Interobserver reproducibility of a hybrid three-tier grading system of papillary nonmuscle invasive urothelial carcinoma: an international Uropathology study.

Abstract

Aims: A hybrid-three tier system with low grade (LG), high grade- G2 (HG-G2), high grade- G3 (HG-G3) has been proposed in recognition of, and to help address, the clinical heterogeneity within high grade WHO 2004/2022. We assessed interobserver reproducibility amongst international uropathologists using this three-tier approach.

Methods and results: Papillary Ta nonmuscle invasive bladder cancer (NMIBC) specimens (n = 30) were selected and graded by two uropathologists and assessed using WHO 2004/2022 and WHO 1973 and categorized as LG (n = 15), HG-G2 (n = 8), HG-G3 (n = 7), and photographed at 10× and 20× magnification. Images were circulated via Survey Monkey to invited uropathologists who determined: (1) that image was LG or HG, and (2) if HG, assigned to G2 or G3. Model-based kappa measure of association was used to assess interrater agreement. Eighteen uropathologists:(eight North American, eight European, two other) assessed 60 images with 1076 gradings for analysis. The kappa value amongst Europeans versus North Americans was 0.663 versus 0.647 for 10× images and 0.682 versus 0.623 for 20× images. At 10×, agreement for LG, HG-G2, and HG-G3 was 74.6%, 63.6%, and 92.0%, and at 20× was 64.3%, 63.9%, and 95.2% respectively.

Conclusion: Three-tier grading of papillary Ta NMIBC had substantial interobserver agreement amongst international uropathologists. The recognition of the HG-G3 case reached the highest concordance. North American uropathologists had comparable kappa scores (substantial agreement) to Europeans, despite being unaccustomed to separating HG cases into G2 and G3, demonstrating three-tier grading could be "quickly" adopted by genitourinary experts if endorsed and required by the relevant bodies in their jurisdiction of practice.