Development and validation of CT-based radiomics to predict the nuclear grade of clear cell renal cell carcinoma

Abstract

To construct a CT-based radiomics model to preoperatively predict the Fuhrman grade of clear cell renal cell carcinoma(ccRCC).173 ccRCC patients from The Cancer Imaging Archive(TCIA) were enrolled in this study. Radiomics features were derived from preoperative CT images of patients with ccRCC. The radiomics signature (Rad-score) was constructed using the least absolute shrinkage and selection operator algorithm. The Rad-score's predictive performance was assessed using the receiver operating characteristic (ROC) curve. Calibration curves were generated to evaluate the radiomics model's calibration. Eight features were selected to construct the Rad-score. The area under the ROC of the Rad-score was 0.800 in the training cohort and 0.710 in the validation cohort. There was high consistency confirmed by the Hosmer-Lemeshow test between the actual and predicted probabilities in the training(p=0.218) and validation (0.067) cohort. Our CT-based radiomics model can preoperatively predict Fuhrman grade in ccRCC patients.