Do regions of interest location and type influence liver stiffness measurement using magnetic resonance elastography?

To assess the variability of liver stiffness measurements using magnetic resonance elastography (MRE) at 1.5T, depending on different approaches of regions of interest (ROIs) drawing.

Fifty consecutive patients with successful liver MRE were included. There were 32 men and 18 women with a mean age of 52±14 (SD) years (range: 20-85 years). MRE was acquired using a gradient recalled-echo MRE sequence. At the level of the portal bifurcation, one observer drawn in the right liver first 3 elliptical ROI and then one free-hand ROI, as large as possible based on the confidence map and the anatomy. Three additional elliptical ROIs were further drawn on the slice above and 3 other on the slice below, for a total of 9 elliptical ROIs. The average value of liver stiffness in the 3 elliptical ROIs of the central slice and the one from the 9 elliptical ROIs were computed. Three liver stiffness values were obtained for each patient from the 3 measurement methods (one free-hand ROI, 3 elliptical ROIs and 9 elliptical ROIs). Inter-method variability was assessed using the intra-class correlation coefficient (ICC) and Bland-Altman analysis.

The variability between the 3 methods was excellent with ICC>0.978 (P<0.0001). The Bland-Altman analysis revealed high agreement between the 3 methods with bias<0.45kPa and limits of agreement<±1.13kPa. The variability was lower when comparing a large free-hand ROI and the 3-elliptical ROIs, than when comparing the 9-elliptical ROIs to one of the other methods.

Our results show that the variability between the 3 methods of ROI drawing and placement is very low.