Skip to main content
. Author manuscript; available in PMC: 2014 Aug 23.
Published in final edited form as: Magn Reson Med. 2012 Dec 27;70(5):1460–1469. doi: 10.1002/mrm.24563

Figure 8. IVIM parameters as a function of organ type.

Figure 8

Estimates of D (a, b) and F (c, d) derived from the LS (a, c) and BLN (b, d) error models were measured without gating in multiple ROI’s in liver, spleen, kidney, and paraspinous muscles of the back. Median and 25–75% quantiles are shown above. There were significant differences between the right lobe of the liver and each other organ type, including the left lobe, for both parameters from both error models (P<0.01, except for one case). The variance of D was significantly lower for BLN estimates than LS estimates for each organ type (P<0.05), except for the left lobe of liver (P=0.12) and for muscle (P=0.61). Variance of D was reduced in the left lobe of the liver, but not by a statistically significant amount. Estimates of D from the LS (e) and BLN (f) error models are plotted as a function of α for ROI’s in each organ type (right liver lobe: blue, left liver lobe: cyan, spleen: magenta, kidney: orange, muscle: red). Every variety of gating for three voxel size variant scans is included. In liver, an inverse relationship between D and α, similar to that predicted by [Eq. A11], can be appreciated for LS estimates, which is reduced though not eliminated for BLN estimates.