Prognostic Factors and Treatment Outcomes in 93 Patients with Uterine Sarcoma from 4 Centers in Turkey

Uterine sarcomas account for 1% of female genital tract cancers and 2-5% of uterine malignancies (Major et al., 1993; Tavassoli & Devilee, 2003). This heterogenous group of tumors derives from uterine mesodermal tissue and because of their rarity and histopathological diversity, there is a lack of consensus on prognostic factors and optimal treatment. Uterine sarcomas were classified into 4 main types: malignant mixt mullerian tumors (MMMT), accounting 40% of cases, also called carcinosarcomas, leiomyosacomas (LMS, 40% of cases), endometrial stromal sarcomas (ESS, 10-15% of cases) and undiferentiated sarcomas (5-10% of cases). Recently, many authors have proposed that MMMT should be classified as a subtype of endometrial carcinoma, due to the fact that their clinical behaviour looks like carcinoma (lymphatic dissemination pattern and response to platinum based chemotherapy) (Sleijfer et al., 2007). But MMMT behave more aggressively than endometrial


Introduction
Uterine sarcomas account for 1% of female genital tract cancers and 2-5% of uterine malignancies (Major et al., 1993;Tavassoli & Devilee, 2003).This heterogenous group of tumors derives from uterine mesodermal tissue and because of their rarity and histopathological diversity, there is a lack of consensus on prognostic factors and optimal treatment.Uterine sarcomas were classified into 4 main types: malignant mixt mullerian tumors (MMMT), accounting 40% of cases, also called carcinosarcomas, leiomyosacomas (LMS, 40% of cases), endometrial stromal sarcomas (ESS, 10-15% of cases) and undiferentiated sarcomas (5-10% of cases).Recently, many authors have proposed that MMMT should be classified as a subtype of endometrial carcinoma, due to the fact that their clinical behaviour looks like carcinoma (lymphatic dissemination pattern and response to platinum based chemotherapy) (Sleijfer et al., 2007).But MMMT behave more aggressively than endometrial carcinomas and they have been still included in most retrospective studies and reviews of uterine sarcomas.In last International Federation of Gynecology and Obstetrics (FIGO) classification system three new classifications have been developed: staging for LMS and ESS, staging for adenosarcomas, staging for carcinosarcomas (Prat, 2009).Carcinosarcomas are staged as endometrial adenocarcinomas.
The aim of this study is to provide retrospective evaluation of clinicopathological characteristics and prognostic factors and treatment outcomes of the patients with uterine sarcoma.

Materials and Methods
This study was designed as retrospective analysis of the patients with uterine sarcoma who were diagnosed and treated from November 2000 to October 2010.The patients were from 4 different centers in Turkey.The main parameters recorded were patient characteristics

1936
(age, menopausal status, body mass index (BMI), parity, patient's history of other malignancies and medical diseases), presenting symptoms, serum level of cancer antigen 125 (CA125), date and type of surgical procedure, presence or absense of residual tumor after surgery, pathological data of tumor (histological type, size, grade, mitotic index, results of peritoneal cytological examination, presence or absence of lymphovascular space invasion (LVSI), myometrial invasion and lymph node involvement), date of recurrence, treatment after recurrence, date of last medical examination and date of death.Histopathological classification of uterine sarcomas were according to WHO classification (Tavassoli & Devilee, 2003): MMMT (Carcinosarcoma), LMS, ESS, other histopathological types.Some patients were diagnosed after FIGO 2009 staging system therefore we used this staging system in this study and we restaged the patients diagnosed before that time according to FIGO 2009 staging system.Relaps free survival (RFS) was calculated as the time in months from the date of diagnosis to either the date of recurrence or the date of last followup.Overall survival was calculated as the time in months from the date of diagnosis to either the date of death or the date of last follow-up.CA125 levels greater than 35 U/mL were considered as positive.
Data were analyzed using SPSS version 11,5.Survival rates were calculated using the Kaplan-Meier method and survival curves were compared using log-rank test.
Variables showing significant differences after univariate analysis were evaluated with multivariate analysis by using Cox regression analysis.Hazard Ratios (HR) and 95% Coinfidence Interval (CI) were calculated for each variable.P value less than 0.05 was considered significant.
Eighty-nine patients underwent primary surgical treatment and tumor cells were completely removed in 74 patients.Surgical procedures are listed in Table 2.In MMMT group, 16 of 25 patients (64%) underwent pelvic and/or paraaortic lymph node dissection.Of the 89 patients who underwent primary surgical treatment, 68 patients received adjuvant therapy: 34 patients received only adjuvant chemotherapy, 11 patients received only radiotherapy, 2 patients received only hormonotherapy and 21 patients received sequential chemotherapy and radiotherapy (Table 2).
Median OS was 56 months (95% CI, 22.5-89.5 months).Cumulative RFS and OS is shown in Figure 1.OS after 1, 2, 5 years were 84.7%, 78%, 49.4% respectively.OS did not differ significantly between histological types (p= 0.917) (Figure 1).Univariate analysis of survival rates showed that age ≥60 years (p=0.002),elevated CA 125 levels (p <0.001), presence of residual tumor after surgery (p <0.001), high grade tumor (p=0.015),tumor size ≥10cm (p=0.010),presence of necrosis (p=0.024),advanced FİGO stage (p=0.006)were significantly associated with poor OS.However, parity, BMI, menopausal status, presence of comorbid diseases, type of primary operation and adjuvant therapy, peritoneal cytology (benign or malign), number of mitosis, presence or absence of LVI and lymph node involvement were not significantly associated with OS.Cumulative OS for each stage and for each histological type of uterine sarcoma is shown in Figure 1.
Multivariate analysis showed that age ≥60 years (p=0.004)and high grade tumor (p=0.035) were significantly associated with poor RFS.Patients who were administered adjuvant treatment with sequential chemotherapy and radiotherapy had longer RFS time (p=0.002).On analysis of RFS, p values for stage II versus stage I, for stage III versus stage I and stage IV versus stage I were 0.021, 0.146 and 0.007 respectively.Multivariate analysis of OS showed that only 2 factors, age ≥60 years (p=0.014)and high grade tumor (p=0.045)(p=0.019,HR:5.18,)and adjuvant sequential chemotherapy and radiotherapy after surgery (p=0.020,HR: 0.149, 95% CI: 0.03-0.74).
Our three patients had personal history of other malignancies, Benito et al and Koivisto-Korander et al have been reported higher occurence of previous cancer in this group of patients (Koivisto-Korander et al., 2008;Benito et al., 2009).Patients in MMMT group were mostly postmenopousal (88%), whereas rate of postmenapousal patients in LMS and ESS group were lower (45.1% and 22.2% respectively).This result is convenient with other studies (Benito et al., 2009;D'Angelo & Prat, 2010).Among our patients, the rate of having medical disease in MMMT group was also higher than that of LMS and ESS group (68% versus 40.7% and 33.3%, respectively).
Total hysterectomy and bilateral salphingooferectomy is usually reported as the most effective treatment for uterine sarcomas (Sleijfer et al., 2007;Gaducci et al., 2008Tsikouras et al., 2008).Addition of lympadenectomy to this procedure is indicated for MMMT group, because of high incidence of lymph node metastasis which was reported as 15-21% (Ali & Wells, 1993;Sartori et al., 1997;Menczer et al., 2005;Temkin et al., 2007;Gaducci et al., 2008).For localized leiomyosarcoma, the incidence of involvement of lymph node is rare, therefore lympadenectomy is not recommended (Gaducci et al., 2008).In our study, 16 patients (62.5%) in MMMT group underwent lymphadenectomy procedure and lymph node metastasis was found in 37.5% of them.But we found no significant effect of lymphadenectomy or any other type of primary surgical procedure on survival rates.
Our multivariate analysis showed that histological  .2012.13.5.1935 Prognostic Factors andTreatment Outcomes in Patients with Uterine Sarcoma in Turkey types of uterine sarcomas had no significant effect on survival.In literature, there are different findings about prognostic value of histological type.Most autors found no significant difference in clinical outcome according to histological type except low grade ESS (Salazar & Dunne, 1980;Echt et al., 1990;Wolfson et al., 1994).In some studies, it was reported that MMMT had poorer prognosis than other types (Ghaemmaghami et al., 2008;Benito et al., 2009).Conversely, Olah et al. reported that prognosis of LMS was poorer than that of MMMT (Olah et al., 1992).In our study, although there was no statistical difference between OS of different histological types; in numerical value, survival rate of LMS at 1 year-period was higher than that of MMMT (87.7% vs 72.4%), but after 5 yearperiod survival rate of MMMT was better than that LMS (66.9% vs 41.9%).RFS and OS rates for each histological type of uterine sarcoma are shown in Table 3. OS rates of MMMT tend to be same after 2 and 5 years, while that of LMS were decreasing.In the study of Benito et al. there was somewhat similar results about this aspect: OS rates of MMMT group after 2, 5, 10-year period were same (26%), while that of LMS decreasing over time, 72%, 42%, 21% respectively (Benito et al., 2009).It might be suggested that almost all deathts in MMMT group occur within 2 years, then survival rates tend to be stable until 5-10 years.But making such a conclusion requires further larger-long term studies.
As reported in other studies (Chavenic et al.,1999;El Husseiny et al., 2002;Kokawa et al., 2006;Koivisto-Korander et al., 2008;Tsikouras et al., 2008), in our study most patients were diagnosed at early stages (stage I-II, 63.4%).Most of the patients in LMS group and ESS group were diagnosed at stage I-II (74.1% and 88.9% respectively), while MMMT group was mainly diagnosed at advanced stages (stage III-IV, 64%).We used the new FIGO staging system.Therefore, in terms of stage, comparison of our study with previous studies may not be rationale.Some of the previous studies have reported that tumor stage is most significant prognostic factor for uterine sarcomas (Gaducci et al., 2008;Park et al., 2008;Benito et al., 2009;D'Angelo & Prat, 2010;Sharma et al., 2011).In our study, there were some conflicting findings about stage.Survival results of patients at different stages were shown in Table 4. Survival results of stage III group was better than that of stage II after 2 years period.This finding may be explained by 2 factors: firstly our stage III group was mostly composed of patients with MMMT (70%) and survival results of MMMT group was better than that of LMS group after 2 years.Secondly according to new staging system, some of LMS and ESS patients at stage II and stage III of previous system were included in stage I and stage II respectively.When we analyzed only LMS group to eradicate the effect of MMMT group on survival, stage III group continued to deteriorate the results.There was no significant effect of tumor stage on survival after multivariate analysis.
In a review (D'Angelo & Prat, 2010), it was concluded that the role of adjuvant therapy on survival of uterine sarcomas was uncertain, adjuvant radiation therapy might be useful in improving local control.Some studies showed improved outcome by administrating adjuvant chemotherapy with different regimens (Nagell et al., 1986;Odunsi et al., 2004;Gaducci et al., 2008;Matoda et al., 2011), whereas some studies did not show (Omura et al., 1985;Hempling et al., 1995).In a Gynecologic Oncology Group (GOG) study including 65 patients with uterine MMMT, adjuvant chemotherapy with ifosfamide and cisplatin was associated with 7-year progression-free survival and OS of 54% and 52% respectively (Sutton et al., 2005).In a phase II study, pelvic radiation "sandwiched" between chemotherapy was reported as efficacious treatment (Einstein et al., 2012).In a retrospective study of 49 patients with uterine MMMT, adjuvant sequential chemotherapy and radiotherapy significantly decreased mortality rate compared to adjuvant chemotherapy alone (Menczer et al., 2005).Our study suggested that adjuvant sequential chemotherapy and radiotherapy for uterine sarcomas improved RFS and had no effect on OS.When we analyzed only LMS group, adjuvant treatment with sequential chemotherapy and radiotherapy was associated with better OS.
S o m e s t u d i e s r e p o r t e d t h a t t u m o r g r a d e (Ghaemmaghami et al., 2008), tumor size (George et al., 1986;Rovirosa et al., 2002;Benito et al., 2009), depths of myometrial invasion (Rovirosa et al., 2002;Sagage et al., 2004;Park et al., 2008), LVSI (Major et al., 1993;Rovirosa et al., 2002;Park et al., 2008) had significant effect on prognosis of uterine sarcomas.In our study grade had significant effecet on RFS and OS.Age and grade were found to be independent prognostic parameters for our LMS group, too.
In conclusion, our study confirms that uterine sarcomas are rare malignancies with poor prognosis, the prognosis is poor even at early stages and adjuvant treatment with sequential chemotherapy and radiotherapy improves

Figure
Figure 1.Relapse-Free Survival (A) and Overall Survival (B) for all Uterine Sarcomas; Overall Survival for Each Sarcoma Stage (p=0.516)(C) and for Each Sarcoma Subtype (p=0.917)(D).

Table 2 . Surgical Management and Adjuvant Treatmentª
MMMT, malignant mixed Mullerian tumor; ESS, endometrial stromal sarcoma; HT, hysterectomy; BSO, bilateral salpingooophorectomy; USO, unilateral salpingo-oophorectomy, VAC, vincristin+actinomycin-D+cyclophosphamide, ªValues are given as number of patients, b All sarcomas, c Other histologic, of MMMT group was higher than that of other Tumor characteristics are shown in Table