Prophylactic oophorectomy is chosen by some women at high risk of ovarian cancer due to inherited predisposition. Unfortunately, this surgery is not 100% effective in preventing intra-abdominal carcinomatosis that histologically resembles ovarian cancer. To determine the incidence of post-oophorectomy carcinomatosis and to quantify the effectiveness of preventive surgery, a multicenter study is ongoing between the National Cancer Institute (NCI), Creighton University, and the United Kingdom. The prospective incidence of malignancy, especially of tissues derived from coelomic epithelium (primarily ovary, fallopian tube, and peritoneum), was compared between women of similar genetic risk who have or have not undergone oophorectomy. Analysis of 12 NCI families has been completed. Prospective observation ran from the date of family ascertainment until the date of cancer incidence, death, or December 31, 1992. Approximately 1600 person-years of observation occurred among 346 first-degree relatives of a breast or ovarian cancer case patient for women who had not undergone oophorectomy. Eight ovarian cancers occurred, compared with two carcinomatosis cases during 460 person-years of observation among 44 oophorectomized women. Compared with Connecticut Tumor Registry data adjusted for age, race, and birth cohort, there was an approximately 24-fold excess of ovarian cancer among non-oophorectomized women and a 13-fold excess of "ovarian" cancer among oophorectomized women, though this difference was not statistically significant. The confidence intervals around these numbers were large, and a collaborative analysis will be required to determine whether this apparent protective effect is real.