Site of metastatic recurrence impacts prognosis in patients with high-grade upper tract urothelial carcinoma

Purpose: Metastatic recurrence occurs in over 25% of upper tract urothelial carcinoma patients treated with radical nephroureterectomy. While metastatic recurrence suggests poor prognosis, the impact of the specific site of recurrence on prognosis is not well documented. Materials and methods: We retrospectively analyzed 188 patients who underwent radical nephroureterectomy for high-grade, node-negative upper tract urothelial carcinoma at our institution from 2003 to 2018 without receiving neoadjuvant or adjuvant chemotherapy. Competing-risks survival analysis was performed to evaluate the cumulative incidence and predictors of metastatic recurrence. The Kaplan-Meier method and log-rank test were used to estimate and compare recurrence site-specific survival probabilities following metastatic recurrence. Cox regression analyses were performed to assess site-specific prognoses. Results: Of the 188 patients, 47 (25%) developed metastatic recurrence over a median follow-up of 30 months (interquartile range: 10.5–58.5 months). The 1- and 2-year cumulative incidences of metastatic recurrence were 13.6% and 23.6%, respectively. On multivariable analysis, lymphovascular invasion was significantly predictive of metastatic recurrence (subhazard ratio: 2.6, P = 0.01). Of the 47 patients who developed recurrence, 38 (80.9%) died over a median follow-up of 10 months (interquartile range: 5–20 months). Metastatic recurrence was most common in the lungs (n= 13, 28%) and at multiple sites (n= 14, 30%). Median time to recurrence was shorter for recurrences at multiple sites (6.5 months) and those in the liver (13 months) and bone (18 months) compared to other sites. Patients who recurred in the liver (hazard ratio: 6.3, P = 0.007), bone (hazard ratio: 4.9, P = 0.02), and multiple sites (hazard ratio: 4.6, P = 0.01) had significantly worse prognosis compared to those who recurred in lymph nodes. Statistical significance persisted after adjusting for treatment with salvage therapy. Conclusions: A significant proportion of high-grade upper tract urothelial carcinoma patients recur systemically after radical nephroureterectomy. Lymphovascular invasion is a predictor of metastatic recurrence and may inform decisions regarding perioperative chemotherapy. Hepatic and osseous recurrences have relatively quicker onset and less favorable prognosis compared to other sites. These findings may benefit future efforts to develop recurrence site-specific treatment plans and highlight the necessity of subsequent endeavors to explore the genetic associations of recurrence in upper tract urothelial carcinoma.