The natural history of serous borderline tumors (SBTs) of the ovary varies considerably. A group of investigators have proposed that a small subset of SBTs with a micropapillary architecture and an allegedly higher incidence of invasive peritoneal implants should be designated "micropapillary serous carcinomas." Based on the overall favorable prognosis of the nonmicropapillary SBTs, these investigators have recommended abandoning the borderline category of serous tumors, restricting them to benign (benign and typical SBTs) and malignant types; other investigators, however, are in favor of retaining the original grouping, designating borderline tumors with a micropapillary pattern as such instead of designating them carcinomas. We have reviewed the clinicopathologic records of 137 patients with ovarian SBTs and obtained follow-up information on 106 of them ranging from 1 to 18 years (mean 7 years). Of the 21 patients with stage I tumors who had conservative surgical treatment, only two (9.5%) were subsequently found to have tumor in the contralateral ovary. Both were successfully managed by reoperation alone. Forty-five stage I patients had procedures that included bilateral oophorectomy, and two of them (4.4%) had a pelvic recurrence, which was fatal in one patient (whose tumor had been understaged) and occurred on multiple occasions in the other patient, finally transforming into invasive carcinoma; that patient survived. Of the 45 stage II-IV patients, only the six (13%) with invasive implants had an unfavorable outcome: three died of tumor (from 7 to 9.3 years), and the other three were alive with progressive disease from 5 to 10 years. Solid epithelial nests or small papillae surrounded by clefts and micropapillary architecture were found more often in invasive than in noninvasive implants. However, the only feature specifically associated with a poor outcome was obvious destructive invasion of the underlying tissue. Among the 137 SBTs, we identified 18 cases of serous borderline tumors with a micropapillary pattern (SBT-MP) (so-called "micropapillary carcinoma") and 20 cases of SBT with microinvasion (SBT-Minv) (three of which were also micropapillary). We compared the two groups of tumors with the remaining 102 cases of typical SBTs (which lacked micropapillary pattern and microinvasion). Of the 17 patients with SBT-MP and follow-up data, only the one patient with invasive implants had an unfavorable outcome; similarly, of the two patients with SBT-Minv and an unfavorable outcome, one had invasive implants and the other had been incompletely staged. SBTs have a very favorable prognosis, but complete surgical staging and prolonged follow-up are advised because pelvic recurrence and occasionally transformation to invasive carcinoma may occur. Designation of SBTs as "atypical proliferative tumors" is not recommended because it discourages complete surgical staging and follow-up. Advanced stage tumors with noninvasive implants are common, characteristically behave in a benign fashion, and can be safely treated conservatively. The rare SBTs associated with invasive implants are almost always fatal. SBT-MP and SBT-Minv are much closer in their biologic behavior to SBTs than to serous carcinomas. The micropapillary pattern alone does not imply an unfavorable prognosis; only micropapillary tumors associated with invasive implants behave aggressively.