Initial CT-based Three-dimensional Radiomics Of Lung Predicting Acute Kidney Injury In Patients With Acute Paraquat Intoxication

【Objective】The purpose of this study was to investigate the value of initial CT-based three-dimensional radiomics of lung predicting the acute kidney injury（AKI）in patients with acute paraquat poisoning（APP）.【Methods】A total of 50 patients with APP confirmed by Clinical treatment base of poisoning in Yunnan Province were enrolled in this retrospective study.Initial chest computed tomography images and clinical baseline data of these patients were collected.They were divided into AKI group（28 patients）and non-AKI group（22 patients）based on the diagnostic criteria for AKI.Then randomly assigned all patients to a training group and a validation group by a ratio of 1:1,and then the radiomics features were extracted by LIFEx from the whole lung.Least absolute shrinkage and selection operator（LASSO）regression were used to select the features and establish the radiomics score.The prediction performance of the radiomics score was assessed mainly by the ROC curves analysis.【Results】The area under the ROC curve of operation was 0.825 and 0.798,respectively,in the training and validation group.The sensitivity and specificity were 81.82%,57.89%,85.71% and 100%,respectively.【Conclusion】In this study,we established and verified a radiomics score based on initial whole-lung CT images within 12 hours of admission,which can be used to predict acute kidney injury in patients with acute paraquat poisoning.