Application Of Enhanced CT Radiomics In The Prediction Of Grade Of Clear Cell Renal Carcinoma And Differential Diagnosis From Renal Angiomyolipoma Without Visible Fat

Part1: Application of enhanced CT radiomics in predicting the grade of clear cell renal carcinomaObjective:Predicting pathological nuclear grading of clear cell renal carcinoma(CCRCC)using enhanced CT radiomics combined with machine learning.Methods:141 patients with cc RCC confirmed by pathology were retrospectively analyzed.There were 88 patients with low grade I-II and 53 patients with high grade III-IV according to WHO/ISUP grade.Interested in arterial and venous phase images manual sketch tumor area,the original image and wavelate transform image extract 1706 radiomics features,cases divided into training set and test set,test set data after Z-score normalization and by PCC space dimension reduction,using analysis of variance filter top 17 F value characteristic of the logistic regression model is set up respectively,both for five-fold the cross validation,selection of cross validation set the highest AUC model as the best model,and to verify the effectiveness test set.The model was validated internally by Bootstrap method to obtain the receiver operating characteristic(ROC)curve,and the area under the ROC curve(AUC)value was calculated to evaluate the diagnostic effectiveness of the model.The clinical characteristics were analyzed to determine the independent predictors and the clinical model was established.The comprehensive model was established by multi-factor binary Logistic regression using the selected independent predictors and the predicted values of the radiomics model.The nomogram based on combined with clinical factor and radiomics features.The test of Hosmer-Lemeshow was used to evaluate the fitness of the line chart.Decision curve analysis was applied for clinical use.Results:The AUC values of the best radiomics model in train cohort and test cohort were0.889(95%CI0.803-0.957)and 0.821(95%CI0.699-0.917),respectively.The AUC of the comprehensive model in the train cohort and test cohort is 0.891(95%CI 0.823-0.960)and 0.823(95%CI 0.798-0.912).Decision curve analysis verified the clinical usefulness of the predictive nomogram.Conclusion:The combination of clinical factors and radiomics features has a strong ability to predict pathological nuclear grade of cc RCC.Part2: Application of enhanced CT radiomics for differentiation of renal angiomyolipoma without visible fat from homogeneous clear cell renal cell carcinomaObjective:Using enhanced CT radiomics for differentiation of renal angiomyolipoma without visible fat from homogeneous clear cell renal cell carcinoma.Methods:Retrospective analysis was performed on 48 cases of homogeneous clear cell renal cell carcinoma and renal angiomyolipoma without visible fat confirmed by pathology.28 cases of homogeneous clear cell renal cell carcinoma and 22 cases of renal angiomyolipoma without visible fat were confirmed by postoperative pathology.Interested in arterial and venous phase images manual sketch tumor area,the original image and wavelet transform image extract 1706 radiomics features,data after Z-score normalization and by PCC space dimension reduction,using analysis of variance to screen the characteristics of the top 6 F value LR model respectively,five-fold cross validation,selecting cross validation set the highest AUC model as the best model.The model was validated internally by Bootstrap method to obtain the receiver operating characteristic(ROC)curve,and the area under the ROC curve(AUC)value was calculated to evaluate the diagnostic effectiveness of the model.The clinical features were analyzed to determine the predictors and the clinical model was established.The comprehensive model was established by multi-factor binary Logistic regression based on the predicted values of the selected clinical factors and the radiomics model.The nomogram based on combined with clinical factor and radiomics features.The test of Hosmer-Lemeshow was used to evaluate the fitness of the line chart.Decision curve analysis was applied for clinical use.Results:The AUC of the best radiomics model verified by Bootstrap method was0.836(95%CI 0.701-0.927).The AUC value of the comprehensive model was0.869(95%CI 0.740-0.949),and the calibration curve of the nomogram showed good consistency.Decision curve analysis verified the clinical usefulness of the predictive nomogram.Conclusion:The combination of clinical factors and radiomics features has a strong ability to distinguish renal angiomyolipoma without visible fat from homogeneous clear cell renal cell carcinoma...