Neville F. Hacker
Endometrial carcinoma is the most common malignancy of the female genital tract in the Western world and the fourth most common cancer in women after breast, lung, and colorectum. Developing countries and Japan have incidence rates four to five times lower than Western industrialized nations, with the lowest rates being in India and south Asia (1).
In the United States, it is anticipated there will be 42,160 new cases and 7,780 deaths from the disease in 2009 (2). Black women have a 40% lower risk of developing the disease but a 54% greater risk of dying from it, mainly because of late diagnosis (3).
Two different clinicopathological subtypes of endometrial cancer are recognized: the estrogen-related (type I, endometrioid), and the non-estrogen-related (type II, nonendometrioid). Each subtype has specific genetic alterations, with endometrioid tumors showing microsatellite instability and mutations in PTEN, PIK3CA, K-ras, and CTNNBI (β-catenin), although nonendometrioid (predominantly serous and clear cell) tumors exhibit p53 mutations and chromosomal instability (4).
Approximately 80% of newly diagnosed endometrial carcinomas in the Western world are endometrioid in type (4). Any factor that increases exposure to unopposed estrogen (e.g., estrogen-replacement therapy, obesity, anovulatory cycles, estrogen-secreting tumors) increases the risk of these tumors, whereas factors that decrease exposure to estrogens or increase progesterone levels (e.g., oral contraceptives or smoking) tend to be protective (1).
The average age of patients with endometrioid cancer is approximately 63 years, and 70% or so are confined to the corpus at the time of diagnosis. Their 5-year survival is approximately 83% (5). By contrast, the average age of patients with nonendometrioid cancer is 67 years, and at least half have already spread beyond the corpus at the time of diagnosis. Their 5-year survival is approximately 62% for clear cell carcinomas and 53% for papillary serous cancers (5).
Endometrial cancer may occasionally develop after radiation treatment for cervical cancer. In such cases, the majority are diagnosed at an advanced stage of disease and are high-risk histological subtypes. Their prognosis is poor but does not appear to be significantly worse when compared to patients with high-stage, high-grade sporadic cancers (6).
Screening of Asymptomatic Women
The ideal method for outpatient sampling of the endometrium has not yet been devised, and no blood test of sufficient sensitivity and specificity has been developed. Therefore, mass screening of the population is not practical. However, screening for endometrial carcinoma or its precursors is justified for certain high-risk people, including those shown inTable 10.1.
Table 10.1 Patients for Whom Screening for Endometrial Cancer Is Justified
Only approximately 50% of women with endometrial cancer have malignant cells on a Papanicolaou (Pap) smear (7). However, compared with patients who have normal cervical cytologic findings, patients with suspicious or malignant cells are more likely to have deeper myometrial invasion, higher tumor grade, positive peritoneal cytologic findings, and a more advanced stage of disease (8).
The appearance of normal-appearing endometrial cells in cervical smears taken in the second half of the menstrual cycle or in postmenopausal women is controversial. Montz reported endometrial histology from 93 asymptomatic postmenopausal women receiving hormonereplacement therapy with normal endometrial cells on a Pap smear. Eighteen patients (19%) had abnormalities identified, including seven endometrial polyps, seven cases of simple hyperplasia (one with atypia), three cases of complex hyperplasia (one with atypia), and one endometrial carcinoma (9). A recent Dutch study of 29,144 asymptomatic postmenopausal women reported that when normal endometrial cells were found in the cervical smear, the prevalence rate of (pre-) malignant uterine disease was significantly higher (6.5%) as compared to smears without these cells (0.2%), resulting in a relative risk of 40.2 [95% confidence interval (CI) 9.4-172.2] (10). If morphologically abnormal endometrial cells are present, then approximately 25% of women have endometrial carcinoma (11).
Guidelines from the Bethesda system recommended that benign endometrial cells (BECs) should be reported in women age 40 and older. Beal et al. reported that BECs were rarely associated with significant endometrial pathology in asymptomatic premenopausal women and that these women may not need further evaluation (12). By contrast, Moroney et al., using liquidbased cytology, reported that 2.1% of asymptomatic premenopausal women with BECs had significant endometrial pathology (13).
The unsatisfactory results obtained with cervical cytology are the result of the indirect sampling of the endometrium, and several commercially available devices have been developed to allow direct sampling (e.g., Pipelle, Gyno Sampler, Vabra aspirator). A satisfactory endometrial biopsy specimen also may be obtained in the office with a small curette such as a Novak or Kevorkian (Fig. 10.1). All of these office techniques for endometrial sampling cause the patient some discomfort, and in approximately 8% of patients it is not possible to obtain a specimen because of a stenotic os. This failure rate increases to approximately 18% for women older than 70 years of age (14).
A metaanalysis reported that the Pipelle was the best device, with detection rates for endometrial cancer in postmenopausal and premenopausal women of 99.6% and 91%, respectively (15). The sensitivity for the detection of endometrial hyperplasia was 81%. The specificity for all devices was 98%.
In the 1990s, transvaginal ultrasonography, with or without color-flow imaging, was investigated as a screening technique. Mean thickness of the endometrial strip was measured as 3.4 ± 1.2 mm in women with atrophic endometrium, 9.7 ± 2.5 mm in women with hyperplasia, and 18.2 ± 6.2 mm in women with endometrial cancer (16). In a large, multiinstutional study of 1,168 women, all 114 women with endometrial cancer and 95% of the 112 women with endometrial hyperplasia had an endometrial thickness of 5 mm or more (17). A metaanalysis reported that 4% of endometrial cancers would be missed using transvaginal ultrasonography for the investigation of postmenopausal bleeding, with a false positive rate as high as 50% (18).
Tamoxifen increases the risk of endometrial cancer twofold to threefold (19) and produces a sonographically unique picture of an irregularly echogenic endometrium that is attributed to cystic glandular dilatation, stromal edema, and edema and hyperplasia of the adjacent myometrium (20). Routine ultrasonic surveillance of asymptomatic women ontamoxifen is not useful because of its low specificity and positive predictive value. Canadian workers studied 304 women on tamoxifen as therapy for breast cancer. Even using an endometrial thickness cutoff of 9 mm, the positive predictive value for the detection of endometrial cancer was only 1.4% (21).
Figure 10.1 Devices used for sampling endometrium. Top, Metal curette, e.g., Kevorkian (pictured); Bottom, Flexible plastic endometrial sampler, e.g., Pipelle (Unimar), Endocell (pictured) (Wallach Surgical Devices, Inc.). From Hillard PJA. Benign Diseases of the Female Reproductive Tract. In: Berek JS, ed. Berek & Novak's Gynecology. 14th ed. Philadelphia: Lippincott Williams & Wilkins, 2007.
Patients taking tamoxifen should be informed of the increased risk of endometrial cancer and told to report any abnormal bleeding or spotting immediately. Any bleeding or spotting must be investigated by biopsy. A retrospective review of tamoxifen treated women who underwent dilatation and curettage found that uterine cancer was found only in those with vaginal bleeding (20).
Endometrial carcinoma should be excluded in all patients shown in Table 10.2. Ninety percent of patients with endometrial cancer will have abnormal vaginal bleeding, most commonly postmenopausal bleeding, and the bleeding usually occurs early in the course of the disease. The usual causes of postmenopausal bleeding are shown in Table 10.3.Intermenstrual bleeding or heavy prolonged bleeding in perimenopausal or anovulatory premenopausal women should arouse suspicion.
The diagnosis may be delayed unnecessarily in these women because the bleeding is usually ascribed to “hormonal imbalance.” A high index of suspicion also is needed to make an early diagnosis in women younger than 40 years of age.
Occasionally, vaginal bleeding does not occur because of cervical stenosis, particularly in thin, elderly, estrogen-deficient patients. In some patients with cervical stenosis, a hematometra develops, and a small percentage have a purulent vaginal discharge resulting from a pyometra.
Table 10.2 Patients in Whom a Diagnosis of Endometrial Cancer Should Be Excluded
Table 10.3 Etiology of Postmenopausal Bleeding
Physical examination commonly reveals an obese, hypertensive, postmenopausal woman, although approximately one-third of patients are not overweight. Abdominal examination is usually unremarkable except in advanced cases when ascites may be present and hepatic or omental metastases may be palpable. Occasionally, a hematometra appears as a large, smooth midline mass arising from the pelvis.
On pelvic examination, it is important to inspect and palpate the vulva, vagina, and cervix to exclude metastatic spread or other causes of abnormal vaginal bleeding. The uterus may be bulky, but often it is not significantly enlarged. Rectovaginal examination should be performed to evaluate the fallopian tubes, ovaries, and cul-de-sac. Endometrial carcinoma may metastasize to these sites or, alternatively, coexistent ovarian tumors such as a granulosa cell tumor, thecoma, or epithelial ovarian carcinoma may be noted.
All patients suspected of having endometrial carcinoma should have an endocervical curettage and an office endometrial biopsy. A histologically positive endometrial biopsy allows the planning of definitive treatment.
Because there is a false negative rate of approximately 10%, a negative endometrial biopsy in a symptomatic patient must be followed by a fractional curettage under anesthesia. A diagnosis of endometrial hyperplasia on endometrial biopsy does not obviate the need for further investigation.
Hysteroscopy is often performed in conjunction with curettage and may identify some small bleeding polyps that would otherwise have been missed. There has been speculation that fluid hysteroscopy may facilitate the abdominal dissemination of malignant cells, but there is no evidence that it has any impact on the disease-free survival (22,23).
While the patient is under anesthesia, careful bimanual rectovaginal examination is performed, a weighted speculum is placed in the vagina, and the cervix is grasped with a tenaculum. The endocervical canal is curetted before cervical dilatation, and the tissue placed in a specially labeled container. The uterus then is sounded, the cervix dilated, and the endometrium systematically curetted. The tissue is placed in a separate container so that the histopathologic status of the endocervix and endometrium can be determined separately.
Routine preoperative investigations for early stage endometrial carcinoma are shown in Table 10.4. If a fractional curettage has not been performed, then an endocervical curettage should be performed to evaluate the endocervix.
Nonroutine tests are sometimes indicated, particularly for more advanced cases or high-risk histologies on curettage. A colonoscopy should be performed if there is occult blood in the stool or a recent change in bowel habits because concomitant colon cancer occasionally occurs, particularly if there is a family history of bowel cancer. A pelvic and abdominal computed tomographic (CT) scan may be helpful to determine the extent of metastatic disease in the following circumstances:
Table 10.4 Routine Preoperative Investigations for Early Stage Endometrial Carcinoma
However, it has limited usefulness in determining the depth of myometrial invasion or the presence of nodal disease (24,25). Magnetic resonance imaging (MRI) was evaluated as a tool for preoperative staging in a National Cancer Institute cooperative study (26). Eighty-eight patients from five participating hospitals were entered in the study. For evaluating the depth of myometrial invasion, the overall accuracy was 66%, but the imaging was considered adequate for the evaluation of paraaortic lymph nodes in only 8% of the cases. Subsequent studies have shown an 83.3% accuracy (100 of 120 cases) for differentiating deep from superficial myometrial invasion (27) and a positive predictive value of 89.8% for the detection of cervical involvement (28). Hence, MRI may help differentiate between low- and high-risk patients.
Positron emission tomography (PET) has only recently been evaluated for the preoperative assessment of patients with endometrial cancer. Although its sensitivity for the detection of extrauterine lesions (excluding retroperitoneal nodes) was somewhat superior to CT or MRI (83.3% vs. 66.7%), its utility is limited by its inability to identify lymph nodes less than 1 cm in diameter (29).
Elevated CA125 levels have been demonstrated to correlate with advanced stage of disease and positive lymph node status (30).
In 1988, the Cancer Committee of the International Federation of Gynecology and Obstetrics (FIGO) replaced the old clinical staging system (Table 10.5) with a surgical staging system for endometrial cancer (Table 10.6). For updated FIGO surgical staging tables (2008), see table 10.6A on page 666. Previously, the disease was staged clinically, based on examination under anesthesia, sounding the uterus, and a limited number pf preoperative investigations. The change was mainly in response to the Gynecologic Oncology Group (GOG) studies, which demonstrated the high incidence of lymph node metastases in high-risk cases (31,32).
As increasing experience with the surgical staging of endometrial cancer has been reported, it seems apparent that there is no need to perform systematic lymphadenectomy in low-risk cases (grade 1 or 2 endometrioid tumors confined to the inner half of the myometrium) (33). These patients require only removal of palpably suspicious nodes.
For high-risk cases (grade 3, serous or clear cell histologies, stages IC or II disease), a systematic pelvic lymphadenectomy should be performed, with at least removal of any clinically suspicious paraaortic lymph nodes (33).
Endometrial carcinoma spreads by the following routes:
Table 10.5 1971 FIGO Clinical Staging for Endometrial Carcinoma
Direct Extension Direct extension is the most common route of spread, and it results in penetration of the myometrium and eventually the serosa of the uterus. The cervix and fallopian tubes and ultimately the vagina and parametrium may be invaded. Tumors arising in the upper corpus may involve the tube or serosa before involving the cervix, whereas tumors arising from the lower segment of the uterus involve the cervix early. The exact anatomic route by which endometrial cancer involves the cervix has not been clearly defined, but it probably involves a combination of contiguous surface spread, invasion of deep tissue planes, and lymphatic dissemination (34).
Transtubal Dissemination The presence of malignant cells in peritoneal washings and the development of widespread intraabdominal metastases in some patients with early stage endometrial cancer strongly suggest that cells may be exfoliated from the primary tumor and transported to the peritoneal cavity by retrograde flow along the fallopian tubes.
Lymphatic Dissemination Lymphatic dissemination is clearly responsible for spread to pelvic and paraaortic lymph nodes. Although lymphatic channels pass directly from the fundus to the paraaortic nodes through the infundibulopelvic ligament, it is rare to find positive paraaortic nodes in the absence of positive pelvic nodes. However, it is quite common to find microscopic metastases in both pelvic and paraaortic nodes, suggesting simultaneous spread to pelvic and paraaortic nodes in some patients. This is in contrast to cervical cancer, where paraaortic nodal metastases are always secondary to pelvic nodal metastases.
It seems likely that vaginal metastases also result from lymph-vascular spread. They commonly occur in the absence of cervical involvement, excluding direct spread as the mechanism, and may occur despite preoperative sterilization of the uterus with intracavitary radiation, excluding implantation of cells at the time of surgery as the mechanism (35). In a study of 632 patients with stage I endometrial cancer managed with hysterectomy at the Mayo Clinic between 1984 and 1996, Mariani and colleagues reported that grade 3 histology and lymphovascular invasion were significant predictors of vaginal relapse, whereas depth of myometrial invasion was not (36).
Table 10.6 1988 FIGO Surgical Staging for Endometrial Carcinoma
Hematogenous Spread Hematogenous spread most commonly results in lung metastases, but liver, brain, bone, and other sites are involved less commonly.
Although stage of disease is the most significant prognostic variable, a number of factors have been shown to correlate with outcome in patients with the same stage of disease. These prognostic variables are summarized in Table 10.8. Knowledge of them is essential if appropriate treatment programs are to be devised.
Table 10.7 Carcinoma of the Endometrium: Distribution by Surgical Stage for Patients Treated in 1999 to 2001
Age appears to be an independent prognostic variable. The GOG reported 5-year relative survival rates of 96.3% for 28 patients no older than 40 years of age, 87.3% for 261 patients 51 to 60 years, 78% for 312 patients 61 to 70 years, 70.7% for 119 patients 71 to 80 years, and 53.6% for 23 patients older than 80 (p = 0.001) (37). All patients had clinical stage I or occult stage II disease. Using proportional hazards modeling of relative survival time, and taking 45 years of age as the arbitrary reference point, the relative risks for death from disease were as follows: 2.0 at 55 years, 3.4 at 65 years, and 4.7 at 75 years of age.
A study of 51,471 patients from the Surveillance, Epidemiology, and End Results (SEER) data base of the National Cancer Institute in the United States demonstrated that patients 40 years and younger had an overall survival advantage compared with women older than 40 years, independent of other clinicopathological prognosticators (38).
Japanese workers have reported menopausal status to be an independent prognostic variable for early endometrial cancer but not for patients with advanced disease (39).
A retrospective review of 388 patients treated at the Mayo Clinic recorded an uncommon histologic subtype in 52 patients (13%). There were 20 adenosquamous, 14 serous papillary, 11 clear cell, and 7 undifferentiated carcinomas (40). In contrast to the 92% survival rate among patients with endometrioid carcinoma, the overall survival rate for these patients was only 33%. At the time of surgical staging, 62% of the patients with an unfavorable histologic subtype had extrauterine spread of disease.
Zaino et al. (41) investigated the prognostic significance of squamous differentiation in 456 patients with typical adenocarcinomas and 175 women with areas of squamous differentiation who had been entered into a GOG clinicopathologic study of stage I and II disease. They reported that the biologic behavior of these tumors reflected the histologic grade and depth of invasion of the glandular component. Although prognostically valuable information was provided by dividing these tumors into adenoacanthomas and adenosquamous carcinomas, more information was gained when they were stratified by the histologic grade of the glandular component. Zaino et al. (41) recommended that the terms adenoacanthoma and adenosquamous carcinoma be replaced by the simple term adenocarcinoma with squamous differentiation.
Papillary serous carcinomas have a poor prognosis even in the absence of deep myometrial invasion or lymph node metastasis (4,42,43,44,45). They disseminate widely, with a particular predilection for recurrence in the upper abdomen (46,47). The mechanisms that have been proposed to explain the characteristic intraabdominal dissemination of these tumors include transtubal spread, vascular-lymphatic invasion, and multifocal disease. Sherman et al. (42) made the interesting observation that “intraepithelial serous carcinoma” was present in the endocervix in 22% of their cases, in the fallopian tube in 5%, on the surface of the ovary in 10%, and on peritoneal surfaces or omentum in 25%. Serous papillary elements are often admixed with endometrioid carcinomas, but a serous component of 25% will portend a poor prognosis (42).
Table 10.8 Prognostic Variables in Endometrial Cancer Other than FIGO Stage
In contrast to the slow, estrogen-driven pathway leading to the biologically more indolent endometrioid carcinoma, a rapid, p53-driven pathway appears to lead to the aggressive serous (4,45) and clear cell carcinomas (4).
Clear cell carcinomas represent fewer than 5% of endometrial carcinomas, although clear cell elements are commonly present in papillary serous tumors (42). Vascular space invasion is more common in these lesions (46). In a review of 181 patients with clear cell endometrial carcinoma treated between 1970 and 1992, Abeler et al. (47) reported 5- and 10-year actuarial disease-free survival rates of 43% and 39%, respectively. Two-thirds of the relapses were outside the pelvis, most frequently in the upper abdomen, liver, and lungs.
SEER data from 1988 to 2001 were used to compare uterine papillary serous (n = 1,473), clear cell (n = 391), and grade 3 endometrioid carcinomas (n = 2,316) (44). Serous and clear cell carcinomas occurred in older patients, and were diagnosed at a more advanced stage. They represented 10%, 3%, and 15% of endometrial cancer, respectively, but accounted for 39%, 8%, and 27% of cancer deaths.
When papillary serous or clear cell carcinomas are limited to the curettings, with no adverse features in the hysterectomy specimen, prognosis may not be impaired (48).
Squamous cell carcinomas of the endometrium are rare. In a review of the literature, Abeler and Kjorstad (49) estimated that the survival rate for patients with clinical stage I disease was 36%.
Histologic Grade and Myometrial Invasion
There is a strong correlation between histologic grade, myometrial invasion, and prognosis. The GOG reported the surgicopathologic features of 621 patients with stage I endometrial carcinoma (32). The frequency of positive pelvic and paraaortic nodal metastases in relation to histologic grade and depth of myometrial invasion is shown in Tables 10.9 and 10.10. When grade 1 carcinomas were confined to the inner third of the myometrium, the incidence of positive pelvic nodes was less than 3%, whereas when grade 3 lesions involved the outer third, the incidence of positive pelvic nodes was 34%. For aortic nodes, the corresponding figures were less than 1% and 23%, respectively.
Local recurrence at the vaginal vault can usually be prevented by prophylactic vault brachytherapy, but the risk of distant metastases in relation to histologic grade and myometrial invasion is shown in Table 10.11 (50).
Vascular Space Invasion
Vascular Space Invasion Vascular space invasion appears to be an independent risk factor for recurrence and for death from endometrial carcinoma of all histologic types(36,51,52,53). Aalders et al. (51) reported recurrences and deaths in 26.7% of patients with stage I disease who had vascular space invasion compared with 9.1% of those without vessel invasion (p = 0.01). Abeler et al. reviewed 1,974 cases of endometrial carcinoma from the Norwegian Radium Hospital and reported an 83.5% 5-year survival rate for patients without demonstrable vascular invasion compared with 64.5% for those in whom invasion was present (52). A Japanese study reported that lymph-vascular space invasion and the number of positive paraaortic lymph nodes were independent prognostic factors for patients with stage IIIC endometrial cancer (54).
Table 10.9 Grade, Depth of Invasion, and Pelvic Nodal Metastasis of Endometrial Carcinoma
Table 10.10 Grade, Depth of Invasion, and Aortic Nodal Metastasis of Endometrial Carcinoma
The overall incidence of lymph-vascular invasion in stage I endometrial carcinoma is approximately 15%, although it increases with increasing myometrial invasion and decreasing tumor differentiation. Hanson et al. (55) reported vascular space invasion in 5% of patients with invasion limited to the inner one-third of the myometrium compared with 70% of those with invasion to the outer one-third. Similarly, it was present in 2% of grade 1 carcinomas and 42% of grade 3 lesions. Ambros and Kurman (56), using multivariate analysis, reported that only depth of myometrial invasion, DNA ploidy, and vascular-invasion-associated changes correlated significantly with survival for patients with stage I endometrioid adenocarcinomas. Vascular invasion-associated changes were defined as vascular invasion by tumor, or the presence of myometrial perivascular lymphocytic infiltrates, or both. In the GOG study, vascular space invasion carried a relative risk of death of 1.5 (37).
Peritoneal Cytologic Results
The significance of a positive peritoneal cytologic result is controversial (57). The incidence of positive cytologic findings in stage I disease is shown in Table 10.12. Positive washings are most common in patients with grade 3 histologic type, metastases to the adnexae, deep myometrial invasion, or positive pelvic or paraaortic nodes(32,57,58,59,60,61,62).
The GOG study reported by Morrow et al. (63) analyzed 697 patients with information on peritoneal cytologic results and adequate follow-up. Disease recurred in 25 of 86 patients (29.1%) with positive washings, compared with 64 of 611 patients (10.5%) with negative washings. The authors noted, however, that 17 of the 25 recurrences were outside the peritoneal cavity. The GOG estimated that the relative risk of death for patients with positive cytologic washings was increased threefold (37).
Table 10.11 Clinical Stage I Endometrial Carcinoma: Distant Metastases versus Histologic Grade and Myometrial Invasiona
In a review of the literature concerning patients with clinical stage I endometrial cancer, Milosevic et al. (64) reported positive peritoneal cytology in 8.3%, 12.1%, and 15.9% of patients with grades 1, 2, and 3 histologic types, respectively. Superficial and deep myometrial invasion were associated with positive washings in 7.6% and 17.2% of the cases, respectively. They concluded that the poor prognosis associated with malignant washings was largely a reflection of other adverse prognostic factors.
Kadar et al. (62) studied 269 patients with clinical stage I and stage II endometrial cancer and reported that if the disease was confined to the uterus, then positive peritoneal cytologic results did not influence survival. If the disease had spread to the adnexa, lymph nodes, or peritoneum, then positive peritoneal cytologic findings decreased the survival rate from 73% to 13% at 5 years, but all recurrences were at distant sites. Others have also reported no prognostic significance when the disease was confined to the uterus (65,66,67,68). Grimshaw et al. reported that 70% of patients with endometrial cancer and positive peritoneal cytology had extrauterine disease at the time of surgery, and the prognosis for these patients was impaired (66). Contrary findings were reported from Duke University, where Havrilesky et al. showed that positive cytology was an independent poor prognostic factor for patients with stages I to IIIA disease (69). Saga et al. from Japan also found positive peritoneal cytology to be an independent poor prognostic factor after reviewing 307 surgically staged patients with disease confined to the uterus (70).
Takeshima et al. (61) studied 534 patients with endometrial cancer to assess the prognostic significance of positive peritoneal washings. They concluded that they were not an independent negative prognostic indicator but potentiated other prognostic indicators. They felt that patients with positive peritoneal cytology in the absence of other adverse prognostic factors did not warrant upstaging.
Table 10.12 Incidence of Positive Peritoneal Cytology in Clinical Stage I Endometrial Carcinoma
These same workers placed a tube in the abdomen to allow peritoneal irrigation in 50 patients with early stage endometrial cancer and positive peritoneal smears detected at surgery (65). Washings were obtained via the tube 7 and 14 days postoperatively. Persistence of positive peritoneal cytology was observed in only five of 50 patients (10%), and four of these patients had adnexal metastases completely resected. They concluded that malignant cells found in the peritoneal cavity generally have a low malignant potential and that only malignant cells from special cases, such as patients with adnexal metastases, may be capable of independent growth.
Hormone Receptor Status
In general, mean estrogen receptor (ER) and progesterone receptor (PR) levels are inversely proportional to histologic grade (71,72,73,74). However, ER and PR content have been shown to be independent prognostic indicators for endometrial cancer; that is, patients whose tumors are positive for one or both receptors have longer survival than patients whose carcinoma lacks the corresponding receptors (71,72,73,75). Liao et al. (72) reported that, even for patients with lymph node metastases, the prognosis was significantly improved if the tumor was receptor positive. PR appears to be a stronger predictor of survival than ER and, at least for the ER, the absolute level of the receptors may be important: The higher the level, the better the prognosis (76).
Nuclear grade is a significant prognostic indicator (76). Christopherson et al. (77) found nuclear grading to be a more accurate prognosticator than histologic grade.
The FIGO grading system takes into account the nuclear grade of the tumor, and “nuclear atypia” inappropriate for the architectural grade raises the grade by 1. However, there is great variability in the literature regarding the criteria for nuclear grading, and intraobserver and interobserver reproducibility of nuclear grading are poor (78).
In an analysis of 142 patients with clinical stage I endometrial carcinoma, Schink et al. (79) reported tumor size as an independent prognostic factor. Lymph node metastases occurred in 4% of the patients with tumors no more than 2 cm in diameter, 15% with tumors greater than 2 cm in diameter, and 35% with tumors involving the entire uterine cavity. The incidence of nodal metastases in relation to tumor size and depth of invasion is shown in Table 10.13.
DNA Ploidy and Other Biologic Markers
Approximately one-fourth of patients with endometrial carcinomas have aneuploid tumors, which is a low incidence compared with many other solid tumors, including ovarian and cervical carcinomas. However, patients with aneuploid tumors are at significantly increased risk of early recurrence and death from disease (56,80,81).
In a prospective study of 174 patients, Susini et al. reported a 10-year survival probability of 53.2% for patients with DNA aneuploid tumors, compared with 91.0% for patients with DNA diploid tumors. By multivariate analysis, DNA aneuploid type was the strongest independent predictor of poor outcome, followed by age and stage (81). The GOG estimated the relative risk to be 4.1 for disease-related death for patients with aneuploid tumors (82).
A number of genetic mutations have also been shown to have prognostic significance in endometrial cancer. Greek workers reported that loss of beta-catenin expression was a strong, independent predictor of a poor prognosis, whereas loss of PTEN was associated with a worse prognosis for patients with early stage disease (83). The p53 mutation correlated with increased stage, lymph node metastases, and nonendometrioid histology in univariate analysis but was not an independent prognostic factor in multivariate analysis (83). Increasing expression of matrix metalloproteinases (MMPs) (84), nuclear bcl-2 expression (85), and Ki-67 expression (86) also have prognostic significance. The clinical implications of these biologic markers are not yet clear.
Table 10.13 Incidence of Lymph Node Metastasis in Endometrial Cancer by Tumor Size and Depth of Myometrial Invasion
Table 10.14 Clinical Stage II Carcinoma of the Endometrium: Comparison of Treatment Methods
Method of Treatment
In contrast to cervical cancer, patients with endometrial cancer treated with hysterectomy alone or hysterectomy and radiation do significantly better than those treated with radiation alone This appears to be related to the inability of radiation therapy effectively to eliminate disease in the myometrium (87,88). Grigsby et al. (87) reported on 116 patients with stage II endometrial carcinoma. Ninety were treated with combined radiation and surgery, whereas 26 received radiation alone. The results of treatment are shown inTable 10.14.
Classic teaching has been that endometrial hyperplasias represent a continuum of morphologic severity; the most severe form, termed atypical adenomatous hyperplasia orcarcinoma in situ, was considered the immediate precursor of endometrial carcinoma (89,90). Since the mid-1980s, this continuum concept has been challenged. Independent studies by Kurman et al. (91) and Ferenczy et al. (92) have suggested the following:
Ferenczy et al. (92) suggested that the term endometrial hyperplasia be used for any degree of glandular proliferation devoid of cytologic atypia and the term endometrial intraepithelial neoplasia for lesions with cytologic atypia. Using similar criteria in a long-term follow-up study of 170 patients with endometrial hyperplasia, Kurman et al. (91) reported a 1.6% risk of progression to carcinoma in patients devoid of cytologic atypia, compared with a 23% risk in patients with cytologic atypia.
Subsequently, Ferenczy and Gelfand (93) reported 85 menopausal women with endometrial hyperplasia. Sixty-five patients had no cytologic atypia, and 84% of this group responded to medroxyprogesterone acetate (MPA). Four (6%) had recurrent hyperplasia after discontinuing the MPA, and none developed carcinoma, with a mean follow-up of 7 years. By contrast, 20 patients had cytologic atypia, and only 50% responded to MPA. Recurrent hyperplasia developed in five (25%), and adenocarcinoma in five (25%). The World Health Organization (WHO) classification of endometrial hyperplasia is shown in Table 10.15
Although the studies of Kurman and Ferenczy are important, the reproducibility of the diagnosis has been questioned. In a GOG study, a panel of three expert gynecologic pathologists reviewed outside slides from 306 patients referred with a diagnosis of atypical endometrial hyperplasia. The majority panel diagnosis was adenocarcinoma in 29%, cycling endometrium in 7%, and nonatypical hyperplasia in 18% of cases (94). Interobserver variation was also common.
Hysterectomy slides from the same GOG study were reviewed by the study pathologists, and 289 hysterectomy specimens were available for review (95). Concurrent endometrial carcinoma was present in 42.6% (123 of 289 specimens). Of these, 30.9% (38 of 123 specimens) had myometrial invasion, and 10.6% (13 of 123 specimens) had invasion to the outer half of the myometrium. Even when the study panel consensus diagnosis was less than atypical endometrial hyperplasia, 14 of 74 women (18.9%) still had carcinoma in the hysterectomy specimen.
Table 10.15 World Health Organization Classification of Endometrial Hyperplasia
Most women with endometrial hyperplasia respond to progestin therapy. Patients who do not respond are at a significantly increased risk of progressing to invasive cancer and should be advised to have a hysterectomy. Patients who are unlikely to respond can be identified on the basis of cytologic atypia. A suggested scheme of management is outlined inFigure 10.2.
Ferenczy and Gelgand reported on 85 women with endometrial hyperplasia treated with medroxyprogesterone (93). The 65 patients who had no cytologic atypia received 10 mg daily for 14 days per month, whereas the 20 patients with cytologic atypia received 20 mg daily. Both groups had endometrial sampling every 3 months. In the group without cytologic atypia, 14% had persistence of their hyperplasia, but none progressed to carcinoma. Of the 20 patients with cytologic atypia, 50% had persistence of their hyperplasia, and 25% developed frank carcinoma at a mean of 5.5 years after initiation of hormonal therapy.
The type of progestin used does not appear to be important, the optimal dosage has not been investigated, and the regimens advocated are essentially arbitrary (96). High doses ofprogestins are often better tolerated than low doses. The main side effects are weight gain, edema, thrombophlebitis, and occasionally hypertension. The incidence of venous thrombosis and embolism may also be slightly increased.
Other approaches to hormonal therapy include levonorgestrel-releasing intrauterine devices (97), the use of danazol in a dose of 400 mg daily for 3 months (98), and the combined use of gonadotropin-releasing hormone (GnRH) analogues and progestins (99).
Treatment of Endometrial Cancer
The cornerstone of treatment for endometrial cancer is total abdominal hysterectomy and bilateral salpingo-oophorectomy, and this operation should be performed in all cases whenever feasible. In addition, many patients require some type of adjuvant radiation therapy to help prevent vaginal vault recurrence and to sterilize disease in lymph nodes. It is difficult to document that radiation actually improves survival rates, but both the GOG study (100) and the postoperative radiation therapy in endometrial carcinoma (PORTEC) trial (101) of surgery versus surgery plus adjuvant pelvic radiation for patients with intermediate to high-risk stage I endometrial cancer showed an improved disease-free survival rate for the radiation-treated group. Neither trial showed an improvement in overall survival because of the ability to salvage most pelvic recurrences in the surgery-only arm with radiation therapy. There may be a survival benefit for patients with high-grade, deeply invasive tumors, but currently there are no prospective data to demonstrate this (102).
With the increasing emphasis on surgicopathologic staging, a more individualized approach to adjuvant radiation is now possible.
Figure 10.2 Management of endometrial hyperplasia.
Microscopic cervical involvement (positive endocervical curettage) is often designated (unof-ficially) as stage II occult disease. For practical purposes, if the cervix is not hard or expanded, such patients can be managed in the same way as patients with stage I disease.
Stage I and Stage II Occult
A recommended treatment plan is shown in Fig. 10.3.
The initial approach for all medically fit patients should be total abdominal hysterectomy and bilateral salpingo-oophorectomy. Removal of a vaginal cuff is not necessary. The adnexa should be removed because they may be the site of microscopic metastases. In addition, patients with endometrial carcinoma are at increased risk for ovarian cancer. Such tumors sometimes occur concurrently (103). Surgical staging, including lymphadenectomy, should be performed in those patients listed in Table 10.16 The use of laparoscopically assisted vaginal hysterectomy is addressed in Chapter 21.
Figure 10.3 Management of patients with stage I and occult stage II endometrial carcinoma. TAH, total abdominal hysterectomy;
Table 10.16 Endometrial Carcinoma Stages I and Occult II: Patients Requiring Surgical Staging
The laparotomy is best performed through a lower midline abdominal incision, particularly in the obese patient. This incision allows easy access to the upper abdomen, including the omentum and paraaortic lymph nodes. A Pfannenstiel incision is commonly used for patients with grade 1 or 2 tumors and a normally sized uterus An alternative approach is to use a transverse, muscle-dividing incision (e.g., the Maylard or Cherney), as discussed in Chapter 20. This incision also gives reasonable access to the upper abdomen.
After the abdomen is opened, peritoneal washings are taken with 50 dL normal saline solution. Thorough exploration of the abdomen and pelvis is performed, with particular attention to the liver, diaphragm, omentum, and paraaortic nodes. Any suspicious lesions are excised or biopsied.
The uterus is grasped with clamps that encompass the round and ovarian ligaments and the fallopian tube. After the round ligaments are divided, the incision is carried anteriorly around the vesicouterine fold of peritoneum and posteriorly parallel and lateral to the infundibulopelvic ligaments. With a narrow Deaver retractor in the retroperitoneum providing gentle traction cephalad in the direction of the common iliac vessels, the iliac vessels and ureter are displayed. With the retroperitoneum displayed, the pelvic lymph nodes can be visualized and palpated, and any enlarged nodes can be removed.
With each ureter under direct vision, the infundibulopelvic ligaments are divided and tied. The bladder is dissected off the front of the cervix, and then the uterine vessels are skeletonized and divided at the level of the isthmus. Straight Kocher clamps are used to secure the cardinal and uterosacral ligaments. The uterus, tubes, and ovaries are removed, and the vaginal vault is closed. The pelvic peritoneum is not closed, and it usually is not necessary to place drains in the pelvis. The sigmoid colon is placed in the pelvis to help exclude loops of small bowel. A vertical abdominal wound is best closed with a continuous Smead-Jones type of internal retention suture, using a long-acting, absorbable suture such as Maxon or PDS.
The decision to undertake surgical staging is usually based on the histopathology from the uterine curettings, the gross findings on opening the uterus on the operating table, and possibly a frozen section of the resected uterus.
A relatively poor correlation has been reported between the grade of cancer on curettings or biopsy and the final grade in the resected uterus, presumably because of a sampling error in the diagnostic procedure. The poorest correlation is for grade 1 tumors, where 20% to 40% may be upgraded after evaluation of the hysterectomy specimen (104,105).
Our practice is to open the specimen on the operating table to determine the need for surgical staging in patients with grade 1 or 2 tumors (Figs. 10.4 and 10.5). All patients with grade 3 tumors (Fig. 10.6), serous papillary, or clear cell carcinomas are surgically staged.
For grade 1 tumors, gross examination fairly accurately predicts depth of myometrial invasion. In an analysis of 113 patients with surgical stage I endometrial carcinoma, Goff and Rice (106) reported that macroscopic examination of the fresh specimen correctly predicted depth of invasion in 55 of 63 grade 1 lesions (87.3%), 24 of 37 grade 2 lesions (64.9%), but only 4 of 13 grade 3 lesions (30.8%). Franchi et al. also concluded that gross inspection of the opened uterus was a reliable and inexpensive approach after evaluating 403 endometrial cancers and noting an accurate prediction of depth of invasion in 344 cases (85.3%) (107).
Tumor diameter should also be taken into account when determining the need for surgical staging, and we use this particularly for grade 2 lesions. Schink et al. (79) reported a 22% incidence of lymph node metastases for grade 2 tumors greater than 2 cm in diameter (7 of 32). None of 19 grade 2 tumors less than 2 cm in diameter had nodal metastases.
Figure 10.4 A small fundal grade 1 endometrial carcinoma. This patient does not require surgical staging.
Figure 10.5 A grade 2 endometrial carcinoma occupying most of the corpus. A patient such as this should undergo surgical staging.
Figure 10.6 Survival for patients with (A) low-risk endometrial cancer and (B) high-risk endometrial cancer versus extent of lymphadenectomy. Reproduced with permission from Chan J et al. Therapeutic role of lymph node resection in endometrioid corpus cancer: A study of 12,333 patients. Cancer 2006; 107:1823-1828.
If doubt exists regarding the need for surgical staging, intraoperative frozen section can be obtained, but this is inaccurate in distinguishing superficial from deep myometrial invasion in 5% to 10% of cases (106,107,108). If the final histopathology is worse than was anticipated intraoperatively, the prognosis will not be impaired if external-beam pelvic radiation is given on the basis of histologic grade and depth of myometrial invasion, as long as any enlarged pelvic or paraaortic nodes have been resected as part of the standard management for all patients.
No preoperative scan is able to detect micrometastases in lymph nodes, so if accurate surgical staging is to be obtained, then full pelvic lymphadenectomy should be performed on all patients who meet the criteria in Table 10.16. Sampling will only lead to inaccurate information (109).
In an analysis of 11,443 patients registered on the SEER data base between 1990 and 2001, Chan et al. reported that the removal of 21 to 25 lymph nodes significantly increased the probability of detecting at least one positive lymph node in endometrioid uterine cancer (110). Lutman et al. also reported that the pelvic lymph node count was an important prognostic variable for patients with FIGO stages I and II endometrial carcinoma and high-risk histology (111). On the other hand, in an analysis of 5,556 patients with low-risk endometrioid endometrial carcinoma from the SEER data base, Chan et al. were unable to find any survival advantage regardless of the extent of the lymphadenectomy (33) (Figure 10.6).
The dissection should include removal of common iliac nodes and of the fat pad overlying the distal inferior vena cava. If full pelvic lymphadenectomy is considered inadvisable because of the patient's general medical condition, which is uncommon, then resection of any enlarged pelvic nodes should be performed.
Management of Paraaortic Lymph Nodes
Although some authors recommend systematic paraaortic lymphadenectomy on all high-risk patients (112) or in patients with two or more positive pelvic lymph nodes (113), this is major surgery for a group of patients who are usually elderly and obese. An extensive paraaortic lymphadenectomy significantly increases operating time and blood loss and also increases postoperative morbidity, particularly lower limb lymphedema. The latter occurs in 20% of patients in our experience (114).
Lymphedema is a lifelong affliction, which is often complicated by recurrent episodes of cellulitis. To avoid progressive deterioration of the condition, regular massage and use of surgical stockings are essential, and both become progressively more burdensome, particularly for elderly patients. In our opinion, primary prevention of lymphedema by selective use of pelvic lymphadenectomy and avoidance of systematic paraaortic lymphadenectomy is highly desirable.
The GOG data (63) suggested that patients with positive paraaortic nodes were likely to have:
Nomura et al. retrospectively reviewed 841 patients with endometrial cancer who underwent their initial surgery at Keio University Hospital in Japan (115). In a multivariate analysis, the clinicopathologic factor most strongly related to paraaortic nodal metastasis was pelvic lymph node metastasis. Among 155 patients who underwent systematic pelvic and paraaortic lymphadenectomy, 96.2% (101 of 105 cases) had negative paraaortic nodes when the pelvic nodes were negative. However, when the pelvic nodes were positive, 48% (24 of 50 cases) also had positive paraaortic nodes. These findings are consistent with those from the Mayo Clinic (112).
In addition to the lymphadenectomy, an omental biopsy is also performed as part of the surgical staging because occult omental metastases may occur, particularly in patients with grade 3 tumors or deeply invasive lesions (116). The omentum should be carefully inspected, along with all peritoneal surfaces, and any suspicious lesions excised. If the omentum appears normal, then a generous biopsy (e.g., 5 × 5 cm) should be taken.
Our current approach is to resect any enlarged paraaortic nodes and give extended-field radiation to all patients with any grossly positive pelvic nodes or more than one microscopically positive pelvic node.
Sentinel Node Biopsy
Sentinel node identification has been investigated in a number of solid tumors, the hypothesis being that if one or more sentinel nodes are negative, the remainder of the regional nodes will be negative, so complete lymphadenectomy can be avoided (117). Lymphatic mapping is performed by infecting tracers around the tumor and identifying the draining node(s). Usually, a blue dye and a radioactive tracer are used, and the best results are achieved when both techniques are used together. Technetium-99 is the most commonly used radioactive substance because of its short half-life (6 hours).
Table 10.17 Sentinel Lymph Node Involvement in Endometrial Cancer
Three approaches have been used for sentinel node identification in endometrial cancer: (i) injection into the cervix, (ii) injection around the tumor via a hysteroscope, and (iii) injection into the subserosal myometrium at the fundus.
Results of some preliminary studies are shown in Table 10.17. Injection around the tumor would seem the most logical approach, but data are scant, and studies are still addressing feasibility and standardization of technique. Although detection of at least one sentinel node is reported to occur in as many as 100% of cases, bilateral pelvic sentinel node detection and paraaortic sentinel node detection both occur in less than 50% of cases. Sentinel node identification is still entirely experimental for endometrial cancer.
Role of Lymphadenectomy
Pelvic lymphadenectomy, with or without paraaortic lymphadenectomy, plays an important role in the surgical staging of endometrial cancer and thus provides more accurate prognostic information. The therapeutic role of lymphadenectomy is less well understood, but its ability to modify adjuvant therapy is being increasingly accepted.
The therapeutic value of pelvic lymph node dissection was investigated by Kilgore et al. from Birmingham, Alabama, who reported on 649 surgically managed patients with adenocarcinoma of the endometrium (126). Two hundred twelve patients had multiple-site pelvic node sampling (mean number of nodes 11), 205 had limited-site sampling (mean number of nodes 4), and 208 had no node sampling. The decision regarding lymph node sampling was surgeon dependent, and prognostic features—including tumor grade, depth of invasion, adnexal metastasis, cervical involvement, and positive cytologic findings—were equally distributed among the three groups. All patients had adjuvant radiation therapy based on traditional prognostic factors. With a mean follow-up of 3 years, patients undergoing multiple-site pelvic node sampling had a significantly better overall survival (p = 0.0002) as well as a better survival for both low-risk and high-risk groups (low-risk, p = 0.026; high-risk, p = 0.0006).
The authors concluded that their results strongly suggested a therapeutic benefit to lymphadenectomy, but confirmation of this must await randomized studies. If there is a therapeutic benefit, it must surely be related to the resection of bulky, positive nodes, which are unlikely to be sterilized with external-beam radiation therapy.
A randomized trial to evaluate the therapeutic effects of lymphadenectomy was conducted by the U.K. Medical Research Council's ASTEC study. The results were published recently (127). No therapeutic benefit was found. However, the number of lymph nodes resected was insufficient to show a benefit in at least one third of patients, paraaortic node dissection was not required, and subsequent radiation was not tailored to the nodal status, so it is difficult to draw any definitive conclusions.
The feasibility of using the results of pelvic lymphadenectomy to modify adjuvant radiation therapy has been addressed in several nonrandomized trials and will be discussed under adjuvant radiation.
In selected patients with marked obesity and medical problems that place them at high risk for abdominal operations, vaginal hysterectomy should be considered.
Peters et al. (128) reported a 94% survival rate among 56 patients with stage I endometrial carcinoma who underwent vaginal hysterectomy. Seventy-five percent had grade 1 lesions, and 32 patients received adjuvant radiation, mainly brachytherapy.
Japanese workers reported on 171 patients aged 70 years and older, 128 (75%) of whom were treated with vaginal hysterectomy and 43 of whom underwent abdominal hysterectomy (129) The 10-year disease-specific survival rates were 83% and 84%, respectively (p = ns). Patients in the vaginal hysterectomy group had significantly shorter operating times, less blood loss, and shorter postoperative stays. Severe complications occurred in 5.4% of the vaginal and 7.0% of the abdominal procedures. Perioperative mortality was zero after vaginal hysterectomy and 2.3% after abdominal hysterectomy. The researchers concluded that vaginal hysterectomy should be considered the elective approach for the treatment of elderly patients with endometrial cancer.
Laparoscopically assisted vaginal hysterectomy is increasingly being used for the management of endometrial cancer, particularly in obese patients. Use of the laparoscope facilitates removal of the adnexae and the pelvic lymph nodes.
The use of adjuvant radiation is decreasing for patients with endometrial cancer. For high-risk patients, surgical staging has allowed the therapy to be better tailored to the needs of the individual patient. The options for postoperative radiation are as follows:
Patients with stage IA or IB, grade 1 or 2 tumors have an excellent prognosis, and no adjuvant radiation is necessary for this group. Canadian workers reported 227 such patients who were followed without radiation, and the 5-year relapse-free survival rate was 95% (130). Elliot et al. (131) from Australia treated 308 patients with grade 1 or 2 lesions confined to the inner third of the myometrium with hysterectomy alone. There were ten vaginal recurrences (3.2%), eight at the vault, and one each in the middle and lower third. The Danish Endometrial Cancer Group (DEMCA) prospectively followed 641 patients with grade 1 and 2 tumors with no more than 50% myometrial invasion (stages IA and IB) who were treated by total abdominal hysterectomy and bilateral salpingo-oophorectomy without adjuvant radiation (132). With follow-up of 68 to 92 months, the disease-free survival rate was 93% (596 of 641). Fanning and colleagues compared surgery and adjuvant radiation with surgery alone for patients with stage I, grade 2 adenocarcinomas of favorable histologic subtype and less than one-third myometrial invasion (133). The 5-year survival rate for surgery and radiation was 94% (128 of 136), and the recurrence rate was 2.2% (3 of 136). The 5-year survival rate for the surgery-alone group was 98% (51 of 52), and the recurrence rate was 1.9% (1 of 52).
If patients are treated without adjuvant therapy, then they must be followed carefully so that vault recurrences can be diagnosed early or when they are eminently curable(132,133,134,135). The diagnosis of recurrence is sometimes first suspected when adenocarcinoma cells are seen on a routine vault smear.
Vaginal brachytherapy significantly reduces the incidence of vault recurrence. With high dose-rate therapy, treatment can be accomplished as an outpatient, and the morbidity is low. A recent study from the University of Virginia reported that compared to observation, postoperative brachytherapy improved survival at a cost of $65,900 per survivor (136).
Several reported studies have demonstrated low recurrence rates at the vaginal vault and pelvic sidewall using vault brachytherapy without external pelvic radiation after pelvic lymphadenectomy with negative nodes. Table 10.18 reveals an incidence of vault recurrence of 0.6% and an incidence of pelvic sidewall recurrence of 1.1% following this approach in 1,135 mainly high-risk patients with stage I and stage II occult disease. Data from the Mayo Clinic would suggest that high-dose-rate brachytherapy should also be considered in low-risk cases with extensive vascular space invasion (36).
Table 10.18 Vault Brachytherapy for Stage I or Occult Stage II Endometrial Cancer after Pelvic Lymphadenectomy
The exact place of vaginal brachytherapy alone must await randomized, controlled trials. If lymph node sampling only has been performed, then it may be safer to use external pelvic radiation because of the increased risk of pelvic sidewall recurrence.
External Pelvic Irradiation
With an increasing number of patients in cancer centers having pelvic lymphadenectomy as part of their primary surgery, the indications for external pelvic irradiation are decreasing. Patients with negative pelvic nodes generally receive vault brachytherapy alone, whereas patients with bulky positive pelvic nodes or more than one microscopically positive pelvic node are better treated with pelvic and paraaortic radiation. External pelvic radiation is a reasonable option for high-risk patients who have not undergone surgical staging but have a negative pelvic and abdominal CT scan and a normal serum CA125 titer.
The GOG reported results of a randomized study of adjuvant pelvic radiation after complete surgical staging for patients with intermediate-risk endometrial carcinoma (100). Eligible patients had surgical stages IB, IC, IIA (occult), or IIB (occult) disease and were randomized to receive either no additional therapy or 5,040 centiGray of external pelvic radiation therapy. There were 390 eligible patients in the study, and median follow-up was 56 months.
The 2-year progression-free survival rate was significantly higher in the group receiving adjuvant radiation (96% vs. 88%; p 0.004). However, overall survival rates were not significantly different because there were more pelvic or vaginal recurrences in the no-treatment arm (17 vs. 3), and these were often effectively treated with second-line therapy.
A European randomized trial (the PORTEC study) of surgery and postoperative external pelvic radiation (46 Gy) versus surgery alone for patients with stage I endometrial cancer was published in 2000 (101). Eligible patients were those with stage IC, grade 1; stage IB or IC, grade 2; or stage IB, grade 3 disease. Patients with serous papillary or clear cell carcinoma were also eligible. Surgery consisted of total abdominal hysterectomy and bilateral salpingo-oophorectomy without lymphadenectomy. A total of 715 patients from 19 radiation oncology centers were randomized.
Actuarial 5-year overall survival rates were similar in the two groups: 81% (radiotherapy) and 85% (controls) ( p = 0.31). Endometrial-cancer-related death rates were 9% in the radiotherapy group and 6% in the control group ( p = 0.37). Treatment-related complications occurred in 25% of the radiotherapy patients and in 6% of the controls ( p = 0.0001). Grade 3-4 complications were seen in eight patients, of which seven were in the radiotherapy group (2%). Two-year survival after vaginal recurrence was 79% in contrast to 21% after pelvic recurrence or distant metastases. Survival after relapse was significantly better ( p = 0.02) for patients in the control group. After multivariate analysis, investigators concluded that postoperative radiotherapy was not indicated in patients with stage I endometrial cancer younger than 60 years and in patients with grade 2 tumors with superficial invasion.
A subsequent 10-year follow-up of the PORTEC trial revealed that the 10-year locoregional relapse rates were 5% after radiation and 14% for controls ( p <0.0001), and the 10-year overall survival rates were 66% and 73%, respectively ( p = 0.09) (146). Endometrial-cancer-related death rates were 11% (RT) and 9% (controls) ( p = 0.47). The researchers concluded that radiotherapy was indicated for patients who were unstaged surgically but had high-risk features because of the significant locoregional control benefit.
To clarify the effect of postoperative external-beam pelvic radiotherapy in patients with early endometrial cancer, Johnson and Cornes from the United Kingdom performed a metaanalysis of data from five randomized trials. Pelvic lymphadenectomy was performed in only two of the five trials. The authors concluded that adjuvant external-beam pelvic radiation should not be used for low (IA or IB, grade 1) or intermediate risk (IB, grade 2) cancer but was associated with a 10% survival advantage for high-risk (stage IC, grade 3) disease (147).
External irradiation appears to be as effective as vaginal brachytherapy for sterilizing micrometastases at the vaginal vault; thus, there seems to be no reason to give both external and vault irradiation after surgery because morbidity will be significantly increased. Weiss and colleagues treated 61 women with stage IC endometrial cancer with postoperative pelvic radiation without vaginal brachytherapy. No patient developed a vaginal recurrence (148).
Indications for external pelvic radiation are shown in Table 10.19
Risk factors for pelvic lymph node metastases portend a lower but significant risk of paraaortic metastases, and failure rates of 15% to 20% in the paraaortic area have been reported for patients receiving pelvic radiation only (149). Approximately 50% of patients with positive pelvic nodes will have positive paraaortic nodes (115). Paraaortic lymph node metastases are associated with an increasing number of pelvic lymph node metastases and with bilateral pelvic nodal involvement (150).
Using what is now considered relatively unsophisticated radiation techniques and a dose between 45 and 50 Gy, approximately 40% of patients with positive paraaortic nodes may achieve long-term disease-free survival with extended-field radiation therapy (151). Mariani et al. reported that none of 11 patients with positive paraaortic nodes failed in the paraaortic area after adequate lymphadenectomy (defined as removal of five or more paraaortic nodes) and extended-field radiation (112).
Our current indications for extended-field radiation are shown in Table 10.20.
Table 10.19 Indications for External Pelvic Radiation in Patients with Stages I and II Occult Endometrial Cancer
Table 10.20 Indications for Extended-Field Radiation Therapy in Patients with Endometrial Cancer
Whole-abdominal radiation has been used for many years in selected patients with omental, adnexal, or peritoneal metastases that have been completely resected, but the GOG recently reported a randomized phase III trial of WAR versus chemotherapy in patients with stage III or IV endometrial carcinoma having a maximum of 2 cm of postoperative residual disease. There were 396 assessable patients. Irradiation dosage was 30 Gray (Gy) in 20 fractions, with a 15-Gy boost. Chemotherapy consisted of doxorubicin 60 mg/m2 and cisplatin50 mg/m2 every 3 weeks for seven cycles, followed by one cycle of cisplatin (152).
The study showed a significant improvement in 2-year disease-free survival for the chemotherapy arm, 58% versus 46% ( p > 0.01). At 60 months, and adjusting for stage, 55% of chemotherapy patients were predicted to be alive compared to 42% of patients receiving radiation. Chemotherapy was also more toxic and probably contributed to the deaths of eight patients (4%) compared to five (2%) on the radiation arm. In the chemotherapy arm, distant failure was reduced from 18% to 10%, but there was no difference in pelvic or abdominal failure rates in the two arms.
In a nonrandomized study of 180 patients with surgically staged III and IV endometrial cancer treated with WAR with pelvic, plus or minus paraaortic boost, the GOG reported 3-year recurrence-free survival rates of 29% for endometrial and 27% for serous and clear cell carcinomas, respectively (153). No patient with gross residual disease survived. Severe toxicity included bone marrow depression in 12.6%, gastrointestinal toxicity in 15%, and hepatic toxicity in 2.2% of patients.
The group at Duke University reported on 356 patients with advanced stage endometrial cancer. Postoperatively, whole-abdominal radiation alone was used in 48% (n = 171), chemotherapy alone in 29% (n = 102), and chemotherapy plus radiation in 23% (n = 83) of patients. After adjusting for age, grade stage, and debulking status, there was a significant survival benefit for combined therapy compared to either modality used alone (154).
A retrospective view of 86 patients from the University of Minnesota with peritoneal spread of endometrial carcinoma reported a recurrence rate of only 16% for patients with stage IIIA disease and only one peritoneal site of spread after treatment with WAR (155). Six percent of patients required surgical intervention for small-bowel obstruction.
Martinez and colleagues reported 10-year survival data on a nonrandomized, prospective trial of whole-abdominal radiation with a pelvic boost in patients with stages I to III endometrial cancer who were considered at high risk for intraabdominopelvic recurrence (156). There were 132 patients treated between 1981 and 2001, including 89 (68%) with stage III disease and 58 (45%) with serous papillary or clear cell histology. The 5- and 10-year cause-specific survival for patients with serous papillary or clear cell tumors was 80% and 74%, respectively. Chronic grade 3 or 4 gastrointestinal toxicity was seen in 14% of patients, and 2% developed grade 3 renal toxicity.
The above studies suggest an advantage to combining radiation with chemotherapy in patients with advanced-stage endometrial cancer, but we would still consider using wholeabdominal radiation in the circumstances shown in Table 10.21.
Table 10.21 Indications for Whole-Abdominal Radiation in Patients with Endometrial Cancer
Although the role of progestins in the management of patients with advanced and recurrent endometrial cancer has been established, they have not been shown to be of value in an adjuvant setting (157,158,159,160). In a randomized study of 1,148 patients with clinical stage I or II endometrial cancer at the Norwegian Radium Hospital, death resulting from intercurrent disease, particularly cardiovascular disease, was actually more common in the progesteronetreated group (p = 0.04) (159). In 461 high-risk patients, a tendency toward fewer cancerrelated deaths and a better disease-free survival rate in the treatment group was observed, but crude survival was unchanged. It was concluded that further studies were needed in high-risk patients but that the evidence suggested that prophylactic progestin therapy was not likely to be a cost-effective approach for patients with endometrial cancer unless the patient had a highrisk, receptor-positive tumor.
An Australian, New Zealand, and United Kingdom trial of 1,012 patients with high-risk disease showed more relapses in the control group, but no difference in survival (160). Patients received medroxyprogesterone acetate 200 mg twice daily for at least 3 years or until recurrence. Steroid receptor status had no influence on outcome in either arm.
The value of adjuvant systemic therapy in patients with high-risk early stage endometrial cancer is still controversial.
Four randomized trials have been conducted that evaluated the efficacy of chemotherapy in the adjuvant setting. The oldest trial (GOG 34), using single-agent doxorubicin, did not show any benefit in a study of 192 patients with clinical stage I or II (occult) disease who had one or more risk factors for recurrence after surgical staging (161). Maggi et al. conducted a randomized controlled trial in 345 high-risk endometrial cancer patients comparing 5 cycles of cisplatin, doxorubicin, and cyclophosphamide with external pelvic radiation. In a multivariate analysis, the researchers reported no difference between therapies in terms of progression-free or overall survival (162).
A Japanese multicenter randomized trial compared whole-pelvic irradiation (WPI) with three or more cycles of cyclophosphamide, doxorubicin, and cisplatin (CAP) chemotherapy in 385 evaluable patients with stages IC to IIIC endometrioid adenocarcinoma (“intermediate risk”; 60% stage IC, 15% grade 3) (163). At a median follow-up of 5 years, there were no significant differences in progression-free (WPI 83.5%.5 vs. CAP 81.8%) or overall survival (85.3% vs. 86.7%). In a subgroup analysis of “high to intermediate risk” cases [stage IC > 70 yrs; stage IC grade 3; stage II or stage IIIA (cytology), n = 120], a survival benefit for CAP was suggested (163).
Hogberg et al. presented the results of a European trial of radiation alone versus adjuvant chemotherapy before or after radiation in 382 patients with stage I, II, IIIA (positive peritoneal cytology only), or IIIC disease who had high-risk factors for recurrence (one or more of deep myometrial invasion, nondiploid DNA, or serous, clear cell, grade 3, anaplastic histology) (164). Chemotherapy was not standardized and could be doxorubicin and platinum (AP); paclitaxel, doxorubicin, and platinum (TAP); paclitaxel and platinum(TP); or paclitaxel, cisplatin, and epirubicin. The study suggested an improvement in progression-free survival with chemotherapy (7% improvement at 5 years, p = 0.03), but survival data were too early to draw any conclusions (164).
These data reinforce the need for adequately powered trials with a single adjuvant chemotherapy treatment in endometrial cancer. Increased pelvic relapse rates have been reported when using adjuvant chemotherapy alone in patients with high-risk or advanced stage disease. The recently opened PORTEC 3 study is a phase III randomized trial comparing chemoradiation and adjuvant chemotherapy with pelvic radiation in patients with high-risk endometrial cancer, and it should establish the role, if any, of concurrent chemoradiation and adjuvant chemotherapy.
Clinical Stage II
When both the cervix and the endometrium are clinically involved with adenocarcinoma, it may be difficult to distinguish between a stage IB adenocarcinoma of the cervix and a stage II endometrial carcinoma. Histopathologic evaluation is not helpful in the differentiation of these two conditions, and the diagnosis must be based on clinical and epidemiologic features. The obese, elderly woman with a bulky uterus is more likely to have endometrial cancer, whereas the younger woman with a bulky cervix and a normal corpus is more likely to have cervical cancer.
The lack of randomized, prospective studies precludes dogmatic statements about the optimal mode of therapy, but modern management favors primary surgery, with adjuvant radiation tailored to the surgical findings.
A large retrospective Italian study reported on 203 patients who underwent primary surgery for stage II endometrial cancer (165). Simple hysterectomy was performed in 135 patients (66%) and radical hysterectomy in 68 (34%). Adjuvant radiation was given to 66 of 111 patients (59%) with stage IIA disease and to 67 of 92 patients (73%) with stage IIB. Survival rates were 79% in the simple hysterectomy group and 94% in the radical hysterectomy group at 5 years, and 74% and 94% at 10 years, respectively ( p = 0.05). Although adjuvant radiation reduced locoregional recurrence, there was no significant difference in survival.
Surveillance, Epidemiology, and End Results data were used in the United States to determine whether primary treatment with simple or radical hysterectomy, with or without adjuvant radiation, altered disease-related survival for patients with FIGO stage II endometrial cancer (166). Cases diagnosed between 1988 and 1994 were analyzed, and included 555 patients (60%) undergoing simple hysterectomy and 377 patients (40%) undergoing radical hysterectomy. The 5-year cumulative survival rates for patients who received surgery alone was 84.4% with simple hysterectomy and 93% with radical hysterectomy ( p = 0.05). There was no significant survival difference for adjuvant radiation versus no radiation in either arm. The researchers concluded that radical hysterectomy was associated with better survival when compared with simple hysterectomy for FIGO stage II corpus adenocarcinoma.
A multicenter study from the United States evaluated 162 patients with surgical stage II endometrial cancer (167). An extrafascial hysterectomy was performed in 75% of patients and a radical hysterectomy in 25%. At least ten nodes were removed in more than 90% of cases. A significantly better 5-year disease-free survival was seen in patients undergoing radical hysterectomy (94%) compared with extrafascial hysterectomy (76%) ( p = 0.05). Adjuvant radiation did not improve survival.
Our current approach to patients with stage II endometrial carcinoma is to perform primary surgery and surgical staging, provided the patient is medically fit.
The surgery is as follows:
Postoperatively, adjuvant radiation is individualized. If lymph nodes are negative, then no adjuvant radiation is given. Patients with one nodal micrometastases in the pelvis receive external pelvic radiation, whereas those with multiple positive pelvic nodes or grossly positive pelvic nodes are given extended-field, external-beam therapy. Patients with upper abdominal disease have cytoreductive surgery.
Surgical Stage IIIA
Patients with stage IIIA endometrial cancer include those with positive peritoneal washings, those with tumor involving the uterine serosa, and those with disease involving the tubes or ovaries. All macroscopic disease can be removed in these patients. Patients with endometrioid histology and positive washings only have a favorable prognosis, with or without adjuvant therapy (168). Adjuvant chemotherapy with or without pelvic radiation is appropriate when the disease involves the adnexae or uterine serosa.
Surgical Stage IIIB
The only study specifically of patients with surgical FIGO stage IIIB endometrial cancer was reported by Nicklin and Petersen in 2000 (169). Isolated vaginal metastases are very uncommon, and only 14 patients (0.7%) could be identified out of 1,940 patients with endometrial cancer treated at the Queensland Centre for Gynaecological Cancer from January 1982 to December
Surgical Stage IVA
Endometrial cancer extending only into the bladder or rectal mucosa is very uncommon. In the latest FIGO annual report, only 49 of 7,990 patients (0.6%) with endometrial cancer had stage IVA disease (5). Treatment must be individualized but would require some type of modified pelvic exenteration, with or without pelvic radiation or chemotherapy.
Surgical Stage IVB
Stage IVB endometrial carcinoma is uncommon, and results of therapy are in general poor. However, an occasional patient is seen with a well-differentiated adenocarcinoma that has metastasized because of prolonged patient or physician delay or because cervical stenosis has prevented the appearance of abnormal bleeding. Such tumors usually contain ER and PR, and prolonged survival may occur with progestin therapy before or after total abdominal hysterectomy, bilateral salpingo-oophorectomy, and possibly radiation therapy.
In a series of 83 patients reported by Aalders et al. (170) from the Norwegian Radium Hospital, the lung was the main site of extrapelvic spread, with 36% of patients having lung metastases. Treatment of stage IV disease must be individualized but usually involves a combination of surgery, radiation therapy, or chemotherapy.
There may be a role for cytoreductive surgery, although data are limited to small, retrospective studies. The largest series, from workers in Baltimore, reported results from 65 patients with stage IVB endometrial cancer. Optimal cytoreduction, defined as residual tumor 1 cm diameter, was accomplished in 36 patients (55.4%), whereas 29 patients (44.6%) underwent suboptimal resection (171). Median survival was 34 months in the optimal group compared with 11 months in the suboptimal group ( p = 0.0001). Patients with no macroscopic residual disease had a median survival of 40.6 months. Similar results have been reported from the Netherlands in a smaller series (172).
In making a decision to undertake primary surgery in a patient with advanced endometrial cancer, both the location and the extent of metastatic disease must be taken in account. In the study by Goff and colleagues, factors that influenced a decision not to undertake cytoreductive surgery included the presence of lung metastases, bladder invasion, clinical involvement of the pelvic sidewall, bone metastases, and liver metastases (173).
A major objective of therapy should be to try to achieve local disease control in the pelvis and to palliate bleeding, discharge, pain, and fistula formation.
Hormonal therapy and chemotherapy for patients with advanced and recurrent endometrial cancer is discussed later in the chapter.
Special Clinical Circumstances
Endometrial Cancer Diagnosed after Hysterectomy
This situation is best avoided by the appropriate investigation of any abnormal vaginal bleeding preoperatively and by routinely opening the excised uterus in the operating room so that the adnexae can be removed and appropriate staging performed if unsuspected endometrial cancer is discovered.
When the diagnosis is made during the postoperative period, the following investigations are recommended:
If the CA125 level is elevated or if the PET or CT scan reveals lymphadenopathy or other evidence of metastatic disease, then laparotomy is usually indicated.
If all investigations are negative, then our approach is as follows.
because of the risk of subsequent ovarian cancer. This is particularly important if there is any family history of breast, ovarian, or colon cancer (Lynch II syndrome).
Synchronous Primary Tumors in the Endometrium and Ovary
This uncommon circumstance occurs much more commonly in young women (174). In at least half of the cases, both endometrial and ovarian tumors are of the endometrioid type, and distinguishing between primary and metastatic lesions may be difficult.
Israeli workers reported that 62% of cases with simultaneous tumors of the endometrium and ovary could be differentiated from metastatic tumors by distinct immunohistochemical expression of ER and PR (p = 0.0006), and 32% could be differentiated by distinct immunostaining for bcl-2 ( p = 0.03) (175).
The Gynecologic Oncology Group reported 74 cases, 23 (31%) of whom had microscopic spread of tumor in the pelvis or abdomen (103). Sixty-four patients (86%) had endometrioid tumors in both sites, and endometriosis was found in the ovary in 23 patients (31%). Patients with tumor confined to the uterus and ovary had a 10% probability of recurrence within 5 years, compared with a 27% probability for those with metastatic disease ( p = 0.006). Similarly, patients with no more than grade I disease at either site had an 8% probability of recurrence within 5 years, compared with a 22% probability for those with a higher grade in either the ovary or the endometrium ( p = 0.05).
Treatment should be determined on the premise that each represents a primary lesion and many require surgery only without adjuvant chemotherapy or radiation (176).
Endometrial Carcinomas in Young Women
Approximately 5% of endometrial cancers occur in women aged 40 years or younger. The majority are associated with a history of chronic anovulation and are usually well-differentiated tumors. A minority occur in association with the hereditary nonpolyposis colorectal cancer syndrome in which case a more variable histologic spectrum occurs (177).Adenocarcinomas of the endometrium occasionally occur in very young women (30 years of age or younger), usually in association with the polycystic ovarian syndrome. Approximately 90% of the lesions are well differentiated and limited to the endometrium (178).
Fertility preservation is often a concern for these young women, and several small series have reported regression of the carcinoma in approximately 80% of cases with a variety ofprogestins (179,180,181,182,183). An MRI is desirable pretreatment to exclude significant myometrial invasion, and the tumors should have grade 1 histology and be PR positive. In spite of successful conservative management, only some 40% of these patients will carry a successful pregnancy.
A 3-month trial of megestrol acetate orally 160 to 320 mg/day or medroxyprogesterone acetate 200 to 500 mg/day is the usual approach. Using megestrol acetate, Gotlieb and colleagues reported a complete response in all 13 patients treated, within a mean period of 3.5 months. Six patients (46%) recurred 19 to 358 months later, four of whom responded to a second course of progestins (179). Taiwanese workers reported complete remission in eight of nine patients (89%) using a combination of megestrol acetate and tamoxifen(180). One patient failed to respond but achieved complete remission after a change from tamoxifen to a gonadotropin-releasing hormone analogue. Four (50%) of the responders later developed recurrent endometrial cancer.
Adequate imaging of the ovaries is important before any decision is made regarding conservative management. In our review of 254 patients with endometrial cancer at the Royal Hospital for Women in Sydney, synchronous ovarian malignancies were found in five of 17 patients (29.4%) younger than 45 years of age compared with 11 of 237 patients (4.6%) older than 45 (p > 0.001). Three other younger patients (17.7%) had secondary ovarian involvement (174). In a series of 102 women aged 24 to 45 years who underwent hysterectomy for endometrial cancer in Los Angeles, 26 (25%) were found to have a coexisting epithelial ovarian tumor; 23 were classified as synchronous primaries, and three were metastases (184).
In the combined series (179,180,181,182,183), approximately 40% of patients who initially respond will recur. Therefore, it is reasonable to recommend hysterectomy once childbearing has been completed to avoid the need for ongoing hormonal manipulation and surveillance with transvaginal ultrasonography. Given the significant incidence of ovarian involvement (174,184) and the efficacy of modern hormonal therapy, there seems little justification for ovarian preservation, unless for psychological reasons.
Endometrial Carcinoma after Endometrial Ablation
With increasing use of endometrial ablation as an alternative to hysterectomy for some women with dysfunctional uterine bleeding unresponsive to hormonal therapy, there have been several reports of the subsequent development of endometrial cancer.
Valle and Baggish (185) reviewed eight case reports and cautioned about the need for proper patient selection. They recommended that all patients should have a preablation biopsy showing a normal endometrium and that patients with persistent hyperplasia unresponsive to hormonal therapy should be recommended for hysterectomy. They also suggested that if endometrial ablation is performed in high-risk patients because of medical contraindications to laparotomy, then vigorous follow-up, including periodic ultrasonography and endometrial sampling, is required. Hysteroscopy with biopsies of the endometrium should be done if bleeding occurs.
Follow-up after management of endometrial cancer should be negotiated with the patient, but our own policy is to alternate visits with the referring gynecologist. Visits are scheduled every 3 months for the first year, every 4 months for the second year, and every 6 months until 5 years. Approximately 10% of recurrences occur beyond 5 years (186), so patients should be told to report early with any abnormal bleeding or other symptoms.
Surveillance gives valuable psychological support to patients and allows accurate collection of data. At each visit, a relevant history should be taken, and any suspicious symptoms investigated appropriately. We do not perform any routine radiological studies, but a vault smear is important to perform on all women who have not had adjuvant radiation.Bristow and colleagues reported that routine vaginal cytology detected an asymptomatic isolated vault recurrence in only two of 377 patients (0.5%), and that it was not cost effective (187). However, most of their patients had received postoperative radiation and would not benefit from early detection of vault recurrence.
Recurrent Endometrial Cancer
According to figures reported in the 26th Annual Report on the Results of Treatment in Gynecological Cancer, approximately 22% of patients treated for endometrial cancer die within 5 years (5) (Table 10.17).
Serum CA125 levels are usually elevated in patients with recurrent disease, particularly if the recurrence is intraperitoneal (188). Pastner et al. (189) reported that none of six patients with an isolated vaginal recurrence had elevated levels, but false positive values may occur in the presence of severe radiation injury of the bowel.
The large series of 379 patients with recurrent disease reported by Aalders et al. (186) from the Norwegian Radium Hospital provides some important information, although management protocols have changed somewhat since that report (190). Local recurrence was found in 50% of the patients, distant metastases in 29%, and simultaneous local and distant metastases in 21%. The median time from primary treatment to detection of recurrence was 14 months for patients with local recurrence and 19 months for those with distant metastases. Thirty-four percent of all recurrences were detected within 1 year and 76% within 3 years of primary treatment. At the time of diagnosis, 32% of all patients were free of symptoms, and the diagnosis was made on routine physical or radiologic examination. For patients with local recurrence, 36% were free of symptoms, 37% had vaginal bleeding, and 16% had pelvic pain.
Isolated Vaginal Recurrence
Isolated vaginal metastases are the most amenable to therapy with curative intent. Before undertaking treatment for an apparent localized recurrence, a PET or CT scan should be obtained to exclude systematic spread.
In the Danish endometrial cancer study in which low-risk patients were followed without radiation, 17 vaginal recurrences were reported, and 15 of these (88.2%) were salvaged with radiation therapy. By contrast, none of seven pelvic recurrences was salvaged (132).
A multiinstitutional study in the United States identified 69 patients with surgical stage I endometrial cancer who were treated without adjuvant radiation and developed an isolated vaginal recurrence (135). Of these, ten (15%) were diagnosed initially with stage IA disease. Histologically, 22 patients (32%) had grade 1 disease, 26 (38%) had grade 2, and 21 (30%) had grade 3. Radiation therapy salvaged 81% of the patients, although 18% died from a subsequent recurrence.
A five-year disease-free survival of 68% for 50 patients was reported from St. Louis with a low rate of complications. Median time to recurrence was 25 months (range 4 to 179 months) (190). On multivariate analysis, age, histologic grade, and size of recurrence were significant predictors of overall survival. All patients who had grade 3 disease were dead by 3.6 years from the time of recurrence.
High-dose-rate brachytherapy usually combined with external-beam therapy, has been reported in a series of 22 patients from Canada (191). After a median follow-up of 32 months, all patients had locoregional control. One developed a distant metastasis and died from disease.
For bulky lesions (>4 cm diameter), surgical resection before radiation may improve local control. Laparotomy has the advantage of allowing a thorough exploration of the pelvis and abdomen to exclude other metastatic foci.
If the patient has had prior pelvic radiation, exploratory laparotomy with a view to some type of pelvic exenteration, offers the only possibility for cure.
Systematic Recurrence: Role of Surgery
Surgery—usually combined with radiation, chemotherapy, or hormonal therapy—may play a role in selected patients with recurrent endometrial cancer, particularly if all residual disease can be resected.
A study of 35 patients undergoing salvage cytoreductive surgery at Johns Hopkins Medical Center reported complete cytoreduction in 23 patients (66%) (192). These patients had a median survival of 39 months compared to 13.5 months for patients with gross residual disease. On multivariate analysis, salvage surgery and residual disease status were significant and independent predictors of postrecurrence survival. Similar conclusions were drawn from a smaller series of patients from Memorial Sloan-Kettering Cancer Center (193).
Patients with a long disease-free interval (>2 years) and an isolated recurrence at any site (e.g., lungs, liver, or lymph nodes) should be considered for surgical resection as long as the patient is medically fit and the surgery is technically feasible.
Role of Hormonal Therapy
Progestational agents have been used successfully as treatment for patients with advanced or recurrent endometrial cancer. Although parenteral administration has been used, oral administration is equally effective (194,195).
The Gynecologic Oncology Group randomized 299 patients with advanced or recurrent endometrial cancer to receive either 200 mg per day or 1,000 mg per day of oralmedroxyprogesterone acetate (196). Among 145 patients receiving the low-dose regimen, there were 25 complete (17%) and 11 partial (8%) responses, for an overall response rate of 25%. For the 154 patients receiving the high-dose regimen, there were 14 complete (9%) and 10 partial (6%) responses, for an overall response rate of 15%. Median survival durations were 11.1 months and 7.0 months, respectively, for the low-dose and high-dose regimens.
The GOG concluded that 200 mg per day of MPA was a reasonable initial approach to the treatment of advanced or recurrent endometrial cancer, particularly for patients whose tumors were well differentiated or PR positive. Patients with poorly differentiated or PR negative tumors had only an 8% to 9% response rate (196).
If an objective response is obtained, then the progestogen should be continued indefinitely. Some responses may be sustained for several years. Side effects from progestins are usually minor and include weight gain, edema, thrombophlebitis, headache, and occasionally hypertension. There is an increased risk of thromboembolism.
The nonsteroidal antiestrogen tamoxifen has also been used to treat patients with recurrent endometrial cancer. It is a first-generation selective estrogen response modulator (SERM) and inhibits the binding of estradiol to uterine ER, presumably blocking the proliferative stimulus of circulating estrogens. Responses are usually seen in patients who have previously responded to progestins, but an occasional response may occur in a patient who is unresponsive to them (197,198). Tamoxifen may be administered orally in a dose of 10 to 20 mg twice daily and continued for as long as the disease is responding. In a review of the literature, Moore et al. (199) reported a pooled response rate of 22% for single-agenttamoxifen.
The third-generation SERM arzoxifene has been evaluated in 29 patients with advanced or recurrent endometrial cancer (200). The drug was administered orally in a dose of 20 mg per day, and toxicity was minimal. There were nine responses (31%) and a median duration of response of 13.9 months, the longest reported in a phase II trial of this patient population. Arzoxifene warrants further evaluation for patients with advanced and recurrent endometrial cancer.
Role of Cytotoxic Chemotherapy
Cytotoxic chemotherapy for endometrial cancer is given with palliative intent, and responses are generally of short duration. Many women with endometrial cancer are elderly and have other comorbidities such as obesity, diabetes mellitus, and cardiovascular disease. They may have had pelvic radiation, which can limit hematological reserve. All of these factors have to be taken into consideration when making treatment recommendations, but chemotherapy should be considered in patients with a good performance status.
The most active drugs are the platinum agents, taxanes, and anthracyclines. Response rates are in the order of 30%, but the duration of response is usually only measured in months(201,202,203,204).
The response to second-line therapy is generally poor. The best response rates are with taxanes, and the GOG has reported paclitaxel to have a 35% response rate in previously untreated women (205) and a 27% response rate in previously treated patients. In the latter group, the median duration of response was 4.2 months, and the median overall survival was 10.3 months. Topotectan was also studied as a second-line agent by the GOG, but the response rate was only 9% (206).
There have been a number of randomized trials of combination versus single agent chemotherapy, and there have two recent systematic reviews of chemotherapy for metastatic endometrial cancer. Both concluded that combination chemotherapy with doxorubicin and cisplatin resulted in higher response rates than doxorubicin alone (207,208). The combination was associated with response rates in the order of 40% with progression-free survivals of 5 to 7 months. The addition of paclitaxel to either of the above regimens resulted in a higher response rate (57% versus 37%) and a small survival advantage. However, the toxicity is excessive, and there have been a number of treatment-related deaths (209).
The combination of carboplatin and paclitaxel has been evaluated in a nonrandomzed setting, and response rates as high as 67%, with 29% complete responses, have been reported. Toxicity is acceptable, the median progression-free survival has been as high as 14 months, and overall survival has been approximately 26 months (210,211). This regimen is commonly used in the community to treat patients with metastatic endometrial cancer. The GOG is comparing carboplatin and paclitaxel with paclitaxel, doxorubicin, and cisplatin(TAP), and the results of this study will be important for the further development of chemotherapy for endometrial cancer.
Recently, the GOG reported the results of a randomized, phase III trial of whole-abdominal radiation versus doxorubicin-cisplatin in advanced endometrial cancer (152). To be eligible, the maximum size of residual disease had to be 2 cm diameter, and there were 388 evaluable patients. There was a significantly better progression-free and overall survival in the chemotherapy arm, but approximately 55% of all patients recurred.
Newer targeted treatments are also being investigated, including sorafenib and bevacizumab, as well as traztusumab in HER-2 amplified tumors (15% to 30% of serous and highgrade endometrial cancers have evidence of amplification). One of the most active new agents (temsirolimus) is directed against the mammalian target of rapamycin (mTOR), and a response rate of 26% has been reported in chemonaive patients (212).
Uterine papillary serous carcinomas are histologically similar to ovarian serous tumors, but the reported response rate to cisplatin-containing combination chemotherapy has been inconsistent, with some studies suggesting a lower response (213). However, there are data to suggest high response rates to carboplatin or cisplatin and paclitaxel. Rodriguez et al. (214) reported a complete response in three of 13 patients (23%) and a partial response in eight of 13 (62%) to various platinum combinations, including cisplatin or paclitaxel in three patients. Median duration of response was 7.5 months (range 1 to 30 months).
Our experience at the Royal Hospital for Women would justify further study of platinum-based chemotherapy in patients with serous endometrial carcinomas (215). All six patients with stages I or II disease given adjuvant platinum-based chemotherapy were tumor free with a mean follow-up of 31.6 months (range 12 to 68 months) and four of 12 (33%) with stages III to IV disease remained tumor free with a mean follow-up of 22.5 months.
There have been a number of similar retrospective phase II reports (216,217,218). Kelly et al. studied 75 patients with stage I serous carcinomas and reported that platinum-based chemotherapy improved both disease-free and overall survival (216). They concluded that patients with stage IA disease could be observed but recommended concomitant platinum-based chemotherapy and vaginal vault radiation for all other patients with stage I disease.
The impact of adjuvant chemotherapy in patients with uterine serous papillary carcinomas should be clarified by the PORTEC 3 study, which is in progress.
Particularly for younger women, hormone-replacement therapy is an important issue after treatment for endometrial cancer. Patients with stage I disease have a good prognosis, and protection against osteoporosis and quality of life issues are important. Although it has been frequently stated that estrogen-replacement therapy is contraindicated in patients who have had endometrial cancer, Creasman et al. (219) have challenged this concept. In a nonrandomized study, they reported no deleterious effect from estrogen given to 47 patients with stage I endometrial cancer compared with 174 patients with similar risk factors who did not receive estrogen. In fact, the estrogen group experienced a significantly longer diseasefree survival.
Our practice is to offer patients daily conjugated estrogens (Premarin) 0.625 mg or tibilone (Livial).
Although individual institutions may report superior results, the most comprehensive survival data are provided in the Annual Report on the Results of Treatment in Gynecological Cancer. Survival by FIGO surgical stage for the years 1999 through 2001 are shown in Table 10.22. Survival by histological grade is shown in Table 10.23. The significance of histologic grade is highlighted by the fact that patients with stage II, grade 1 and 2 tumors have a better prognosis than patients with stage I, grade 3 lesions.
Survival in relation to grade and depth of myometrial invasion for stage I disease is shown in Table 10.24, and the poor prognosis associated with papillary serous and clear cell carcinomas is shown in Table 10.25.
Table 10.22 Carcinoma of the Corpus Uteri: Patients Treated from 1999 to 2001; Survival Rates by FIGO Surgical Stage (n = 7,990)
Table 10.23 Carcinoma of the Corpus Uteri: Patients Treated from 1999 to 2001; Survival Rates for Surgical Stages I and II by Histologic Grade
Table 10.24 Carcinoma of the Corpus Uteri: Patients Treated from 1999 to 2001; Survival Rates in Stage I by Surgical Stage and Grade of Differentiation (n = 5,095)
Table 10.25 Carcinoma of the Corpus Uteri: Patients Treated from 1999 to 2001; Survival Rates by Histologic Type (n = 8,033)
Uterine sarcomas are rare mesodermal tumors that account for approximately 3% of uterine cancers (220). They are a heterogeneous group of tumors, and thus individual experience with each lesion is limited. The main lesions are leiomyosarcomas, carcinosarcomas, and endometrial stromal sarcomas, the first two tumors having a higher incidence in black women in the United States (221). Treatment protocols are not standardized, and there are few controlled studies evaluating different therapeutic approaches. Most subgroups behave in an aggressive manner and have a poor prognosis, with high rates of local recurrence and distant metastases (222).
Pelvic radiation is thought to predispose to the subsequent development of uterine sarcomas (223). Zelmanowicz et al. (224) reported that endometrial carcinomas and malignant mixed müllerian tumors have a similar risk factor profile, which is compatible with the hypothesis that the pathogenesis of these two tumors is similar.
Criteria for the histopathological classification of sarcomas has been changing, and such lesions should be reviewed by an expert gynecologic pathologist. Much less emphasis is placed on mitotic counts than was previously the case.
Mesodermal derivatives from which sarcomas may arise include uterine smooth muscle, endometrial stroma, and blood and lymphatic vessel walls. Uterine sarcomas can be divided basically into two types:
They also may be subdivided into homologous and heterologous tumors, depending on whether the malignant mesodermal elements are normally present in the uterus. Malignant smooth muscle and stroma represent homologous elements, whereas malignant striated muscle and cartilage represent heterologous elements.
There is a new FIGO staging system for Uterine Sarcomas (Table 10.26) on page 667. More accurate prognostic information is obtained by surgical staging, particularly for carcinosarcomas.
Smooth Muscle Tumors
Leiomyosarcomas, which must be distinguished from the cellular leiomyomas and atypical leiomyomas (see Chapter 5), occur most commonly in the 45- to 55-year age group and account for approximately 30% of uterine sarcomas. In the SEER study of 1,396 patients, the median age of the patients was 52 years (225).
Leiomyosarcomas usually arise de novo from uterine smooth muscle, although rarely they may arise in a preexisting leiomyoma. A subset of myomas with a deletion of a specific portion of chromosome 1 has both a specific cellular morphology and a genetic-transcription profile similar to those of leiomyosarcomas. Thus cytogenetics may help identify these exceptions to the rule, and allow prediction of malignant progression (226). A review of 1,432 patients undergoing hysterectomy for presumed fibroids at the University of Southern California revealed leiomyosarcoma in the hysterectomy specimens in ten patients (0.7%). The incidence increased steadily from the fourth to the seventh decades of life (0.2%, 0.9%, 1.4%, and 1.7%, respectively) (227).
Most leiomyosarcomas are accompanied by pain, a sensation of pressure, abnormal uterine bleeding, or a lower abdominal mass. Rapid enlargement of a fibroid is a possible sign of malignancy. A few patients may have signs of metastatic disease such as a persistent cough, back pain, or ascites. On physical examination, it is impossible to distinguish leiomyosarcomas from large leiomyomas or from other uterine sarcomas. Pap smears are unrewarding, and uterine curettings are diagnostic for only the 10% to 20% of tumors that are submucosal (228). Diagnosis usually is not made before surgery.
Intravenous leiomyomatosis is a rare, relatively benign uterine smooth muscle tumor in which much of the tumor is present in (and may arise from) veins (229). It may extend as rubbery cords beyond the uterus into the parametrium or occasionally into the vena cava. Some patients may survive for prolonged periods in spite of incomplete resection of diseased tissue. High levels of ER and PR are present in some tumors, and regression may occur after menopause.
Leiomyomatosis peritonealis disseminata is a condition in which numerous nodules of histologically benign smooth muscle are present on peritoneal surfaces (230). It is frequently associated with a term pregnancy or with the use of oral contraceptives, and regression may occur after termination of pregnancy.
Benign metastasizing leiomyoma is a rare disorder characterized by a histologically benign smooth muscle tumor that originates in the uterus and spreads elsewhere, usually to the lungs. Controversy exists regarding whether lung lesions represent metastases of a benign uterine primary tumor or synchronous or metachronous development of an independent lung lesion. Optimal therapy is also unclear, but surgical resection and hormonal therapy are generally recommended (231).
The only treatment of any proven curative value for the frankly malignant leiomyosarcomas is surgical excision. This typically involves total abdominal hysterectomy and bilateral salpingo-oophorectomy, although in young patients it is reasonable to preserve the ovaries, particularly if the tumor has arisen in a fibroid (232).
Lissoni et al. (233) reported eight young patients with a diagnosis of leiomyosarcoma after myomectomy who were followed conservatively. All were nulliparous, and all had no evidence of disease on ultrasonography, hysteroscopy, chest radiography, and pelvic and abdominal CT scan or MRI. The mean mitotic count of the leiomyosarcomas was 6 per 10 high-power field (HPF), with a range of 5 to 33. With a median follow-up of 42 months, three live births were recorded, but one patient recurred and died.
For leiomyosarcomas the GOG study of 59 patients reported positive lymph nodes in only 3.5% of patients, positive washings in only 5.3%, and adnexal involvement in only 3.4% (233). For 71 patients with leiomyosarcoma confined to the uterus and/or cervix, the Memorial Sloan-Kettering group reported ovarian metastases in two patients (2.8%) and lymph node metastases in none. Three of 37 patients (8.1%) with gross extrauterine disease had positive nodes, and all were clinically suspicious (234). Wu et al. reported no pelvic or paraaortic lymph node involvement in 21 patients who underwent complete surgical staging (235). The SEER study reported lymph node metastases in 23 of 348 patients (6.6%) who underwent lymphadenectomy, but performance of a lymphadenectomy was not associated with improved survival (224), because most lymph node metastases are found in patients with advanced disease.
Spread of leiomyosarcomas is mainly hematogenous, so surgical staging appears to be of less importance for these tumors.
A possible role for secondary cytoreduction for leiomyosarcomas was reported from the Johns Hopkins Medical Center. In a recent review of 128 patients with recurrent uterine leiomyosarcoma, researchers reported prolonged survival in a select group who had a prolonged progression-free survival and an isolated site of recurrent disease amenable to complete resection. Neither chemotherapy nor radiation therapy was beneficial (237).
A large study of 1,042 patients with uterine sarcomas recorded in the Cancer Registry of Norway from 1956 to 1992 reported no change in the 5-year survival rate after the introduction of chemotherapy into the treatment protocols (238).
The most active drugs include doxorubicin, cisplatin, ifosfamide, paclitaxel, docetaxel, and gemcitabine. Unfortunately, most responses are partial and of short duration.
In the GOG trials, leiomyosarcomas had a 25% (7 of 28) overall response rate to doxorubicin (239). Liposomal doxorubicin (Doxil) showed no advantage over historical results withdoxorubicin (240). There was a 17.2% (6 of 35) partial response rate to ifosfamide (241) and a 3% (1 of 33) partial response rate to cisplatin (242). For paclitaxel, there were three complete responses (9%) whereas eight patients (24%) had stable disease for at least two courses of treatment (243). The combination of gemcitabine plus docetaxel demonstrated activity in patients with metastatic leiomyosarcomas, the GOG reporting objective responses in 15 of 42 patients (35.8%) overall. The complete response rate was 4.8%, and the partial response rate was 31%, with an additional 11 patients (26.2%) having stable disease (244).
The combination of docetaxel and gemcitabine has also been reported to be active as second-line therapy for patients with metastatic leiomyosarcoma. In a study of 48 patients, 90% of whom had progressed following doxorubicin-based chemotherapy, Hensley et al. reported a complete response in three patients (8.8%) and a partial response in 15 patients (44.1%) for an overall response of 53% (95% CI 35% to 70%) (245). An additional seven patients (20.6%) had stable disease. Fifty percent of patients who were treated previously with doxorubicin had a response. The median time to progression was 5.6 months.
Because of the propensity for early hematogenous spread, adjuvant chemotherapy after hysterectomy to eliminate micrometastases is an attractive concept. However, in a randomized GOG study of doxorubicin after total abdominal hysterectomy and bilateral salpingooophorectomy for stage I or II uterine sarcoma, neither survival nor progression-free interval was prolonged by the adjuvant chemotherapy (246).
A study of 27 patients from the Massachusetts General Hospital also reported that use of adjuvant chemotherapy after optimal surgery did not decrease the rate of recurrence for patients with leiomyosarcomas (247). Future studies should evaluate the use of docetaxel and gemcitabine.
Postoperative external-beam pelvic radiation is of no proven benefit in terms of survival, although it has generally been thought to improve tumor control in the pelvis (248). However, the recent randomized clinical trial of adjuvant external pelvic radiation versus observation for patients with nonmetastatic leiomyosarcomas revealed no benefit even in terms of local control (249).
Summation of Management of Leiomyosarcomas
The only proven benefit for patients with leiomyosarcomas is total abdominal hysterectomy. Ovaries should normally be removed in postmenopausal women. Surgical staging appears to be of no value, and neither does adjuvant chemotherapy. These patients should be entered onto clinical trials of new therapeutic agents, particularly docetaxel andgemcitabine. Pelvic radiation appears to offer no benefits in terms of local control or survival.
Endometrial Stromal Tumors
Endometrial stromal tumors are divided into two major categories: benign endometrial stromal nodules and endometrial stromal sarcomas. The division of endometrial stromal sarcomas into low-grade and high-grade categories has fallen out of favor, and the term endometrial stromal sarcoma is now considered best restricted to neoplasms that were formally referred to as “low-grade” endometrial stromal sarcomas (250). High-grade tumors without recognizable evidence of a definite endometrial stromal phenotype are now termed endometrial sarcomas (251) or undifferentiated uterine sarcomas. Mitotic counts are no longer used to differentiate high-grade from low-grade lesions. Endometrial stromal sarcomas constitute 15% to 25% of uterine sarcomas (252).
Most patients are in the age range of 42 to 53 years. More than half the patients are premenopausal, and young women and girls may be affected. Abnormal vaginal bleeding is the most common presenting symptom, and abdominal pain and uterine enlargement may occur (253). Although they may be intramural, most endometrial stromal sarcomas involve the endometrium, and uterine curettage usually leads to diagnosis.
Endometrial stromal sarcomas have infiltrating margins and demonstrate venous and lymphatic invasion. Although their behavior is relatively indolent, late recurrences and distant metastases may occur (252,253). The most frequent sites of recurrence for patients with stage I disease are the pelvis and abdomen (252,253). Prolonged survival and even cure are not uncommon after surgical resection of recurrent or metastatic lesions.
Undiffentiated uterine sarcomas behave aggressively compared with endometrial stromal sarcomas. In the original series reported by Evans, six of the seven patients died of disease between 10 and 34 months after diagnosis (251).
Total abdominal hysterectomy and bilateral salpingo-oophorectomy, with radical cytoreductive surgery for extrauterine involvement, has been the standard recommendation for endometrial stromal sarcomas (254).
Preservation of the ovaries may be an option for premenopausal women with stage I disease. In a multiinstitutional case-control study, Li et al. reported no differences in progression-free or overall survival when 12 premenopausal patients who did not undergo bilateral salpingooophorectomy were matched with 24 controls (255). Twenty-two available tumors demonstrated positivity for both estrogen and progesterone receptors.
There are limited data available on the incidence of lymph node metastases in endometrial stromal sarcomas. Goff et al. reported five low-grade and five high-grade endometrial stromal tumors, seven of which had lymph node sampling. None of the seven had positive nodes (256). Ayhan et al. reported eight cases of endometrial stromal sarcoma (two low-grade and six high-grade). Four had lymph node sampling, and none had lymph node metastases (257). Riopel et al. reported positive nodes in five of 15 patients (33%) with low-grade endometrial stromal sarcomas. In two of the five patients, nodal metastases were found years later at the time of recurrence; of the three patients who had lymph node metastases at their initial presentation, two had gross extrauterine involvement (258).
Patients with a late recurrence of an endometrial stromal sarcoma benefit from aggressive cytoreductive surgery to remove all macroscopic disease if possible.
Chemotherapy or Hormonal Therapy
Endometrial stromal sarcomas usually contain estrogen and progesterone receptors (255), and patients with advanced and recurrent disease respond well to progestins or aromatase inhibitors such as letrozole (259). Chu et al. reported eight patients with recurrent endometrial stromal sarcoma who were treated with progestin therapy. Complete responses were seen in four patients (50%), and three others (38%) had stable disease. There are many anecdotal case reports of hormonal therapy, and the responses can be very durable. There are no data to support prophylactic hormonal therapy (260)
Patients with a previous history of low-grade endometrial stromal sarcoma should not receive estrogens or tamoxifen (259) because there may be stimulation of growth, although the estrogen and progesterone of normally functioning ovaries does not seem to be a problem (255,260).
There are limited data on the role of systemic chemotherapy in women with endometrial stromal sarcoma. However, chemotherapy could be considered in selected patients who progress on hormonal therapy as well as those with metastatic undifferentiated endometrial sarcomas. There are anecdotal case reports of response to a variety of agents, and similar drugs to those used in carcinosarcomas would be a reasonable option. A recent report from China suggested that a combination of ifosfamide, epirubicin, and cisplatin had activity in patients with undifferentiated endometrial sarcomas (261).
Radiation therapy is not appropriate for endometrial stromal sarcomas. It may decrease pelvic recurrence without improving survival for patients with endometrial sarcomas.
Summation of Management for Endometrial Stromal Tumors
Patients with endometrial stromal sarcomas should undergo total abdominal hysterectomy, bilateral salpingo-oophorectomy, and resection of any extrauterine disease. Preservation of ovaries is an option for premenopausal patients with stage I disease. Patients with advanced disease respond well to progestins or aromatase inhibitors. Late recurrences are not infrequent, and aggressive secondary cytoreduction to remove all macroscopic disease may result in long-term survival.
Patients with undifferentiated uterine sarcomas should undergo total abdominal hysterectomy and bilateral salpingo-oophorectomy. Pelvic radiation will decrease pelvic recurrence without improving survival.
Malignant mixed mesodermal tumors (MMMTs) or carcinosarcomas usually occur in an older age group, most patients being postmenopausal (234). The frankly malignant variants grow rapidly and usually are accompanied by postmenopausal bleeding, pelvic pain, a palpable lower abdominal mass, or symptoms of metastatic disease. Most patients have an enlarged uterus, and the tumor protrudes through the cervical os like a polyp in approximately half the patients (220). Uterine curettage usually detects malignant tissue in the uterus, although determination of the exact nature of the tumor may require histologic examination of the entire specimen.
There is now convincing evidence that most, but not all, uterine carcinosarcomas are monoclonal tumors and really metaplastic carcinomas (262). The behavior of these tumors is similar to that of high-grade endometrioid adenocarcinomas. Metastases are usually from the carcinomatous element, and the sarcomatous element is believed to be derived as a result of dedifferentiation of the carcinomatous component (263). Heterologous differentiation, including rhabdomyosarcomotous differentiation, is not uncommon in MMMTs, and occasionally pure heterologous sarcomas have been reported (264).
Carcinosarcomas should undergo surgical staging in the same manner as high-grade endometrial carcinomas. The Gynecologic Oncology Group reported a clinicopathologic study of 301 carcinosarcomas in 1993 (234). Adnexal metastases were present in 12% of patients, lymph node metastases in 18%, and positive peritoneal washings in 21%. No omental biopsies were taken.
A Californian study of 62 patients with carcinosarcoma apparently confined to the uterus reported occult metastases in 38 patients (61%) (265). Adnexal metastases were present in 23% of patients, positive pelvic nodes in 31%, positive paraaortic nodes in 6%, omental involvement in 13%, and positive peritoneal washings in 29%.
Our experience in Sydney suggests that very good survival rates can be obtained in patients with carcinosarcomas if they are subjected to surgical staging, postoperative radiation based on the surgical findings, and adjuvant chemotherapy with cisplatin and epirubicin (266). In a study of 38 patients with disease clinically confined to the uterus, 9 patients (24%) were upstaged to III or IV. Positive lymph nodes were present in 13%, adnexal metastases in 11%, and omental metastases in 8%.
There are a number of active agents, including cisplatin, carboplatin, anthracyclines, ifosfamide, and paclitaxel. As these are relatively uncommon tumors, the studies have generally been small and patient accrual has been slow. For cisplatin (50 mg/m2 every 3 weeks), the GOG reported a complete response rate of 8% and a partial response rate of 11% among 63 patients with advanced or recurrent mixed mesodermal tumors who had received no previous chemotherapy (242).
Ifosfamide is also an active agent for mixed mesodermal sarcomas, the GOG demonstrating nine responses among 28 patients (31.2%) (267). A small improvement in progression-free survival was noted with the addition of cisplatin to ifosfamide in a phase III GOG trial, but the added toxicity may not justify use of this combination (268).
The GOG also evaluated paclitaxel in 44 patients with carcinosarcoma of the uterus. Four patients (9.1%) had a complete response, and four (9.1%) had a partial response (269).
More recently, the GOG reported the results of a randomized study comparing ifosfamide alone versus ifosfamide and paclitaxel in 179 patients with advanced uterine carcinosarcomas (270). They reported a response rate of 29% with ifosfamide alone and 45% with the combination. The odds of response stratified by performance status were 2.21 times greater for the combination arm (p = 0.017). The median progression-free and overall survivals for ifosfamide compared with ifosfamide and paclitaxel were 3.6 versus 5.8 months and 8.4 versus 13.5 months, respectively.
Adjuvant Chemotherapy for Early Stage Disease
In a nonrandomized study, Peters et al. (271) reported 17 patients with high-risk clinical stage I uterine stromal sarcomas or mixed mesodermal tumors who were treated with six cycles of cisplatin 100 mg/m2 and doxorubicin 40 to 60 mg/m2 every 3 to 4 weeks. Fourteen of the patients had invasion to the outer third of the myometrium, seven had documented lymph node metastases, and five had positive peritoneal washings. With a median follow-up of 34 months, there were only four recurrences, giving a projected 5-year survival rate of 75%.
This experience is similar to our own data from the Royal Hospital for Women (266) with cisplatin and epirubicin, together with tailored radiation based on the surgical findings.
The GOG reported an overall 5-year survival of 62% for 65 patients with completely resected clinical stage I or II disease treated with adjuvant ifosfamide and cisplatin. No postoperative radiation was given, and the authors commented that pelvic relapse remained problematic (272).
Wolfson et al. recently reported the results of a relatively large randomized clinical trial comparing whole-abdominal radiation with three cycles of cisplatin, ifosfamide, and mesna(CIM) in 201 patients with stages I to IV uterine carcinosarcomas (273). The estimated crude probability of recurring within 5 years was 58% for WAR and 52% for CIM. Adjusting for stage and age, the recurrence rate was 29% lower for patients receiving CIM. They concluded that the results of the trial favored chemotherapy although the differences were small.
Several nonrandomized studies have suggested that radiotherapy reduces local recurrence (274,275,276). In a report of 300 patients treated at the M.D. Anderson Hospital between 1954 and 1998, Callister et al. reported that radiotherapy reduced the risk of local recurrence from 48% to 28% and prolonged the time to distant relapse but did not improve overall survival (276). The recent EORTC randomized study of stages I and II uterine sarcomas confirmed that external pelvic radiation decreases pelvic relapse but does not improve overall survival for carcinosarcomas (249). There were 91 patients with carcinosarcomas in the study.
Based on the limited available data, our current approach is to treat patients with negative nodes with vault brachytherapy. We give whole-pelvic radiation to patients who have had no lymph node dissection or who have one positive pelvic node and give extended-field radiation to patients with multiple positive pelvic nodes or positive paraaortic nodes.
Summation of Management of Carcinosarcomas
Current evidence would suggest that carcinosarcomas of the uterus should undergo full surgical staging and resection of any gross metastatic disease. Postoperative radiation should be tailored to the operative findings. Adjuvant chemotherapy with either cisplatin and epirubicin or an ifosfamide-based combination may be beneficial based on small phase II studies, but confirmation requires a randomized phase III trial.
The frankly malignant uterine sarcomas usually have a poor prognosis. Surgical stage is the most important prognostic variable, with younger age (less than 50 years) also significant in a Japanese multiinstitutional study (277).
In the Mayo study of 208 patients with leiomyosarcoma (232), the median disease-specific survival for 130 patients with stage I disease was 7.8 years, whereas it was 3.7 years for 13 patients with stage II disease, 2.3 years for 18 patients with stage III, and 1.3 years for 41 patients with stage IV. Thirty-three patients with a tumor <5 cm diameter had a median survival of more than 30 years, whereas 128 patients with tumors larger than 5 cm diameter had a median survival of 3.5 years. The significance of tumor size confirmed the findings from a large Nordic study (278). Adjuvant chemotherapy or radiation therapy did not seem to be of any benefit (232). If the leiomyosarcoma arises in a benign fibroid, the prognosis appears to be better (228).
A recent analysis of 1,396 leiomyosarcomas identified from the SEER database from 1988 to 2003 reported 5-year disease specific survivals of 75.8%, 60.1%, 44.9%, and 28.7% for stages I, II, III, and IV, respectively (224). Independent prognostic factors on multivariate analysis were age, race, stage, grade, and primary surgery. Oophorectomy or lymphadenectomy had no impact on survival.
For 301 mixed mesodermal tumors, the GOG reported a recurrence rate of 53% (homologous 44%; heterologous 63%) (234). Factors significantly related to progression-free interval by univariate analysis were adnexal spread, lymph node metastases, tumor size, vascular space invasion, depth of myometrial invasion, positive peritoneal washings, histologic grade, and cell type. On multivariate analysis, the significant prognostic factors were adnexal spread, lymph node metastases, cell type, and cell grade.
Better survivals have been reported in smaller studies, including our own, following surgical staging, tailored radiation, and adjuvant chemotherapy (266,271). A randomized trial to determine the role of adjuvant chemotherapy after surgical staging and tailored postoperative radiation would be justified.
Endometrial stromal nodules are benign but can usually only be distinguished from endometrial stromal sarcomas after hysterectomy. For younger women wishing to preserve fertility, a combination of diagnostic imaging and hysteroscopy may be useful in monitoring the growth of the lesion, and local excision has been successful in occasional cases (279).
Endometrial stromal sarcomas are relatively indolent tumors with a tendency to very late recurrence, one-third to one-half of patients recurring as many as 30 years after treatment (253,260).
Chang et al. (253) reported that for endometrial stromal sarcomas, stage and mitotic index were both independent predictors of overall and disease-free survival, but when only stage I patients were considered, mitotic index disappeared from the Cox model. These authors placed most high-grade sarcomas into the undifferentiated sarcoma category on the basis of anaplastic cells that had mitotic indices in excess of 20 per 10 HPF.
Endometrial sarcomas are highly aggressive tumors with a very poor prognosis (251). Li et al. reported a recurrence rate of 24% (9 of 37) for stage I endometrial stromal sarcomas compared to 73% (8 of 11) for endometrial sarcomas (261).