Mitotic activity in noninvasive papillary urothelial carcinoma: its value in predicting tumor recurrence and comparison with the contemporary 2-tier grading system

Abstract

Measures of mitotic activity predict behavior of noninvasive papillary urothelial carcinoma of the urinary bladder, but it is unclear what role these should have in tumor grading. In this article, we compare measures of mitotic activity to contemporary tumor grading, specifically in their association with recurrence of noninvasive papillary urothelial carcinoma. The study uses a retrospective cohort of 199 tumors from 124 patients. Mitotic activity was treated as a categorical variable (mitotic-inert, mitotic-low, or mitotic-high). Evaluating only first-occurrence tumors, recurrence was more frequent in mitotic-high (hazard ratio [HR], 8.8; P < .0001, Cox model) and mitotic-low tumors (HR, 3.7; P = .017) compared with mitotic-inert tumors, when controlling for treatment with intravesical bacillus Calmette-Guerin, age, and sex. Recurrence was likewise more frequent in high-grade tumors (HR, 3.1; P = .00019, Cox model) compared with low-grade tumors, controlling for these factors. However, mitotic group, but not tumor grade, was significantly associated with recurrence in a multivariate Cox model including mitotic group, tumor grade, and treatment status (HR, 6.5 [P = .0025] for mitotic-high versus reference; HR, 3.7 [P = .018] for mitotic-low versus reference). Frailty models including both first-occurrence and recurrent tumors showed similar results. Isolating the analysis to first occurrence, low-grade tumors, recurrence was more frequent in mitotic-high (HR, 6.8; P = .0044, Cox model) and mitotic-low (HR, 3.4; P = .027) tumors compared with mitotic-inert tumors. The findings indicate that mitotic activity is associated with behavior of noninvasive papillary urothelial carcinoma and may be valuable as an adjunct to the contemporary grading system.