Impact of Histological Variants on Clinical Outcomes of Patients with Upper Urinary Tract Urothelial Carcinoma.

International audiencePURPOSE: We investigated the clinical and prognostic impact of variant histologies on upper tract urothelial carcinoma outcomes after radical nephroureterectomy. MATERIALS AND METHODS: Data on 1,648 patients with upper tract urothelial carcinoma treated with radical nephroureterectomy without preoperative chemotherapy or radiotherapy were reviewed for histological differentiation and variants. We analyzed differences between pure upper tract urothelial carcinoma and upper tract urothelial carcinoma with variant histology, and differences in the histological variants using different stratifications. RESULTS: A total of 398 patients (24.2%) had histological upper tract urothelial carcinoma variants. The most common variants were squamous cell and glandular differentiation in 9.9% and 4.4% of cases, respectively. Histological variants were associated with advanced tumor stage, tumor multifocality, sessile tumor architecture, tumor necrosis, lymphovascular invasion and lymph node metastasis compared to pure upper tract urothelial carcinoma (p ≤0.031). On univariable analysis variant histology was associated with disease recurrence (p = 0.002) and cancer specific mortality (p = 0.003). In 174 patients treated with adjuvant chemotherapy there was no difference in disease recurrence or survival between variant histology and pure upper tract urothelial carcinoma (p = 0.42 and 0.59, respectively). On multivariable analysis adjusted for the effects of standard clinicopathological characteristics variant histology was not associated with either end point. CONCLUSIONS: Almost 25% of patients with upper tract urothelial carcinoma treated with radical nephroureterectomy harbored histological variants. Variant histology was associated with features of biologically aggressive upper tract urothelial carcinoma. While variant histology is associated with worse outcomes on univariable analysis but this effect did not remain significant on multivariable analysis