Ovarian serous borderline neoplasm with noninvasive implants traditionally have been considered to be nonaggressive tumors associated with an excellent prognosis. However, in our experience, recurrences commonly develop as patients are followed over many years. Eighty cases of advanced-stage ovarian serous borderline tumor with noninvasive implants were identified; the minimum follow-up period for these cases was 5 years or until the death of the patient. The following cases were excluded: patients treated by cystectomy, patients who died of other causes, patients who developed other tumors, and patients who had as the only positive material after resection of the primary borderline neoplasm a tumor detected on a second look or third look operation. Hematoxylin and eosin-stained slides from the original ovarian tumor and the staging biopsies were reviewed in all cases. Slides of the recurrent tumor were available in all cases except for 2 in which the diagnosis was established clinically. The presence or absence of a micropapillary/cribriform pattern and microinvasion in the ovarian tumor was recorded. Follow-up was obtained from the patients' charts. Fischer exact test was used for statistical analysis. The patients' ages ranged from 17 to 67 years (median 36 y). Seventy-three patients were treated by total abdominal hysterectomy and bilateral salpingo-oophorectomy. Seven patients were treated by total abdominal hysterectomy and unilateral salpingo-oophorectomy. The International Federation of Gynecology and Obstetrics stage was as follows: stage II (29 cases), stage III (50 cases), and stage IV (1 case). After surgery, 58 patients were treated with chemotherapy, 7 with radiotherapy, and 1 with hormonal therapy. The follow-up ranged from 5 to 31 years (median 15.7 y). Thirty-five patients (44%) developed recurrences. Only 10% of the patients had a recurrence in less than 5 years, 19% had their recurrences between 5 and 10 years, 10% between 10 and 15 years, and 5% more than 15 years after resection of the primary neoplasm. The only statistically significant feature associated with recurrence was the presence of a micropapillary/cribriform pattern, although this pattern was present in only 26% of the cases that recurred. Of the 35 patients who had a recurrence, 2 were diagnosed clinically, both are alive with progressive disease at 1 and 5 years after the diagnosis of the recurrence; 6 had recurrent serous borderline tumors, all are without evidence of disease with a follow-up ranging from 7 to 18 years after resection of the ovarian borderline tumor (median 14 y); and 27 patients subsequently developed low-grade serous carcinoma, 7 are alive with progressive disease with a follow-up ranging from 10 to 29 years (median 15 y) and 20 died of disease between 3 to 25 years after resection of the ovarian borderline tumor (median 16 y). In summary, the true recurrence rate of ovarian serous borderline tumors with noninvasive implants can only be obtained through a long follow-up. In this group of patients, 77% and 34% of the subsequent tumors developed 5 years and 10 years after diagnosis of the ovarian tumor, respectively. Histologic examination of the recurrent tumor is important in determining further therapy and prognosis for these patients; all patients who recurred with borderline tumor are without evidence of disease, whereas 74% of the patients who recurred with low-grade serous carcinoma died of disease. We propose that patients be followed for a minimum of 10 years to evaluate for recurrences and for 20 years to evaluate for survival.