Sensitivity of the intraoperative assessment of myometrial invasion in patients undergoing hysterectomy for endometrial cancer

Objective: The objective of this study was to determine the ability of the surgeon to distinguish between deep and superficial myometrial invasion on gross inspection as compared to frozen section. Methods: All patients undergoing hysterectomy for endometrial cancer were eligible for this prospective study carried out at a single institution. After removal of the uterus, the specimen was bi-valved and the depth of myometrial invasion was determined grossly by the attending surgeon (gross). The depth of myometrial invasion by visible tumor was measured in millimeters and then described as “superficial” or “deep” if the invasion did not or did extend into the outer half of the myometrial thickness, respectively. The specimen was sent for frozen section evaluation of tumor grade and depth of invasion (frozen). We calculated the sensitivity of each method, gross and frozen, to correctly identify deep myometrial invasion, the specificity to correctly identify superficial or no invasion, and the accuracy, defined as the number of true positives plus true negatives divided by the total number of cases. The final pathologic assessment of myometrial invasion (final) was considered the gold standard for comparison. Results: Between 4/4/2011 and 4/9/2014, a total of 220 patients underwent hysterectomy by the division of gynecologic oncology in the San Diego area of Kaiser Permanente for a preoperative diagnosis of endometrial cancer. Of these, 196 had complete specimen information (gross, frozen, and final). On final, 35 specimens had deep invasion (17.9%). The sensitivity of gross to detect deep invasion was 54.3% and 77.1% for frozen. Specificity for gross and frozen was 93.8% and 98.8%, respectively. Gross inspection failed to identify deep myometrial invasion in 16 (45.7%) of the cases, frozen failed to identify deep myometrial invasion in 8 (22.9%) of the cases, and both gross and frozen failed to identify deep myometrial invasion in 7 (20.0%) of the cases. Conclusions: Frozen section demonstrates greater accuracy and sensitivity in the intraoperative determination of deep myometrial invasion. This should be considered if intraoperative assessment is used to determine staging or treatment decisions.