Although the microscopic grade of epithelial ovarian carcinoma is recognized as an important prognostic factor, the grading systems are still controversial. A cytohistological grading system is described using the sum of the scores (from 1 to 4) of eight histological and cytological features. This grading is applied to the initial surgical specimens of 37 patients seen at UCLA Center for the Health Sciences; it is also used for evaluation of pathologic specimens obtained at subsequent laparotomies. The evolution of the score at subsequent laparotomies (stable, increasing, or decreasing) is correlated with the type of therapy, response to treatment, and prognosis. An increasing score at the second-look laparotomy is a poor prognostic factor with a significantly shorter patient survival (p less than 0.05). The tumor score at second-look laparotomy is significantly increased in cases of macroscopically positive operation (p less than 0.05) and tumor progression (p less than 0.001). The type of therapy has no significant effect on the tumor score. A multivariate analysis shows that a decreasing score is related to a lower mitotic activity. Conversely, an increasing score is related to a change in tumor pattern. This study demonstrates the prognostic value of our grading system, which needs to be further validated by prospective studies.