Purpose To prospectively assess agreement and repeatability of magnetic resonance (MR) elastography liver stiffness measurements across imager manufacturers, field strengths, and pulse sequences. Materials and Methods This prospective cross-sectional study was approved by the institutional review board; informed consent was obtained from all subjects. On the basis of an a priori power calculation, 24 volunteer adult subjects underwent MR elastography with four MR imaging systems (two vendors) and multiple pulse sequences (two-dimensional [2D] gradient-echo [GRE] imaging, 2D spin-echo [SE] echo-planar imaging, and three-dimensional [3D] SE echo-planar imaging). Each sequence was performed twice in each patient with each imaging system. Intraclass correlation coefficients (ICCs) were used to assess agreement and repeatability. P < .05 was considered indicative of a statistically significant difference. Results Pairwise ICCs were 0.67-0.82 and 0.62-0.83 for agreement between pulse sequences across manufacturers (n = 4) and field strengths (n = 5), respectively. ICCs were 0.45-0.90 for pairwise agreement between sequences while fixing manufacturer and field strength (n = 8). Test-retest repeatability across the various manufacturer, field strength, and pulse sequence combinations (n = 10) was excellent (ICCs, 0.77-0.94). The overall ICC for all manufacturer, field strength, and sequence combinations (n = 10) was 0.68 (95% confidence interval [CI]: 0.55, 0.82). ICC according to field strength was 0.78 (95% CI: 0.67, 0.88) at 1.5 T (n = 5) and 0.64 (95% CI: 0.49, 0.78) at 3.0 T (n = 5). ICCs according to vendor were 0.83 (95% CI: 0.73, 0.91) (n = 4) and 0.65 (95% CI: 0.51, 0.79) (n = 6). Average patient level variance was 0.042 kPa, with a coefficient of variation of 10.7%. Conclusion MR elastography is a reliable method for assessing liver stiffness, with small amounts of variability between imager manufacturers, field strengths, and pulse sequences. RSNA, 2016.

Purpose To evaluate the correlation between ultrasonographic (US) point shear-wave elastography (SWE) and magnetic resonance (MR) elastography liver shear-wave speed (SWS) measurements in a pediatric population and to determine if US data dispersion affects this relationship. Materials and Methods Institutional review board approval was obtained for this HIPAA-compliant investigation; informed consent and patient assent (as indicated) were obtained. Patients (age range, 0-21 years) undergoing clinical liver MR elastography between July 2014 and November 2015 were prospectively enrolled. Patients underwent two-dimensional gradient-recalled-echo 1.5-T MR elastography with point SWE performed immediately before or immediately after MR elastography. Spearman rank correlation coefficients were calculated to assess the relationship and agreement between point SWE and MR elastography SWS measurements. Uni- and multivariate logistic regression were performed to identify predictors of US data dispersion, with the best multivariate model selected based on Akaike information criterion. Results A total of 55 patients (24 female) were enrolled (mean age, 14.0 years ± 3.9 (standard deviation) (range, 3.5-21.4 years). There was fair correlation between point SWE and MR elastography SWS values for all patients (ρ = 0.33, P = .016). Correlation was substantial, however, when including only patients with minimal US data dispersion (n = 26, ρ = 0.61, P = .001). Mean body mass index (BMI) was significantly lower in patients with minimal US data dispersion than in those with substantial US data dispersion (25.4 kg/m ± 7.8 vs 32.3 kg/m ± 8.3, P = .003). At univariate analysis, BMI (odds ratio, 1.12; 95% confidence interval [CI]: 1.03, 1.21; P = .006) and abdominal wall thickness (odds ratio, 2.50; 95% CI: 1.32, 4.74; P = .005) were significant predictors of US data dispersion. In the best multivariate model, BMI was the only significant predictor (odds ratio, 1.11; 95% CI: 1.03, 1.20; P = .009). Conclusion Point SWE and MR elastography liver SWS measurements correlate well in patients with a BMI of less than 30 kg/m and minimal US data dispersion; increasing US data dispersion is directly related to a higher BMI. RSNA, 2016.