Development of a new outcome-prediction model for patients with urothelial carcinoma of the bladder prior to cystectomy.
360 Background: Prognostic models for patients undergoing cystectomy are spares and mostly rely on histopathological parameters. The present study investigated clinical prognostic indicators that are available prior to cystectomy for developing a risk stratification model. Methods: Charts were reviewed from 279 bladder cancer patients that were treated with radical cystectomy and bilateral lymphadenectomy. Of those 194 were eligible for study entry due to availability of necessary data. None of the patients had neoadjuvant therapy. Various preoperative parameters were analyzed and optimum cut-off points were identified using ROC curves. A risk stratification model was developed based on multivariate analysis. Analyzed factors included serum-creatinine, hemoglobin, white blood cells, C - reactive protein (CRP), age, gender, and body mass index. Results: In multivariate analysis, preoperative serum-creatinine (>93 umol/l), white blood cells (>10.2 tsd/ul), CRP (>3 mg/l) and BMI (<18,5 kg/m2 or ≥30 kg/m2) were independent predictor of poor cancer-specific survival (CSS). A new scoring model was developed, consisting of those four parameters. The stratification model showed significant differences based on Kaplan-Meier analysis for the whole cohort as well as for subgroup analysis (pT2-4, Nx, M0) (p<0.001). CSS after 36 (60) months for low, intermediate and high risk group was 82.6% (76.2%), 37.7% (30.9%) and 22.2% (16.7%), respectively (p<0.0001). Conclusions: Our scoring model based on preoperative clinical parameters can discriminate bladder cancer patients prior to cystectomy in respect of CSS. This might help physicians to recommend treatment options such as neoadjuvant therapy. The scoring model needs external validation considering prospective study design and discriminating patients who received neoadjuvant chemotherapy.