Chemotherapy of Endometrial Cancer

The therapy of disseminated or recurrent carcinoma of the endometrium with cytotoxic chemotherapy is currently disappointing.^{1, 2} Most clinical studies of efficacy are confounded by the mixture of histologic types of endometrial carcinoma, including the particularly virulent serous papillary and clear cell carcinomas, that contaminate the sample population and by the frequent concomitant administration of hormonal therapy to which approximately 15% of patients may respond. In general, response rates to chemotherapy for endometrial carcinoma have been low and duration of response has been short.³ The problem of chemotherapy administration in this disease is important because endometrial cancer is common in an aging population, and the frequency of anaplastic endometrioid carcinoma that is advanced at presentation or that recurs after surgery and radiation is now much greater than it was 20 years ago.

Apart from these discouraging considerations there has been reluctance to use cytotoxic chemotherapy as primary treatment in patients who are frequently elderly, may have cardiovascular and cerebrovascular problems, and may be diabetic. With the realization that the response to progestins among poorly differentiated tumors is no better than to alkylating or cycle-specific agents has come a reevaluation of the use of these drugs. This chapter describes the available information on chemotherapy for endometrial cancer and tries to confine discussion to studies of endometrioid endometrial cancer, as far as this is possible from the published data. Serous and clear cell endometrial cancers and mixed müllerian tumors of the uterus present a completely distinct therapeutic challenge that merits separate discussion.

Donovan⁴ reviewed the literature in 1974 and found that 15 chemotherapy agents had been used in 126 patients, with an overall response rate of 27%. No data were cited on the differentiation of the tumors. This overall response rate was better than the response rate of progestins in poorly differentiated tumors.

More recently, two major reviews of the use of cytotoxic chemotherapy in endometrial carcinoma have been written, one by Moore and associates in 1991¹ and the other by Muss in 1994.² There appears to be some unanimity of practice in that the preferred treatment of disseminated or recurrent endometrial cancer not amenable to focal irradiation is the combination of cisplatin and doxorubicin (Adriamycin) with or without cyclophosphamide. This practice is likely to be modified by the finding that paclitaxel is significantly effective as a single agent in endometrial carcinoma.⁵

In the last 20 years there has been a more focused effort to identify active single agents in endometrial cancer. Most of these tumors have been high grade. For a particular chemotherapy drug to be regarded as effective, it should induce response in at least 15% of patients in the study.⁵

Anthracyclines

Doxorubicin has emerged as the most efficacious medication shown by data from three disease-specific trials^{6, 7, 8} (Table 1 ^{6, 7, 8, 9, 10}). Among 161 patients there was a 12% complete response rate and a 26% overall response rate. Epirubicin, an anthracycline analogue, also appears to be active,⁹ but so far it has not found acceptance as a component in combination chemotherapy regimens. Idarubicin¹¹ and mitoxantrone^{12, 13} appear to have no activity.

TABLE 1. Single-Agent Chemotherapy in Endometrial Cancer: Anthracyclines

DOX = Doxorubicin; Epir = epirubicin

Cisplatin

The second most effective single agent appears to be cisplatin,¹⁴ and possibly also carboplatin^{15, 16} (Table 2 ^{14, 15, 16, 17, 18, 19, 20}). The Gynecologic Oncology Group studies^{18, 19} found that cisplatin appeared to be more effective in patients not previously exposed to cytotoxic chemotherapy, and in fact Seski and colleagues²⁰ found that with a dose greater than 50 mg/m², the drug had to be discontinued in one third of the patients because of dose-related toxicity. There appears to be little value in exceeding this dose in patients with endometrial cancer. In patients who had no previous cytotoxic chemotherapy, the complete response rate for single-agent cisplatin chemotherapy was 6% with an overall response rate of 29%, compared to 4% and 29%, respectively, for carboplatin. Cisplatin remains the preferred agent because of its lack of hematologic toxicity.

TABLE 2. Single-Agent Chemotherapy in Endometrial Cancer: Cisplatin and Carboplatin

*Second-line chemotherapy
CBDCA = carboplatin; CDDP = cisplatin

5-Fluorouracil

5-Fluorouracil (5-FU) is a third agent that shows activity in endometrial cancer, with a response rate of 23% among 43 patients.²¹ More frequently, 5-FU has been used as a component of combination chemotherapy either with melphalan^{22, 23} or lately with leucovorin.²⁴

Cyclophosphamide

Cyclophosphamide is a drug with long-time use in metastatic or recurrent endometrial cancer, usually in combination regimens. It was initially administered with Adriamycin and then with added cisplatin when this medication became available. There is, however, evidence that as a single agent it is not effective. In 1976, DeVita and colleagues²⁵ found a 21% response rate of single-agent cyclophosphamide among 33 patients from five broad-based phase II studies. However, Horton and co-workers⁶ found no response in 19 patients using a dose of cyclophosphamide 600 mg/m². A phase III study by the Gynecologic Oncology Group⁸ failed to show a therapeutic advantage when cyclophosphamide was added to cisplatin and doxorubicin. The efficacy of cyclophosphamide in endometrial cancer, therefore, remains in doubt.

Paclitaxel (Taxol)

More recently there has been great interest in the use of paclitaxel in endometrial cancer. In a phase II study by the Gynecologic Oncology Group,²⁶ there were 4 complete and 6 partial responses in 28 patients given 250 mg/m² of Taxol over a 24-hour period administered every 3 weeks. The progression-free interval and overall survival, however, were not given. This 35% response rate for Taxol is better than that of other drugs used as single agents and is promising evidence to support its use in combination with cisplatin and perhaps doxorubicin also.

Other Single-Agent Chemotherapy

Ifosfamide is another drug that has been examined for use in endometrial cancer. Barton and associates²⁷ had 2 responders among 16 patients, and Sutton and colleagues²⁸ found an 8.6% response rate in 23 patients with previously untreated endometrial cancer. Hexamethymelamine (altretamine) has only limited usefulness, partly because toxicity precluded prolonged administration. Seski and co-workers²⁹ found a 30% response rate with a medial duration of response of 3.5 months, while a Gynecologic Oncology Group study³⁰ had 3 patients (7%) with an objective response among 34 patients. None of the other drugs examined so far has been recognized as useful in endometrioid endometrial cancer at this time. This includes etoposide³¹ and methotrexate.³²

Right from the time that Adriamycin and then cisplatin became available, they were given in combination in recurrent or initially metastatic endometrial carcinoma. However, there have been few prospective trials comparing single agents with combination chemotherapy or comparing one combination with another.

Melphalan and 5-Fluorouracil, With or Without a Progestin

Among the first combination chemotherapy regimens reported was melphalan with 5-FU and the progestin medroxyprogesterone. Cohen and associates³³ obtained an objective response in six of seven patients with advanced or recurrent disease previously treated with surgery and radiation. Five of the patients had grade 2 and 3 histology, and the duration of response ranged from 16 to more than 64 months, a truly remarkable result considering that the patients with the longest duration of response had grade 3 histology. Piver and colleagues³⁴ treated 50 analogous patients using the same regimen and found a 48% overall response rate and a 20% complete response rate. In that study, however, the median progression-free survival was only 5 months.

Adriamycin, Cyclophosphamide, and 5-Fluorouracil

The other combination chemotherapy initially reported on was Adriamycin, cyclophosphamide, and 5-FU together with medroxyprogesterone.³⁵ Unfortunately, the grade of tumor was not included. Only one patient had prior chemotherapy with melphalan and 5-FU. All seven patients showed a response, and this was sustained for more than 3 months in four patients. Perhaps some of the response was due to the progestin; this is a problem in assessing results in all patients with endometrial cancer given cytotoxic chemotherapy and hormonal therapy simultaneously.

The Mount Sinai group reported a further 29 patients treated with the same combination and obtained a 44% objective response rate with a median duration of survival of 11 months.³⁶ By 1984 a Gynecologic Oncology Group study reported by Cohen and co-workers³⁷ showed a 16% complete and a 22% partial response rate among 77 patients. A further 131 patients given melphalan, 5-FU, and medroxyprogesterone had a 17% complete and a 19% partial response rate. The grade of the tumor was not stated in the manuscript, nor was the duration of response, so in this study it is difficult to assess the effect of omitting the progestin.

Doxorubicin and Cyclophosphamide

In 1977 Muggia and associates showed a response in six of eight patients with recurrent, high-grade endometrial carcinoma treated with doxorubicin and cyclophosphamide (Table 3).^{8, 38, 39} They suggested that cytotoxic therapy may have a place in the primary treatment of advanced or recurrent high-grade endometrial cancer in which progestins are unlikely to be effective. Campora and colleagues⁴⁰ obtained a 46% response rate with the same medications with a median survival of 10 months. The most important study of the combination of doxorubicin with cyclophosphamide was performed by the Gynecologic Oncology Group,⁸ who compared doxorubicin alone with doxorubicin and cyclophosphamide. The objective response rate with combined cyclophosphamide-doxorubicin was 66%, compared to a 35% response rate for doxorubicin alone. The median progression-free survival was 6.2 months for the combination and 3.9 months for doxorubicin alone. The median survival for both groups was 9 months. These paradoxical data support the practice of giving combination therapy rather than single-agent chemotherapy. It would appear that a relatively inactive single agent such as cyclophosphamide may enhance the effectiveness of an active single agent when given in combination with it. Intellectually this is difficult to accept, and supportive in vitro studies would clearly be of interest. Nevertheless, the majority of investigators currently continue to use cyclophosphamide in combination with doxorubicin and cisplatin.

TABLE 3. Doxorubicin-Cyclophosphamide in Endometrial Cancer

CTX = cyclophosphamide; DOX = doxorubicin

Cisplatin and Doxorubicin

Tropé and co-workers⁴¹ reported on the use of doxorubicin and cisplatin in 19 patients with recurrent endometrial cancer and 1 patient with initial metastatic cancer, all treated primarily with this combination. The dose of each drug was 50 mg/m², and treatment cycles were every 4 weeks. There was an objective response rate of 60%, with two complete remissions and a median survival period for those with partial remission of greater than 11 months. Seltzer and associates⁴² gave the same doses in nine patients, but administered the drug every 3 weeks, and these patients had significant toxicity. There were two partial responses and one complete response. Deppe and colleagues⁴³ treated 19 patients with the same doses, also every 3 weeks, and had 7 patients with a partial response, an overall response rate of 36%, and a median survival of 36 months. There was no significant toxicity in this study. Pasmantier and co-workers⁴⁴ achieved a 92% response rate in 12 patients who had received no prior chemotherapy and a 50% response rate in 4 patients who had received prior chemotherapy. The median survival was 10 months. The drugs were administered every 28 days on an outpatient basis. These encouraging findings were substantiated by a prospective randomized study of the Gynecologic Oncology Group reported by Thigpen and associates.¹⁰ In 223 patients, there was a 35% response rate with single-agent doxorubicin and a 67% response rate with doxorubicin plus cisplatin. However, the progression-free survival was only 6 months and there was no survival advantage for the combination.

Cisplatin, Doxorubicin, and Cyclophosphamide

All other reported trials of doxorubicin and cisplatin also include cyclophosphamide (Table 4 ^{45, 46, 47, 48, 49}) and frequently a progestin, usually megestrol (CAP+M). Hoffman and colleagues⁵⁰ had 4 patients with a complete response and 1 with a partial response among 15 patients, giving an overall response rate of 33%. The median survival was 15 months for responders and 5 months for nonresponders, but the median progression-free survival for responders was only 7 1/2 months. All 15 patients had endometrioid endometrial carcinoma, 7 grade 3, and 6 grade 2. The dose of doxorubicin was 30 mg/m², cisplatin 50 mg/m², and cyclophosphamide 250 mg/m², escalated to 500 mg/m² in the absence of toxicity. Turbow and co-workers⁴⁷ had 9 responders (47%) among 19 patients treated with CAP. Two of these were complete responses, and seven were partial responses. Seventeen patients had adenocarcinoma and three had serous papillary tumors. Twelve of the patients had grade 3 tumors. These authors found better survival with good performance status and believed the cyclophosphamide added toxicity and no therapeutic benefit. Lovecchio and associates⁴⁹ evaluated CAP+M in 15 patients and found a 60% response rate with a tumor-free progression interval of 8 months and a mean survival of 12 months. Hancock and colleagues⁴⁶ used CAP in 18 patients with a 56% objective response rate; 5 (28%) of these patients had a complete response. All of these patients had hormone-resistant tumors. Neither of the last two reports discussed tumor grade or tumor type. Edmonson and co-workers⁴⁸ treated 16 patients with progestin-refractory endometrial cancer with CAP chemotherapy given every 4 weeks. Five patients (31%) had a partial response, and two patients survived for 2 years. This Mayo Clinic study concluded that more effective agents need to be found to make this therapy worthwhile. Burke and associates⁵¹ reported on a CAP study from the MD Anderson Cancer Center in which 42 patients with advanced and 60 patients with recurrent endometrial cancer were studied. The drugs were given every 28 days. Of 87 patients with measurable disease, 12 (14%) had a complete clinical response and 27 (31%) a partial response. Both groups responded similarly, but the median duration of response was only 4.8 months. It was concluded that enthusiasm for this therapy should be tempered by the brief duration of response. Fung and colleagues⁵² reported on 44 patients treated with CAP-medroxyprogesterone; 15 had measurable disease with a 53% response rate. The median survival for the whole group was 31 months. The volume of tumor was considered to be the factor most important in determining survival. Histologic grade did not predict response, although many patients had low-grade tumors. Dunton and co-workers⁵³ treated 25 patients with CAP chemotherapy every 3 weeks and found a 47% response rate among 17 evaluable patients. The group included serous papillary tumors and nine grade 3 endometrioid carcinomas. The 2-year disease-free survival was 20%.

TABLE 4. Cyclophosphamide, Doxorubicin, and Cisplatin, With or Without Megestrol, in Endometrial Cancer

CTX = cyclophosphamide; CDDP = cisplatin; DOX = doxorubicin; NS = not significant; q 4 = every 4 weeks.

In 1994 two reports appeared in the literature advocating the use of adjuvant chemotherapy with CAP as primary treatment in advanced endometrial carcinoma. Smith and associates⁵⁴ assessed the effect of adjuvant CAP in stage IC and II carcinoma of the endometrium with more than 60% myometrial invasion and in completely resected stage III and IV carcinomas before planned radiation therapy. Of the 47 women, 8 had grade 2 and 29 grade 3 histology; however, the series included 8 patients with serous papillary tumors, 2 with clear cell carcinomas, and 6 with adenosquamous histologies. The 2-year progression-free survival was 72.5% for nonserous-histology cancers and 22% for serous papillary cancers. The authors concluded that CAP chemotherapy followed by radiation appeared to offer survival advantage for patients with endometrioid histology when compared with historic controls.

Burke and colleagues⁵⁵ treated 62 patients with endometrial cancer at high risk for recurrence by virtue of (1) more than 50% myometrial invasion, (2) grade 3 histology, (3) completely resected extrauterine tumor, and (4) clear cell and serous histology. Recurrence occurred in 48% of 29 patients with extrauterine disease and in 24% of 33 patients with disease confined to the uterus. The median time of recurrence was 13 months. The authors admitted that CAP chemotherapy did not prevent distant failure but did appear to provide a survival advantage.

Although methotrexate as a single agent³² had not found favor in endometrial cancer, Long and co-workers⁵⁶ treated 30 patients with advanced or recurrent carcinoma of the endometrium with methotrexate, vinblastine, doxorubicin, and cisplatin. Regression of tumor was noted in 67% of patients, with a complete response in 8 patients (7%). The median overall survival was 10 months, and the median survival of responders was 11 months. There was significant toxicity, and there were two treatment-related deaths. The authors nevertheless stated in their discussion that this combination compared favorably with cisplatin and doxorubicin alone. In most investigators' hands, however, the two drugs doxorubicin and cisplatin have little severe toxicity, and their conclusion does not appear justified.

These data show that cytotoxic chemotherapy is effective in high-grade endometrioid endometrial cancer that is either recurrent or extensive at the time of diagnosis. It is interesting that the administration of prior pelvic radiation, which many of the patients presenting with recurrence have received, does not appear to affect the administration of the cytotoxic chemotherapy or prejudice response, which seems to be the same as in patients receiving primary chemotherapy for advanced disease. The response rates appear to exceed those achieved primarily with progestational agents, and this is likely to be most true for poorly differentiated tumors. Combination chemotherapy appears preferable to the use of single agents, and it is likely that cisplatin with doxorubicin and paclitaxel will be the preferred combination. If the use of Adriamycin or Taxol, is precluded by cardiac damage, then 5-FU and leucovorin are acceptable alternatives. The combination of 5-FU and cisplatin has not yet been explored.

Advocacy for the use of cytotoxic chemotherapy needs, however, to be tempered by the recognition that the duration of response is brief and may be so brief as not to justify the administration of these necessarily toxic drugs to an elderly patient population, who frequently already have poor quality of life and in whom the chemotherapy does not confer significant survival advantage. The use of granulocyte colony-stimulating factor is just beginning to be explored, but it is unlikely to change the response duration significantly. If the performance status of the patient is good and there is desire for therapy, then the administration of chemotherapy appears justified, provided the patient and her family recognize the limitations of treatment.^{56, 57} Performance status, rather than age, is the critical determinant. Besides considering the requirements of managed care and the financial aspects of significantly expensive treatment with a very brief duration of benefit, both the physician and the patient need to be sure that the disadvantages of such therapy do not outweigh the advantages of such a brief prolongation of life.