OBJECTIVE:The objective of the current study was to compare, in a single center experience, the discriminating accuracy of two prognostic models to predict the outcome of patients surgically treated for conventional renal cell carcinoma (RCC).
PATIENTS AND METHODS: We retrospectively evaluated the clinical and pathological data of 100 patients surgically treated for RCC between 1998-2008 at our institution. For each patient, prognostic scores were calculated according to two models: the University of California Los Angeles integrated staging system (UISS) and the Stage, Size, Grade, and Necrosis (SSIGN) developed at the Mayo Clinic. The prognostic predictive ability of models was evaluated using receiver operating characteristic (ROC) curves.
RESULTS: The median follow-up was 62 months (range 12-120). All clinical and pathological features that compound the algorithms were significantly associated with death from RCC in univariate and multivariate setting. The 5-year cancer-specific survival (CSS) according to the SSIGN score were 95% in the '0-2' category, 88% in '3-4', 60% in '5-6', 37% in '7-9' and 0% in the '> or = 10' group (long-rank p value < 0.001); according to the UISS the 5 yr CSS probabilities in non-metastatic patients were 100% in low, 80% in intermediate and 54% in high-risk groups; in metastatic patients, the respectively CSS were 40% in low and 25% in high-risk groups (long-rank p value < 0.001). The area under the ROC curve was 0.815 for the SSIGN score and 0.843 for the UISS (p = 0.632).
CONCLUSION: In our series the SSIGN and UISS discriminated well, without relevant differences. Currently both algorithms represent usefuls clinical tools that allow risk assessment after surgical treatment of RCC. We encourage the uro-oncologist to begin to routinely rely on them in real-life practice.
Written by:
Martella O, Galatioto GP, Necozione S, Pomante R, Vicentini C. Are you the author?
Division of Urology, Giuseppe Mazzini Hospital, Teramo, Italy.
Reference: Arch Ital Urol Androl. 2011 Sep;83(3):121-7.
PubMed Abstract
PMID: 22184835
