Prognostic factors for overall survival (OS) in patients with metastatic renal cell carcinoma (RCC) treated with vascular endothelial growth factor (VEGF)-targeted agents: Results from a large multicenter study

Abstract

5041 Background: Prognostic factors (PF) for OS have yet to be fully defined for patients with metastatic RCC in the era of VEGF-targeted therapy. This study identifies PFs in this population and updated survival and validation results are presented. Methods: Baseline characteristics and outcomes on anti-VEGF-naïve metastatic RCC patients were collected from three US and four Canadian centers. Using a Cox proportional hazards model, 3 risk categories for predicting survival were identified on the basis of 6 pretreatment clinical features. Results: Six-hundred forty-five patients were included. The median (m) OS was 22 months (95% CI: 20.0–24.8) with a median follow-up of 25 months. Patients were treated with sunitinib (n = 396), sorafenib (n = 200) or bevacizumab (n = 49); 33% had prior immunotherapy. Four of the five PFs previously identified by MSKCC were independent predictors of short survival, including hemoglobin below the lower limit of normal (LLN) (p < 0.0001), corrected calcium above the upper limit of normal (ULN) (p = 0.0006), Karnofsky performance status <80% (p < 0.0001) and time from initial diagnosis to initiation of therapy ULN (pULN (p = 0.012) were independent adverse PFs. Patients were assigned one point for each poor PF and were segregated into three risk categories: favorable-risk (0 PFs, n = 133) median OS (mOS) 37.0 months; intermediate-risk (1 - 2 PFs, n = 292) mOS 28.5 months; and poor-risk (3–6 PFs, n = 139) mOS 9.4 months (log rank p < 0.0001). This model produced a c-index of 0.74 and the bootstrap procedure confirmed good internal validity. The discriminatory ability of the model and its parameter estimates were not affected after adjusting for prior use of immunotherapy or the type of anti-VEGF drug used. Conclusions: These data validate components of the MSKCC model with the addition of platelet and neutrophil counts. This model derived from a large population can be incorporated into patient care and clinical trials of VEGF-targeted agents. [Table: see text]