Efficiency of bladder wash cytology during surveillance of non-muscle invasive bladder cancer: Analysis in a contemporary patient cohort
Bladder wash cytology (BWC) is widely used as an adjunct investigation for non-muscle invasive (NMI) bladder cancer (BC) surveillance. Traditionally, a high specificity has been reported. However, more recent data revealed a less optimistic performance of BWC particularly for low-grade BC. We aimed to assess the performance of BWC in daily clinical practice in a contemporary patient cohort followed for NMIBC.
We retrospectively analyzed 2064 BWC&[prime]s derived from 315 patients followed for NMIBC in a tertiary care academic center between 2003 and 2013. All patients were followed using a combination of cystoscopy and BWC. Patients with either positive cystoscopy or BWC underwent bladder biopsy. Patients with both negative cystoscopy and cytology were followed. All cytologies were performed by a specialized pathologist. BWC was considered positive if malignant cells were reported. Sensitivity (SE), and specificity (SP) were calculated overall, for low-grade (LG) and high-grade (HG) tumors. The same analysis was performed with an extended positive cytology (defined as presence of either suspicious, atypical or malignant cells). The influence of postoperative BCG treatment on BWC performance was also assessed (Breslow-Day test).
Patients were followed for pTa (206 patients (66%)), pT1 (99 patients (31%)) or pure CIS (10 patients (3%)) BC. LG and HG disease was found in 159 (50%) and 156 patients (50%), respectively. Mean number of BWC performed per patient was 5.6 (± 4.5). A total of 95 recurrences were detected. Overall SE and SP were 0.16 and 0.99, respectively. For LG disease, the calculated numbers were 0.13 and 1, for HG disease 0.2 and 0.99. If extended positive cytology was used the calculated SE and SP were 0.54 and 0.94 overall, 0.35 and 0.95 for LG and 0.72 and 0.93 for HG tumors. BWC performance was influenced by postoperative BCG treatment only if extended positive cytology was used (p=0.003).
Our analysis revealed large differences in SE depending on the applied criteria for a positive test. Extended criteria increase the SE, while keeping the SP almost constant. Overall, the benefit of BWC remains questionable, especially in the surveillance of NMIBC.