MP67-16 DETERMINING THE BEST COMPLETE BLOOD COUNT-BASED BIOMARKER FOR ONCOLOGIC OUTCOMES FOLLOWING RADICAL CYSTECTOMY FOR UROTHELIAL CARCINOMA OF THE BLADDER

Abstract

INTRODUCTION AND OBJECTIVES: Our group has previously demonstrated that pre-treatment neutrophil-to-lymphocyte ratio (NLR) is a predictor of oncologic outcomes following radical cystectomy (RC). Others have demonstrated that several other pre-treatment complete blood count (CBC)-based markers also have prognostic value, such as monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), haemoglobin (Hb) level. Our objective was to determine which combination of these biomarkers has the best predictive value for oncologic outcomes in patients undergoing RC for bladder cancer (BC). METHODS: Our institutional database of patients undergoing RC (1992-2012) was analyzed. The following biomarker values were obtained (or calculated) based on CBCs performed pre-RC, or before neo-adjuvant chemotherapy where applicable: Hb level, NLR, MLR and PLR. NLR and MLR were log-transformed. The outcome measures were recurrence-free survival (RFS), cancer-specific survival (CSS), and overall survival (OS). The 4 CBC-based biomarkers, along with age, gender, Charlson Score, T-stage, and N-stage, were considered for inclusion in predictive multivariable Cox-regression models. Models that minimized the Aikeke Information Criterion (AIC), representing models that have the best predictive ability while simultaneously minimizing the number of predictors, were selected. RESULTS: Our cohort included 418 patients with a median follow up of 58.4 months. All CBC-based biomarkers were significant predictors of RFS, CSS and OS on univariate analyses. NLR, T-stage and N-stage were selected for inclusion in all 3 models. Hb and age were selected for inclusion in models predicting CSS and OS, while Charlson Score was selected for inclusion in the model for OS. NLR was the only CBC-based biomarker with criteria for inclusion into the model predicting RFS (HR per 1-log unit increase 1⁄41.46, 95%CI1⁄41.13-1.87, p1⁄40.003). NLR was also independently predictive of CSS (HR per 1-log unit increase 1⁄41.43, 95% CI1⁄41.09-1.90, p1⁄40.01) and OS (HR per 1-log unit increase 1⁄41.50, 95% CI1⁄41.07-2.09, p1⁄40.02). Hb was also independently predictive of CSS (HR per 1g/L increase1⁄40.93, 95%CI1⁄40.89-0.97, p<0.001) and OS (HR per 1g/ L increase 1⁄40.91, 95%CI1⁄40.89-0.94, p<0.001). CONCLUSIONS: Among all the CBC biomarkers studied in our dataset, NLR was found to be the best marker for predicting RFS, while NLR and Hb were the best markers for predicting CSS and OS. NLR and Hb are cost-effective and efficient biomarkers for predicting oncologic outcomes for BC following RC.