**Congress:**

# Which number of b-value is better for accurate calculation of virtual MR elastography?

**ECR 2019**

**Poster Number:**

**Type:**

**Keywords:**

**Authors:**

__Y. Matsumoto__

^{1}, M. Harada

^{2}, Y. Kanazawa

^{2}, M. Otomo

^{2}, D. E. Amgalan

^{1}, G. ariunbold

^{1}, S. takamastu

^{3}, Y. Yamashita

^{3};

^{1}Tokushima-shi/JP,

^{2}Tokushima/JP,

^{3}Otawara-shi/JP

**DOI:**

# Conclusion

The purpose of this study was to investigate the effect of the IVIM model and the number of b-value on accurate calculation of the virtual MRE.
In this study,
the results of the linear regression between the estimated IVIM parameter map and the measured shear stiffness depended on IVIM models and the number of b-value.
The result suggest that the estimated IVIM parameter map not only has information about the shear stiffness obtained from non-Gaussian diffusion,
but also contains several factors which are unwanted for the virtual MRE.
Thus,
these unwanted factors in the base images,
used to calculate virtual shear stiffness such as the *ADC*,
*D**,
*D* and *α* map,
should be taken into consideration.
For example,
the probability distribution function (PDF) of the base images is important and when the SNR of the base image is inadequate,
the PDF in the liver is described by a Rayleigh distribution and becomes asymmetric [12,
13].
The mean and mode of the PDF are thus not consistent in this asymmetric distribution as shown in figure 9.

Several factors may have influenced the result of the virtual MRE. Therefore, a robust method, which can separate the non-Gaussian diffusion having component of shear stiffness of the liver tissue, is needed for virtual MRE.