| Abstract: |
Objectives: To evaluate if preoperative symptom classification could refine prediction of outcomes for patients with clinically localized upper-tract urothelial carcinoma (UTUC) managed by radical nephroureterectomy (RNU). Methods: Data on 654 patients with localized UTUC who underwent RNU were reviewed. Preoperative symptoms were classified as incidental (S1), local (S2), and systemic (S3). Clinical and pathologic data were compared between the cohorts. Kaplan-Meier analyses and Cox proportional hazard modeling were used to determine recurrence-free and cancer-specific survival amongst the symptom cohorts. Results: Symptom classification was S1 in 213 (33%) patients, S2 in 402 (61%), and S3 in 39 (6%). S3 symptoms were associated with advanced pathology, including higher stage, grade, and lymph node (LN) positivity. Five and 10-year recurrence-free and cancer-specific survival estimates were similar for patients with S1 and S2 symptoms (P = 0.75 and 0.58, respectively), but was worse for patients with S3 symptoms (P < 0.001 for both). On multivariate analysis adjusting for final pathologic stage, grade, and LN status, S3 symptoms were not an independent predictor of recurrence (HR 1.44, P = 0.19) or death due to disease (HR 1.66, P = 0.07). Addition of symptom classification, however, increased the accuracy of a model consisting of stage, grade, and LNs for prediction of recurrence-free and cancer-specific survival by 1.4% and 1.3%, respectively (P < 0.001 for both). Conclusions: Local symptoms do not confer worse prognosis compared with patients with incidentally detected UTUC. However, systemic symptoms are associated with worse outcomes despite apparently effective RNU. Patients with systemic symptoms may harbor micrometastatic disease and could potentially benefit from a more rigorous metastatic evaluation or perioperative chemotherapy regimens. © 2011 Elsevier Inc. |
