ObjectiveTo determine value of texture analysis based on bi-phasic enhanced CT images in diagnosis of acute pancreatitis (AP) with acute renal injury (AKI).MethodsA total of 62 patients with clinically proven AP including 39 patients with AKI and 23 patients without AKI were analyzed retrospectively. The region of interest (ROI) was chosen at the axial CT-enhanced images of bilateral kidneys using the ITK-Snap software and the texture analysis was performed by the Analysis-Kinetics (A.K.) analysis software. Using the Analysis of Variance, Mann-Whitney U test, Spearman correlation analysis and LASSO regression to reduce the features dimension, and screening out the textures by the logistic regression. The receiver operating characteristic (ROC) curve was established to determine the diagnostic performance of the features.ResultsIn the total of 396 image histological features originally extracted from the texture analysis, 6 features were finally screened out through the dimensionality reduction, involving the Haralick correlation, Inertia, Mean value, Cluster prominence, Short run high grey level emphasis, and Surface area. The area under curve (AUC), threshold, sensitivity, specificity, and accuracy in diagnosing of AP with AKI respectively was 0.926, 0.619, 89.4%, 71.4% and 82.7% by the Haralick correlation; which respectively was 0.790, 0.665, 59.6%, 82.1%, 68.0% by the Inertia; which respectively was 0.983, 0.662, 89.4%, 100%, 93.3% by the Mean value; which respectively was 0.903, 0.696, 80.9%, 85.7%, 82.7% by the Cluster prominence; which respectively was 0.980, 0.778, 76.6%, 100%, 85.3% by the Short run high grey level emphasis; which respectively was 0.819, 0.604, 78.7%, 75.0%, 77.3% by the Surface area.ConclusionTextures of contrast-enhanced CT images have better resolving ability and higher accuracy in diagnosis of AP with AKI and diagnostic efficiency of Mean value is the best.