Patient-derived AVATAR mouse models to predict prognosis in advanced renal cell carcinoma.

551 Background: Anti-angiogenic (AA) drugs are the cornerstone of first-line (FL) treatment in renal cell carcinoma (RCC). While several mechanisms of resistance to AA have been described, second line (SL) therapies are disappointing with short PFS reported. A recent strategy makes use of individual patient-derived tumor models in animals for drug testing simultaneously with that patient’s FL and SL treatments (AVATARs). The predictive value of each specific AVATAR for that same original patient will be key in order to predict SL treatment response in the clinic trying to help in decision making in this setting. Methods: We generated a unique panel of patient-derived mouse models of RCC based on the orthotopic implantation of primary renal tumor biopsies or metastasis directly obtained from patients. Sunitinib FL treatment followed by different SL treatments are evaluated in each mouse model. Response to treatments in these models and molecular profiling in search of predictive factors of response/resista...