Is there value to routine oncologic surveillance after radical cystectomy? Comparative outcomes of symptomatic versus asymptomatic recurrence.

Abstract

421 Background: Reported rates of adherence to post-radical cystectomy (RC) surveillance guidelines in real-world practice have been as low as 9%, in part reflecting a nihilistic view held by many of the value of routine follow-up. Indeed, conflicting data exist regarding the outcomes of patients with cancer recurrence detected by scheduled surveillance versus symptom-directed evaluation. Herein, we assessed comparative outcomes of patients with symptomatic recurrence (SR) versus asymptomatic recurrence (AR) after RC. Methods: We reviewed our Institutional Registry of RC patients to identify patients with cancer recurrence following RC. Presenting symptoms in the SR cohort included pain, constitutional symptoms, gastrointestinal symptoms, and hematuria/voiding symptoms, whereas AR was defined as recurrence detected on routine surveillance in the absence of symptoms. Baseline demographic and clinical characteristics were compared between study groups using chi-square and t-test. Kaplan-Meier and Cox survival analyses were performed to compare cancer-specific survival (CSS) and overall survival (OS) between AR and SR groups. Results: Of 3822 patients who underwent RC from 1980-2018 (with a median follow-up after RC of 2.4 years (IQR 1.1,5.5), a total of 1100 were subsequently diagnosed with recurrence, including 311 (28.3%) with AR and 789 (71.7%) with SR. Median time from RC to recurrence was longer in the AR group (13.2 months) than in the SR group (10.8 months; p = 0.01). Presenting symptoms included pain (70.2%), constitutional symptoms (50.7%), gastrointestinal symptoms (23.3%), and urinary symptoms (23.3%). Median follow-up after recurrence was 2.4 years (IQR 1.1-5.5), during which time 997 patients died, including 840 who died of bladder cancer. Compared to patients with SR, patients with AR had a longer median CSS (54.5 months vs 27.3 months, p < 0.001) and OS (43.0 months vs 25.8 months, p < 0.001). On multivariable Cox proportional hazards models adjusting for demographic and clinical factors, SR was associated with a significantly increased risk of cancer-specific (hazard ratio [HR] 1.66 [95% confidence interval 1.41-1.96], p < 0.0001) and all-cause mortality (HR 1.48 [1.23-1.71], p < 0.0001). Conclusions: SR after RC is associated with worse oncologic outcomes than post-RC recurrence detected by routine surveillance. As such, continued surveillance is warranted, while further study is needed to determine the optimal follow-up regimen balancing patient and disease risks.