The usefulness of transient elastography, acoustic-radiation-force impulse elastography, and real-time elastography for the evaluation of liver fibrosis

Background/Aims Several noninvasive methods have recently been developed for the evaluation of liver fibrosis. The accuracy of transient elastography (TE), acoustic-radiation-force impulse (ARFI) elastography, and real-time elastography (RTE) in predicting liver fibrosis were evaluated. Methods Seventy-four patients who had undergone a liver biopsy within the previous 6 months were submitted to evaluation with TE, ARFI, and RTE on the same day. Results There were significant correlations between fibrosis stage and liver stiffness measurement (LSM) using the three tested methods: TE, r2=0.272, P=0.0002; ARFI, r2=0.225, P=0.0017; and RTE, r2=0.228, P=0.0015. The areas under the receiver operating characteristic curves (AUROC) for the diagnosis of significant fibrosis (≥F2, Metavir stage) by TE, ARFI, RTE, TE/platelet count (PLT), velocity of shear wave (Vs)/PLT, and elasticity score (Es)/PLT were 0.727, 0.715, 0.507, 0.876, 0.874, and 0.811, respectively. The AUROC for the diagnosis of cirrhosis by TE, ARFI, RTE, TE/PLT, Vs/PLT, and Es/PLT were 0.786, 0.807, 0.767, 0.836, 0.819, and 0.838, respectively. Comparisons of AUROC between all LSMs for predicting significant fibrosis (≥F2) produced the following results: TE vs. RTE, P=0.0069; ARFI vs. RTE, P=0.0277; and TE vs. ARFI, P=0.8836. Applying PLT, the ability of each LSM to predict fibrosis stage significantly increased: TE/PLT vs. TE, P=0.0004; Vs/PLT vs. ARFI, P=0.0022; and Es/PLT vs. RTE, P<0.0001. However, the ability to predict cirrhosis was not enhanced, combining LSM and PLT. Conclusions TE and ARFI may be better methods for predicting significant liver fibrosis than RTE. This predictive ability increased significantly when accounting for platelet count. However, all of the measures had comparable efficacies for predicting cirrhosis.