Analysis Of Prognostic Factors And Establishment Of The Relapse Prediction Model For Patients With Renal Cell Carcinoma After Surgery

Objective: To observe the clinical pathological characteristics of renal cell carcinoma(RCC), investigate the relevant prognostic factors in patients after surgery and establish a postoperative relapse prediction model for patients with clear cell RCC(ccRCC), so as to guide the postoperative further treatment and improve the survival rate and the quality of life.Methods: The clinical pathological features and prognosis of 1328 RCC patients after radical or partial nephrectomy between January 2006 and December 2013 in the second hospital of Tianjin Medical University was analyzed retrospectively. 1214patients(91.4%) have been followed up. Univariate comparisons of survival analysis used the Kaplan-Meier method and Log-rank test. COX regression model method was used in the multivariate analysis.The same methods were used to analysis the 1034 patients with nonmetastatic ccRCC that have been completely followed up. Harrell’s concordance index was used to assess the prognostic accuracy of the new model.Results: 1. Of 1214 patients, the median follow-up was 41 months( range,4~109 months), the average follow-up was 44 months(range, 4-109 months). The cancer- specific survival(CSS) rates after 1 year, 3 years, 5 years, 7 years were 96.7%,90.4%, 86.0%, 74.6% respectively. The progression-free survival(PFS) rates after 1year, 3 years, 5 years, 7 years were 94.9%, 88.6%, 83.8%, 80.6% respectively.Univariate analysis of survival showed that the features of age, physical examination found tumors hematuria, fever, tumor cachexia, haemoglobin, preoperative platelet count, preoperative lymphocyte absolute value, preoperative neutrophils to lymphocytes ratio(NLR), alkaline phosphatase levels, ECOG score, tumor size,surgical procedure, tumor stage(2010), Fuhrman grade, vein tumor thrombus, renal capsular invasion, histological tumor necrosis, microvascular invasion,urinary-collecting system invasion, lymphatic metastasis, distant metastasis and postoperative adjuvant therapy were associated with CSS. Multivariate analysis showed that the features of age, physical examination found tumor, hematuria, tumor size, preoperative platelet count, tumor stage(2010), Fuhrman grade, microvascular invasion, renal capsular invasion, histological tumor necrosis, distant metastases wereindependent associated with CSS.2. Of 1034 patients with ccRCC,the median follow-up was 39 months(range,4~109 months). The Relapse-free surival(RFS) ates after 1 year, 3 years, 5 years, 7years were 94.8%, 88.6%, 83.4%, 80.4%. Univariate analysis showed that the features of age, physical examination found tumors, fever, tumor cachexia,lymphocyte absolute value, NLR, haemoglobin, preoperative platelet count, alkaline phosphatase, tumor size, tumor stage(2010), Fuhrman grade, vein tumor thrombus,histological tumor necrosis, renal capsular invasion and microvascular invasion were associated with RFS for ccRCC patients. Multivariate analysis showed that the features of age, tumor stage(2010), tumor size, Fuhrman grade, histologic tumor necrosis, microvascular invasion were independent associated with RFS for cc RCC patients. These factors were assigned to develop a new model. The risk of relapse were divided into three groups: low-risk group(0-4), intermediate-risk group(5-7),high-risk group(≥8). The difference among the prognosis of patients with three groups was statistically significance(P<0.001). These 5-year RFS rates were 95.8%,74.5% and 24.6% respectively. Using this sample, the accuracy of Leibovich model was 0.784. The accuracy of the new model validation was 0.796.Conclusions: 1. Age, physical examination found tumors, hematuria, tumor size,preoperative platelet count, tumor stage(2010), Fuhrman grade, microvascular invasion, histological tumor necrosis, renal capsular invasion and distant metastasis at the surgery are independent predictors of CSS. Vein tumor thrombus and regional lymph node metastasis have a significant effect on prognosis of RCC, but are not independent prognosis factors.2. In patients with ccRCC, age, tumor size, microvascular invasion, tumor stage(2010), Fuhrman grade and histologic tumor necrosis are significant independent predictors of RFS after surgery. Based on these indicators, we established a prediction relapse model after surgery for patients with ccRCC and verified the model. Its prognosis accuracy was better. In this sample, the prediction accuracy of this model was higher than the Leibovich model. Although the new model has better clinical value in the Chinese population, but also needs further clinical research to verify its accuracy.