In order to confront the question whether localized and diffuse ovarian cancers are the extremes of the same disease or two dif­ferent entities, we compared anamnestic, intraoperative and laboratory findings, as well as the outcome. Sixty-four patients with invasive epithelial ovarian cancer were treated in our department between 1984-1991. After excluding patients with incomplete data, 59 women were finally included in the analysis. The patients were divided retrospectively into two groups according to the extent of the disease observed during the primary surgery. Disease confined to the ovaries was defined as localized form, and any other extraovarian involvement was defined as diffuse form. Twenty patients had localized (Group I), and 39 had diffuse disease (Group II). Comparing the two groups, no difference was observed concerning patient age, the age of menarche and menopause, gravity and parity, the percentage of smoking and hypertensive women, and the histologic type of the ovarian tumor. On the other hand, significant differences were obtained concerning the positive family history of cancer in localized versus diffused disease (10% versus 41 %) (p < 0.04), the size of the tumor involving the ovaries (16.2 ± 5.4 cm versus 7.0 ± 3.1 cm) (p < 0.0001), the unilateral as opposed to bilateral disease (p < 0.000 1), the volume of ascitic fluid (p < 0.002) and the histologic grade (p < 0.02). Laboratory parameters, including CA-125, serum albumin, total protein and lactate dehydrogenase differed significantly comparing the localized and dif­fused disease (p < 0.04, p < 0.0001, p < 0.003, p < 0.05, respectively). The median survival rate was significantly higher (p < 0.001) in patients with localized versus diffused disease. Our results support the concept that localized and diffused forms of ovarian cancer are probably two different entities.