Rates of lymph node invasion and their impact on cancer specific mortality in upper urinary tract urothelial carcinoma.

PURPOSE: To analyze lymph node invasion (LNI) rates according to tumor characteristics and to test the impact of LNI and its extent on cancer specific mortality (CSM) in surgically-treated non metastatic urothelial upper urinary tract carcinoma (UTUC) patients. MATERIALS AND METHODS: Within the SEER database (2004-2014), we identified 2098 patients with histologically confirmed non-metastatic urothelial carcinoma of renal pelvis or ureter who underwent NU with LND. LNI rates stratified according to tumor location and stage were described. Kaplan-Meier plots illustrated CSM rates according to LNI and its extent. Multivariable Cox regression models (MCRMs) tested the effect of LNI and its extent on CSM. RESULTS: Of 2098 UTUC patients, who underwent nephroureterectomy with lymph node dissection, 646 (33%) had LNI. The median number of removed lymph nodes was 3 [Interquartile range (IQR): 1-7]. The median number of positive lymph nodes in patients, who harbored LNI was 1 (IQR:1-3). LNI rates according to tumor location were, respectively, 23.6% for ureteral and 36.5% for renal pelvis tumors. LNI rates according to tumor stage were 9.6, 18.0, 38.7 and 63.9%, for respectively, T1, T2, T3 and T4 UTUC. In MCRMs, LNI achieved independent predictor status for higher CSM (HR 3.00; p < 0.001). Finally, in MCRMs, number of positive lymph nodes defined as the 75th percentile (n ≥ 3) achieved independent predictor status for higher CSM (HR 1.37; p = 0.04). CONCLUSIONS: LNI in non-metastatic UTUC patients is the most important determinant of CSM. Number of positive lymph node is independently associated with higher CSM. In consequence, lymph node dissection can provide extensive prognostic information.