Development and external validation of a local pelvic recurrence risk score after radical cystectomy: identifying the ideal candidate for adjuvant radiation clinical trials.

Objective: To develop a risk stratification tool for locoregional recurrence (LR) after radical cystectomy (RC). LR confers a poor prognosis and current risk stratification tools are lacking, representing a critical knowledge gap in facilitating trial design for adjuvant local therapy.

Methods: We reviewed our institutional Cystectomy Registry to ideurontify factors associated with LR. A risk score was created using the multivariable regression coefficients.

Results: Among 1256 patients who underwent RC, 227 experienced LR, with median time to LR of 11 months, resulting in a 2- and 5-year LR risk of 17% and 22%, respectively. Median follow up was 4.4 years after LR. On multivariable analysis, pT-stage (HR=1.72), pN+ (HR=1.90), and number of lymph nodes removed (HR=0.98 per node) were independently associated with LR (all p<0.01). Positive ureteral/urethral margin (HR=1.52) and positive radial margin (HR=1.58) were associated with LR with p=0.08 and p=0.07, respectively. Our risk score stratified patients by risk of 5-year LR from 2% to 52% with c-index of 0.73. External validation in a European cystectomy cohort (N=614) demonstrated comparable performance to the development cohort with c-index of 0.75 at 3 years.

Conclusions: We developed a risk score based on post-cystectomy pathologic variables which effectively stratified patients' risk of subsequent LR. The risk score had similar performance in the internal development and external validation cohorts, despite the cohorts being geographically distinct with different LR event rates. This risk score may therefore be considered in designing adjuvant therapy trials (e.g., radiation, immunotherapy) for high-risk patients after RC.