How efficient is acoustic radiation force impulse elastography for the evaluation of liver stiffness?

BACKGROUND In chronic liver diseases, a correct estimation of the severity of liver fibrosis is important for recommendations regarding the treatment. Nowadays, evaluation of fibrosis is done by noninvasive methods such as biochemical scores and transient elastography instead of liver biopsy. The lack of sensitivity to detect fibrosis, because of its heterogeneity is a drawback of liver biopsy (LB). OBJECTIVES To compare transient elastography (TE) and acoustic radiation force impulse (ARFI) for the evaluation of liver stiffness (LS), against percutaneous LB. PATIENTS AND METHODS Our study comprised of 223 subjects; 52 without fibrosis (38 volunteers and 14 patients with F0 on LB), 36 with F1, 40 with F2, 26 with F3 and 69 with liver cirrhosis (46 with LB and 23 with signs of cirrhosis). For each patient we performed in the same session 10 TE and 5 ARFI measurements. The median values were calculated. RESULTS A strong linear correlation (Spearman rho = 0.870) was found between TE and fibrosis (P < 0.0001); there was also a weaker correlation between ARFI and fibrosis (Spearman rho = 0.646; P < 0.0001). TE measurements were also correlated with ARFI measurements (Spearman rho = 0.733, P < 0.0001). The best test for predicting significant fibrosis (F ≥ 2) was TE with a cut-off value of 7.1 kPa (AUROC 0.953). For ARFI, the cut-off value was 1.27 m/s-area under ROC curve (AUROC): 0.890, sensitivity (Se) of 88.7%, specificity (Sp) of 67.5%, positive predictive value (PPV) of 64.5%, and negative predictive value (NPV) of 90% (P = 0.0044). For predicting cirrhosis (F = 4), the optimum cut-off values were 14.4 kPa for TE (AUROC: 0.985, Se: 95.6%, Sp: 94.7%, PPV: 89.2%, NPV: 98%) and 1.7 m/s for ARFI (AUROC: 0.931, Se: 93%, Sp: 86.7%, PPV: 73.6%, NPV: 96.9%) (P = 0.0102). CONCLUSIONS LS evaluation by means of ARFI is not superior to TE for the assessment of liver fibrosis. ARFI is an accurate test for the diagnosis of cirrhosis.