Clinical and pathologic predictors of survival in patients with endometrial stromal sarcoma

Endometrial stromal sarcoma (ESS) is a rare uterine neoplasm and the clinical and pathologic factors that predict outcomes are poorly understood. We conducted an institutional retrospective review of patients with the diagnosis of ESS between January 1990 and April 2012. Demographic and clinical features, treatment data and outcomes were collected. Overall survival (OS) and progression free survival (PFS) were estimated using the Kaplan-Meier method. In 37 patients with ESS, 3 clinicopathologic factors were associated with OS in a multivariate model— age (hazard ratio (HR) 1.05, 95% confidence interval (CI) 1.01-1.09, p=0.03), FIGO stage IV versus stage I disease (HR 4.05, 95% CI 1.11-14.8, p=0.03), and estrogen receptor status (HR 0.11, 0.02-0.69, p=0.02). Although the relationship between adjuvant therapy and OS was not significant, we demonstrate an association between adjuvant therapy and improved PFS in patients with ESS. Our observations suggest that advanced age and clinical stage are associated with worse outcomes OS and PFS in patients with ESS, while ER expression may be a marker of


INTRODUCTION
Uterine sarcoma is a rare gynecologic tumor arising from the myometrium or connective tissue of the uterus and accounting for 3-7% of all uterine malignancies (1). Endometrial sarcoma is a subtype of uterine sarcoma that represents between 7-17% of uterine sarcomas (2,3). Historically, endometrial sarcoma has been classified into 2 subtypes based on histological characteristics and mitotic rate-low grade endometrial stromal sarcoma (LGESS) and high grade endometrial stromal sarcoma (HGESS) (4). More recently, oncogenetic studies have demonstrated that the majority of LGESS contain the JAZF1/JJAZ1 gene fusion while this fusion is absent in the more aggressive HGESS (5). Based on differences in tumor genetics and natural history, the World Health Organization revised their classification of endometrial sarcoma into two categories-HGESS as undifferentiated endometrial sarcoma (UES) or high-grade undifferentiated uterine sarcomas (HGUS) with little or no evidence of endometrial stromal differentiation and LGESS as endometrial stromal sarcoma (ESS). The rarity of these tumors has led to difficulty in firmly establishing important prognostic factors and understanding the optimal treatment strategies, especially in those patients with advanced stage disease.
ESS frequently expresses estrogen receptor (ER) and progesterone receptor (PR) with a pooled analysis demonstrating ER expression in 40-100% of cases and PR expression in 60-100% of cases. The relationship between hormone receptor status and outcomes is uncertain, but ER or PR expression makes these tumors potential targets for hormonal therapy (6). Proestrogen states including obesity and the use of hormone replacement therapy have been speculated to be harmful in patients with ESS, and obesity has been evaluated with varying results as a risk factor for carcinogenesis and prognosis for endometrial sarcoma (7,8).
While hysterectomy is the treatment of choice for early stage EES, bilateraloophorectomy (BSO) is also typically recommended for patients given the hormonally responsive nature of the tumor although whether this imparts a survival advantage remains unclear (9)(10)(11). The role of adjuvant therapy including radiation, hormonal, or chemotherapy remains controversial in ESS with two retrospective series showing that adjuvant radiotherapy has been shown to improve disease free survival but not overall survival (OS) in ESS while another study demonstrated the absence of recurrence in patients with ESS given adjuvant hormonal or radiation therapy (12)(13)(14).
Most of the data regarding the clinical and pathologic variables that may predict prognosis and the benefits of adjuvant therapy are based on small retrospective studies (12)(13)(14)(15)(16).
In a population-based analysis from the Surveillance, Epidemiology, and End Results (SEER) database from 1988 to 2003, Chan et al reported on the relationship between clinical and pathologic variables and disease-specific survival (DSS), demonstrating a relationship between older age, black race, advanced stage, higher grade, lack of primary surgery and nodal metastasis and poorer DSS (17). Yoon et al reported a multicenter study in which 114 patient with ESS in which stage, expression of ER and PR and nodal metastasis were significantly associated with overall survival (OS) (18). In this study we review our experience over the past 22 years treating using the log-rank test. All tests were two-sided and a p value <0.05 was considered significant.
Data were analyzed using R version 2.15.2. The study was approved by the Institutional Review Board (IRB).

RESULTS
A total of 64 patients were identified through the Moffitt Cancer Center tumor registry with a diagnosis code of ESS. Of these, 27 patients were excluded for incomplete records (n=7, incomplete clinical or staging information, or no follow-up data) or a confirmed pathologic diagnosis other than ESS (n=20) resulting in 37 evaluable patients. The clinical and pathologic features of these patients are summarized in Table 1 Table 3 illustrates the univariate analysis of adjuvant treatments and survival in ESS. There was a trend towards improvement in OS for those receiving any adjuvant therapy (p=0.09).

DISCUSSION
Uterine sarcomas are rare tumors with clinical heterogeneity making prospective or controlled trials difficult, and as such much of the data related to the diagnosis is derived from retrospective case series (12)(13)(14). We sought to add to the body of literature regarding this diagnosis by summarizing our experience with ESS over a 22-year period with particular attention to the clinical and pathologic features important for prognosis. We also sought to describe a potential role for adjuvant therapy in high-risk patients.
Our study supports previous observations demonstrating worse overall survival in older patients with ESS. In a large analysis of 831 women with ESS from the Surveillance, Epidemiology and End Results (SEER) database, Chan et al found that age over 52 years was associated with a worse OS disease-specific survival (DSS) (P=<0.001) (17). Our study also demonstrated a relationship between age and outcome with each year increase in age adversely impacting overall survival (OS). As with many age-related disease outcomes, advanced age likely serves as surrogate for comorbid conditions and functional status limitations that are related to outcome but cannot completely be accounted for in multivariate models. While containing fewer patients treated at a single institution than the data reported by Chan et al, our data adds to the available literature by demonstrated the relationship between age and both OS and PFS in a more contemporary series of patients.
Obesity is another a suspected risk factor for endometrial sarcoma given the elevated estrogen levels associated with obesity and the peripheral conversion of androgens to estrogens by adipose tissue. One case control study of patients with all subtypes of uterine sarcomas demonstrated an association with obesity and an increased risk of death, and a pooled analysis showed a significant risk endometrial sarcoma in obese patients (7,8). The majority of the women diagnosed with ESS were overweight or obese with a median BMI 27 kg/m 2 , but we did not detect a relationship between overweight and obesity and outcomes including PFS or OS.
We are not able to draw definitive outcomes regarding obesity as a risk factor for ESS or worse outcomes, but given the small sample size it is possible that a true relationship was not detected.
Large multicenter collaborative studies are needed to further assess the relationship between obesity and disease incidence and outcomes in patients with ESS.
ESS can strongly express ER and PR, but some tumors are ER and PR negative (6,(19)(20)(21). The relationship between hormone receptor status and outcomes is unclear, but some data suggest that ER status is associated with improved outcomes. In one study of patients with all subtypes of uterine sarcoma, ER status correlated with improved median OS and was a strong prognostic factor (22). The mechanism by which ER and PR status may improve OS is unclear.
There may be an intrinsic difference in tumor biology, or the difference may be related to the use of targeted hormonal therapy in the adjuvant or metastatic setting. Our results confirm previous observations suggesting that ER positive tumors may have unique tumor biology as the relationship between receptor status and outcomes remained significant after adjusting for the use of hormonal therapy. ER and PR status should be reported in future research of ESS and survival analysis should be stratified based on these two groups to further assess if ER expression may be a prognostic marker in ESS.
While total hysterectomy is considered standard of care for localized ESS, controversy remains over whether extending hysterectomy to include bilateral salpingoopherectomy impacts survival (9)(10)(11)(23)(24)(25). In our analysis TAH-BSO was not associated with improved PFS or OS, possibly due to the small number of patients undergoing TAH alone. Since ESS may be an estrogen responsive malignancy TAH-BSO remains the recommendation for treatment of patients with ESS, but ovarian sparing surgery can be considered in younger patients (<35 years) with small tumors (2-3cm) (6).
Pelvic lymphadenectomy in the setting of clinically node negative is typically not recommended. A large multi-institutional review and National Cancer Institute Surveillance, Epidemiology, and End Results Program (SEER) study both concluded that there was no benefit to complete surgical staging in both ESS and UES (26,27). We describe an 8% incidence of nodal disease in patients with ESS. At our center therapeutic lymph node dissection remains the standard in patients with radiographic or clinical evidence of nodal disease, but elective lymph node dissection is not routinely performed.
In order to potentially improve outcomes in patients with ESS the use of adjuvant therapy has been explored. Options for adjuvant therapy include radiotherapy, chemotherapy, or hormonal therapy with these modalities being used alone or in combination. In a study of 22 patients, Kim el al concluded adjuvant therapy of any type had no benefit on disease-free survival in stage I ESS (9), while Diesing et al reported that survival was improved in six of eleven patients with UES and ESS who received adjuvant therapy (28). The only randomized clinical trial to study adjuvant radiotherapy for uterine sarcomas was completed by the European Organization for Research and Treatment of Cancer (29). In this study 28 patients with UES and ESS (of 224 uterine sarcomas total) were randomized to postoperative radiation or observation, and there was no survival benefit detected in any histological subtype, and no local control benefit in patients with UES and ESS (29). The results of these studies are limited by small sample sizes.
To date no randomized trial have been completed to determine the impact of adjuvant hormonal therapy on outcomes in hormone receptor positive UES or ESS, though retrospective studies have assessed the role of adjuvant hormonal therapy (10,21,30). In a multi institutional study of 43 patients with UES and ESS there was a decreased recurrence rate with post-operative progestin therapy while another study of 31 women with stage III to IV ESS and UES found that adjuvant progesterone therapy decreased the risk of disease progression from 75% to 20% (6,10). The role of adjuvant chemotherapy in ESS is questionable but treatment with doxorubicin or ifosfamide-based treatments have shown response rates in patients with ESS (31,32). A study of 127 patients with advanced stage uterine sarcoma (37 patients had ESS or UES) there was a benefit from adjuvant chemotherapy though results specific to ESS were not reported (33).
Data from our series demonstrates a relationship between the use of any adjuvant therapy and improved PFS with a trend towards improved OS. These results must be interpreted within the context of the small patient numbers, and the inherent limitation selection bias given a nonrandomized study design with regards to adjuvant therapy. Another limitation of the study is that given the small sample size and very long study period, we cannot account for changes in outcomes due to time trends. Finally, it is uncertain whether or not the results of this single institution series are generalizable to other centers. We recommend that until a more definitive clinical trial on the role of adjuvant therapy in ESS is completed, that adjuvant therapy be used on a case-by-case basis in patients with high-risk ESS (i.e. those with suboptimal surgery, those with advanced age, or other risk factors for recurrence).

CONCLUSION
ESS is a rare disease that is difficult to study in a prospective manner at a single center.
As with other rare tumors, collaborative multi-institutional and even international series and studies are needed to further define the natural history and optimal treatment strategy in ESS. In the absence of molecular pathologic data for analysis in this study, our observations suggest that factors such as hormone receptor status and age should be considered when discussing adjuvant therapy with ESS patients. These factors should best be studied in the context of clinical trials at institutions with experience in the management of ESS, and until more data is available adjuvant therapy should be considered in high-risk patients.