Purpose: The purpose of this study was to construct a simple and powerful prognostic index (PI) of epithelial ovarian cancer, the PIEPOC.
Patients and methods: In a retrospective review, data from 768 women with stage III or IV epithelial ovarian cancer from 24 institutions in Japan were evaluated for clinical features predictive of overall survival. A PI and risk groups to predict overall survival after initial surgery were developed using the proportional hazards regression model.
Results: Of six factors, the four prognostic factors that remained independently significant in the analysis of a training sample (538 randomly selected patients) were age, performance status (PS), histologic cell type, and residual tumor size. From the regression function, we derived a PI = 1 (if age 70 and above) + 1 (if PS 1 or 2) + 2 (if PS 3 or 4) + 1 (if mucinous or clear-cell) + 2 (if residual size 0.1 cm and above). Patients were classified into three risk groups (PIEPOC): low risk (PI 0-2), intermediate risk (PI 3), and high risk (PI 4-6). The PIEPOC was equally predictive in a validation sample (n = 230), identifying three groups (5-year survival: 0.67 in low, 0.43 in intermediate, 0.17 in high risk).
Conclusion: Our proposed PI, the PIEPOC, was predictive in our patient population and may have utility in clinical practice. Prospective studies would be needed to confirm the prognostic predictive ability of the PIEPOC for patients with advanced epithelial ovarian cancer.