Reliability Criteria of Two-Dimensional Shear Wave Elastography: Analysis of 4277 Measurements in 788 Patients

BACKGROUND & AIMS: Two-dimensional shear wave elastography (2D-SWE) is an accurate method for the non-invasive evaluation of liver fibrosis. We aimed to determine the reliability criteria and the number of necessary reliable measurements for 2D-SWE. METHODS: 788 patients with chronic liver disease underwent liver biopsy and 2D-SWE examination in three centers. The 4277 2D-SWE measurements performed were 2:1 randomly divided into derivation (n = 2851) and validation (n = 1426) sets. Reliability criteria for a 2D-SWE measurement were defined in the derivation set from the intrinsic characteristics given by the device (mean liver stiffness, standard deviation, diameter of the region of interest), with further evaluation in the validation set. RESULTS: In the whole population of 4277 measurements, AUROC for bridging fibrosis was 0.825 +/- 0.006 and AUROC for cirrhosis was 0.880 +/- 0.006. Mean stiffness and coefficient of variation (CV) were independent predictors of bridging fibrosis or cirrhosis. From these two parameters, new criteria were derived to define a reliable 2D-SWE measurement: stiffness <8.8 kPa, or stiffness between 8.8-11.9 kPa with CV <0.25, or stiffness >= 12.0 kPa with CV <0.10. In the validation set, AUROC for bridging fibrosis was 0.830 +/- 0.013 in reliable measurements vs 0.667 +/- 0.031 in unreliable measurements (P < .001). AUROC for cirrhosis was 0.918 +/- 0.014 vs 0.714 +/- 0.027, respectively (P < .001). The best diagnostic accuracy for a 2D-SWE examination was achieved from three reliable measurements. CONCLUSIONS: Reliability of a 2D-SWE measurement relies on the coefficient of variation and the liver stiffness level. A 2D-SWE examination should include three reliable measurements according to our new criteria.

Automated summary

This study aimed to determine the reliability criteria and the number of necessary reliable measurements for 2D-SWE, and found that mean stiffness and coefficient of variation were independent predictors of bridging fibrosis or cirrhosis. According to the new criteria, a 2D-SWE examination should include three reliable measurements.