To identify prognostic factors for untreated ovarian cancer, DNA ploidy, proliferative index (P.I.) and epidermal growth factor receptor (EGFR) expression were analyzed in a prospective series of 40 patients with ovarian cancer and 7 patients with borderline malignant ovarian tumor followed up for 5 years or more (median, 77 months). The frequency of aneuploid cells was 53.8% (21/39) in ovarian cancer and 14.3% (1/7) in borderline malignancy. There was no significant association between DNA ploidy and the clinicopathologic findings, in which aneuploid ovarian cancer was more common among advanced tumors. The S-phase fraction and P.I. value were higher in the patients with aneuploid tumors (p = 0.076). EGFR expression was detected in 76.9% (30/39) of ovarian cancers and 42.9% (3/7) of borderline malignant ovarian tumors, and the mean EGFR level was 5.8 +/- 12.1 (range: 0-49.5) and 28.3 +/- 71.1 (range: 0-189.4) fmol/mg protein, respectively. There was no correlation between EGFR expression and DNA ploidy, P.I., and clinicopathologic findings analyzed. The 5-year survival rate in patients with aneuploid tumors was significantly worse in patients with ovarian cancer (p = 0.0165, log-rank test). No significant relationship was shown between P.I., EGFR expression, and 5-year survival. Cox multivariate analysis showed that DNA ploidy, P.I., and EGFR expression are not associated with the risk of death (p = 0.5917, p = 0.9924, and p = 0.6840, respectively), although clinical stage shows a significant relationship (p = 0.0027). Our data showed that DNA ploidy is significantly related to the prognosis by univariate analysis, but DNA ploidy, P.I., and EGFR expression were not independent prognostic factors for the untreated ovarian cancer.