Interobserver discrepancy using the 1998 World Health Organization/International Society of Urologic Pathology classification of urothelial neoplasms: practical choices for patient care.

Abstract

Morphological classifications designed by experts to stratify neoplasms according to biological potential must define categories that are reproducible among practitioners or the schemes actually create the heterogeneous populations that they seek to avoid. The application of the 1998 World Health Organization/International Society of Urologic Pathology scheme for urothelial neoplasms was studied in a community practice setting. We documented interpretive discrepancies for each category of neoplasm and determined whether a period of pathologist education may have a positive effect on the frequency of discrepant interpretations. The results suggest that patients may benefit from modifying the classification system. A consecutive series of specimens was divided into learning and study sets that were each independently examined by 3 pathologists. Specimens in the learning set were interpreted without previous structured education, while those in the study set were interpreted immediately after intensive education. Interpretations for each specimen were compared and interpretive discrepancies were analyzed. Case distribution after education was similar among the pathologists but interpretations for any particular specimen often differed. The level of interpretive discrepancies varied according to the morphological similarity among categories in the classification scheme and was not necessarily decreased by education. When pathologists were required to discriminate between papillary urothelial neoplasm of low malignant potential and low grade carcinoma, the discrepancies were 50% after education compared with 39% before education. In contrast, there were no discrepancies when the discrimination was between papillary urothelial neoplasm of low malignant potential and high grade carcinoma or carcinoma in situ. Eliminating categories with poor reproducibility markedly improved the likelihood of unanimous agreement among practitioners but a probably irreducible level of 10% discrepancies remained. The 1998 World Health Organization/International Society of Urologic Pathology classification of urothelial neoplasms requires certain discriminations that cannot be reliably made by practitioners. Modifying the scheme to create categories of low grade neoplasm and high grade carcinoma would markedly increase its practical value to patients without significantly altering patient care.