Intraoperative frozen section diagnosis of depth of myometrial invasion in endometrial adenocarcinoma.

From January 1981 to December 1988, a prospective trial was conducted on 216 consecutive patients undergoing hysterectomy for FIGO stage I endometrial adenocarcinoma to determine the accuracy of intraoperative frozen section (IFS) diagnosis of depth of myometrial invasion. IFS and D&C diagnoses were compared to the permanent section diagnosis to determine their accuracy. Patients with an IFS or D&C diagnosis of grade 3 carcinoma, deep myometrial invasion, cervical invasion, or adnexal involvement were considered high risk for paraaortic nodal metastasis, and paraaortic lymphadenectomy was performed. IFS correctly diagnosed the depth of myometrial invasion in 194 of 204 cases (95%). The sensitivity of IFS diagnosis of deep invasion was 87%, the specificity was 99%, the positive predictive value was 98%, and the negative predictive value was 94%. Grade, subtype, cervical invasion, and adnexal involvement were also accurately diagnosed. Based on IFS and D&C diagnosis, paraaortic lymphadenectomy was performed in 32% of patients, while 68% were considered low risk and were spared paraaortic lymphadenectomy. Eight patients (5%) were incorrectly diagnosed as low risk and did not have paraaortic lymphadenectomy performed. Sixty-five of seventy-three (90%) patients considered high risk on permanent section had paraaortic lymphadenectomies performed on the basis of IFS and D&C diagnosis. Inaccurate IFS diagnosis of depth of myometrial invasion can occur when tumor involves the uterine isthmus or cornua and when tumor invades areas of adenomyosis.