Systematic restaging was performed prospectively in 100 patients referred to the Ovarian Cancer Study Group—comprised of investigators from the following: Mayo Clinic, Rochester, Minnesota; M. D. Anderson Hospital, Houston, Texas; Roswell Park, Buffalo, New York; and the Medicine Branch of the National Cancer Institute—with a diagnosis of “early” (stage la-lib) ovarian cancer. Of the 100 patients referred with a diagnosis of early ovarian cancer, 68 patients were reexamined at the referring institution. The 32 patients who did not have the second procedure were not operated on because they were upstaged by other procedures (9 patients), were not medically suitable for the second surgery (4 patients), or refused the second surgical procedure (19 patients). The 68 patients who had an initial surgical procedure before referral and subsequently had detailed staging and a comprehensive surgical reexamination form the group used to document the frequency of undetected residual disease. Sixty-one (90 per cent) of these women were said to be completely free of disease after their initial surgery. The staging work-up at the four institutions included a complete history, physical examination, and pelvic examination. Routine blood chemistries included 12-factor automated chemical analysis, complete blood cell count, and platelet count. An IVP and lymphangiogram were routinely performed on patients unless treatment delay precluded these procedures. Proctoscopy and cystoscopy were performed when clinically indicated. Peritoneoscopy with abdominal washings for cytological study was performed at some institutions and used to upstage some patients. However, all patients with normal findings from peritoneoscopy went on to a staging laparotomy unless there were medical contradictions or the patient refused. In addition, complete review of the material from the referring hospital, clinical and surgical notes, and the pathological material including sections of the primary ovarian tumor were obtained in all instances. If indicated, studies to complete the examination included pelvic and abdominal computed tomographic scan and ultrasound, gallium scan, barium enema examination, and upper gastrointestinal tract series. One hundred women were referred to member institutions with a diagnosis of stage la-lib epithelial ovarian cancer after having completed initial clinical and surgical staging. Ninety-two of these women were believed to be free of disease by their referring physician. However, the clinical staging that had been performed before referral was limited. Whereas a majority of patients had a pelvic examination (96 per cent) and a chest roentgenogram (83 per cent), a minority of patients had an IVP (42 per cent), barium enema examination (34 per cent), and abdominal roentgenogram (24 per cent) or upper Gl tract series (12 per cent). More aggressive staging procedures were rarely employed in these 100 women. The surgical incision used in the original staging examination was adequate to examine the pelvis in most instances but not extensive enough to allow thorough inspection of the abdominal contents. Fifty-eight per cent of patients were staged through a midline subumbilical incision and 17 per cent had an even more limited exploration. Only 25 per cent of the patients had a median or paramedian incision that extended above the umbilicus in a manner that would allow the surgeon full access to the abdominal cavity. Despite the limited surgical access, the vast majority of patients (92 per cent) were said to be free of disease at referral. Sixty-nine per cent of the patients had a total abdominal hysterectomy and bilateral salpingo-oophorectomy with or without omentectomy. However, the other 31 per cent of the patients had lesser resections but only 4 of these 31 patients were thought to have any residual disease despite more limited surgical resection. Four patients had no tumor resection. Three had only ascites sampled that was positive for adenocarcinoma, and only 1 was referred after visual inspection only. Of the 68 patients reexamined at the member institutions, 61 were thought by their referring physicians to be free of disease. Laparotomies in these patients showed a substantial amount of residual disease. Substantial changes in the final stage occurred as a result of the comprehensive restaging performed during this study. Changes in the final stage occurred in virtually every category but were more common in patients initially thought to be stage II (43 per cent) than in those initially thought to be stage I (28 per cent). A total of 31 per cent of patients were upstaged as a result, and of these, 23 (77 per cent) of 31 were advanced to stage III. Upstaging occurred in patients with each histological grade of tumor.