Survival and impact of clinical prognostic factors in surgically treated metastatic renal cell carcinoma.

Background: The survival impact of metastasectomy for metastatic renal cell carcinoma (mRCC) is still an active research field, particularly in the multimodal/targeted therapy era. Objective: To determine the survival impact of clinical prognostic factors and their application to stratification of patients according to their prognosis so clinicians may be aided in their management of mRCC. Design, setting, and participants: Retrospective, bi-institutional cohort study of 109 consecutive patients (71 male and 38 female; median age: 62 yr (range: 25-82 yr) with renal cell carcinoma (RCC) who underwent partial or radical nephrectomy and at least one metastasectomy for mRCC. Intervention: Metastasis resection from various anatomic sites with the aim of completely removing detected lesions. Outcome measurements and statistical analysis: Univariable and multivariable Cox regression models were used to analyse the impact of clinical prognostic factors on cancer-specific survival (CSS). Kaplan-Meier analysis with the log-rank test was used to compare CSS. Receiver operating characteristic (ROC) analysis was performed to test accuracy of prognostic groups. The α error for statistical significance was set at 0.05. Results and limitations: Multivariable analysis revealed that primary tumour T stage ≥3 (hazard ratio [HR]: 2.8; p < 0.01), primary tumour Fuhrman grade ≥3 (HR: 2.3; p < 0.03), nonpulmonary metastases (HR: 3.1; p < 0.03), disease-free interval ≤12 mo (HR: 2.3; p < 0.058), and multiorgan metastases (HR: 2.5; p < 0.04) were independent pretreatment prognostic factors. Leuven-Udine (LU) prognostic groups based on these covariates were created and analysed with Kaplan-Meier and log-rank tests. The 2- and 5-yr CSS were significantly different; the respective group A CSS rates were 95.8% and 83.1%; group B, 89.9% and 56.4%; group C, 65.6% and 32.6%; and group D, 24.7% and 0% (p < 0.0001). ROC analysis on the accuracy of prognostic grouping revealed respective areas under the curve of 0.87 and 0.88 at 2 and 5 yr. Main limitations to present study are the retrospective design and the presence of different metastasis sites. Conclusions: LU prognostic groups could be considered an accurate clinical tool to stratify patients according to prognosis and aid clinicians in the management of mRCC.