Utilisation of preoperative imaging for muscle-invasive bladder cancer: a population-based study

Objective To test the hypotheses that: (i) use of preoperative imaging for muscle-invasive bladder cancer (MIBC) conforms to practice guidelines; (ii) preoperative imaging, through more accurate staging is associated with improved outcomes. Patient and Methods In this population-based cohort study, records of treatment were linked to the Ontario Cancer Registry to identify all patients with MIBC treated with cystectomy from 1994 to 2008. Utilisation of chest, abdomen-pelvis and bone imaging were evaluated. Trends were evaluated over time. Logistic regression was used to analyse factors associated with utilisation. Cox model analyses were used to explore associations between imaging and survival. Results In all, 2 802 patients with MIBC underwent cystectomy during 1994-2008. Over the three 5-year study periods there was an increase in the proportion of patients having preoperative: chest X-ray (55%, 64%, 63%, P < 0.001), computed tomography (CT) of the chest (10%, 10%, 21%, P < 0.001), bone scan (30%, 34%, 36%; P = 0.04) and CT/magnetic resonance imaging/ultrasonography abdomen/pelvis (80%, 87%, 90%, P = 0.001). Use of chest imaging was associated with age (odds ratio [OR] 1.24-1.59 compared with the youngest age group), N-stage (OR 0.79 for the NX group compared with the N+ group), surgeon volume (OR 0.47-0.53 compared with the highest volume quartile) and geographic region (OR 0.47-2.19 compared with the largest region). Use of bone scan was associated with N-stage (OR 0.57 for the NX group compared with the N+ group) and geographic region (OR 0.71-1.34 compared with the largest region). In adjusted analyses, we found that patients who did not have preoperative chest imaging had inferior overall survival (OS), hazard ratio (HR) 1.12 (95% confidence interval [CI] 1.01-1.25) but not cancer specific survival (CSS), HR 1.09 (95% CI 0.97-1.22); those who did not have preoperative bone scan had inferior OS (HR 1.11, 95% CI 1.01-1.22) and CSS (HR 1.09, 95% CI 1.01-1.25). Survival in the abdomen and pelvis imaging group was not evaluated due to lack of a suitable control group. Conclusion In routine clinical practice there is considerable variation in use of preoperative chest, body, and bone imaging. Preoperative chest and bone imaging is associated with improved outcomes; this association probably reflects better patient selection for cystectomy.