Histologic grading of urothelial papillary neoplasms: impact of combined grading (two-numbered grading system) on reproducibility

Abstract

The clinical management of tumor patients is often strongly infuenced by the tumor grade. The presence of heterogeneity is well recognized in a variety of tumors. Overall grade is based on highest grade area identified within a tumor. Urothelial carcinoma often contains different histological grades within the same tumor. This study investigates the impact of a combined grading system on the reproducibility of papillary urothelial neoplasms. A set prepared for an earlier study consisting of ten cases of each category (papillary urothelial neoplasm of low malignant potential (PUNLMP), LGPUC, and HGPUC) was used. Agreement between pairs of pathologists was evaluated using kappa statistics for the combined scoring system. Interobserver agreement was fair to substantial as reflected by kappa values ranging from 0.24 to 0.74 (mean kappa = 0.43). The combined scores of 2 and 3 which included PUNLMP showed the lowest degree of agreement and when this category was excluded from the analysis, interobserver agreement increased significantly (mean kappa = 0.65; ranging from 0.43 to 0.92) in terms of combined scores of 4, 5, and 6. PUNLMP has been shown to be the least reproducible component of a combined scoring system even among experienced observers. Exclusion of PUNLMP from grading scheme seems to improve interobserver variability.