Prospective evaluation of a molecular marker panel for prediction of recurrence and cancer-specific survival after radical cystectomy

Background Retrospective studies demonstrated that cell cycle-related and proliferation biomarkers add information to standard pathologic tumor features after radical cystectomy (RC). There are no prospective studies validating the clinical utility of markers in bladder cancer. Objective To prospectively determine whether a panel of biomarkers could identify patients with urothelial carcinoma of the bladder (UCB) who were likely to experience disease recurrence or mortality. Design, setting, and participants Between January 2007 and January 2012, every patient with high-grade bladder cancer, including 216 patients treated with RC and lymphadenectomy, underwent immunohistochemical staining for tumor protein p53 (Tp53); cyclin-dependent kinase inhibitor 1A (p21, Cip1) (CDKN1A); cyclin-dependent kinase inhibitor 1B (p27, Kip1); antigen identified by monoclonal antibody Ki-67 (MKI67); and cyclin E1. Intervention Every patient underwent RC and lymphadenectomy, and marker staining. Outcome measurements and statistical analysis Cox regression analyses tested the ability of the number of altered biomarkers to predict recurrence or cancer-specific mortality (CSM). Results and limitations Pathologic stage among the study population was pT0 (5%), pT1 (35%), pT2 (19%), pT3 (29%), and pT4 (13%); lymphovascular invasion (LVI) was seen in 34%. The median number of removed lymph nodes was 23, and 60 patients had lymph node involvement (LNI). Median follow-up was 20 mo. Expression of p53, p21, p27, cyclin E1, and Ki-67 were altered in 54%, 26%, 46%, 15%, and 75% patients, respectively. In univariable analyses, pT stage, LNI, LVI, perioperative chemotherapy (CTx), margin status, and number of altered biomarkers predicted disease recurrence. In a multivariable model adjusting for pathologic stage, margins, LNI, and adjuvant CTx, only LVI and number of altered biomarkers were independent predictors of recurrence and CSM. The concordance index of a baseline model predicting CSM (including pathologic stage, margins, LVI, LNI, and adjuvant CTx) was 80% and improved to 83% with addition of the number of altered markers. Conclusions Molecular markers improve the prediction of recurrence and CSM after RC. They may identify patients who might benefit from additional treatments and closer surveillance after cystectomy.