Abstract
PURPOSE: Systemic inflammatory responses, which are defined in terms of the Glasgow prognostic score (GPS), have been reported to be independent predictors of unfavorable outcomes in various human cancers. We assessed the utility of the GPS as a predictor of intravesical recurrence after radical nephroureterectomy (RNU) in upper urinary tract carcinoma (UTUC). MATERIALS AND METHODS: We collected data for 147 UTUC patients with no previous history of bladder cancer who underwent RNU from 2004 to 2012. Associations between perioperative clinicopathological variables and intravesical recurrence were analyzed by using univariate and multivariate Cox regression models. RESULTS: Overall, 71 of 147 patients (48%) developed intravesical recurrence, including 21 patients (30%) diagnosed with synchronous bladder tumor. In the univariate analysis, performance status, diabetes mellitus (DM), serum albumin, C-reactive protein, GPS, and synchronous bladder tumor were associated with intravesical recurrence. In the multivariate analysis, performance status (hazard ratio [HR], 2.33; 95% confidence interval [CI], 1.41-3.85; p=0.001), DM (HR, 2.04; 95% CI, 1.21-3.41; p=0.007), cortical thinning (HR, 2.01; 95% CI, 1.08-3.71; p=0.026), and GPS (score of 1: HR, 6.86; 95% CI, 3.69-12.7; p=0.001; score of 2: HR, 5.96; 95% CI, 3.10-11.4; p=0.001) were independent predictors of intravesical recurrence. CONCLUSIONS: Our results suggest that the GPS as well as performance status, DM, and cortical thinning are associated with intravesical recurrence after RNU. Thus, more careful follow-up, coupled with postoperative intravesical therapy to avoid bladder recurrence, should be considered in these patients.