High-Dose Chemotherapy and Autologous Stem-Cell Transplantation for Ovarian Cancer: An Autologous Blood and Marrow Transplant Registry Report

Advanced-stage epithelial ovarian cancer is often chemosensitive but is not usually curable with conventional-dose therapies (1). New paclitaxelplatinum chemotherapy combinations have improved expected survival of patients with advanced disease. However, these regimens do not increase rates of pathologic complete remission and seem unlikely to cure substantially more patients than standard cyclophosphamide regimens (2). The prognosis of women with advanced ovarian cancer remains poor. Although clinical trials of double-dose chemotherapy in advanced ovarian cancer do not indicate increased cure rates, considerable in vitro and in vivo data support the importance of dose intensity in treating ovarian cancer, including trials of high-dose cisplatin delivered intraperitoneally (3-7). Ovarian cancer shares many features with hematologic malignant conditions, which benefit from dose-intensive therapy. Such features include almost universal initial chemosensitivity and occasional cures with conventional therapy despite large tumor bulk (1, 2). A favorable doseresponse curve is also seen for various agents, and many drug combinations exhibit synergy in vitro (3-6). Consequently, there is considerable interest in using high-dose therapy with autologous hematopoietic stem-cell support (autotransplantation) for ovarian cancer. Although conventional salvage therapy for ovarian cancer produces responses in 15% to 25% of women with platinum-sensitive disease (8-12), early trials of autotransplantation report response rates as high as 75% for drug-resistant tumors (13-15). However, duration of response is short (on average, 6 to 7 months) in patients who receive transplants for refractory disease. Recent studies of transplantation focus on patients with relapsed chemosensitive disease. Stiff and colleagues (16) reported a median progression-free survival of 19 months and a median survival of 29 months in 20 women with relapsed platinum-sensitive tumors and minimal tumor burden (maximal diameter<1 cm). These results are particularly encouraging because these women had received a median of two previous conventional chemotherapy regimens. Past studies of conventional therapy in this setting indicate median survivals of only 16 to 20 months (8-12). In a recent study, Legros and coworkers (17) report a 5-year survival rate of 60% in 53 patients who received autotransplants after completing induction therapy for advanced ovarian cancer. Because 5-year survival rates of 20% to 30% are common in unselected groups of women with advanced-stage disease who were receiving standard platinum-based chemotherapy (1, 2), the study by Legros and colleagues suggests that autotransplantation may be beneficial during first remission. Although early studies of autotransplantation for advanced ovarian cancer are encouraging, most reported series are small, especially those that focus on patients with minimal previous therapy. No randomized trials have been reported, and although several are under way, accrual is slow. To better assess the effectiveness of high-dose chemotherapy and autotransplantation, we analyzed 421 women who were reported to the Autologous Blood and Marrow Transplant Registry (ABMTR) from 1989 to 1996. We determined progression-free survival and survival for the group as a whole and identified prognostic factors related to outcome of transplantation. Methods Autologous Blood and Marrow Transplant Registry The ABMTR is a voluntary organization of more than 200 institutions, located primarily in the United States, Canada, and Central and South America, that perform autotransplantation. Participating centers report data on consecutive autotransplantations to a statistical center at the Medical College of Wisconsin, Milwaukee. The ABMTR defines autotransplantation as treatment with a dose of chemotherapy that is high enough to warrant autologous bone marrowderived or blood-derived hematopoietic stem-cell support. The treating institution decides whether a particular regimen requires such support. The ABMTR began collecting data in 1992. Data were collected retrospectively for persons receiving autotransplants between 1989 and 1992 and prospectively thereafter. Participating centers register consecutive autotransplantations for all diseases. According to data collected in the Centers for Disease Control and Prevention National Hospital Discharge Surveys (18, 19), approximately 50% of autotransplantations in North America are registered with the ABMTR. Participating centers are required to register all consecutive transplantations, and compliance is monitored by on-site audits. Computerized error checks, physician review of submitted data, and on-site audits of participating centers ensure data quality. The ABMTR research program, which includes collection and analysis of transplantation data, has been approved by the institutional review board of the Medical College of Wisconsin. The ABMTR collects data at two levels: registration and research. Registration data include disease type, age, sex, pretransplantation disease stage and responsiveness to chemotherapy, date of diagnosis, graft type (bone marrowderived or blood-derived stem cells), high-dose conditioning regimen, post-transplantation disease progression and survival, development of a new malignant condition, and cause of death. Information on progression or death is requested every 6 months. All ABMTR centers contribute registration data. Transplantation centers that have the required data management support collect research data for consecutive registered patients on comprehensive report forms. Submitted information encompasses pre- and post-transplantation clinical information, such as tumor size and pathologic characteristics, sites of disease, all treatments for ovarian cancer, bulk of disease at transplantation, clinical status (including cardiac, pulmonary, renal, and liver function), doses of high-dose therapy, blood or marrow graft treatment, and sites of post-transplantation progression. Demographic characteristics and survival rates in women with ovarian cancer are similar in the registration and research databases. Patients Registration data were collected on all patients with ovarian cancer who were treated in participating centers between 1 January 1989 and 31 December 1996. Fifty-seven North American transplantation centers submitted comprehensive research data for 421 of all 513 eligible registered patients. According to our analysis, the study group of 421 patients and the 92 patients for whom only registration data were available did not differ significantly in demographic features and survival rates. We focused on the 421 women for whom comprehensive data were available, including 100 patients from Loyola University in Maywood, Illinois, who were previously described in a less extensive analysis (20). Median follow-up of survivors was 29 months (range, 3 to 82 months). Statistical Methods To compare patient and treatment characteristics, we used the chi-square test for categorical variables and the KruskalWallis test for continuous variables. Probabilities of 100-day mortality (death in the first 100 days from any cause), progression-free survival, and overall survival were calculated by using the KaplanMeier product-limit estimate. Treatment-related death (death without progression) was calculated by using the cumulative incidence estimate (21). A pointwise test was used for univariate comparisons. Variables were tested in univariate analyses for their association with death and treatment failure (the inverse of progression-free survival) by using Cox proportional-hazards regression; these analyses also examined whether the proportional hazards assumption was met for each covariate, using the time-dependent covariate method. Variables considered were age (above or below the median), Karnofsky performance status (<90% vs. 90%), disease stage and histologic characteristics at diagnosis, response to initial therapy, number of previous chemotherapy regimens, pretransplantation remission state, pretransplantation bulk of disease, chemosensitivity, platinum sensitivity, interval between diagnosis and transplantation, high-dose regimens, graft type (bone marrow or blood stem cells), and year of transplantation. For the final multivariate model, forward stepwise selection was used to select variables; a P value less than 0.05 was considered statistically significant. First-order interactions of variables in the final model were also checked. Using a time-dependent covariate approach, we checked all variables to ensure that the assumption of the Cox model was valid. A score test (22) was used to determine whether any center-specific effects required adjustment. Chemosensitivity was defined as partial response (>50% decrease in the size of measurable disease) or complete response to the last chemotherapy regimen given in the 6 months before transplantation. Platinum sensitivity was defined as a partial or greater response to the last platinum-based chemotherapy regimen given in the 6 months before transplantation or an unmaintained remission after the completion of platinum-based therapy 6 or more months before transplantation. Patients with optimal debulking were defined as those who had tumors less than 1 cm in maximal diameter after initial surgery. Progression of disease after transplantation was defined as clinical or radiographic evidence of new or progressive disease or a CA 125 value greater than 50 IU/mL and increasing. Role of the Funding Sources The funding sources had no role in the collection, analysis, or interpretation of the data or in the decision to submit the paper for publication. Results Patient characteristics are shown in Tables 1 and 2. More than 80% of women had stage III or IV disease at diagnosis, which is somewhat higher than the percentage observed in unselected surveys of ovarian cancer. However, histologic characteristics, grade, type