AB090. Establishment of prediction model of progression after radical nephrectomy or partial nephrectomy for Chinese patients with clear cell renal cell carcinoma

OBJECTIVE: To develop an algorithm to predict progression to relapse or metastases after radical or partial nephrectomy for Chinese patients with localized clear cell RCC (ccRCC), so as to guide the postoperative treatment. METHODS: The clinical and pathological features and prognosis of 1,034 localized ccRCC patients between January 2006 and December 2013 in the second hospital of Tianjin Medical University was analyzed retrospectively. Univariate comparisons of survival analysis used the Kaplan-Meier method. COX regression model method was used in the multivariate analysis. Harrell’s concordance index was used to assess the prognostic accuracy of the new model. RESULTS: The median follow-up was 39 months (range 4-109 months). Relapse or metastases occurred in 129 patients. The relapse-free survival (RFS) rates after 1 year, 3 years, 5 years, 7 years were 94.8%, 88.6%, 83.4% and 80.4% respectively. In these 1,034 ccRCC patients, 130 patients underwent partial nephrectomy and 904 patients underwent radical nephrectomy. There was no difference in RFS rates between group of partial nephrectomy and group of radical nephrectomy (P=0.061). Multivariate analysis showed that the features of age, tumor size, symptoms at presentation, preoperative platelet count, tumor stage [2010], Fuhrman grade, histologic tumor necrosis were independent predictors associated with RFS for ccRCC patients. A scoring algorithm to predict progression to relapse or metastases after patients underwent partial or radical nephrectomy for ccRCC was developed using the regression coefficients from the multivariate analysis. In this model, the score was calculated as 2 (for pT1b), 3 (for pT2), 4 (for pT3 and pT4), 1 (for age ≥58), 1 (for tumor size ≥5.5 cm), 1 (for symptoms at presentation), 2 (for platelet count ≥400×10(9)/L), 1 (for grade 3), 3 (for grade 4), 1 (for histologic tumor necrosis) and 0 otherwise. The risk of progression to relapse or metastases was divided into three groups: low-risk group (score 0-3), intermediate-risk group (score 4-7) and high-risk group (score ≥8). The estimated 5-year RFS rates were 96.9%, 72.7% and 13.1% respectively in these three groups (P<0.001). Using this sample, the C-index of Leibovich model was 0.784. The C-index of the new model validation was 0.802. CONCLUSIONS: N patients with ccRCC, age, tumor size, presenting symptoms, preoperative platelet count, tumor stage [2010], Fuhrman grade and histologic tumor necrosis are significant independent predictors of RFS after surgery. Based on these indicators, we established a scoring algorithm that can be used to predict disease progression after surgery for Chinese patients with ccRCC.