The aim of this study was to compare the Memorial Sloan-Kettering Cancer Center (MSKCC) and the Cleveland Clinic Foundation (CCF) models of classification of aRCC patients. In addition, the model developed from the pivotal trial of temsirolimus and those proposed by Motzer et al. in 2004, Escudier et al., Heng et al., Choueiri et al. and Bamias et al. were examined. An observational, retrospective study of patients starting first-line systemic therapy was conducted between 2008 and 2011. The variables used to evaluate the classification models were median overall survival (mOS) and median progression-free survival (mPFS). The comparison of different classification models was performed by comparing the area under the ROC (Receiver Operating Characteristic) curve (AUC) for time-dependent variables proposed by Heagerty. Eighty-eight patients were included. When the different models were compared, it was found that although based on the mOS, the Escudier model had better short-term (1-year) prognostic value, followed by the Heng model; in the long term, the models that presented a higher prognosis capacity were the Hudes and CCF models, closely followed by the Heng model. In addition, the Heng model had a slightly higher predictive ability than the other models. Based on the results, and in line with the European society for medical oncology (ESMO) guidelines, it appears that the model of Heng could be the best model to classify patients with aRCC and combines good short- and long-term prognostics while possessing better predictive ability and a more equal distribution of patients.
Keywords: Classification; Predictive; Prognostic; Renal cell cancer.