Interobserver Variability in the Interpretation of Tumor Cell Necrosis in Uterine Leiomyosarcoma

Abstract

On the basis of the most recent World Health Organization classification, distinction of leiomyosarcoma (LMS) from leiomyoma is based on the presence of the following morphologic criteria: (1) nuclear atypia; (2) mitotic index; and (3) tumor cell necrosis (TCN). Unlike ischemic-type necrosis, which may be seen in benign and malignant smooth muscle tumors (SMTs), TCN is thought to be found only in LMS. The distinction between these 2 types of necrosis can be challenging, especially during the early stages, when necrotic foci are small, or when overlapping features are identified. The aim of this study is to assess the interobserver variability in the interpretation of TCN in uterine LMS. Thirty-four LMS cases were retrieved, and a representative hematoxylin and eosin slide showing 1 area of necrosis was selected from each case. Pathologists from 6 different institutions subspecializing in gynecologic pathology performed a blinded, independent review of the slides. Using the current World Health Organization criteria for assessment of TCN, they had to classify the necrotic foci into: (1) TCN; (2) no TCN; or (3) indeterminate for TCN. Agreement among panelists was categorized as: full-all pathologists in agreement; partial-4 or 5 pathologists in agreement; no agreement-≤3 pathologists placing the case into the same category. Full agreement regarding the presence or absence of TCN was reached in 12 cases (35%) (7 thought to show TCN); partial agreement in 16 (47%); and no general consensus was obtained in 6 (18%). Overall, the level of agreement was moderate (κ=0.436). In 8 of 34 instances (23.5%), ≥1 pathologist made a diagnosis of "TCN" and ≥1 pathologist made the diagnosis of "no TCN" for the same slide. The number of cases diagnosed as "indeterminate for TCN" by each pathologist ranged from 0 to 10 with a mean of 5.8. In 20 cases, at least 1 pathologist diagnosed "indeterminate for TCN" (59%), at least 2 and 3 were undecided in 10 (29%) and 4 (12%) cases, respectively, and 4 pathologists diagnosed "indeterminate for TCN" in 1 instance. When excluding foci of necrosis diagnosed as "indeterminate" by any pathologist, disagreement occurred in 2/14 (14%) cases. From these results we conclude that the level of interobserver agreement among expert gynecologic pathologists in the assessment of TCN in uterine SMTs is only moderate. These results further reiterate the importance of assessing for both nuclear atypia and mitotic activity when differentiating between benign and malignant SMTs and not relying solely on the presence of TCN.