摘要
Abstract
Objective To investigate the value of three-dimensional speckle tracking imaging(3D-STI)combined with blood biochemical indices in predicting death for hospitalized patients with COVID-19.Methods The admission data were collected from 165 patients with COVID-19 diagnosed and hospitalized from November 2022 to February 2023.LASSO re-gression was employed to screen the indicators affecting the death of patients with COVID-19.The data were divided into a training set and a test set at a ratio of 7:3,and the prediction model was built by five times of 10-fold cross-validation on the training set.The nomogram was established for prediction.The model was evaluated in terms of the area under the re-ceiver operating characteristic(ROC)curve,accuracy,sensitivity and specificity.The decision curve analysis(DCA)was performed on the test set to determine the actual benefits of the model in clinical decision-making.Results Five key varia-bles including global principal strain,global longitudinal strain,albumin,serum creatinine and lactate dehydrogenase were selected by LASSO regression for model construction.The training set and test set included 115 and 50 patients,respective-ly,with no difference in baseline data.The prediction on the training set and test set showed,the area under ROC curve of 0.909 and 0.882,the accuracy of 0.835(0.754,0.898)and 0.840(0.709,0.928),the sensitivity of 0.739 and 0.800,and the specificity of 0.960 and 0.900,respectively.The DCA results indicated that the prediction model had considerable potential clinical benefits.Conclusion The nomogram model constructed by three-dimensional spot tracking imaging com-bined with blood biochemical indices has a high value in predicting the risk of death in hospitalized patients with COVID-19,and can guide the early intervention of adverse outcomes.关键词
三维斑点追踪成像/血生化指标/新冠肺炎/死亡预测Key words
Three-dimensional speckle tracking imaging/Blood biochemical index/COVID-19/Prediction of death
