Estrogen receptor and progesterone receptor status and their prognostic implications in pelvic high-grade serous carcinomas: A retrospective study

Background: There is mounting evidence that ovarian, tubal and peritoneal high-grade serous carcinoma (HGSC) share common origin. It was also suggested that extrauterine HGSCs may originate from endometrium. The aim of this study is to compare the estrogen receptor (ER) and progesterone receptor (PR) status among pelvic HGSCs, and to analyze the prognostic role of ER and PR in pelvic HGSCs patients, and the prognostic factors in patients with different ER or PR status Methods: In total, 283 patients diagnosed with ovarian, tubal, peritoneal, and uterine HGSC were retrospectively analyzed. All patients’ diagnosis were reviewed by a panel of gynecologists and pathologists strictly according to criteria based on lesion distribution. Results: Patients in endometrial group were older than ovarian (60.1±8.0 vs. 54.1±8.3, p=0.000) and tubal (60.1±8.0 vs. 55.8±9.5, p=0.008) group. A higher proportion of ovarian group presented with advanced stage disease than fallopian and endometrial group (73.7% vs. 47.2%, p=0.000 and 73.7% vs. 47.4%, p=0.002, respectively). PR positivity rate was much lower in peritoneal compared to ovarian group (25.0% vs. 65.6%, p=0.000). There was no difference in survival rates among four groups. Although ER and PR were not prognostic factors for 5-year overall survival (OS) and progression-free survival (PFS), the prognostic factors were different in patients of distinct ER/PR status. More chemotherapy cycles was a protective prognostic factor in ER(+) or PR(+) patients but not in ER(-) or PR(-) patients. P53 mutation was adverse prognostic factor for OS in PR(-) patients but in PR(+) patients. Conclusions: PR positivity rate were much lower in peritoneal compared to

HGSC. Although ER and PR were not prognostic factors in pelvic HGSCs, prognostic factors for survival were different in patients of different ER/PR status.
Endometrial HGSC comprises 10~15% of endometrial cancer but account for more than 40% of endometrial cancer-related death due to its aggressive behavior and poor prognosis (3).
Although HGSC occurs in ovary and peritoneum, serous cells were not contained in these organs natively (4). The effort to identify a precursor lesion within an ovary has never been successful (5), nor has been reported in peritoneum. Fallopian tube contain native serous cells (6). Evidence has accumulated favoring a tubal origin of ovarian and peritoneal HGSCs in the past two decades. Serous tubal intraepithelial carcinoma (STIC), which was considered the precursor lesion of ovarian, tubal, and peritoneal HGSCs, was detected in 50% of female BRCA mutations carriers (6) and 35-70% of sporadic ovarian and peritoneal HGSCs (7). Mutational analysis of these HGSCs with concurrent STIC showed identical P53 mutations in the vast majority of cases (8).
There were few studies have compared the estrogen receptor (ER) and progesterone receptor (PR) expression among pelvic HGSCs. In the present study, we propose to compare the ER and PR status of pelvic HGSCs after systematically reviewing and conforming the cases to diagnosis criteria based on lesion distribution strictly, to inspect the hypothesis that pelvic HGSCs have common origin.
The role of ER or PR in the prognosis of ovarian HGSC patients was of dispute. ER expression was concluded to have no impact on survival outcomes of patients in a large multicenter consortium study (10), but was reported to be associated with shorter overall survival (OS) in another study (11). PR expression was reported to be associated with improved disease-specific survival (DFS) in ovarian HGSC (12).
Endometrial HGSC was classified as type II endometrial cancer and was hormonally independent(10). The prognostic role of ER and PR in tubal and peritoneal HGSC patients were unclear. In the present study, we also tried to analyze the prognostic role of ER and PR in pelvic HGSCs patients, and the prognostic factors in patients with different ER or PR status.

Methods
This is a retrospective study approved by the ethics committee of the OB/GYN Hospital of Fudan University. The inclusion criteria were: (1) patients who underwent primary surgery between January 2000 and December 2012; (2) histological diagnosis of HGSC of ovary, fallopian tube, peritoneal, or uterine endometrium. The exclusion criteria were: (1) patients did not receive primary surgery in this hospital; (2) patients received neoadjuvent chemotherapy or radiotherapy before the primary surgery; (3) patients with concurrent genital or extra-genital malignancy; (4) patients had HGSC mixed with other tumor constituents such as endometrioid carcinoma or sarcoma; (5) cases with incomplete clinical documents; (6) cases cannot be assigned according to conventional diagnostic criteria (described in detail in the next paragraph).
Patients were assigned to ovarian, tubal, peritoneal, or endometrial HGSC group taking into account clinical history, surgical findings, imaging and pathological findings. The reviewing process was executed by a panel of two gynecologists and two pathologists. The diagnostic criteria for ovarian carcinoma (OC) were: (1) the disease was mainly located in the ovaries; (2) if fallopian tube was involved, the lesion should be intruded form the external serosa. Patients were assigned to tubal carcinoma (FC) if: (1) the disease was mainly located in the fallopian tube and arose from the endosalpinx; (2) the ovaries were normal or involved with lesion smaller than the tubal mass. Patients were assigned to peritoneal carcinoma (PC) according to the Gyncologic Oncology Group criteria(13): (1) both ovaries were in normal size or enlarged for benign disease; (2) the extent of pelvic disease was greater than involving the ovaries; (3) microscopically, the ovaries showed either no tumor or tumor confined to surface, with stroma infiltration less than 5mm×5mm. Patients were diagnosed with endometrial carcinoma (EC) if: (1) HGSC was found in the uterine endometrium; (2) if the tube or ovary was involved, the lesion should be contiguous to the endometrial lesion.
In the study period, there were 171, 53, 21, and 38 cases assigned to OC, FC, PC, and EC group, respectively ( Figure 1). There were 37 cases can not be assigned for the following reasons: (1) the lesions at ovaries and fallopian tubes can not be distinguished for they were jointed or wrapped together, 13 cases; (2) there were independent lesions at tubal fimbria and ovary, 10 cases; (3) there were independent lesion located at the fallopian tube and the ovary; the lesion of tube was in the luminal and not at the fimbria, 5 cases; (4) simultaneously detected lesion at the ovary and the endometrium, but the tube was not involved, 9 cases.
Patient demographic data such as menstrual information, previous history, family history, surgical treatment, chemotherapy data, follow-up of vital status, and cause of death were recorded. Reproductive duration was defined as age at natural menopausal minus menarche age for postmenopausal patients, and as age at  Spearman's correlation analysis was used to assess the correlation between variables and variables with p<0.05 were included in the logistic regression model.
OS and progression-free survival (PFS) was calculated from the date of primary surgery to death and recurrence, respectively, or the last disease-free visit. Survival analysis was performed using Kaplain-Meier and Cox regression model. All p values reported are two-tailed and a p<0.05 was considered significant.

FIGO stage
All the patients of PC group were of FIGO stage III, thus the ratio of advanced disease (FIGO stage III and IV) were higher than the other three groups. A higher proportion of OC group presented with advanced stage disease that FC and EC group (73.7% vs. 47.2%, p = 0.000 and 73.7% vs. 47.4%, p = 0.002, respectively) ( Table 1).

Patient history and family history
More patients of EC group had a history of tubal ligation (28.9% vs. 11.7%, p = 0.011), which make major contribution to differed tubal surgery rate between two groups (31.6% vs. 14.0%, p = 0.016). No difference was found in uterine surgery history, ovary surgery history, tumor history of the patients themselves and their first-degree relatives ( Patients of OC group received more chemotherapy cycles than PC (median 6.0 vs. 3.0, p = 0.000) and EC (median 6.0 vs. 5.0, p = 0.000) group. Patients of FC group received more chemotherapy cycles than PC group (median 6.0 vs. 3.0, p = 0.017) ( Table 3).

Progesterone receptor
The PR positivity rate of OC, FC, PC and EC group was 65.6%, 50.0%, 25.0%, and 50.0%, respectively. PR positivity rate was higher in OC compared to FC and PC group (p = 0.046 and p = 0.000, respectively)( Table 4).  (Table 4). However, Ki-67 immunostaining was not correlated with group but correlated with parity (p = 0.035), natural menopause age (p = 0.013), reproductive years (p = 0.012), and P53 mutation status (p = 0.034). After entering into linear regression model, none of these factors was an independent risk factor for Ki-67 immunostaining.

P53 mutation
Normally, p53 expression was sparsely positive in nucleus, and was categorized as score 1 in the present study. P53 mutation can manifest as either negative (nonsense mutation, score 0) or diffuse positive (missense mutation, score 2, 3, and 4). In 268 patients with P53 information in the present study, there were 223 (83.2%) cases present with P53 mutation. The mutation rate of P53 was 86.4%, 84.3%, 60.0%, and 80.0% in OC, FC, PC, and EC group, respectively. There was no difference of mutation rate among these groups (Table 4).

Survival results
The rate of loss of follow-up was 13.1% (37/283) in the present study. There were 14.0% (24/171), 11.3% (6/53), 9.5% (2/21), 13.2% (5/38) patients were loss of follow-up in OC, TC, PC and EC group, respectively. The rate of loss of follow-up was not different among groups. There were 153 cases of disease-specific deaths in total. The 5-year OS rate of OC, TC, PC and EC group was 39.5%, 38.7%, 44.9%, and 53.5%, respectively. PFS rates were calculated among patients who received satisfactory debulking surgery and were sensitive to platinum-based chemotherapy  Table 6).

Discussion
In the present study, the clinical characteristics were distinct among pelvic HGSCs such as age, FIGO stage, and serum CA125 level. EC patients were older than OC and FC patients (61 vs. 54 and 61 vs. 56 years, respectively). Another study also reported patients of serous EC were older than serous OC and FC patients (68 vs. 62 and 68 vs. 63 years, respectively)(4). However, the patients in that study were 7 years older in average than that of our study. The deviance may be originated from race or other factors, which is worthy of further investigation. A study based on the Surveillance, Epidemiology, and End Results (SEER) database found that tubal HGSC had better survival rates compared to ovarian, peritoneal, and uterine HGSCs(4). However, patients who did not receive surgery were included in the study and chemotherapy information of the included patients was absent.
Moreover, pelvic HGSCs usually presents with metastatic and bulky multifocal disease, making the original site difficult to determine(14). But differentiation of primary site based on central clinical and pathology review was absent in that study. In the present study, we found no difference in survival among four groups. However, no difference of survival does not necessarily indicate pelvic HGSCs were of identical origin. In fact, disputes with respect to the tubal origin of PC and the endometrium-origin of extra-uterine HGSCs still exist.
Clinically, it was reported that 4.5% BRCA mutation carriers who had concurrent STIC developed peritoneal HGSC disease after risk-reducing salpingo-oophorectomy (RRSO)(15), which indicated that prophylactic RRSO failed to protect toward PC.
Epidemiologically, the risk factors for serous PC differed from those for serous OC or FC in several key areas. Having a term pregnancy and breast-feeding was associated with decreased risk of OC and FC, but not of PC(16). Having a firstdegree relative with breast or ovarian cancer was associated with 70% increased risk of OC, but not of PC(16). On the contrary, obesity was associated with 2-fold increased risk of PC, but not of OC or FC(16). Genetically, over-expression of HER-2 has been consistently reported to be more frequent in PC compared to OC(17). In the present study, we also found that PR positivity rate was significantly lower in PC Although ER or PR was not independent prognostic factor for survival in pelvic HGSCs in the present study, we found that prognostic factors were different in patients with different ER or PR status. More chemotherapy cycles was a protective prognostic factor for 5-year OS in ER(+) or PR(+) patients but not in ER(-) or PR(-) patients. In PR(-) patients but not in PR(+) patients, P53 mutation was an adverse prognostic factor for OS.
P53 mutation in the present study was assessed by immunohistochemical staining.
However, a study reported that immunohistochemical staining (defining P53 mutation as either 0% or 60-100% positive cells) correctly identified 94% of sequencing-confirmed P53 mutations cases(20). The authors suggested that immunohistochemical analysis is a robust surrogate for sequencing for inferring the presence of P53 mutation( 20).
The current study has inherent limitations to retrospective studies, and was carried out in a single institute with a relative small sample size. However, to our knowledge, this is the first study comparing the ER and PR status among pelvic HGSCs, and also the first study trying to compare the above parameters on the basis of systematic clinical and pathological reviewing. Moreover, this is the first study figuring the prognosis factor of pelvic HGSC patients with different ER or PR status. We found that PR status was different among pelvic HGSCs. Although ER or PR was not an prognostic factor for survival, the prognostic factors were different in patients of distinct ER/PR status. Specifically, more chemotherapy cycles was a protective prognostic factor in ER(+) or PR(+) patients but not in ER(-) or PR(-) patients. P53 mutation was adverse prognostic factor for OS in PR(-) patients but not in PR(+) patients.

Conclusions
The PR positivity rate was much lower in peritoneal HGSC compared to ovarian HGSC patients, which cast doubts on the hypothesis that peritoneal HGSC have a common origin with ovarian HGSC. The prognosis was not different among pelvic HGSCs. Although ER and PR were not prognostic factors for survival, the prognosis factors for survival were different in patients of distinct ER/PR status. More chemotherapy cycles was a protective prognostic factor in ER(+) or PR(+) patients

Consent for publication
Not applicable.

Availability of data and material
All data supporting this study are included in this article. Please contact the corresponding author for data requests.

Competing interests
The authors declare that they have no competing interests.        Figure 1 Patients included in the study OC: ovarian carcinoma; FC: fallopian tubal carcinoma; PC: peri Survival curves for overall survival (OS) and progression-free survival (PFS) of four groups (A