Endometrial cancer, which is the most common malignancy of the female genital tract, is mostly curable by surgery and/or radiotherapy. 1 However, of the approximately 40% of patients with either recurrent or metastatic disease, only about one third respond to conventional single-agent therapy with progestins. 2 Progestins have been used in the management of advanced endometrial carcinoma for more than 4 decades with some success, but their exact mechanism of action is unknown. 2,3 It is most probable that as with other hormones, they exert their effect by means of a receptor. 3 Progesterone receptors are known to be present in human endometrial cancer, 3 and there may also be an approximate correlation between the histologic grade of the endometrial cancer and the level of progesterone receptors and clinical response to progestin therapy. 4
Tamoxifen is an antiestrogen now widely used for the treatment of hormone-sensitive breast cancer. Estrogen receptors are known to be present in human endometrial cancer, 3 and tamoxifen has been shown to inhibit the binding of estradiol to uterine estrogen receptors. 5 A short-term administration of tamoxifen can cause a significant increase in the progesterone receptor levels in postmenopausal women with endometrial cancer. 6 Both of the receptors are present in about 80% of well-differentiated endometrial carcinoma, 7 although the proportion may decrease with loss of differentiation. 8
Thus, a trial was initiated to explore whether the combination of an antiestrogen and a progestin, namely, tamoxifen and megestrol, was additive or synergistic in improving the management of advanced endometrial carcinoma.
The primary goal of the study was to assess the effect of adding tamoxifen to standard progestin therapy of megestrol in terms of response and survival in the treatment of metastatic or inoperable endometrial carcinoma. Other objectives were to evaluate the toxicity of the combination hormonal therapy and to investigate the relationship of tumor histopathology to response and prognosis.
MATERIALS AND METHODS
Selection and Registration
Eligibility was limited to patients with histologically confirmed diagnosis of recurrent or metastatic endometrial carcinoma that was no longer responsive to surgery or radiotherapy, but that had not been exposed to prior hormonal therapy. The disease must have been measurable, meaning that abdominal, pelvic, or other masses must have been defined in at least two dimensions by palpation or by computed tomography scan, but not by sonography; ascites and pleural effusion alone were not considered measurable. Eligible patients also had to have Eastern Cooperative Oncology Group performance status of 0, 1, or 2 as well as adequate bone marrow, renal, and hepatic functions, unless abnormalities were caused by metastatic disease. Results of estrogen and progesterone receptor studies were not known on any of these patients, because they were not routinely obtained at the time this study was performed.
Patients were randomized at the Central Randomization Desk at the Eastern Cooperative Oncology Group Operations Office. When the megestrol arm was closed to patient entry, subsequent patients were directly assigned to receive megestrol and tamoxifen. Written informed consent was required of all patients before treatment assignment, and the study was approved by the institutional review boards of the participating institutions.
The schedule of treatment A consisted of 80 mg megestrol taken orally twice a day, whereas the schedule of treatment B consisted of 80 mg megestrol and 10 mg of tamoxifen also taken orally twice a day. After a minimum duration of 2 months of therapy, treatment was discontinued in the event of disease progression or relapse. Tamoxifen was to be discontinued if severe thrombocytopenia (platelets ≤ 50,000/mm3) developed or if hypercalcemia (serum calcium >11 mg/dl) persisted for 1 week.
The distribution of patient characteristics between the two treatment arms was examined by Fisher exact test for balance. Fisher exact test was also used to analyze contingency tables of response. Survival curves were computed by Kaplan-Meier estimation 9 and were assessed by the log-rank test 10 for differences. Stepwise analyses to identify simultaneously influential prognostic variables were based on the likelihood ratio test, using logistic regression 11 for response analysis and proportional hazards regression 12 for survival analysis.
A total of 66 patients were entered, of whom 4 (6%) were ineligible: 1 had inadequate renal and hepatic functions, 1 had no measurable disease, and 2 had inadequate pretreatment evaluation. As summarized in Table 1, the distribution of on-study patient characteristics shows no serious imbalance between the two treatment arms. Thus, homogeneity of the patient population is a reasonable assumption despite uneven accrual between the two treatment arms because of early termination of the megestrol arm. The ages on the megestrol arm ranged between 56 and 81 years with a median age of 68 years, whereas the ages on the megestrol and tamoxifen arm ranged from 52 to 82 years with a median of 65 years. All cases were of postmenopausal status either naturally, surgically, or because of irradiation. Time since menopause ranged from 1 to 50 years on the megestrol arm and from 1 to 37 years on the megestrol and tamoxifen arm, with medians of 12 and 15 years, respectively.
Overall, mild and moderate toxicities were experienced by 40% of the patients receiving megestrol, and 57% of the patients receiving the combination, of which hematologic toxicities were the most frequent. In addition, severe to life-threatening toxicities were seen in 5% of the patients on the combination arm, including one case of pulmonary embolism.
Measurement of treatment effect was determined according to Eastern Cooperative Oncology Group criteria for tumor response 13 and is summarized in Table 2. The response rates yield 95% confidence intervals of (6%, 44%) and (9%, 34%), respectively. The probability of a false negative finding on the megestrol arm, that is, of observing a response rate of 20% when the true rate is 30% or greater, is less than 0.238. Likewise, the probability of observing a response rate of 19% on the megestrol/tamoxifen arm when the true rate is 30% or greater is less than 0.080. There was no impact of any of the pretreatment patient characteristics on the response to treatment, including the primary tumor cell differentiation.
Because the protocol was revised to be only a phase II study of megestrol and tamoxifen when the megestrol arm was eliminated because of poor accrual, minimal power is available for detecting statistically significant treatment differences.
Figure 1 shows the survival curves by treatment. The survival times on the megestrol arm ranged from 11 days to 60.6 months, with a median of 12.0 months, whereas the survival times on the megestrol and tamoxifen arm ranged from 48 days to 46.6 months, with a median of 8.6 months. The median survival time of ambulatory patients is 11.3 months compared to a median time of 3.5 months among the nonambulatory patients. Weight loss in the previous 6 months (p = 0.056), prior chemotherapy (p = 0.093), and well to moderately differentiated primary tumor (p = 0.083) are also marginally related to survival. However, initial performance status (p = 0.018) is the only covariate that retains significance in the stepwise proportional hazards regression of survival. The difference in survival between ambulatory and nonambulatory patients is illustrated in Figure 2.
Both megestrol and the combination of megestrol and tamoxifen demonstrated modest activity against endometrial cancer with response rates of 20% and 19%, respectively. The overall survival was slightly shorter for the combination, with a median time of 12.0 months on the megestrol arm and a median time of 8.6 months on the megestrol and tamoxifen arm. Although the small sample size per arm did not suffice to provide enough power for a comparative evaluation, the addition of tamoxifen to megestrol does not appear to offer clinical advantage in the hormonal therapy of advanced endometrial cancer. Megestrol only induced mild and moderate toxicities, whereas more toxic complications were experienced on the megestrol and tamoxifen arm, including a life-threatening case of pulmonary embolism. Among the on-study patient characteristics, only initial performance status retains significance in the stepwise proportional hazards regression of survival. Nonambulatory performance status is significantly associated with decreased survival; the median survival time of nonambulatory patients is 3.5 months compared to the median survival time of 11.3 months among ambulatory patients.
Our results with megestrol are similar to those reported by Kohern 14 with response observed in 10% of poorly differentiated endometrial carcinoma. A number of phase II studies of tamoxifen, mostly given to patients after failure of progesterone treatment, are reported in the literature 15–18 with no activity 16,17 to 20%18 and 30% response rates. 15 A sequential administration of tamoxifen for 5 days followed by medroxyprogesterone for 20 days given to 20 patients with metastatic poorly differentiated endometrial carcinoma produced an objective complete response in one patient. 19 Thus, adding tamoxifen to progestins either sequentially or given concomitantly as in our study appears to yield no therapeutic advantage.
1. Glassburn JR. Carcinoma of the endometrium. Cancer 1981; 48: 575–81.
2. Davis TE. Hormonal therapy of endometrial cancer
. In: Rose DP, ed. Endocrinology of cancer. Vol 1. Boca Raton, Florida:CRC Press,
3. Gurpide E. Hormone receptors in endometrial cancer
. Cancer 1981; 48: 638–41.
4. Martin PM, Rolland PM, Gammerre M, et al. Estradiol and progesterone receptors in normal and neoplastic endometrium: correlation between receptors, histopathologic examinations and clinical response under progestin therapy. Int J Cancer 1979; 23: 23321–29.
5. Lunan CB, Green B. 3-H-Estradiol uptake in vivo by human uterine endometrium: effect of tamoxifen
. Clin Endocrinol 1974; 3: 465–80.
6. Mortel R, Levy C, Wolff FP, et al. Female sex steroid receptors in postmenopausal endometrial carcinoma and biochemical response to antiestrogen. Cancer Res 1981; 41: 1140–7.
7. McCarty KS Jr, Barton TK, Fetter BF, et al. Correlation of estrogen and progesterone receptors with histologic differentiation in endometrial adenocarcinoma. Am J Pathol 1979; 96: 171–83.
8. Greasman WT, McCarty KS Sr, Barton TK, et al. Clinical correlates of estrogen and progesterone binding proteins in human endometrial adenocarcinoma. Obstet Gynecol 1980; 55: 363–70.
9. Kaplan EL, Meier P. Nonparametric estimation of incomplete observations. J Am Stat Assoc 1958; 53: 457–81.
10. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–70.
11. Cox DR. Analysis of binary data. London: Chapman and Hall, 1970.
12. Cox DR. Regression models and life tables. J R Stat Soc Series B 1987; 4: 187–202.
13. Oken MM, Creech RM, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982; 5: 649–55.
14. Kohern EI. Gestogens and endometrial carcinoma. Gynecol Oncol 1986; 4: 398–411.
15. Hold I, Salimtschick M, Mouridsen HT. Tamoxifen treatment
of advanced endometrial carcinoma. Eur J Gynaecol Oncol 1983; 4: 83–87.
16. Slavik M, Petty WM, Blessing JA, et al. Phase II clinical study of tamoxifen
in advanced endometrial adenocarcinoma: a Gynecologic Oncology Group study. Cancer Treat Rep 1984; 68: 809–11.
17. Edmonson JH, Krook JE, Hilton JF, et al. Ineffectiveness of tamoxifen
in advanced endometrial carcinoma after failure of progestin treatment
. Cancer Treat Rep 1986; 70: 1019–20.
18. Quinn MA, Campbell JJ. Tamoxifen
therapy in advanced/recurrent endometrial carcinoma. Gynecol Oncol 1989; 32: 1–3.
19. Kline RC, Freedman RS, Jones LA, et al. Treatment
of recurrent metastatic poorly differentiated adenocarcinoma of the endometrium with tamoxifen
and medroxyprogesterone acetate. Cancer Treat Rep 1987; 71: 327–8.