Scrape cytology for intraoperative evaluation of lymph nodes in gynecologic cancer

To compare scrape cytology with frozen section in accuracy and time required for preparation and interpretation and assess accuracy and surgical effect of scrape cytology as the only method of intraoperative evaluation of lymph nodes from women with gynecologic cancers.

Between February 1998 and March 1999, 242 lymph nodes removed from 60 women with gynecologic cancers were submitted for scrape cytology alone (34 women, 177 lymph nodes) or with frozen section (26 women, 65 lymph nodes). Results of scrape cytology and time required for preparation and interpretation were compared with those from frozen section. Accuracy and proportion of incorrect surgical decisions based on combined evaluation by scrape cytology and frozen section were compared with those from scrape cytology alone. For statistical analysis, we used Student t test, chi2, and Cohen's kappa statistics. Two-tailed P < .05 was considered significant.

There was excellent agreement between scrape cytology and frozen section (kappa = .78 with 95% confidence interval .495, 1.065). Scrape cytology had a higher accuracy than frozen section and required significantly less time (95.8% versus 87.5% and 14 +/- 3.2 versus 25 +/- 5.1 minutes, P < .005). The accuracy and proportion of incorrect surgical decisions based on combined evaluation were similar to those based on cytology alone (91.7% for both and 7.7% versus 8.8%, respectively).

Scrape cytology is as accurate and faster than frozen section in intraoperative evaluation of lymph nodes from women with gynecologic cancers. Adding frozen section to scrape cytology did not seem to improve accuracy or reduce incorrect surgical decisions.