Agreement between transurethral resection of bladder tumor and radical cystectomy pathology in patients with bladder cancer subtype histology: A retrospective cohort study.

Introduction and objectives: Transurethral Resection of Bladder Tumor (TURBT) is a diagnostic staging procedure for bladder cancer (BC). Its pathologic interpretation may be limited by cautery artifact, lack of spatial orientation of tumor specimens, inter-pathologist variance in identifying subtypes, and sampling bias. Accurately identifying subtype histology (SH) on TURBT is critical for clinical decisions. We compared the agreement between TURBT and radical cystectomy (RC) pathology in patients with SH BC.

Methods: We examined TURBT and RC pathology of patients who underwent RC at our institution. We included patients with pure SH and mixed histologies in either TURBT or RC specimens. Cohen's kappa coefficient was used to determine the degree of agreement between TURBT and RC.

Results: From 1135 RC performed, 650 (57%) patients had SH in either TURBT or RC; 225 patients were (y)pT0 at the time of RC and 36 patients had rare histologies, leaving 389 patients for analysis. 172 (44%) patients had an exact match between TURBT and RC. We found a high level of agreement between TURBT and RC in pure non-UC histology (kappa range: 0.82-0.98). In contrast, we found substantial (sarcomatoid; kappa 0.70), moderate (squamous, glandular, plasmacytoid, small cell/neuroendocrine; kappa range: 0.42-0.55) and fair (micropapillary; kappa 0.38) concordance between TURBT and RC in patients with UC mixed with SH.

Conclusions: We found variable levels of agreement of SH detection between TURBT and RC. Agreement was high in pure non-UC histology. Further, we found that NAT, completeness of TURBT, and >50% SH at TURBT are associated with the persistence of SH at RC. Future efforts are needed to develop reproducible diagnostic tools for accurate characterization of SH in UC.