pN, more advanced stage, higher Ki-67, and higher HER2 positivity and PR negativity than ER^high cases [7,8,9,14,16]. Our results are consistent with those of previous studies, as ER^low tumors presented with more advanced pT, pN, and anatomic stages, higher histological grade and Ki-67, and negative/low PR and positive HER2 statuses than ER^high tumors. Furthermore, ER^low cases were more associated with lymph node metastasis, advanced pN and anatomic stage, and positive PR and HER2 statuses than ER^neg cases. These findings are partially consistent with the observations of previous studies in which ER^low tumors frequently presented at a younger age, lower grade, more lymphovascular invasion, PR positivity, HER2 positivity, lower Ki-67, and lower basal marker expression than ER^neg tumors [7,8,9]. As shown in Table 1 and as reported by Yu et al. [16], the difference in clinicopathological characteristics between ER^low and ER^high tumors was greater than the difference between ER^low and ER^neg tumors.

Regarding survival outcomes for ER^low tumors, a previous large cohort study reported that ER^low tumors (n = 250) had poorer distant recurrence-free survival (DRFS), recurrence-free survival (RFS) and overall survival (OS) than ER^high tumors (n = 7,764) regardless of endocrine therapy, but DRFSs, RFSs and OSs were similar to ER^neg tumors (n = 1,625) [13]. Poon et al. [8] reported that the DFS of ER^low tumor (n = 54) was significantly worse than that of ER^high tumors (n = 1,266), but comparable to that of ER^neg tumors (n = 503). We also observed that DFS was poorer for ER^low tumors than ER^high tumors, and no significant differences were observed between DFSs of ER^low and ER^neg tumors. However, other studies have reported contradictory results. Fei et al. [7] reported that ER^low tumor (n = 97) had better RFS and disease-specific survival (DSS) than ER^neg tumors (n = 1,100), and no significant differences were observed between RFSs and DSSs of ER^low and ER^high (n = 2,982) tumors. These diverse clinical outcomes may be explained by different clinicopathological characteristics of ER^low tumors in the study population, such as PR positive rates (20%–75%) and proportions of patients who received hormone therapy (20%–66%) or chemotherapy (49%–97%) [7,8,13]. The HER2 positivity rate and the proportion of patients who received anti-HER2 therapy would also affect survival outcomes. In Korea, trastuzumab was first approved for metastatic HER2-positive breast cancer in September 2006, and its use was expanded in July 2009 as an adjuvant treatment for HER2-expressing node-positive breast cancer. Approximately a year later, it was also approved for use in node-negative breast cancer patients with tumors > 1 cm. Therefore, we performed a survival analysis on patients who underwent surgical treatment after July 2009, when trastuzumab administration was possible in the adjuvant setting. In this cohort (n = 1,759), ER^low tumors had poorer DFS than ER^high or ER^neg tumors. The worst survival in patients with ER^low tumors is probably due to advanced anatomic stage compared to those with ER^high or ER^neg tumors.

In addition to our study, a recent Korean study [9] also showed that adjuvant hormone therapy did not have an effect on DFS in patients with ER^low tumor, and that patients with ER^high tumors showed clear endocrine responsiveness. These results support the notion that hormone therapy has limited benefits for ER^low breast cancer patients. Cai et al. [15] recently reported that short-term endocrine therapy for 2–3 years might be an alternative for patients with ER^low breast cancer instead of standard 5 years of treatment because there was no significant difference in DFS between the 2 groups. We did not observe differences in DFS of patients with ER^low according to hormone therapy, although the comparison was made with a small number of patients. A meta-analysis of 6 studies also revealed a lack of benefit of endocrine therapy in patients with ER^low tumors because patients with ER^low tumors who received endocrine therapy showed a prognosis similar to those without endocrine therapy and those with ER^neg tumors who received endocrine therapy [16].

There are several limitations to this study. First, it is limited by its retrospective nature and the small number of patients (n = 46) with an ER^low tumor, which prevented the subgroup analysis according to adjuvant treatment options and pathologic features of ER^low tumors. However, to the best of our knowledge, the present study was conducted in the largest cohort of Korean IBCs (n = 2,309) with a long median follow-up of 99 months to elucidate the clinicopathological characteristics and survival outcomes after treatment of ER^low tumors. Second, most patients with an ER^low tumor received adjuvant chemotherapy or anti-HER2 therapy, which may have influenced the clinical outcomes of ER^low tumors, although it should be noted that all patients received standard treatment at the time of diagnosis. We suggest a large nationwide multicenter study to overcome these limitations and provide detailed information on the biological characteristics of ER^low breast cancer.

In summary, ER^low breast cancers are heterogeneous and have distinct clinicopathological features that differentiate them from ER^high and ER^neg breast cancers. Patients with ER^low tumors have poorer DFSs than those with ER^high tumors. Furthermore, no benefit from hormone therapy was observed in the ER^low group compared to the ER^high group, although this study was limited by the small sample size of the ER^low group.