Abstract: Objective To explore the clinical application value of ultrasound elastography in non-invasive evaluating anti-tuberculosis drug-induced liver injury and predicting the risk of it to avoid the aggravation of patients' condition. Method Fifty-six patients hospitalized at the Third People's Hospital of Shenzhen from January to December 2020 were selected and divided into two groups: anti-tuberculosis group with liver injury (n=26) and anti-tuberculosis group without liver injury (n=30). Thirty healthy volunteers were enrolled in this study as the normal control group (n=30). All subjects underwent two-dimensional ultrasound examination, ultrasound elastography and blood biochemical test. ROC curve were used to analyze the efficacy of ultrasound elastography in the diagnosis of anti-tuberculosis drug-induced liver injury. Logistic regression analysis was used to screen the independent risk predictors of anti-tuberculosis drug-induced liver injury. Result ① The values of spleen length, liver STE and liver STQ in the anti-tuberculosis group with liver injury were significantly higher than those in the control group and the anti-tuberculosis group without liver injury. ② The sensitivity of liver STE, liver STQ and splenic STE in the diagnosis of anti-tuberculosis drug-induced liver injury were 88.46%, 80.77% and 69.23% respectively, and the specificity was 96.77%, 86.89% and 73.68%. The areas under the curve were 0.978, 0.894 and 0.739. ③ Logistic multivariate regression analysis showed that liver STE was an independent predictor of anti-tuberculosis drug-induced liver injury. Conclusion Ultrasound elastography, especially liver STE, has clinical value for the risk prediction and diagnosis of anti-tuberculosis drug-induced liver injury.

Key words: Anti-tuberculosis drug-induced liver injury, Ultrasound elastography, Diagnostic efficacy, Risk prediction