PERITONEAL WASHING CYTOLOGY IN GYNECOLOGIC CANCERS - LONG-TERM FOLLOW-UP OF 355 PATIENTS

Background: Microscopic evaluation of cells washed from the peritoneal cavity during surgery for gynecologic tumors is used to detect subcli nical intraperitoneal metastases from these tumors. The prognostic sig nificance of this test, however, has been questioned. Purpose: Stressi ng histologic correlation and pitfalls in interpretation, we previousl y reported that the sensitivity of intraoperative peritoneal washing c ytology was lower than was suggested earlier. This study evaluates the clinical utility of this test in the long-term follow-up of our patie nts. Methods: Staging (International Federation of Gynecology and Obst etrics [FIGO], 1971) and follow-up information was available for 355 u nselected patients with primary tumors who had peritoneal washings per formed during initial surgery at University Hospital-Stony Brook, NY, during the period from 1980 through 1989. There were 135 patients with endometrial carcinomas, 112 with ovarian carcinomas, 99 with cervical carcinomas, and 16 with borderline (i.e., of low malignant potential) ovarian tumors. The median follow-up of the patients was 57 months (r ange, 0.154 months). Follow-up data were obtained from the Tumor Regis try at University Hospital-Stony Brook. Survival differences were dete rmined by Kaplan-Meier analysis and were evaluated by two-tailed logra nk test. Results: Peritoneal washing cytology was positive at initial surgery for 120 (33.8%) of 355 patients, including 90 (80.4%) of 112 p atients with ovarian carcinomas, five (31.2%) of 16 patients with bord erline ovarian tumors, 17 (12.6%) of 135 patients with endometrial car cinomas, and eight (8.7%) of 92 patients with cervical cancers. For 20 3 patients with stage I tumors, the peritoneal cytology was positive i n 29.4% of the patients with ovarian carcinomas, 18.2% with borderline ovarian tumors, 6.1% with endometrial carcinomas, and 5.2% with cervi cal carcinomas. By use of peritoneal histology as the standard, perito neal cytology was highly specific (98.1%) but less sensitive (82.9%) i n detecting intraperitoneal involvement. For patients with stage I tum ors, 80.0% with ovarian carcinomas, 83.3% with endometrial carcinomas, and 100% with cervical carcinomas who showed positive cytology died o f their cancer, compared with 25.0% with ovarian carcinomas, 13.0% wit h endometrial carcinomas, and 21.9% with cervical carcinomas who showe d peritoneal cytology. Four patients with stage I tumors had positive peritoneal cytology but negative peritoneal histology. Of these patien ts, three (two with ovarian carcinoma and one with cervical carcinoma) died of their cancer, whereas one patient with a borderline ovarian t umor was free of disease at the last follow-up. Survival analysis indi cated that peritoneal washing cytology stratified for stage provides b etter prognostic information for each primary cancer site studied than does stage alone. All patients with borderline ovarian tumors were al ive at last follow-up, regardless of disease stage or peritoneal statu s. Conclusions: Regardless of FIGO stage, positive peritoneal washing cytology predicted poor prognosis for women with epithelial tumors of the genital tract, except for patients with borderline ovarian tumors. Patients in whom peritoneal cytology was the only evidence of intrape ritoneal spread were few, but the disease in such patients was associa ted with poor outcome. Implications: Strict adherence to specialized c ytologic criteria in peritoneal washing cytology allows for results th at are highly predictive of survival. This information may be useful i n stratifying women in therapeutic trials for treatment of genital tra ct carcinomas.