Inter-platform reproducibility of liver stiffness measured with two different point shear wave elastography techniques and 2-dimensional shear wave elastography using the comb-push technique

ABSTRACT

PURPOSE: The purpose of this study was to compare the technical success rate and reliability of measurements made using three shear wave elastography (SWE) techniques and to assess the inter-platform reproducibility of the resultant liver stiffness measurements. METHODS: This prospective study included 54 patients with liver disease. Liver stiffness (LS) measurements were obtained using 2-point SWE techniques (Virtual Touch Quantification and S-Shearwave) and 2-dimensional (2D) SWE, with transient elastography (TE) serving as the reference standard. The technical success rates and measurement reliability of the three techniques were compared. LS values measured using the three SWE techniques and TE were compared using Spearman correlation coefficients and 95% Bland-Altman limits of agreement. Intra-class correlation coefficients (ICC) were used to analyze the inter-platform reproducibility of LS measurements. RESULTS: The three SWE techniques and TE showed similar technical success rates (P=0.682) but demonstrated significant differences in the reliability of LS measurements (P=0.006) and mean LS measurements (P<0.001). Despite strong correlations (r=0.73-0.94) between SWE systems, various degrees of inter-platform reproducibility (ICC, 0.58-0.92) were observed for the three SWE techniques. The best agreement was observed between S-Shearwave and TE (ICC, 0.92), and the worst agreement was observed between 2D-SWE and TE (ICC, 0.58). In the Bland-Altman analysis, a tendency toward lower LS values with the three SWE techniques than with TE in patients with F3 and F4 disease was observed. CONCLUSION: Significant inter-system variability was observed in LS measurements made using the three SWE techniques. Therefore, LS values measured using different SWE techniques should not be used interchangeably for longitudinal follow-up.