Clinical-pathologic stage discrepancy in patients with bladder cancer treated with radical cystectomy: Associations with clinical variables and survival.

Abstract

248 Background: Radical cystectomy (RC) is the most common treatment for bladder cancer (BC) in the United States. We examined clinical-pathologic stage discrepancy using the National Cancer Data Base. Methods: 16,953 patients with BC treated with RC between 1998 and 2009 were analyzed. Clinical factors associated with stage discrepancy were assessed by multivariable generalized estimating equation models. Survival analysis was conducted for patients treated between 1998 and 2004 (N=7,270) using a Cox proportional hazards model. Results: 41.9% of patients were upstaged at RC while 5.9% were downstaged. Upstaging was more common in females (OR 1.08, p=.04), the elderly (OR 1.26 for age ≥80 vs. 18-59, p=.001), higher tumor grade (OR 2.29 for grade 3-4 vs. grade 1, p<.0001), non-urothelial histology (OR 1.31, p=.002 for squamous and OR 1.26, p=.03 for adenocarcinoma), and with extended lymphadenectomy (OR 1.27 for ≥10 lymph nodes examined vs. 0-9, p<.0001). Downstaging was less common in the elderly (OR 0.50 for age ≥80 vs. 18-59, p<.0001), in Hispanics (OR 0.58, p=.009) and with variant histology (OR 0.55, p=.003 for squamous and OR 0.3, p<.0001 for adenocarcinoma). Receipt of neoadjuvant chemotherapy (CT) was highly associated with downstaging (OR 2.31, p<.0001). 5-year survival by stage is shown in the table. Upstaging was associated with increased 5-year mortality (HR 1.79, p<.0001) as was receipt of CT (HR 1.28, p=.02 for neoadjuvant and HR 1.23, p<.0001 for adjuvant). Extended lymphadenectomy was associated with decreased 5-year mortality (HR 0.82 for ≥10 lymph nodes examined vs. 0-9, p<.0001). Downstaging was not associated with survival (HR 0.88, p=0.17). Conclusions: This study is the largest to date to analyze stage discrepancy and survival in BC patients treated with RC. Upstaging is common and is associated with decreased 5-year survival. These data can be used in pre-operative risk stratification, treatment decision making and comparison with studies of non-operative management. [Table: see text]