Survival Determinants in Patients with Advanced Ovarian Cancer

S. M. Ansell, B. L. Rapoport, G. Falkson, J. I. Raats, C. M. Moeken

Research output: Contribution to journalArticlepeer-review

29 Scopus citations


An analysis was performed on 127 consecutive women with advanced measurable ovarian cancer to evaluate factors predicting for survival. All patients received cis-platinum-based chemotherapy as treatment for stage IIIB to stage IV disease. Eighteen clinical, radiological, and biochemical parameters were subjected to univariate and multivariate analyses. Recursive partitioning and amalgamation (RPA) was used to define prognostic subsets with different survival potentials. In the univariate analysis, factors that predicted for survival were weight loss, histology, stage, number of metastases, presence of ascites, size of the residual tumor, and rate of tumor response. When these significant variables were included in a Cox model, advanced stage of disease, histology other than adenoserous carcinoma, the presence of tumor bulk, and a slow rate of tumor response independently predicted a poorer survival. Using the three disease-related prognostic variables, a RPA model was derived and three groups were identified with median survival times of 76, 28, and 21 months, respectively (P = 0.001). The best survival time of 76 months was seen in patients with stage III, non-bulky, adenoserous ovarian carcinoma. It is concluded that the rate of tumor response is important in predicting the outcome of patients with ovarian cancer. Furthermore, the interactions between prognostic factors are emphasized by the RPA model and a subgroup of patients with a projected 10-year survival of 50% is identified.

Original languageEnglish (US)
Pages (from-to)215-220
Number of pages6
JournalGynecologic oncology
Issue number2
StatePublished - Aug 1993

ASJC Scopus subject areas

  • Oncology
  • Obstetrics and Gynecology


Dive into the research topics of 'Survival Determinants in Patients with Advanced Ovarian Cancer'. Together they form a unique fingerprint.

Cite this