INTRODUCTION — The uterine papillary serous (UPSC) and clear cell types of endometrial cancer, which account for less than 15 percent of all endometrial cancers, are biologically more aggressive than usual endometrioid cancers.

• Epidemiology — Women with UPSC tend to be older (median age 65 to 70) than women with endometrioid cancers, nonobese rather than obese, and parous rather than nulliparous. They are also more common in African-American as compared to Caucasian women.

• Atypical hyperplasia, the precursor of endometrioid cancer, is not a precursor lesion of UPSC, and UPSC tumors are not associated with estrogen excess. It has been proposed that UPSC carcinomas develop from "endometrial intraepithelial carcinoma" (EIC), a lesion related to malignant transformation of the endometrial surface epithelium (such as a benign endometrial polyp), against a background of endometrial atrophy.

• Histology — UPSC tumors have complex papillary architecture, which resembles papillary serous carcinoma of the ovary; psammoma bodies are present in 60 percent. Marked nuclear atypia is always present, and all UPSCs are considered high-grade tumors. Most endometrioid cancers are well differentiated endometrioid adenocarcinomas, which are characterized by a proliferation of back-to-back endometrial glands without intervening stroma.

Clear cell cancers are characterized by tubulocystic, papillary, or solid patterns; psammoma bodies are present in up to 10 percent of cases, most often in the papillary variant. The cells may be clear because of the presence of glycogen. Myometrial invasion occurs in about 80 percent of cases. At least from the standpoint of gene expression, they appear more similar to clear cell cancers arising in other organs (eg, the kidney) than to other uterine cancers including those of the papillary serous variety. Nevertheless, they are treated similar to papillary serous cancers.

• Lymph node metastasis and extrauterine disease are extremely rare with endometrioid cancers that are noninvasive but frequently occur with noninvasive UPSC. Furthermore, In UPSC, tumor grade and depth of myometrial invasion are not predictive of extrauterine spread. By comparison, the incidence of extrauterine spread in endometrioid cancer correlates with both tumor grade and depth of myometrial invasion

In one study, lymph node metastases were found in 36 percent of women with UPSC and no myometrial invasion, compared to 50 and 46 percent of those with inner one-half, and outer one-half invasion, respectively. The corresponding rates of intraperitoneal disease for these three groups were 43, 37, and 35 percent, respectively. Even having as little as 10 percent of the tumor composed of UPSC places the patient at the same risk for metastasis

• Both UPSC and clear cell tumors have a higher propensity for lymphovascular invasion, and intraperitoneal as well as extraabdominal spread, than endometrioid cancers. At the time of presentation, approximately 60 to 70 percent of women with UPSC will have disease spread outside of the uterus

Prognostic implications — As a result of all of these factors, UPSC and clear cell tumors are associated with a poorer outcome than endometrial adenocarcinomas, particularly among African-American women

• In a study of 574 women with stage I-II endometrial cancer reported that UPSC and clear cell types accounted for 12 percent of cases, and they were associated with higher recurrence rates than with adenocarcinomas (38 and 22 versus 9 percent, respectively)

• A study of 60 comprehensively surgically staged patients with node-negative stage I UPSC noted the five-year progression-free survival rate was only 70 percent

• The inferior survival of papillary serous and clear cell cancers compared to other histologies is reflected in the latest survival data from FIGO (the International Federation of Gynecology and Obstetrics), stratified according to histologic type. Even among women with stage I disease, the five year survival rates for UPSC or clear cell cancer were 72 and 80 percent, respectively; the corresponding value for endometrioid cancers was 89 percent.

At least some data suggest that outcomes for clear cell cancers that are confined to the uterus and without extension to the cervix are better than those for similarly staged patients with UPSC

Race and prognosis — As with other histologic types of uterine cancer, African-American women with UPSC have a consistently poorer outcome than do Caucasians, an effect that is incompletely explained by imbalances in social and clinicopathologic factors.

While these results are preliminary, they raise the possibility of future therapeutic strategies that target HER-2/neu (eg, with the anti-HER-2/neu monoclonal antibody trastuzumab [Herceptin] that is currently used for treatment of breast cancer) in women who are predicted to have a poor outcome with standard therapies.

SURGICAL MANAGEMENT — Recurrence patterns (particularly for UPSC) are more similar to that of epithelial ovarian carcinoma (EOC) than they are to endometrial cancer As such, surgical therapy should parallel that of EOC, and serum CA-125 levels provide a good marker of disease activity

Meticulous staging is important to exclude the presence of extrauterine disease, which portends a poor prognosis and influences the postoperative management plan. Approximately 70 percent of patients with UPSC will be found to have macroscopic or clinically occult stage III or IV disease. Therefore, women with UPSC and clear cell cancers should undergo total abdominal hysterectomy and bilateral salpingo-oophorectomy (TAH-BSO), pelvic/paraaortic dissection (and not just sampling, omentectomy, assessment of the peritoneal cavity including pelvic and diaphragmatic cytology, and resection of gross metastases. If possible, these women should be referred to gynecologic oncologists for surgical management.

The strongest predictor of overall survival is the amount of residual disease following surgery. In two series totaling 83 patients, as an example, women with optimal cytoreduction (ie, residual disease 1 cm in maximal diameter) had longer median survival than those left with suboptimal residual disease (26 versus 10 months  and 15 versus 8 months. In another report from the Gynecologic Oncology Group (GOG), there were no long-term survivors among those left with gross residual disease following surgery (see below)

Because of poor results with surgery alone, even for patients with stage I disease, both radiation therapy (RT) and chemotherapy have been added postoperatively in an attempt to improve outcomes. However, the benefit of these approaches as well as the optimal treatment for each stage of disease remain unclear. Randomized controlled trials are not available, and interpretation of the majority of retrospective series is compromised by the use of a variety of adjuvant strategies, and the heterogeneity of the treated population. Most reports include incompletely surgically staged women (with some notable exceptions  whose disease stage ranges from I to IV. This renders the results virtually uninterpretable.

ADJUVANT RT — The role of adjuvant RT in women with UPSC and/or clear cell endometrial carcinomas is controversial. Traditionally, such patients have been treated with adjuvant RT, typically pelvic RT with or without vaginal brachytherapy. Since the initial observation of the propensity of UPSC to relapse in the abdomen however, attention has been primarily focused on adjuvant whole abdominal RT (WART).

Despite the popularity of this approach, no prospective phase III clinical trials have been conducted evaluating the role of WART in UPSC and/or clear cell cancer; as a result, its true benefit remains unproven. Published retrospective series report large variations in disease-free survival and overall survival following WART

• In a series of 28 patients, three-year disease-free and overall survival rates folllowing adjuvant WART were 87 and 87 percent for women with stage I to II disease, and 32 and 61 percent for patients with stage III to IV disease, respectively. Only three women recurred in the upper abdomen.

• Of the 58 women treated with postoperative WART in a second series, 11 of the 21 relapses were in the abdomen or pelvis. Ten-year disease-free and cause-specific overall survival rates were 35 and 74 percent, respectively.

• A single arm study from the GOG (GOG 94) which included 180 surgically staged and optimally debulked women with stage III or IV disease, the three-year disease-free survival and overall survival rates after adjuvant WART were 27 and 35 percent, respectively

The future of WART in stage III to IV UPSC and/or clear cell carcinoma is uncertain given the results of GOG 122. In this trial, optimally debulked (2 cm residual) women with stage III to IVa  endometrial cancer were randomly assigned to WART with a pelvic boost or eight cycles of chemotherapy with cisplatin and doxorubicin. Chemotherapy proved to be superior to WART in terms of two-year overall survival and progression-free survival (PFS). Twenty percent of the enrolled patients had UPSC, and response to treatment was the same for both endometrioid and UPSC tumors. It is likely that few, if any, patients with stage III or IV UPSC and/or clear cell cancer will undergo WART as the sole adjuvant therapy at most institutions in the future.

The benefit of WART is particularly unclear for stage I-II tumors evidenced by the following:

• In a review of 193 women with stage I-II UPSC from nine published studies, abdominal failures occurred in 6 of 68 patients (9 percent) treated with WART versus 10 of 125 patients (8 percent) treated without WART. The use of pelvic and/or vaginal irradiation was beneficial in terms of pelvic control, however.

• In the only prospective study evaluating WART in women with all stages of UPSC (GOG 94), there was only a 35 percent five-year PFS rate for 31 women with surgical stage I UPSC  These data have only been reported in abstract form, although the results in patients with stage III and IV UPSC have been published

There is a stronger rationale to support the use of adjuvant pelvic RT for patients with stage I or II disease. In one review, pelvic recurrences occurred in substantially fewer women who received pelvic RT compared to those who did not (11 versus 27 percent, respectively)

The utility of pelvic RT has been challenged in patients with stage I to II disease who undergo complete surgical staging. At least in one report of 60 women with stage I UPSC after complete surgical staging which included complete lymph node dissection, the recurrence rate was similar among the 20 patients who received adjuvant therapy (12 of whom had pelvic RT) and the 40 who did not (16 versus 17 percent). In particular, pelvic sidewall failures were not observed despite the lack of adjuvant pelvic RT

While an argument can be made for withholding adjuvant pelvic RT for patients with stage I to II disease who undergo complete surgical staging, no degree of lymph node dissection addresses the vaginal cuff. Thus, even complete staging does not obviate the need for vaginal brachytherapy in these patients. In a review of surgical stage I UPSC patients (of whom 48 percent received adjuvant chemotherapy), vaginal recurrences were noted in none of 43 patients who received vaginal brachytherapy versus 6 of 31 patients (19 percent) who did not  These results support the use of vaginal brachytherapy in patients with early stage UPSC who are completely surgically staged.

Even more limited data are available regarding the role of adjuvant RT in patients with clear cell tumors. Unfortunately, most investigators simply group these tumors together with UPSC; as a result, the benefit of adjuvant RT for clear cell tumors has not been adequately defined. The largest study to date focusing on adjuvant RT reviewed outcomes of 38 clear cell patients treated with primary surgery. Pelvic recurrences were seen in 0 of 22 treated patients versus 8 of 16 patients treated without adjuvant RT. Corresponding pelvic recurrence rates for stage I-II patients with and without adjuvant RT were 0 of 16 versus 5 of 6. Of note, although no women received WART, only one (2 percent) failed in the upper abdomen. These results suggest that the optimal adjuvant RT approach for these women is pelvic RT. It remains unclear whether vaginal brachytherapy is sufficient for patients with surgically completely staged clear cell cancer.

Summary — Although popular, there is little evidence to support the utility of WART in patients with UPSC and/or clear cell carcinoma of the uterus, particularly those with early stage disease. In contrast, adjuvant pelvic RT appears to reduce the risk of locoregional failure following surgery.

Vaginal brachytherapy is effective in reducing vaginal vault recurrences and is associated with minimal toxicity. However, it remains unclear if vaginal brachytherapy alone is sufficient therapy for women with completely surgically staged I or II UPSC or clear cell cancer.

ADJUVANT CHEMOTHERAPY — The benefit of adjuvant chemotherapy is also unclear in women with UPSC or clear cell histology; randomized trials limited to this subset have not been conducted. However, as with other endometrial cancers, there is evidence that chemotherapy is emerging as an important component of successful treatment

• In the previously mentioned series of 60 surgically staged patients with stage I UPSC, 7 of 40 women who did not receive adjuvant therapy recurred (17 percent). Despite the lack of RT, none of the disease recurrences involved the pelvic sidewall, and six of the seven involved distant metastases. None of eight patients who received adjuvant chemotherapy failed or died of their disease. Similar results have been noted in another similar series that inclued 68 women with surgically staged I or II UPSC

• As noted above, the superiority of chemotherapy over WART was shown in GOG 122, which randomly assigned women with stage III or IV endometrial cancer debulked to 2 cm of disease to WART with a pelvic boost or eight cycles of chemotherapy with cisplatin and doxorubicin. Twenty percent of the patients enrolled in GOG 122 had UPSC, and response to treatment was the same for both endometrioid and UPSC tumors

• A benefit for adjuvant chemotherapy in patients with stage I UPSC was further suggested in a retrospective review of 74 women with comprehensively surgically staged I UPSC (33 stage IA, 29 stage IB, 12 stage IC) treated at a single institution using a variety of adjuvant strategies over a 16-year period Women with stage IA disease and no residual tumor in the surgical specimen (ie, that the biopsy removed all tumor) had no recurrences regardless of whether they received adjuvant therapy or not. For patients with stage IA disease who had residual tumor in the surgical specimen, none of the seven who received adjuvant cisplatin-based chemotherapy and concomitant vaginal cuff irradiation recurred compared to six of 14 who received no adjuvant treatment. Similarly, none of the 12 women with stage IB disease who received chemotherapy plus RT recurred compared to 10 of the 17 who did not; the corresponding values for stage IC disease were one of five versus four of seven.

Overall, of the 21 patients who received adjuvant chemotherapy, only one recurred (3 percent). In contrast, of the 43 patients who did not receive adjuvant chemotherapy, 20 (47 percent) recurred, 13 of whom had received irradiation.

None of the patients who received vaginal brachytherapy recurred at the vaginal apex, compared to 6 of 31 (19 percent) who did not receive vaginal brachytherapy. These data confirm the findings of others that omission of vaginal brachytherapy is associated with a high rate of vaginal apex recurrence. Given the low morbidity of vaginal brachytherapy, this should be offered to all women with UPSC.

Taken together, these data support the view that chemotherapy is of benefit for all patients with resected UPSC, except perhaps those who have no residual disease remaining after the initial biopsy.

The choice of regimen is unclear. Because UPSC spreads like epithelial ovarian cancer (EOC), there have been many attempts to treat this disease with the same regimens that are effective in EOC. In studies from GOG, response rates of UPSC to doxorubicin/cisplatin, doxorubicin/paclitaxel and doxorubicin/paclitaxel cisplatin were between 37 and 50 percent. However, taxanes appear to be highly active against UPSC as well For this reason, either the combination of paclitaxel and carboplatin or paclitaxel/doxorubicin and cisplatin are recommended for patients who require adjuvant chemotherapy. Randomized trials directly comparing these regimens are now underway, many within the GOG.

Summary — Adjuvant chemotherapy appears to be of benefit for all patients with resected UPSC, except perhaps those who have no residual disease remaining after the initial biopsy. The optimal adjuvant strategy for women with UPSC or clear cell histology remains uncertain. There is a compelling need for prospective studies to establish optimal therapy for such women.

CHEMORADIOTHERAPY — Combined chemotherapy and RT may be more effective than RT alone in the adjuvant setting, as evidenced by the following:

• A pilot study involving 23 women with UPSC noted a higher overall survival in patients treated with RT, paclitaxel and carboplatin combination therapy compared to those treated with RT alone (80 versus 30 percent)

• In a second report of 26 patients treated with sequential cisplatin-based chemotherapy followed by RT for stage I to IV UPSC (13 stage I or II), 14 remained alive and disease-free with a median follow-up of 28 months; 12 of whom had stage I or II disease

Further studies of this approach are warranted.

SUMMARY AND RECOMMENDATIONS — Uterine papillary serous (UPSC) and clear cell cancer differ from the more common endometrioid cancer in a number of ways, including epidemiology, precursor lesions, behavior, and prognostic factors.

Recurrence patterns (particularly for UPSC) are more similar to that of epithelial ovarian carcinoma than they are to endometrial cancer. As such, surgical therapy tends to parallel that for ovarian cancer, and serum CA-125 levels provide a good marker of disease activity. As with ovarian cancer, the strongest predictor of overall survival is the amount of residual disease following surgery.

Treatment recommendations by stage of disease are as follows:

Well staged stage IA disease — For all women with well-staged Ia UPSC, we suggest vaginal brachytherapy. We suggest chemotherapy for women with a good performance status if they have evidence of residual disease remaining after the initial biopsy

Others advocate no additional treatment for women with disease limited to a resected polyp, or those with no residual disease following hysterectomy and a comprehensive staging evaluation

Guidelines from the National Comprehensive Cancer Network (NCCN) suggest pelvic RT with or without vaginal brachytherapy or whole abdominal radiation therapy (WART) with or without vaginal brachytherapy in patients with stage IA disease

Stage IB, IC, and II disease — We suggest postoperative pelvic RT for women with resected stage IB, IC, and II UPSC or clear cell endometrial cancers who are able to tolerate this approach

We also suggest adjuvant chemotherapy for all of these women. In the absence of metastatic disease, we suggest paclitaxel and carboplatin (although paclitaxel, doxorubicin, and cisplatin is an acceptable alternative.

Guidelines from the NCCN suggest one of three approaches: pelvic RT with or without vaginal brachytherapy, WART with or without vaginal brachytherapy, or chemotherapy plus vaginal brachytherapy

Stage III and IV — We recommend chemotherapy when UPSC has spread beyond the uterus . We suggest paclitaxel, doxorubicin, and cisplatin, although paclitaxel and carboplatin is an acceptable alternative

We suggest the use of tumor-directed radiation therapy, plus vaginal brachytherapy, if women are able to tolerate it  We do not suggest the use of concomitant chemoradiotherapy in this setting

Guidelines from the NCCN recommend WART with or without vaginal brachytherapy or chemotherapy in patients with stage III or IV UPSC or clear cell cancer

Incompletely staged patients — Patients who are incompletely surgically staged because of their body habitus or other factors should be presumed to have stage III disease, and treated accordingly (see above).