Can Bladder Cancer Biomarkers from Patients Undergoing Cystectomy Predict the Need for Adjuvant Radiotherapy?

Abstract

There is a high rate of recurrence of muscle-invasive bladder cancer (MIBC) after cystectomy with local recurrence rates of about 40% and up to 50% distant recurrence. Currently, the inability to predict recurrence patterns limits the use of adjuvant radiotherapy. The ability to use biomarkers to predict local vs distant recurrence could individualize adjuvant therapy. This study evaluates the ability of PD-L1, MRE11 and Ki67 labeling index in radical cystectomy specimens to predict recurrence (no recurrence (NR), local-only recurrence (LOR), and distant recurrence (DR)) in those with MIBC. Our institutional radical cystectomy (RC) database, which includes 2405 cases of urothelial MIBC from 1/1992-12/2014, was queried for LOR. This yielded 119 patients. Of these, pathologic confirmation of disease was available on 20 patients. These were case-matched to 20 NR patients and 20 DR patients. Those without tissue for biomarker staining or adequate follow-up data were excluded. Representative tumor specimens were selected based on H&E slides. Normal bladder tissue adjacent to areas of tumor from RC specimens was also evaluated. A tissue microarray (TMA) was built from formalin fixed, paraffin embedded tissue and analyzed for presence of MRE11, PD-L1 and Ki67 antibodies via immunofluorescence and immunohistochemistry (PD-L1 and Ki67 only). We evaluated pathologic specimens of 42 patients (18 NR, 16 LOR, and 8 DR). The median age of the patients was 69 years. 81% had pT3/pT4 disease, 36% were pN0, and 14% had positive margins. Median follow-up was 147 months. Compared to normal bladder tissue, urothelial cancer cells had higher PDL1 H score, Ki67 H score, % Ki67 on IHC, MRE11 positive cells, MRE11 positive cell density, PD-L1 positive cells, and PD-L1 positive cell density. On univariate analysis, positive surgical margins were more likely to have PD-L1 CPS score >10 and a lower mean Ki67 H score, Ki67 percentage on IHC, PD-L1 positive cells, %PD-L1 positive cells and PD-L1 cell density. Patients with DR had a significantly lower Ki67 mean positive cells (IF) than those with LOR (p=0.04). PD-L1 H score, PD-L1 positive cells and % positive PD-L1 cells were significantly correlated with RFS (p= 0.008, 0.018, 0.015, respectively). On multivariable analysis (MVA), none of the three biomarkers were correlated with LOR/DR, RFS, or OS. PD-L1, MRE11, and Ki67 could discriminate cancer cells from noncancer cells but were unable to predict types of recurrence or clinical cancer outcomes on MVA. Further studies with larger sample size are required to examine the relationship between these biomarkers and bladder cancer recurrence.