Treatment of Uterine Sarcomas

Uterine sarcomas are a heterogeneous group of rare tumors that make up about 4% of cancers of the uterine corpus and 1% of all tumors of the female genital tract.^1,2,3,4,5,6 In contrast to endometrial carcinomas, they are among the most lethal gynecologic malignancies. This group of tumors is important because they must be considered in the differential diagnosis of all patients with abnormal uterine bleeding and are clinically challenging with respect to diagnosis and management of primary as well as recurrent disease.²

The three most common histologic types of uterine sarcoma are carcinosarcoma (malignant mixed mullerian tumor) and leiomyosarcoma, each of which make up about 40% of cases, and endometrial stromal sarcoma, which represents 15%.² Other rare sarcomas make up the remaining 5% of tumors from this group. The classification and histologic origin of these tumors are outlined in Table 1 and were covered in the previous chapter.⁷

TABLE 1. Histologic Origin of Common Uterine Sarcomas

The most widely referenced epidemiologic study on uterine sarcomas was conducted by Harlow and coworkers,⁵ who analyzed the population-based registries of the Surveillance, Epidemiology and End Results program of the National Cancer Institute from 1973 to 1981. The incidence rate per million women per year was 8.2 for carcinosarcoma, 6.4 for leiomyosarcoma, 1.8 for endometrial stromal sarcoma, and 0.7 for unclassified sarcomas. There was a marked racial disparity between the incidence rates of carcinosarcoma and leiomyosarcoma, with higher incidences in black women than in white women. The standard incidence ratio for carcinosarcoma and leiomyosarcoma was 2.7 and 1.6, respectively. This data is in contrast to the findings of Schwartz and associates⁸ in Israel, who found a higher incidence in women of European or American origin than in those of Asian or African descent.

There is a distinct age-specific incidence pattern with uterine leiomyosarcoma and carcinosarcoma.⁵ For both blacks and whites, the incidence of uterine leiomyosarcoma reaches a plateau in middle age; however, the incidence of carcinosarcoma gradually increases from age 40 to 44 to age 50 to 54, then steeply increases thereafter.

Epidemiologic risk factors for uterine sarcoma have not been clearly defined, except for prior radiation exposure. An association has been observed between exposure to radiation and development of uterine sarcomas, most of which are carcinosarcomas. The percentage of patients with a history of previous pelvic radiation ranges from 0% to 37%, with a latency period of 1.5 to 27.5 years.^{1,2,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23} Meredith and associates²⁴ analyzed 1,208 cases of uterine corpus cancer from the Ohio State University Registry from 1940 to 1983 to address the relation between a history of previous pelvic irradiation and the development of sarcoma. Of the 30 patients with a history of pelvic irradiation, 5 (17%) had a carcinosarcoma of the uterus, and 4 of these patients died of their malignancy within 7 months. Although there are data to suggest that postirradiation sarcomas of nongynecologic origin are diagnosed at advanced stages and are highly aggressive,²⁵ there are few data to extend this observation to the gynecologic oncology literature. No other epidemiologic factors examined in the literature, including parity (increased²⁶ and decreased²⁷ risk), age at first birth,²⁶ and marital status,²⁸ have consistently demonstrated an effect on risk of developing uterine sarcoma.

Although the natural history and patterns of metastases differ for the three major types of uterine sarcomas, several generalizations can be made regarding their behavior.¹ First, carcinosarcomas, high-grade leiomyosarcomas, and endometrial stromal sarcomas have aggressive growth patterns and a poor survival rate, with most patients dying 1 to 2 years after diagnosis.¹ Second, low-grade leiomyosarcomas and low-grade endometrial stromal sarcomas (endolymphatic stromal myosis) as well as adenosarcomas are less aggressive tumors associated with high cure rates.¹ When these tumors recur, their growth is usually local and indolent.¹

Uterine sarcomas can spread by lymphatic and hematogenous routes²⁹ as well as by local extension and peritoneal spread.³⁰ Several studies have addressed the metastatic pattern of uterine sarcomas. Chen³¹ examined nodal metastases in 20 patients with clinical stage I uterine sarcomas. Fourteen patients had carcinosarcoma, four had leiomyosarcoma, and two had endometrial stromal sarcoma. Of nine patients (45%) with lymph node metastases, six had both para-aortic and pelvic node involvement and three had only pelvic node involvement. A high frequency of association was observed between nodal spread and deep myometrial invasion. Other sites of extrauterine spread were the ovary (two patients) and the serosa of the uterus (one patient). DiSaia and colleagues³² reported on 28 patients with clinical stage I and II carcinosarcoma who underwent total abdominal hysterectomy, bilateral salpingo-oophorectomy, and pelvic and para-aortic lymph node sampling. Ten (35%) of the 28 patients had positive pelvic nodes; 4 of these 10 patients also had para-aortic node metastasis (14.5%). In every instance of nodal involvement, the myometrial invasion was to the middle or outer third.

Rose and associates²⁹ studied the autopsy findings of 73 patients with uterine sarcoma (43 with carcinosarcoma, 19 with leiomyosarcoma, 9 with endometrial stromal sarcoma, and 2 with endolymphatic stromal myosis). The peritoneal cavity and omentum were the most frequently involved sites (59%), followed by the lung (52%), pelvic (41%) and para-aortic (38%) lymph nodes, and liver parenchyma (34%). The presence of lung metastasis was not associated with pelvic or para-aortic nodal metastasis or intraperitoneal disease.

In another review of autopsy findings, Fleming and co-workers³⁰ reviewed the records of 22 patients treated for uterine sarcoma (11 carcinosarcomas, 6 endometrial stromal sarcomas, and 5 leiomyosarcomas) and found that 59% of the patients had lymph node involvement. Forty-five percent of the patients died with disease limited to the pelvis and abdomen. The most common site of disease above the diaphragm was the lung. The authors reviewed autopsy findings of 58 patients from previous reports in the literature^{33,34,35,36,37,38,39,40,41,42,43} (24 with carcinosarcoma, 29 with leiomyosarcoma, and 5 with endometrial stromal sarcoma) and found a similar incidence of nodal involvement (57%) and disease limited to the pelvis and abdomen (31%). In addition, 65.5% of the patients had abdominal as well as distant metastases.

Goff and colleagues⁴⁴ examined nodal metastasis patterns of uterine leiomyosarcomas and endometrial stromal sarcomas. Retroperitoneal lymph node sampling was performed on 15 of 21 women with leiomyosarcoma and 7 of 10 women with endometrial stromal sarcoma. In the women with leiomyosarcoma, 4 of 15 had lymph node metastases; in each of these, there was also disseminated intraperitoneal disease. No lymph node metastases were present in the women with endometrial stromal sarcoma.

Prognostic factors also differ for the three major types of uterine sarcomas. Major and co-workers⁴⁵ reported the Gynecologic Oncology Group (GOG) clinicopathologic study of clinical stage I and II uterine sarcoma that included 59 patients with leiomyosarcoma and 301 with carcinosarcoma. In patients with leiomyosarcoma, lymphvascular space involvement and involvement of the cervix and isthmus were common, whereas lymph node metastases, adnexal metastases, and positive peritoneal cytology were uncommon. The only surgicopathologic finding that correlated with progression-free interval was mitotic index. There were no treatment failures among the three women who had fewer than 10 mitoses per 10 high-power fields (HPF), but 61% of women with 10 to 20 mitoses per 10 HPF and 79% of women with more than 20 mitoses per 10 HPF developed recurrences. Fewer mitoses was also associated with longer survival.

In contrast to patients with leiomyosarcoma, patients with carcinosarcoma had high rates of nodal and adnexal metastases as well as positive peritoneal cytology. Pelvic nodes were involved twice as often as aortic nodes (15% vs. 7.8%), and both nodal groups were involved in 5% of the patients. Surgicopathologic factors related to progression-free interval included adnexal spread, lymph node metastasis, histologic cell type (heterologous vs. homologous), and grade of sarcoma.

Recurrence patterns have been characterized for leiomyosarcomas, carcinosarcomas, and endometrial stromal sarcomas. Berchuck and colleagues⁴⁶ reviewed the Memorial Sloan-Kettering experience with uterine leiomyosarcomas and observed recurrence rates of 83.3%, 68.2%, and 100% for patients with stage I/II disease receiving adjuvant therapy, stage I/II disease not receiving adjuvant therapy, and stage III/IV disease, respectively. In patients with stage I and II disease, 64% recurred in the pelvis and abdomen and 48% recurred in the lung. Similarly, in patients with stage III/IV disease, 50% recurred in the pelvis and lung and 42% recurred in the abdomen. In the GOG study of early-stage sarcoma,⁴⁵ the recurrence rate for uterine leiomyosarcoma was 71%. Only 13.6% of patients recurred first in the pelvis, whereas 40.7% and 16.9% recurred first in the lung and extrapelvic sites, respectively. In another study, 14 isolated lung recurrences were observed in 21 patients with recurrent leiomyosarcoma.⁴⁷ For carcinosarcoma, the overall recurrence rate was 53% in the GOG study of early-stage sarcoma.⁴⁵ Similar percentages of patients recurred first at pelvic and extrapelvic sites (20.9% and 20.6%, respectively), whereas 13.6% recurred first in the lung.

There are fewer data regarding recurrence patterns for endometrial stromal tumors. Berchuck and associates⁴⁸ reviewed the Memorial Sloan-Kettering experience with endometrial stromal tumors, including 22 patients with endolymphatic stromal myosis (low-grade stromal sarcoma) and 9 with endometrial stromal sarcoma. All three patients who presented with advanced endometrial stromal sarcoma died of rapidly progressive disease, whereas two of the six patients with stage I disease developed recurrent disease. Of 15 patients treated for recurrent disease, 6 recurred in the abdomen and pelvis, 6 in the vagina or pelvis, and 3 in the lung.

The most common presenting symptom for all types of uterine sarcomas is abnormal vaginal bleeding, which occurs in 75% to 95% of patients.¹ Pelvic pain occurs in one third of patients with carcinosarcoma and may be associated with the passage of clots or a sarcomatous mass through an effaced cervix.^49,50,51 Pain is also an important, early symptom in patients with leiomyosarcoma.^1,50,52,53 The presenting signs of patients with uterine sarcomas include an enlarging abdominopelvic mass and profuse, foul-smelling vaginal discharge. Although patients with carcinosarcomas and endometrial stromal sarcomas may present with a normal pelvic examination, uterine enlargement is present in 50% to 95% of cases.^1,11,50,51 Likewise, uterine enlargement has been observed in up to 50% of patients with leiomyosarcoma.^1,52,53 In one study of patients with carcinosarcoma, 73% had gross tumor at the cervical os on presentation.¹²

Despite the above signs and symptoms, the diagnosis of uterine sarcoma is frequently made only at the time of hysterectomy for a presumed benign condition.^1,44,54,55 Studies have demonstrated that in most cases preoperative endometrial sampling is not a reliable method for making the diagnosis of uterine sarcoma.⁴⁴ Patients with carcinosarcoma who have bleeding usually have an endometrial biopsy that reveals the poorly differentiated component of the tumor.¹ However, for leiomyosarcomas and endometrial stromal sarcomas, preoperative dilation and curettage of the uterus misses the diagnosis in at least 40% and 20% of cases, respectively.^{44,46,48,53,54,55,56} Schwartz and coworkers,⁵⁷ attempting to define characteristic features of leiomyosarcomas that may preoperatively distinguish benign from malignant myomas, observed that leiomyosarcomas were both broad-based and pedunculated, had no preferential uterine location, and in 95% of cases were either the largest or only mass. They recommended that the largest myoma be monitored most closely during conservative management of a uterus containing leiomyomas. Parker and associates⁵⁸ examined the prevalence of uterine sarcomas in patients operated on for presumed leiomyoma and rapidly growing leiomyoma. Only 1 of 371 women undergoing hysterectomy for an indication of a rapidly growing uterus was found to have a leiomyosarcoma.^1,58

The preoperative evaluation of patients with uterine sarcomas should include a complete blood count as well as serum chemistries. Radiologic studies should include a chest x-ray. Computed tomography can be used to identify patients with extrauterine spread and magnetic resonance imaging can be used to assess depth of myometrial invasion.^59,60 The role of ultrasonography is being studied by several investigators.^61,62,63 Kurjak and coworkers⁶¹ evaluated the role of transvaginal color Doppler in differentiating uterine sarcomas from leiomyomas. They examined 10 cases of uterine sarcomas with respect to color Doppler sonographic patterns and compared this information with 150 normal and 1,850 myomatous uteri. All uterine sarcomas exhibited abnormal tumoral blood vessels. The mean resistance index was significantly lower than that of the normal and myomatous uteri, and the authors concluded that this technique had the potential to distinguish uterine sarcoma from benign uterine lesions.

The International Federation of Gynecology and Obstetrics (FIGO) staging criteria for corpus cancer is commonly used for uterine sarcomas.^1,64 This surgical staging system is illustrated in Table 2.

TABLE 2. Surgical Staging of Uterine Sarcomas

(Creasman WT: New gynecologic cancer staging. Obstet Gynecol 75:287, 1990)

Patients with a preoperative or intraoperative diagnosis of uterine sarcoma should undergo an exploratory laparotomy, total abdominal hysterectomy, and a staging procedure. Thorough surgical staging is important because it will influence subsequent management of the patient. Peritoneal washings should be taken at the beginning of surgery. In patients with carcinosarcomas limited to the uterus by pathologic staging, the cytologic presence of malignant cells in the peritoneal washings is a poor prognostic factor.⁶⁵ The upper abdomen and pelvis should then be explored, with palpation of the pelvic and para-aortic lymph nodes.

All patients should undergo an extrafascial hysterectomy. There are recommendations in the literature that in low-grade endometrial stromal sarcoma, a radical hysterectomy should be performed if the tumor is palpable in the parametrium.^1,66 The ovaries should be retained in premenopausal patients with leiomyosarcoma because this appears to improve the prognosis.^{2,41,67,68,69,70} However, a bilateral salpingo-oophorectomy should be performed in all other patients, including those with low-grade stromal sarcomas, because these tumors may be hormone-dependent or -responsive, and they have a propensity for extension into the parametria, broad ligament, and adnexal structures.

There are no randomized trials addressing the role of pelvic and para-aortic lymph node sampling at the time of laparotomy in patients with uterine sarcomas. For carcinosarcomas, a high percentage of patients with clinical stage I or II disease are upstaged at the time of laparotomy.^{11,12,17,20,71} Thus, it appears reasonable to stage these patients surgically. However, there is no role for lymph node sampling in patients with gross intra-abdominal tumor metastases. There are few data regarding the role of lymph node sampling in patients with leiomyosarcoma and endometrial stromal sarcoma, but it appears that almost all patients with these sarcomas who have lymph node metastases also have evidence of intraperitoneal disease spread.^72,73

The finding of a leiomyosarcoma arising in a myoma specimen after myomectomy in a young patient is rare and in general is associated with a good prognosis. Lurain and Piver² reviewed 5 reports^{4,67,69,74,75} and found that of 22 patients who initially underwent myomectomy for leiomyosarcoma, there was only 1 recurrence, in the series reported by Dinh and Woodruff.⁶⁹ Therefore, it appears that myomectomy is curative in leiomyosarcomas confined to a benign-appearing myoma at the time of surgery.

Unlike other gynecologic malignancies, there is a role for thoracotomy in patients with uterine sarcoma metastatic to the lung. Levenback and coworkers⁷³ reviewed 45 patients whose pulmonary metastases from uterine sarcoma were resected at Memorial Sloan-Kettering. Most of these tumors were leiomyosarcomas (84%), with endometrial stromal sarcomas and carcinosarcomas representing 9% and 7% of the total, respectively. Most of the lesions were unilateral (71%), 70% were greater than 2 cm, and half were isolated lesions. The mean survival of patients with unilateral disease (39 months) was significantly greater than that of patients with bilateral disease (27 months). No single risk factor was identified that could exclude an individual patient from consideration for pulmonary resection.

There is also a role for surgery in the treatment of local and regional recurrences of uterine sarcomas.¹ In the MD Anderson series of 120 patients with carcinosarcomas reported by Spanos and colleagues,⁷⁶ 67 patients developed recurrent disease. In the six patients with lesions deemed suitable for resection (one locoregional, two pulmonary, and three abdominal), five complete and partial responses occurred, for an overall response rate of 83%. Recurrent or metastatic low-grade endometrial stromal sarcoma may also be amenable to surgical excision of pelvic disease or pulmonary metastases.

There are no randomized trials that have specifically addressed the role of radiation therapy in the treatment of uterine sarcomas. Current studies indicate that radiation therapy has a role in the treatment of carcinosarcomas and endometrial stromal sarcomas, but leiomyosarcomas do not appear to respond to radiation therapy.²

Primary irradiation may be used for patients who are inoperable for medical or technical reasons.¹ Several investigators have examined recurrence after primary irradiation therapy in carcinosarcomas.^{21,50,76,77,78} The pelvic recurrence rate ranges from 17% to 75% and the distant recurrence rate from 33% to 100%, which is higher than the recurrence rates observed for surgery alone or surgery combined with radiation therapy. The largest series of patients receiving primary irradiation was reported by Spanos and associates.⁷⁶ Of the 18 patients who underwent primary irradiation (7 in conjunction with chemotherapy), 3 patients had local and regional recurrences and 6 patients had distant recurrences (2 to the lung and supraclavicular nodes, 4 to the abdomen and para-aortic nodes). Primary irradiation for stage I and II disease should include 4500 cGy external beam irradiation to the whole pelvis followed by two intracavitary applications for a total of 4,000 to 5,000 mg/hour to the uterus and 3,500 to 4,000 mg/hour vaginal surface dose.¹

Most studies have found adjuvant pre- or postoperative radiation therapy to be of value in terms of increased survival and decreased pelvic recurrence in patients with localized carcinosarcoma and endometrial stromal tumors, but not leiomyosarcoma. Vongtama and colleagues⁷⁹ reported an overall 5-year survival of 53% in 31 patients receiving either pre- or postoperative irradiation, compared to 40% in 61 patients treated by surgery alone and 29% in 12 patients who received only radiation therapy.² Also, the rate of local treatment failure (i.e., pelvic recurrences) was less than half for the group receiving combined therapy (26%) than for those having surgery alone (57%).²

Other studies have confirmed the value of adding pelvic radiation to surgery for the treatment of stage I and II uterine sarcomas.² Perez and coworkers⁸⁰ noted pelvic recurrence in only 3 of 17 patients (18%) receiving postoperative pelvic irradiation compared to 3 of 6 patients (50%) treated with surgery alone for stage I carcinosarcomas.² Covens and colleagues⁸¹ reported that the number of local recurrences decreased in stage I tumors from 9 of 22 cases in which surgery alone was performed to 0 of 15 cases in which pelvic radiation was added.² In a GOG study,⁸² patients who received radiation therapy to the pelvis after surgery for stage I and II carcinosarcomas had significantly fewer recurrences within the radiation treatment field.²

Data regarding recurrence rates with and without radiation therapy for endometrial stromal sarcomas are limited. Berchuck and coworkers⁴⁸ reported on 25 patients with stage I and II endometrial stromal sarcomas who underwent hysterectomy. Three of these patients received external pelvic radiation with or without intravaginal radiation therapy. None of these patients developed local recurrences, although 1 patient developed a distant recurrence at 29 months. The recurrence rate for the 18 patients who did not receive adjuvant therapy was 56%. Piver and associates⁷⁸ studied 52 patients with endolymphatic stromal myosis. Of 34 patients with stage I disease, 16 (47%) developed local recurrences. However, no local recurrences occurred in the five patients who received adjuvant pelvic radiation therapy.

Radiation therapy thus seems to have a role in the combined treatment of carcinosarcoma and endometrial stromal tumors confined to the pelvis by increasing pelvic control and increasing the disease-free progression interval. This, then, probably has some effect on increasing overall survival, although this has been difficult to prove. The GOG is conducting a phase III study of accelerated hyperfractionated whole abdominal radiation therapy versus ifosfamide and mesna with cisplatin in patients with optimally debulked stage I to IV carcinosarcoma (protocol 150).

Although uterine sarcomas are relatively rare, the dismal survival rates of patients with these tumors has made the development of effective chemotherapy regimens a high priority. Studies by the GOG first identified the differential sensitivity of carcinosarcomas and leiomyosarcomas to drug therapy.⁸³ Because these two cell types appear to respond differently to chemotherapy, they will be discussed separately.

Single-Agent Chemotherapy

Several drugs have been studied as single-agent therapy in advanced or recurrent carcinosarcomas or leiomyosarcomas, including cisplatin,^84,85,86 ifosfamide,⁸⁷ doxorubicin,^83,88 etoposide,^89,90 mitoxantrone,⁹¹ piperazinedione,⁹² and aminothiazole.⁹³

Of the six agents studied in carcinosarcomas, only ifosfamide and cisplatin have shown significant response rates. Ifosfamide is the most active single agent in the treatment of advanced or recurrent carcinosarcomas of the uterus.¹ Sutton and coworkers⁸⁷ conducted a phase II study of ifosfamide and mesna in 30 patients with advanced or recurrent carcinosarcomas who had no prior chemotherapy. Of 28 patients available for response, 5 had complete responses (17.9%) and 4 had partial responses (14.3%), for a total response rate of 32.2%. However, the response duration ranged from 1.4 to 8.6 months, with a median response duration of only 3.8 months.

Cisplatin was first studied in patients with advanced or recurrent carcinosarcomas with prior chemotherapy and radiation therapy. Thigpen and associates⁸⁵ conducted a phase II study of cisplatin in patients with advanced or recurrent carcinosarcomas of the uterus using doses of 50 mg/m². Of 28 patients available for evaluation, 2 complete responses and 3 partial responses occurred, for a total response rate of 17.8%. The duration of the 2 complete responses was 14 and 16 months, and the median duration of the partial responses was 3 months. In this population, 42.9% of patients had received radiation therapy and 89.3% had received prior therapy with doxorubicin or a doxorubicin-containing combination. This prior chemotherapy history is significant, given that second-line chemotherapy for sarcomas has been notably unsuccessful. Gershenson and coworkers⁸⁶ treated 18 patients with advanced or recurrent carcinosarcomas with cisplatin using a higher dose of 100 mg/m². Of 12 patients with measurable disease, 2 had no prior chemotherapy or radiation therapy, 1 had prior radiation therapy, 3 had prior chemotherapy with doxorubicin, and 4 had a combination of doxorubicin and radiation therapy. In this group of patients, 1 complete response (duration, 5.5 months) and 4 partial responses (duration, 5 months each) were observed, for an overall response rate of 42%.

The results of these studies prompted further investigations in patients not previously exposed to cytotoxic agents. Thigpen and associates⁸⁴ treated 63 patients with advanced or recurrent carcinosarcomas with primary cisplatin chemotherapy as part of a larger GOG study of cisplatin and uterine sarcomas. Forty-four percent of the patients had received prior radiation therapy. There were 5 complete responses and 7 partial responses for an overall response rate of 19%. The median duration of response was 6.6 months; the median survival was 16.8 months for complete responders and 9.3 months for partial responders.

Doxorubicin is the most active single agent in the treatment of advanced or recurrent leiomyosarcoma. In the two GOG phase III trials comparing doxorubicin-based chemotherapy in combination with dimethyl-triazeno-imidazole carboxamide (DTIC)⁸³ or cyclophosphamide⁹⁴ in advanced uterine sarcomas, response rates of 25% and 13%, respectively, were observed in patients with leiomyosarcoma treated with doxorubicin alone.⁸³ Patients with leiomyosarcomas also had a significantly longer survival time than the other histologic cell types studied (12.1 vs. 6.0 months). Ifosfamide also has activity in this disease. The GOG conducted a phase II trial of ifosfamide and mesna in patients with advanced or recurrent leiomyosarcomas of the uterus.⁹⁵ Of 35 patients who could be evaluated for response, 6 had partial responses (17.2%). However, no complete response was observed, and the median response duration was only 3.8 months. Thus, ifosfamide appears to have activity comparable to that of doxorubicin as a single agent.⁹⁵

There are few data in the gynecologic literature regarding the use of chemotherapy for endometrial stromal sarcomas. The GOG conducted a phase II study of ifosfamide in women with metastatic endometrial stromal sarcomas previously unexposed to chemotherapy. Of 21 patients who could be evaluated, 3 had complete tumor responses and 4 had partial responses, for an overall response rate of 33.3%.⁹⁶ Other agents have also been found to be active in the treatment of endometrial stromal sarcomas, including doxorubicin,^44,48 the combination of vincristine, actinomycin D, and cyclophosphamide (VAC),⁴⁸ the combination of mitomycin and vinblastine,⁴⁸ chlorambucil,⁹⁷ and the combination of cyclophosphamide, vincristine, doxorubicin, and dacarbazine (CYVADIC).⁹⁷

Combination Chemotherapy

Because carcinosarcomas and leiomyosarcomas are so rare, there is a lack of randomized phase III clinical trials of combination chemotherapy with each tumor specifically.

Two randomized phase III clinical trials with doxorubicin-based chemotherapy have been conducted in patients with uterine sarcomas before investigators separated these two tumor types in studies. As discussed previously, the GOG conducted a randomized phase III study of doxorubicin with and without DTIC in patients with advanced uterine sarcomas.⁸³ No significant differences between the two regimens were identified, and neither histologic type accrued a sufficient number of patients to allow subset analysis.¹ In a second study, the GOG compared doxorubicin with or without cyclophosphamide in patients with advanced or recurrent uterine sarcomas.⁹⁴ The overall response rate was the same for doxorubicin with or without cyclophosphamide (19.2% total response rate); thus, no benefit was demonstrated by adding cyclophosphamide to doxorubicin. When the data were analyzed by cell type, 3 of the 20 patients with carcinosarcomas had complete responses.

In view of the activity of cisplatin and ifosfamide in carcinosarcomas, in early 1989 the GOG activated a phase III trial of ifosfamide and the uroprotector mesna with or without cisplatin in patients with advanced, persistent, or recurrent carcinosarcomas (protocol 108). This study was completed in July 1996, and the results have not yet been published.

Cisplatin has also been studied in the treatment of advanced or recurrent carcinosarcomas in several series in combination with hexamethylmelamine, cyclophosphamide, and doxorubicin,⁹⁸ with DTIC⁹⁹ and doxorubicin,¹⁰⁰ and with etoposide and doxorubicin.¹⁰¹

Given the central role of doxorubicin-based combination chemotherapy in the treatment of softtissue and bone sarcomas,¹⁰² several studies have used combination chemotherapy for the treatment of advanced or recurrent uterine leiomyosarcoma. Azizi and coworkers¹⁰³ treated six patients with metastatic uterine leiomyosarcoma with a combination of vincristine, doxorubicin, and DTIC. Three patients had complete responses and one patient had a partial response. The average duration of response was 15.6 months, with 1 patient free of disease at 24 months. As discussed previously, the GOG conducted two phase III trials with doxorubicin-based chemotherapy in advanced uterine sarcomas and found no benefit to the addition of either cyclophosphamide or DTIC to these regimens.^83,94 Piver and associates¹⁰⁴ used the CYVADIC regimen to treat 26 patients with pelvic sarcomas. Of the 10 patients with uterine sarcomas, the only patient with leiomyosarcoma had a complete response but was dead of disease at 10 months. The GOG conducted a phase II study of doxorubicin in combination with ifosfamide and mesna in previously untreated patients with advanced or recurrent uterine leiomyosarcoma.¹⁰⁵ Of 33 patients, there was 1 complete response and there were 10 partial responses. Although the overall response rate was 33.3%, toxicity was severe, with 13 patients developing grade 4 neutropenia and 2 patients developing grade 4 cardiotoxicity.

The GOG recently reported on the results of a phase II trial of hydroxyurea, DTIC, and etoposide in the treatment of advanced or recurrent uterine leiomyosarcoma.¹⁰⁶ Of 38 patients available for evaluation, there were 2 complete responses and 5 partial responses, for an overall response rate of 18.4%; 6 of the 7 responders had disease outside the pelvis. The response duration was 12.1 months. Thus, this combination had moderate activity in the treatment of advanced or recurrent uterine leiomyosarcoma.

Current treatment recommendations for metastatic or recurrent carcinosarcomas and uterine leiomyosarcomas are outlined in Table 3.

TABLE 3. Chemotherapy for Advanced or Recurrent Uterine Sarcomas

Adjuvant Chemotherapy

The high recurrence rate in stage I and II disease has led investigators to examine the role of adjuvant chemotherapy in patients with uterine sarcomas. The only randomized trial of adjuvant chemotherapy in patients with uterine sarcomas was conducted by the GOG from 1973 to 1982 (protocol 20).¹⁰⁷ One hundred fifty-six patients with stage I or II uterine sarcoma were randomly assigned to adjuvant chemotherapy with doxorubicin for 6 months or no further treatment; pelvic irradiation was optional before randomization. Most patients in the study had stage I disease (82.7% for the observation group and 85.3% for the doxorubicin treatment group). In the doxorubicin treatment group, 33.3% of patients had leiomyosarcoma and 58.7% had carcinosarcomas; in the observation group, 28.4% had leiomyosarcoma and 60.5% had carcinosarcomas. The recurrence rate was similarly high in both groups (41.3% for patients who received adjuvant doxorubicin and 53.1% for those who received no treatment) and was the same with and without pelvic radiation therapy. The progression-free interval and survival were not prolonged in the doxorubicin treatment group. The authors concluded that there was no benefit to adjuvant treatment with doxorubicin in this setting. Similarly, other investigators have failed to demonstrate a benefit to the use of adjuvant chemotherapy in nongynecologic sarcomas.^{108,109,110,111}

However, the role of adjuvant chemotherapy has also been studied in several smaller studies, with more encouraging results. Piver and associates¹¹² conducted a study of adjuvant doxorubicin chemotherapy in 19 patients with stage I uterine sarcoma. The patients treated with doxorubicin had a lower recurrence rate (25% vs. 64%) and a longer median survival time (68 months vs. 38 months) in comparison to the patients treated with surgery alone. Because of the significant responses to CYVADIC in advanced pelvic sarcomas reported by the same authors, the doxorubicin trial was terminated and a study of adjuvant CYVADIC after surgery for stage I uterine sarcomas was initiated without a concurrent control group.¹¹² In the 11 patients studied over a 5-year period, only 2 developed recurrent disease (1 with leiomyosarcoma and 1 with endometrial stromal sarcoma); the 5-year survival rate was 89%.

Two studies using VAC chemotherapy in an adjuvant setting in patients with early-stage disease have been reported. Buchsbaum and coworkers¹¹³ administered VAC for 6 courses over 6 months to 17 patients with stage I or II uterine sarcoma after hysterectomy. Of 10 patients available for evaluation from 3 to 7 years, 4 developed recurrent disease and died, 1 was alive with disease, and 5 were alive and free of disease. Likewise, van Nagell and associates¹¹⁴ conducted a pilot study with adjuvant VAC chemotherapy after hysterectomy in stage I uterine sarcoma. Of 7 patients studied, 2 developed recurrent disease and the remaining 5 were free of recurrent disease 48 to 73 months after treatment. Kohorn and colleagues¹¹⁵ treated five patients with stage I and II carcinosarcomas with radiation, surgery, and adjuvant chemotherapy (four with doxorubicin and cyclophosphamide, one with doxorubicin and cisplatin). Four patients (80%) were disease-free for 38 to 96 months.

Hannigan and coworkers¹¹⁶ reviewed 9 years' experience with early uterine sarcoma at MD Anderson. Sixty-seven patients received no adjunctive chemotherapy. Thirty-four patients received either adjuvant VAC chemotherapy or doxorubicin alone or in combination with vincristine and cyclophosphamide. There was no improvement in survival or disease-free interval in the patients receiving adjuvant chemotherapy.

The GOG recently reported the results of a study of three cycles of adjuvant ifosfamide, mesna, and cisplatin in patients with completely resected clinical stage I or II carcinosarcoma of the uterus.¹¹⁷ There were 65 patients who could be evaluated in the study. The study originally used a 5-day regimen, but this was reduced to 4 days early in the study because of myelotoxicity. Despite this dose reduction, 63.1% of patients experienced grade 3 or 4 neutropenia and 26.2% of patients had grade 3 or 4 thrombocytopenia. With a minimum of 2 years follow-up, 41 patients (63.1%) were progression-free and 48 (73.8%) were alive. In GOG protocol 20, in which either doxorubicin or no adjuvant chemotherapy was given to a similar group of patients, 24 of 44 patients (54.5%) were progression-free and 26 (59.1%) were alive after doxorubicin; 25 of 50 (50%) were progression-free and 31 (62%) were alive after surgery alone.¹⁰⁷ Further maturation of this data is required to assess whether adjuvant ifosfamide, mesna, and cisplatin will improve progression-free interval and survival after primary surgery.

Uterine sarcomas are a heterogeneous group of rare tumors that are problematic with respect to diagnosis and management of primary and recurrent disease. The primary treatment of these tumors is surgical and should include abdominal exploration, hysterectomy, and bilateral salpingo-oophorectomy, as well as lymph node sampling in carcinosarcomas. Radiation therapy seems to have a role in the postoperative treatment of carcinosarcomas and endometrial stromal sarcomas, but not uterine leiomyosarcomas, confined to the pelvis by decreasing pelvic recurrences and thereby increasing the disease-free progression interval and probably overall survival. Although the use of adjuvant chemotherapy is attractive, given the biologically aggressive nature of these tumors as well as their metastatic pattern, its value has not been proven. Chemotherapy with doxorubicin and ifosfamide or cisplatin and ifosfamide combinations is used to treat metastatic or recurrent disease. Active clinical trials currently being conducted by the GOG and other groups can be accessed on the National Cancer Institute's website on the Internet.