External validation of a risk model to predict recurrence-free survival after radical cystectomy in patients with pathological tumor stage T3N0 urothelial carcinoma of the bladder

Abstract

Purpose: Patients with stage pT3N0 urothelial bladder cancer vary in outcome after radical cystectomy. To improve prognosis estimation a model was recently developed that defines 3 risk groups for recurrence-free survival based on pT substaging, lymphovascular invasion and positive surgical margin. We present what is to our knowledge the first external validation of this risk model.

Materials and methods: Analogous to the risk model derivation cohort our study group comprised 472 patients with stage pT3, pN0, cM0 disease without perioperative chemotherapy and with a median followup of 42 months (IQR 20-75). The primary end point was recurrence-free survival. The effect of variables was determined by univariate and multivariate Cox regression analysis, and predictive accuracy was determined by ROC analysis.

Results: Stage pT3aN0 and pT3bN0 cases showed significantly different recurrence-free survival after 5 years (51% vs 29%, p<0.001). In the multivariate Cox model pT3 substage (HR 1.86, p<0.001), lymphovascular invasion (HR 1.48, p=0.002), positive surgical margins (HR 1.90, p=0.030) and patient age with a dichotomy at 70 years (HR 1.51, p=0.001) had an independent effect on recurrence-free survival. In the low (221 patients or 47%), intermediate (184 or 39%) and high (67 or 14%) risk groups the 5-year recurrence-free survival rate was 55%, 45% and 13%, respectively (p<0.001). The concordance index of the risk model to predict recurrence-free survival was 0.64 (95% CI 0.59-0.69).

Conclusions: This user friendly risk model can be recommended to estimate prognosis in patients with stage pT3N0 after radical cystectomy. Patients at high risk showed clearly compromised recurrence-free survival and should be included in adjuvant therapy studies.