Influence of saturation effects on biexponential liver intravoxel incoherent motion

Abstract

Purpose: Studies on intravoxel incoherent motion (IVIM) imaging in the liver have been carried out with different acquisition protocols. The number of acquired slices and the distances between slices can influence IVIM measurements due to saturation effects, but these effects have often been disregarded. This study investigated differences in biexponential IVIM parameters between two slice settings.

Methods: Fifteen healthy volunteers (21-30 years) were examined at a field strength of 3 T. Diffusion-weighted images of the abdomen were acquired with 16 b values (0-800 s/mm² ), with four slices for the few slices setting and 24-27 slices for the many slices setting. Regions of interest were manually drawn in the liver. The data were fitted with a monoexponential signal curve and a biexponential IVIM curve, and biexponential IVIM parameters were determined. The dependence on the slice setting was assessed with Student's t test for paired samples (normally distributed IVIM parameters) and the Wilcoxon signed-rank test (non-normally distributed parameters).

Results: None of the parameters were significantly different between the settings. For few slices and many slices, respectively, the mean values (SDs) for $D$ were $1.21{\mathrm{\mu m}}^2/\mathrm{ms}$ ( $0.19{\mathrm{\mu m}}^2/\mathrm{ms}$ ) and $1.20{\mathrm{\mu m}}^2/\mathrm{ms}$ ( $0.11{\mathrm{\mu m}}^2/\mathrm{ms}$ ); for $f$ they were 29.7% (6.2%) and 27.7% (3.6%); and for $D^{*}$ they were $8.76\cdot {10}^{-2}{\mathrm{mm}}^2/s$ ( $4.54\cdot {10}^{-2}{\mathrm{mm}}^2/s$ ) and $8.71\cdot {10}^{-2}{\mathrm{mm}}^2/s$ ( $4.06\cdot {10}^{-2}{\mathrm{mm}}^2/s$ ).

Conclusion: Biexponential IVIM parameters in the liver are comparable among IVIM studies that use different slice settings, with mostly negligible saturation effects. However, this may not hold for studies that use much shorter TR.

Keywords: IVIM; diffusion; liver; perfusion; saturation effects.