OBJECTIVE: To investigate the efficacy of simplified (two- or three-tiered) Fuhrman grading systems as prognostic indicators in clear-cell renal cell carcinoma (RCC). PATIENTS AND METHODS: By reviewing records, various clinicopathological factors were assessed in 431 patients who received surgical management for clear-cell RCC. A conventional four-tiered Fuhrman grading system was compared with a modified two-tiered grading system (Fuhrman grades I and II were combined as one class, and grades III and IV as another) and also with a three-tiered grading system (only grades I and II were combined). Efficacies of grading systems were assessed via univariate analyses and multivariate models for prediction of cancer-specific survival. RESULTS: In univariate analysis, the four-tiered and three-tiered grading systems showed similar accuracies (76.5 vs 76.2%, P =0.614) for predicting cancer-specific survival, which were greater than that of the two-tiered system (72.5%; both P < 0.05). Of the three grading systems, only the three-tiered system was an independent predictor of cancer-specific survival in multivariate analysis (P = 0.046). When receiver operating characteristic-derived areas under the curve (AUCs) of multivariate models for predicting cancer-specific survivals were assessed, AUCs for models including the three-tiered Fuhrman grading system and the conventional four-tiered Fuhrman grading system were the same (95.3%), followed by that of a model incorporating the two-tiered grading system (95.1%). CONCLUSION: A modified, three-tiered Fuhrman grading system can be considered an appropriate option in the application of a nuclear grading system to the prognostication of clear-cell RCC in both univariate analysis and multivariate model setting.
OBJECTIVE: To investigate the efficacy of simplified (two- or three-tiered) Fuhrman grading systems as prognostic indicators in clear-cell renal cell carcinoma (RCC). PATIENTS AND METHODS: By reviewing records, various clinicopathological factors were assessed in 431 patients who received surgical management for clear-cell RCC. A conventional four-tiered Fuhrman grading system was compared with a modified two-tiered grading system (Fuhrman grades I and II were combined as one class, and grades III and IV as another) and also with a three-tiered grading system (only grades I and II were combined). Efficacies of grading systems were assessed via univariate analyses and multivariate models for prediction of cancer-specific survival. RESULTS: In univariate analysis, the four-tiered and three-tiered grading systems showed similar accuracies (76.5 vs 76.2%, P =0.614) for predicting cancer-specific survival, which were greater than that of the two-tiered system (72.5%; both P < 0.05). Of the three grading systems, only the three-tiered system was an independent predictor of cancer-specific survival in multivariate analysis (P = 0.046). When receiver operating characteristic-derived areas under the curve (AUCs) of multivariate models for predicting cancer-specific survivals were assessed, AUCs for models including the three-tiered Fuhrman grading system and the conventional four-tiered Fuhrman grading system were the same (95.3%), followed by that of a model incorporating the two-tiered grading system (95.1%). CONCLUSION: A modified, three-tiered Fuhrman grading system can be considered an appropriate option in the application of a nuclear grading system to the prognostication of clear-cell RCC in both univariate analysis and multivariate model setting.