Correlation Study On Accurate Evaluation Of Glomerular Filtration Rate In Patients With Clear Cell Renal Cell Carcinoma

Objective:1.To analyze the correlation between the imaging features of contrast-enhanced CT and Gates-GFR（g GFR）obtained from renal ⁹⁹m Tc-diethylene triamine pentaacetic acid(^99mTc-DTPA)dynamic imaging in clear cell renal cell carcinoma.2.To explore the difference of renal GFR before and after operation,and to analyze the accuracy of g GFR in evaluating renal function before operation.3.The formula which is more suitable for evaluating the preoperative GFR of patients with clear cell renal cell carcinoma is selected from 10 commonly used GFR estimation formulas,the factors affecting the accuracy of the formula are discussed,the multi-factor logistic regression model is constructed,and draw the nomogram.Methods:1.The data of 56 patients with unilateral clear cell renal cell carcinoma were analyzed retrospectively.Inclusion criteria:（1）Unilateral clear cell renal cell carcinoma confirmed by pathology;（2）^99mTc-DTPA renal dynamic imaging was performed within1 week before operation and at least 3 months after operation.Exclusion criteria:（1）renal artery stenosis;（2）Renal malformation:such as horseshoe kidney,disc kidney,sigmoid kidney,etc.;（3）Recurrence of tumor.Through the imaging analysis platform of Huiyi Huiying Company,the ROI regions of arterial phase tumors were delineated.The normal and impaired of Gates-GFR of the affected kidneywere classified as the outcome of study（<40ml/min·1.73cm²）.The features were extracted and selected by LASSO dimensionality reduction,and the LR model was established to analyze the ability of imaging features to identify Gates-GFR is normal or impaired.2.20 patients with unilateral partial nephrectomy were collected retrospectively.The inclusion and exclusion criteria agree with the previous.Gates-GFR of impaired kidney and total kidney within 1 week before operation,Gates-GFR of impaired kidney in 3 months after operation and e GFR of total kidney in 6 months after operation were recorded.Paired-Samples T-test were carried out on g GFR of kidney within 1 week before operation and 3 months after operation,as well as g GFR of total kidney within 1week before operation and e GFR within 6 months after operation.3.The clinical data,laboratory and imaging data of 302 patients with clear cell renal cell carcinoma who underwent radionuclide renal dynamic imaging in our hospital were analyzed retrospectively.The inclusion and exclusion criteria agree with the previous.Using ⁹⁹m Tc-DTPA plasma clearance by dual plasma sampling method as reference GFR（r GFR）,the deviation,precision,accuracy and consistency of each formula e GFR were analyzed,and the formula with the highest accuracy was selected.The|e GFR-r GFR|<r GFR×30%is regarded as the accurate estimation of e GFR.The independent predictors affecting the accuracy of e GFR estimation are analyzed by univariate and multivariate Logistic regression analysis,and draw the nomogram.Results:1.Three features including texture features and first-order features are extractedfrom the arterial phase images of enhanced CT.The results show that there is no obvious correlation between tumor imaging features and preoperative g GFR of imapriedkidney（AUC=0.685）.2.There was no significant difference between preoperative g GFR and postoperative g GFR,as well as preoperative r GFR and e GFR after operation（P>0.05）.3.The results of each formula tend to overestimate the actual GFR.Compared with other formulas,Xiangya formula showed better consistency（CCC=0.392）,accuracy（P30mm 69.1）,precision（IQR=19.5）and lower deviation（Bias=10.8）.Therefore,we carried out Logistic regression analysis based on Xiangya formula,and found that there was a significant correlation between Cys C,R score and the accuracy of e GFR estimation.Based on Xiangya formula,the nomogram is drawn according to sex,age,Cys C and R score.ROC curve（training group AUC=0.706,verification group AUC=0.652）and calibration curve show that the prediction model of e GFR applicability has good accuracy.Conclusion:1.Through the correlation analysis between enhanced CT imaging features and g GFR,it is found that there is no obvious correlation between imaging features and preoperative g GFR,which indicates that the enhanced CT imaging features of tumor have no obvious influence on the accuracy of preoperative g GFR.2.Renal ^99mTc-DTPA dynamic imaging can accurately evaluate the preoperative renal function and predict the postoperative residual renal function.3.The e GFR applicability prediction model based on four influencing factors has high accuracy and is helpful for clinicians to make clinical decisions.