International validation of the prognostic value of lymphovascular invasion in patients treated with radical cystectomy

OBJECTIVE To externally validate the prognostic value of lymphovascular invasion (LVI) in a large international cohort of patients treated with radical cystectomy (RC) for urothelial carcinoma of the bladder (UCB). PATIENTS AND METHODS We collected data from 4257 patients treated with RC and pelvic lymphadenectomy for UCB, without neoadjuvant chemotherapy, at 12 centres. LVI was defined as presence of nests of tumour cells within an endothelium-lined space. RESULTS LVI was detected in 1407 patients (33.1%); the proportion of LVI increased with advancing stage, higher grade, soft-tissue surgical margin involvement, and lymph node metastasis (P < 0.001 for all). In standard multivariate models, LVI was associated with both disease recurrence (hazard ratio 1.43, P < 0.001) and cancer-specific mortality (1.45, P < 0.001). In the entire cohort, adding LVI to a base model that included standard features improved only minimally its predictive accuracy for both recurrence and cancer-specific mortality (by 1.1% and 1.2%, respectively). In 3122 patients with negative lymph nodes, LVI remained independently associated with and improved the predictive accuracy of the standard predictors for recurrence (hazard ratio 1.68, P < 0.001; +2.3%) and cancer-specific mortality (1.70, P < 0.001; +2.4%). By contrast, in 1071 node-positive patients, LVI only marginally improved the prediction of cancer-specific recurrence (hazard ratio 1.20, P < 0.001; +0.2%) and survival (1.23, P < 0.001; +0.5%). CONCLUSIONS LVI is strongly associated with clinical outcome in node-negative patients treated with RC. The assessment of LVI might help to identify patients who could benefit from adjuvant therapy after RC. After confirmation in different populations, LVI should be included in the staging of UCB.