Loading [MathJax]/jax/output/PreviewHTML/jax.js
Skip to main page content
U.S. flag

An official website of the United States government

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Oct;76(4):1252-62.
doi: 10.1002/mrm.25980. Epub 2015 Oct 29.

Orientation dependence of microcirculation-induced diffusion signal in anisotropic tissues

Osama M Abdullah  1 , Arnold David Gomez  2   3 , Samer Merchant  2 , Michael Heidinger  4 , Steven Poelzing  5 , Edward W Hsu  2
Affiliations

Orientation dependence of microcirculation-induced diffusion signal in anisotropic tissues

Osama M Abdullah et al. Magn Reson Med. 2016 Oct.

Abstract

Purpose: To seek a better understanding of the effect of organized capillary flow on the MR diffusion-weighted signal.

Methods: A theoretical framework was proposed to describe the diffusion-weighted MR signal, which was then validated both numerically using a realistic model of capillary network and experimentally in an animal model of isolated perfused heart preparation with myocardial blood flow verified by means of direct arterial spin labeling measurements.

Results: Microcirculation in organized tissues gave rise to an MR signal that could be described as a combination of the bi-exponential behavior of conventional intravoxel incoherent motion (IVIM) theory and diffusion tensor imaging (DTI) -like anisotropy of the vascular signal, with the flow-related pseudo diffusivity represented as the linear algebraic product between the encoding directional unit vector and an appropriate tensor entity. Very good agreement between theoretical predictions and both numerical and experimental observations were found.

Conclusion: These findings suggest that the DTI formalism of anisotropic spin motion can be incorporated into the classical IVIM theory to describe the MR signal arising from diffusion and microcirculation in organized tissues. Magn Reson Med 76:1252-1262, 2016. © 2015 Wiley Periodicals, Inc.

Keywords: IVIM; anisotropic blood flow; apparent diffusion coefficient; diffusion-weighted MRI; organized capillary flow; perfusion fraction.

PubMed Disclaimer

Figures

FIG. 1
FIG. 1
Numerical simulations of flow-induced MR signal attenuation. a: Panels correspond to plug flow of varying speeds in tubes that have identical and same orientation as the encoding gradient. The low, moderate, and high flow settings correspond to mean speeds (vm) of 0.32±0.13, 0.64±0.25, and 0.86±0.34 mm/s, respectively, obtained from Gaussian-distributed capillary radii at different flow rates. b: Plug flow of the same speed distribution as the “high flow” case above in tubes whose orientations have varying degree of anisotropy of quasi Gaussian-distributed capillary orientations with standard deviations of 52° (low anisotropy), 44° (mod anisotropy), and 28° (high anisotropy), respectively. c,d: Results obtained for laminar flow of the same volumetric flow rate under the same configurations as in (a) and (b), respectively. The solid lines denote nonlinear least-squares single exponential fits of the noiseless data points, and the dotted lines denote the 95% confidence interval of the fits.
FIG. 2
FIG. 2
a–c: Anisotropy of tissue IVIM parameters in numerical analysis of laminar flow in the same perfectly aligned capillary network as a function of encoding angle θ. The low, moderate, and high flow settings correspond to mean speeds (vm) of 0.32±0.13, 0.64±0.25, and 0.86±0.34 mm/s, respectively, obtained from Gaussian-distributed capillary radii at different flow rates. Behaviors of the vascular signal attenuation for θ = 0° are shown in Figure 1c. Solid lines represent results obtained in the noiseless simulation.
FIG. 3
FIG. 3
a–c: Orientation dependence of tissue IVIM parameters in numerical analysis of same laminar flow distribution (vm = 0.86±0.34 mm/s) in capillary networks of varying degrees of anisotropy. The low, moderate, and high anisotropy settings correspond to quasi Gaussian-distributed capillary orientations with standard deviations of 52°, 44°, and 28°, respectively. Behaviors of the vascular signal attenuation for θ = 0° are shown in Figure 1d. Solid lines represent results obtained in the noiseless simulation.
FIG. 4
FIG. 4
MRI of a representative isolated perfused heart. a–c: Images were obtained from a FLASH scan showing the gross morphology (a), and diffusion-weighted scans encoded in each of readout (horizontal) axis (b), and slice (in–out of page) direction (c). d: The latter is used to delineate the LV circumferential fiber region and bin the pixels into 4 approximately equal-size regions-of-interest according to the circumferential location and the relative absolute angle the fibers made with respect to the readout (horizontal) gradient direction.
FIG. 5
FIG. 5
Experimental diffusion-weighted signal intensities in a representative isolated heart. a,b: Normalized signal intensities over the selected ROIs (mean ± SD) measured for diffusion encoded parallel (a) and perpendicular (b) to the LV circumferential fiber direction are plotted on a semi-log scale. Inverted triangles and open squares correspond to low and normal flow settings, respectively. Solid and dotted lines represent IVIM (two-compartment) and mono-exponential fits, respectively, of the data points.
FIG. 6
FIG. 6
a–c: Anisotropy of experimentally observed IVIM parameters in isolated perfused hearts. The group averaged parameters (mean ± SEM; n = 7) are plotted as a function of the nominal relative encoding angle obtained as illustrated in Figure 4d. Entries with asterisks (*) are significantly different from values observed at |θ| = 0° for each flow setting according to one-way repeated-measures ANOVA, which include under normal flow at |θ| of 60° and 90°, and Dtissue under both normal and low flow at |θ| of 60° and 90°.
FIG. 7
FIG. 7
Arterial spin labeling MRI of a representative isolated heart. Slice-selective and nonselective FAIR images are used to obtain the falsecolor-coded longitudinal relaxation T1ss and T1ns maps, respectively, and subsequently the myocardial blood flow (MBF) map, for both normal and low flow settings.
FIG. 8
FIG. 8
Scatter plot between ASL-based MBF measurements in a ROI spanning whole heart slice (mean ± SEM) versus (aortic) input pressure in each heart. Open squares correspond to normal flow rate while inverted triangles correspond to low flow setting. Linear regression solid line and 95% confidence bands (dotted lines) are overlaid over the data points. Pearson correlation coefficient (r) is reported, along with the P value in top left corner, shows strong correlation between mean ASL-based MBF measurements and the recorded aortic pressures. Note that the SEM error bars are too small to be seen.
FIG. 9
FIG. 9
IVIM parameters graphed as functions of ASL-derived MBF for different encoding angles. Measurements obtained at both low (inverted triangles) and normal (open squares) flow setting are combined in the scatter plots. The Dapp∗ and the other parameters are fitted to a quadratic and straight (solid) line respectively, with the order of the polynomial indicated in the subscripts of the C coefficients. Error bars obtained from IVIM fitting are shown with each data point, and the 95% confidence interval dotted bands of the regression fit are also displayed. The coefficient of determination r2 is reported under each coefficient (with asterisks in 2nd row denote P < 0.05 for linear regression). Units of MBF is mL/min/100 g, Dapp∗ ×10−3 mm2/s, Dtissue normalized units, and VF is % of total volume.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J Chem Phys. 1965;42:288.
    1. Stejskal EO. Use of spin echoes in a pulsed magnetic-field gradient to study anisotropic, restricted diffusion and flow. J Chem Phys. 1965;43
    1. Moseley ME, Cohen Y, Mintorovitch J, Chileuitt L, Shimizu H, Kucharczyk J, Wendland MF, Weinstein PR. Early detection of regional cerebral ischemia in cats: comparison of diffusion- and T2-weighted MRI and spectroscopy. Magn Reson Med. 1990;14:330–3346. - PubMed
    1. Basser PJ, Mattiello J, LeBihan D. MR diffusion tensor spectroscopy and imaging. Biophys J. 1994;66:259–267. - PMC - PubMed
    1. Le Bihan D, Breton E, Lallemand D, Grenier P, Cabani E, Jeantet L. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology. 1986;161:401–407. - PubMed

Publication types

LinkOut - more resources