Introduction:

Locoregional recurrence (LR), defined as recurrence within the soft tissue of the cystectomy bed or pelvic lymph nodes, confers a poor prognosis after radical cystectomy (RC) for urothelial carcinoma. Reported LR incidences vary widely in the literature, and no risk stratification tools are commonly used to predict LR following RC. Such a tool would inform clinical trial eligibility criteria for adjuvant radiation after RC. Herein, we construct and externally validate a LR risk score to identify factors associated with LR following RC.

Methods:

We queried our prospectively maintained RC registry to identify patients with nonmetastatic urothelial carcinoma between 1988-2022 with LR data available. Univariable and multivariable Cox regression models were used to identify factors associated with LR following cystectomy. Using the multivariable regression coefficients, a LR risk score was then developed and then externally validated using data from another insitution.

Results:

In a cohort of 1256 patients, 18% (N=227) experienced LR at a median time of 0.89 years, translating to a 2- and 5-year LR risk of 17% and 22%. On multivariable analysis, increasing pT-stage (HR 1.72) and pN+ (HR 1.90) were associated with increased LR, while more lymph nodes removed were inversely associated with LR (HR 0.98, all p<0.01). Positive ureteral/urethral margin and positive radial margin were both associated with increased LR, albeit with p=0.06 and p=0.07, respectively. A risk score based on regression coefficients stratified patients with increasing risk of 5-year LR from 2% to 52% with Harrell’s c-index of 0.73 (Figure 1A).

Our risk score was then externally validated in a separate European cohort (N=614) in which 7.8% (N=48) of patients had LR. The risk score (Figure 1B) was able to stratify patients with comparable performance to the development cohort with a c-index of 0.75 at 3 years.

Conclusion:

Approximately 10%-20% of patients undergoing RC will experience LR within 5 years, and the majority of these occur within 2 years. A LR risk score based on pT, pN, margin status, and number of lymph nodes removed effectively stratified patients by LR risk. The LR risk score had similar performance in the internal development and external validation cohorts despite the cohorts being in different continents, with different populations and different LR event rates. This supports the robustness and external validity for using this score to create eligibility criteria for clinical trials of adjuvant radiation therapy after RC.

Funding: N/A