Prognostic factors in undifferentiated uterine sarcoma: a subanalysis of the SARCUT study

The aim of this study was to analyze the prognostic factors related to the recurrence rate and overall survival of patients with undifferentiated uterine sarcoma. An international multicenter study involving 43 international centers, the SARCUT study, collected 966 uterine sarcoma cases; among them 39 cases corresponded to undifferentiated uterine sarcoma and where included in the present subanalysis. The risk factors related to the oncological outcomes where analyzed. The median age of the patients was 63 (range 14–85) years. Seventeen (43.5%) patients presented FIGO stage I. The 5-year overall survival (OS) was 15.3% and 12-months disease-free survival (DFS) 41%. FIGO stage I was significantly associated with a better prognosis. In addition, patients who received adjuvant radiotherapy showed significant longer disease-free survival compared to those without adjuvant radiotherapy (20.5 vs. 4.0 months, respectively; p = 0.04) and longer overall survival (34.7 vs. 18.2 months, respectively; p = 0.05). Chemotherapy administration was associated with shorter DFS (HR 4.41, 95% CI 1.35–14.43, p = 0.014). Persistent disease after primary treatment (HR = 6.86, 95% CI 1.51–31.09, p = 0.012) and FIGO stage IV (HR 4.12, 95%CI 1.37–12.44, p = 0.011) showed significant worse prognosis for OS. FIGO stage seems to be the most important prognostic factor in patients with undifferentiated uterine sarcoma. Adjuvant radiotherapy seems to be significantly associated also to a better disease-free and overall survival. On the contrary, the role of chemotherapy administration remains unclear since was associated to a shorted DFS.


Introduction
Uterine sarcoma is an uncommon malignant disease of the female genital tract accounting for 1% of all malignant gynecologic tumors and about 3-9% of all uterine malignancies [1][2][3] with an incidence of 0.36 per 100,000 woman-years in the United States [4]. Sarcoma originates from myometrium or connective tissue of the uterus and has an aggressive course and poor prognosis. Leiomyosarcoma is the most common histological type accounting for 60% of the uterine sarcomas, followed by endometrial stromal sarcoma, undifferentiated uterine sarcoma and adenosarcoma [5]. Endometrial stromal sarcoma is subclassified into 3 categories by the World Health Organization (WHO): endometrial stromal nodules (ESNs), endometrial stromal sarcoma (ESS, historically referred to as low-grade sarcoma), and undifferentiated endometrial sarcoma (UES, historically called high-grade sarcoma). Undifferentiated sarcoma is an exclusion diagnosis when there are no characteristic features of other type sarcomas and such diagnosis is set in 5% of cases [6].
Kurihara et al. [7] proposed to subcategorize undifferentiated uterine sarcomas into pleomorphic with marked nuclear pleomorphism and uniform type with relative nuclear isomorphism. There are distinct differences between those two types of undifferentiated uterine sarcomas using immunohistochemistry where estrogen and progesterone receptors as well as b-catenin and cyclin D1 more frequently expressed among uniform tumors and aberrant p53 staining patterns in pleomorphic ones [7].
Data on prognostic factors is very limited due to the rarity of the disease. Approximately 70% of patients are diagnosed at FIGO stage III-IV, and among them more than 50% at stage IV disease with a median disease-free survival (DFS) of 7-10 months and a median overall survival (OS) of 11-23 months [23].
Our objective was to analyze the prognostic factors of undifferentiated uterine sarcoma patients by means of a subanalysis of the international SARCUT study (SARComa of the UTerus). An international multicenter study involving 43 international  centers and 52 researchers, the SARCUT study, collected  966 uterine sarcoma cases diagnosed from 2001 to 2007 with  a follow-up until 2012; among them 39 cases corresponded to undifferentiated uterine sarcoma and where included in the present subanalysis. Institutional Review Board was first obtained at the coordinating center (#PI-1382), and posteriorly at the rest of participating institutions.

Materials and methods
We included all cases with pathological confirmation of uterine undifferentiated uterine sarcoma and primary treatment performed at the participating center. Patients with primary treatment carried out in other centers were excluded.
In all cases, the same oncological team in each center performed all surgical interventions and management of the patients. External beam radiotherapy (EBRT), brachytherapy and chemotherapy were selectively used postoperatively depending on the Institutional Tumor Board decision, as well as patients' follow-up.
Descriptive patient demographics and characteristics of the patients and tumor stage according to FIGO staging classification were collected [24]. Primary treatment and treatment at recurrence were recorded. Uniform criteria of nomenclature for surgical procedures, pathologic variables, and sites of recurrence were applied. The presence of recurrence, type of therapies, and patient status were collected during the patient's follow-up. Local recurrence was defined as the appearance of tumor in the same location after a minimum disease-free period of 6 months, and distant recurrence when it appeared in a new location after treatment.
Classification of surgical cytoreduction was performed according to Zapardiel and Morrow classification [25].

Statistical analysis
Qualitative variables were described by absolute values and percentages. Quantitative data were described by median and range. For comparison between groups, we used the Mann-Whitney U test and Kruskal-Wallis test. Categorical variables were compared using chi-squared test. Multivariate analysis by logistic regression was performed to determine potential prognostic factors. Survival analysis was performed using Kaplan-Meier curves and log-rank test. All comparisons were 2-tailed, and alpha error was set at 5%. Hazard ratios with 95% confidence intervals were calculated. All data were analyzed using SPSS software version 22.0 (SPSS Inc, Chicago, IL) and SAS 8.0 (SAS Institute Inc. Madrid, Spain).

Results
Thirty-nine patients with undifferentiated uterine sarcoma were included in the study. All baseline characteristics are shown in Table 1.
Nineteen (48.7%) patients presented early stage disease (FIGO stage I-II), while 20 (51.3%) patients presented advanced stage disease (FIGO stage III-IV). Vaginal bleeding was the most frequent symptom accounting in 21 (58.3%) cases. In 14 (35.9%) patients the disease was affecting outside of the uterus, and in 35 (89.7%) cases malignant disease was suspected before treatment was initiated.
The majority of patients underwent total hysterectomy (32 patients-82%) with bilateral salpingoophorectomy by laparotomic approach (32 patients-82%). Two patients had laparascopic surgery (5.1%) as a diagnostic procedure. In patients with tumor spread additional procedures were performed (Table 1). In 5 (12.8%) patients pelvic lymphadenectomy was performed and 2 (5.1%) of them also had a para-aortic lymphadenectomy. One patient had para-aortic lymphadenectomy without pelvic lymphadenectomy. None of those cases had metastatic lymph nodes. Complete resection with residual disease less than 1 mm was achieved in 25 (73.5%) patients.
Among the 13 (33.3%) patients with positive surgical margins, chemotherapy was given in 7 (17.9%) cases. Four (10.2%) of them also received adjuvant pelvic radiotherapy. Three (7.7%) patients with stage II-IV disease did not received any adjuvant treatment despite the involvement of surgical margins. Radiotherapy without chemotherapy was given in 3 (7.7%) patients, all of them with tumoral cut-through. Among the 26 (66.7%) patients with clear resection margins, chemotherapy was administered in 8 (20.5%) patients and 5 (12.8%) of them also received radiotherapy. All these 8 (20.5%) patients had stage I-II disease. Adjuvant radiotherapy was given in 15 (38.4%) patients of this patient subgroup. In addition, we did not find significant differences in the rate of chemotherapy administered between early or advanced stages (42.1% vs. 51.4%, respectively).  Age, tumor size, surgical resection margin, tumor necrosis, lymphovascular space involvement, extrauterine tumor spread and type of cytoreductive surgery were not predictive factors for disease recurrence and OS in our sample.

Discussion
FIGO Stage at diagnosis seems to be the most important prognostic factor in undifferentiated uterine sarcoma. In our study, only 17 (43.5%) patients presented FIGO stage I, corresponding almost to the only chance to get cured from the disease; however, Meurer M et al. observed even a lower rate of early stage disease, between 20 and 34% of patients [23,[26][27][28]. Early diagnosis in undifferentiated uterine sarcoma it is complicate since they account for 0.2% of all uterine malignancies and around 6% of all uterine sarcomas [5]. Occasionally patients present vaginal bleeding, in such cases, an endometrial Pipelle or hysteroscopic biopsy may lead to the diagnosis [26,29]. In our study, 58.3% of patients presented it, maybe because of it 89.7% of cases were suspected before the primary treatment that is very unusual in uterine sarcomas. Additional tools that may lead to the preoperative diagnosis are the ultrasound scan and MRI, which integrated in a validated score such as PRESS Score; uterine sarcoma could be predicted with an 84.1% of accuracy [30]. In addition, MRI and ultrasound may identify tumor necrosis, present in 43.5% of our cases and in 84.6% of Meurer M et al. study [26], and it could help to identify undifferentiated uterine sarcoma [31]. Tumor size and fast growth [32], as well as other factors such as smoking habit, diabetes and obesity, has been identified as risk factors for  uterine sarcoma, but the reality is that they seems to be too unspecific for their use in the clinical setting [33,34]. In fact, in our population the median size was 50 mm (80% of patients presented tumors under 60 mm) compared to 80 mm reported by Meurer et al. [26], and only 2 patients presented previous smoking history.
When the patient is not diagnosed in early stages and the tumor is disseminated in the abdominal cavity, which occurred in 35.9% of our cases, the cornerstone of the treatment seems to be a complete resection of the tumor, including always hysterectomy and bilateral salpingoophorectomy [35]. Our rates of complete resection are comparable to previously reported proportion of 70% [26]. The management of the retroperitoneal area remains unclear [5,15]. We observed some cases where regional lymphadenectomy was performed but in all of them negative results were obtained. The only clear indication for it could be in the rare cases of bulky nodes that would need to be removed to obtain a complete cytoreduction, as well as happened with other organs, which would require isolated resection of metastases [23,36,37].
Adjuvant radiotherapy in the management of undifferentiated uterine sarcoma seems to have a significant positive impact in prognosis, both in DFS and OS. In the study of Malouf et al. only postoperative pelvic radiotherapy with or without brachytherapy correlated with improved DFS (19.1 vs 6.5 months; P = 0.04) and OS (54.5 vs 16.7 months; P = 0.01) [23]. Meurer et al. [26] reported medians for DFS and OS, 46.7 and 39 months, respectively, among those who underwent adjuvant radiotherapy and 41 and 10.3 months for those who underwent adjuvant chemotherapy. In Meurer´s study [26], after multivariate analysis, adjuvant radiotherapy showed to be an independent prognostic factor for both disease-free survival (P = 0.036) and overall survival (P = 0.012). In addition, Ríos et al. reported an absolute local control of the disease (50% of them were long term survivals) after adjuvant radiotherapy [38]. However, when including all tumor types besides undifferentiated uterine sarcomas, such in EORTC 55,874 randomized trial, it did not show any benefit [28,39]. In our study, 16 (41.0%) patients with no residual tumor and 4 (10.3%) patients with positive surgical margin reported in pathology received adjuvant pelvic radiotherapy, with excellent results.
Interestingly, in our data, we found a worse prognosis among patients who received chemotherapy, with a significant poor impact on DFS with a HR 4.41. More than one third (38.5%) of the patients received medical treatment in our study, which is a bit higher compared to other series published reporting from 10 to 30% of cases [23,28,38]. In addition, the use of chemotherapy has been shown to offer a protective effect with a HR 0.13 in other studies [26], which for us makes more sense. A possible explanation of our findings could be due to the FIGO stage distribution, since our rate of early disease is lower (48.8%) compared to those other populations (76.9%) that dramatically impact the prognosis [26], although this is not clear whether chemotherapy adds anything even in early stage disease. Moreover, among all patients with chemotherapy 10 presented a progressive disease so the utility of chemotherapy was really limited and it could explain worse survival among patients receiving adjuvant chemotherapy which is not related to chemotherapy itself rather than an aggressive tumor behavior and advanced tumor stage. Unfortunately, there was no information available as to which cytostatic substances and chemotherapy regimens the patients were treated with in the adjuvant setting and this may also increase the heterogenicity in our study population.
Recently there have been raising interest on hormonal treatment of uterine sarcomas especially those expressing estrogen and progestin receptors. In the study of Gremel et al. [40] even undifferentiated uterine sarcomas expressed estrogen and progestin receptors in 22.7% and 18.2% of cases. Hormonal therapies, including aromatase inhibitors, progestins and gonadotropin-releasing hormone analogs may represent an effective option with fewer side effects when compared to chemotherapy with comparable effectiveness. There are studies showing better survival results for uterine sarcomas expressing estrogen and progestin receptors, but these are limited to low patient proportion with undifferentiated uterine sarcomas [41,42].
Median OS in our study was 20 months what is slightly worse than 32.7 months published by Meurer et al. [26], that could be explained but the higher rate of advanced disease in our study; but very similar to Tanner et al. [23] and Malouf et al. [27] with OS rates of 11.8 and 23 months, respectively. Similar findings were observed for DFS, in our data the median DFS was 15.5 months and Meurer et al. [26] reported 23 months; again, other data were closer to our results reporting median DFS between 9 and 12 months [23,28].
Finally, findings within our study should be considered with caution because there are significant limitations with a low number of patients having a specific morphology, retrospective study nature and heterogenicity of clinical factors that may affect solid and conclusive data.

Conclusions
FIGO stage seems to be the most important prognostic factor in patients with undifferentiated uterine sarcoma. Adjuvant radiotherapy seems to be significantly associated also to a better disease-free and overall survival. On the contrary, the role of chemotherapy administration remains unclear since was associated to a shorted DFS, probably due to the disease characteristics of the population who received it in our study.
Author contributions RM: project development, data collection, manuscript writing, manuscript review and approval; YY: data collection, manuscript writing, manuscript review and approval; RM: data collection, manuscript writing, manuscript review and approval; PA: data collection, manuscript writing, manuscript review and approval; TKM: data collection, manuscript writing, manuscript review and approval; EO: data collection, manuscript writing, manuscript review and approval; MRM: data collection, manuscript writing, manuscript review and approval; OZ: data collection, manuscript writing, manuscript review and approval; IZ: project development, data collection, data analysis, manuscript writing, manuscript review and final approval. As a main author of the paper have been doing statistical calculations myself, as well as article writing. I have been participating with a second largest patient amount in the SARCUT study and thereafter I have beet adviced to do a subanalysis of patients with undifferentiated uterine sarcomas. Before recruiting patients into SARCUT study, I have been particitating in the treatment of some of those patients by performing a surgery.
Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Declarations
Competing interests The authors have no relevant financial or nonfinancial interests to disclose.

Ethics approval
This study is an observational study. The reference Ethics Committee at La Paz University Hospital in Madrid, Spain, approved the study with reference number PI-1382 on October 2012.

Informed consent
Since this is a retrospective observational study, patients signed the informed consent for the treatment of the disease, but it was not required to be included in the study according to the reference Ethics Committee.