1231

On Probing Intravoxel Incoherent Motion in the Heart using Spin Echo versus Stimulated Echo Diffusion Weighted Imaging

Georg Ralph Spinner¹, Christian Torben Stoeck¹, Linda Mathez¹, Constantin von Deuster¹, Christian Federau¹, and Sebastian Kozerke¹

¹Institute for Biomedical Engineering, ETH Zurich, Zurich, Switzerland

Synopsis

Intravoxel Incoherent Motion (IVIM) imaging in the heart remains challenging resulting in large variation of the perfusion surrogates in practice. In the present work, IVIM sensitivity of standard and motion-compensated Spin Echo (SE) and Stimulated Echo Acquisition Mode (STEAM) diffusion-weighted imaging approaches are analyzed using Monte Carlo simulations and perfused porcine heart experiments. An extended IVIM model is proposed to account for the microstructural properties of the myocardium. It is demonstrated that motion-compensated SE sequences provide insufficient perfusion sensitivity. In contrast, STEAM allows delineating hypoperfused myocardium using experimental imaging data.

Introduction

The Intravoxel Incoherent Motion (IVIM)¹ model has been used to assess perfusion surrogates in the heart of healthy volunteers^2–4, patient cohorts^5,6 and in animal models (pigs⁷, dogs⁸). However, IVIM imaging in the heart remains challenging despite technical advances of cardiac diffusion imaging^9–11. In practice, the IVIM perfusion parameters have a relatively high variation compared to diffusion estimates due to the relatively small influence of perfusion on the magnitude signal and the inherent non-linearity of the model. Accordingly, the assumptions of the IVIM model such as anisotropic microcirculation and sufficient directional changes of spins undergoing perfusion require further investigation. Recent cardiac diffusion acquisition strategies using first- and second-order motion compensated spin-echo (SE) sequences (M1¹², M²⁹) prompt for careful considerations concerning their perfusion sensitivity compared to the conventional Stejskal-Tanner experiment (M0¹³).

It is the objective of the present work to systematically investigate IVIM sensitivity of Spin Echo (SE) and Stimulated Echo Acquisition Mode (STEAM) diffusion-weighted imaging approaches using Monte Carlo simulations and perfused porcine hearts.

Theory & Methods

The concept of normalized phase distributions¹⁴ in Equation 1 (trajectory $$${\bf{x}} \left( t \right)$$$, gradient profile $$${{\bf{g}}_h}\left( t \right)$$$ and velocity $$${{\bf{v}}} \left( t \right)$$$):

$$\phi = \gamma \int\limits_0^T {\bf{x}} \left( t \right) \cdot {{\bf{g}}_h}\left( t \right)dt = - \gamma \int\limits_0^T {\left( {\int\limits_0^t {{{\bf{g}}_h}\left( {t'} \right)dt'} } \right) \cdot {\bf{v}}} \left( t \right)dt \qquad (1)$$

and $$$\phi = v\sqrt {bT} {\vartheta _h}$$$ is used in simulations (Figure 1) and extended by histological findings: the capillary segments¹⁵ are scattered along a main direction using a von Mises distribution^16–18. The segment lengths follow a Weibull distribution¹⁶ and flow velocity is set to 1 mm/s^8,19. Magnitude attenuation and phase is determined using 10⁵ simulated spins via Equation 2:

$$F = \left\langle {\exp \left( {i\phi } \right)} \right\rangle \qquad(2)$$

Simulation parameters are: total gradient duration $$$T$$$=105.30ms (SE), 60/44s=1363.60ms (STEAM), gradient lobe duration $$$\delta$$$=10.40/21.25/15.04/3.40ms for M0/M1/M2/STEAM. Zenith angle concentration $$$\kappa$$$=2.5^20,21, maximum zenith angle 0.5223π (corresponds to 6% back-flow¹⁷), gradient azimuth angle zero and length mean/standard deviation $$${\mu _l}/{\sigma _l}$$$=60/40μm¹⁶. A Gaussian Phase Approximation (GPA) is derived after correcting for a non-vanishing net phase.

For the ex-vivo experiment a porcine heart was perfused with 0.9% NaCl solution with a peristaltic pump (44rpm, 340ml/min) imaged on a clinical 1.5T MRI scanner (Philips Achieva, Best, the Netherlands) with two single-loop surface coils (Philips SENSE Flex-M, Best, the Netherlands). The scan parameters were: 230x110mm² FOV, 2.5x2.5x10mm³ resolution, 43 EPI lines single-shot, Local-Look reduced FOV technique²², 3x60/44s=4.091s TR, 148.40ms TE, total gradient durations as above, SPIR fat suppression and triggering to pump rotation (trigger delay 1000ms). The acquired b-values were 10, 25, 50, 100, 150, 200, 250, 300, 400, 600, 800 and 1000s/mm². Six optimized²³ gradient directions were acquired in parallel and anti-parallel direction with two averages. Baseline scans were acquired after deactivating the pump and a delay of 5 minutes. An occlusion experiment was performed by pushing the cannulating tube from the left main artery into the left anterior descending artery (LAD) thereby blocking the circumflex coronary artery (LCX).

Parameter fitting was performed using a two-tensor model (Equation 3) to account for anisotropy (diffusion tensor $$$D$$$, perfusion fraction $$$f$$$, pseudo-diffusion tensor $$$D^*$$$, gradient vector $$${\bf{g}}$$$ and b0 magnitude $$${S_0}$$$):

$$S\left( {b,{\bf{g}}} \right) = {S_0}\left[ {f \cdot \exp \left( { - b{{\bf{g}}^T}{D^*}{\bf{g}}} \right) + \left( {1 - f} \right) \cdot \exp \left( { - b{{\bf{g}}^T}D{\bf{g}}} \right)} \right] \qquad(3)$$

Results

Simulations in Figure 2 indicate that M1 and M2 sequences do not provide sufficient perfusion sensitivity. M0 and STEAM yield magnitude attenuation below 0.5 for b=1000s/mm² for directions offset from the main capillary direction. While a net phase is observed for both sequences, no net phase is detected for M1 and M2 sequences. The GPA holds only for STEAM, while an overestimation of the magnitude attenuation results with SE protocols.

The perfused heart experiments show only minute differences between baseline and perfusion for SE magnitudes (Figure 3) and parameter estimates (Figure 4). In contrast, STEAM yields noticeable differences between baseline and perfusion including a detectable net phase (Figure 3). Using STEAM the area affected by blocking the LCX artery is successfully delineated (Figure 5).

Discussion & Conclusion

Simulations and experimental data presented here demonstrate insufficient IVIM sensitivity of first- and second-order moment compensated (M1, M2) SE sequences while STEAM provides sufficient perfusion sensitivity allowing to delineate ischemic tissue territories even with reduced number of sampled b-values and directions.

It is therefore concluded, that studies of IVIM of in-vivo myocardium shall deploy STEAM based diffusion encoding approaches along with using an extended IVIM model to account for specific myocardial characteristics.

Acknowledgements

The authors would like to thank Dr. Andreas Wetscherek for the helpful discussions about IVIM simulations and Stephen Wheeler for help in pump maintenance. We would also like to thank the team of veterinary doctors for providing the porcine hearts: Dr. Nikola Cesarovic, Dr. Marko Canic, Dr. Miriam Lipiski, Dr. Thea Fleischmann and Dr. Mareike Sauer.

References

1. Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E, Laval-Jeantet M. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology. 1986;161(2):401-407. doi:10.1148/radiology.161.2.3763909.

2. Moulin KK, Croisille P, Feiweier T, et al. In vivo free-breathing DTI and IVIM of the whole human heart using a real-time slice-followed SE-EPI navigator-based sequence: A reproducibility study in healthy volunteers. Magn Reson Med. 2016;76(1):70-82. doi:10.1002/mrm.25852.

3. Spinner GR, von Deuster C, Tezcan KC, Stoeck CT, Kozerke S. Bayesian intravoxel incoherent motion parameter mapping in the human heart. J Cardiovasc Magn Reson. 2017;19(1):85. doi:10.1186/s12968-017-0391-1.

4. Xiang S, Zhang X, Yang S-J, et al. STROBE-A preliminary investigation of IVIM-DWI in cardiac imaging. Medicine (Baltimore). 2018;97(36):e11902. doi:10.1097/MD.0000000000011902.

5. Mou A, Zhang C, Li M, et al. Evaluation of myocardial microcirculation using intravoxel incoherent motion imaging. J Magn Reson Imaging. 2017;46(6):1818-1828. doi:10.1002/jmri.25706.

6. Spinner G, Deuster C Von, Oebel S, Stoeck CT, Manka R. Intravoxel Incoherent Motion Model in the Heart of Patients under Adenosine Induced Stress In order to validate previous IVIM animal studies , patients where measured both during rest and under adenosine induced stress using a slice following second-order. 2016:2016-2018.

7. Deuster C Von, Stoeck CT, Wissmann L, et al. Verification of the intra-voxel incoherent motion (IVIM) model in the porcine heart. In: Proc. Intl. Soc. Mag. Reson. Med. Vol 23. ; 2015:2612.

8. Callot V, Bennett E, Decking UKM, Balaban RS, Wen H. In vivo study of microcirculation in canine myocardium using the IVIM method. Magn Reson Med. 2003;50(3):531-540. doi:10.1002/mrm.10568.

9. Stoeck CT, von Deuster C, Genet M, Atkinson D, Kozerke S. Second-order motion-compensated spin echo diffusion tensor imaging of the human heart. Magn Reson Med. 2016;75(4):1669-1676. doi:10.1002/mrm.25784.

10. Scott AD, Nielles-Vallespin S, Ferreira PF, et al. An in-vivo comparison of stimulated-echo and motion compensated spin-echo sequences for 3 T diffusion tensor cardiovascular magnetic resonance at multiple cardiac phases. J Cardiovasc Magn Reson. 2018;20(1):1-15. doi:10.1186/s12968-017-0425-8.

11. Nguyen C, Fan Z, Sharif B, et al. In vivo three-dimensional high resolution cardiac diffusion-weighted MRI: a motion compensated diffusion-prepared balanced steady-state free precession approach. Magn Reson Med. 2014;72(5):1257-1267. doi:10.1002/mrm.25038.

12. Gamper U, Boesiger P, Kozerke S. Diffusion imaging of the in vivo heart using spin echoes--considerations on bulk motion sensitivity. Magn Reson Med. 2007;57(2):331-337. doi:10.1002/mrm.21127.

13. Stejskal EO, Tanner JE. Spin Diffusion Measurements: Spin Echoes in the Presence of a Time‐Dependent Field Gradient. J Chem Phys. 1965;42(1):288-292. doi:10.1063/1.1695690.

14. Wetscherek A, Stieltjes B, Laun FB. Flow-compensated intravoxel incoherent motion diffusion imaging. Magn Reson Med. 2015;74(2):410-419. doi:10.1002/mrm.25410.

15. Le Bihan D, Turner R. The capillary network: a link between IVIM and classical perfusion. Magn Reson Med. 1992;27(1):171-178.

16. Kaneko N, Matsuda R, Toda M, Shimamoto K. Three-dimensional reconstruction of the human capillary network and the intramyocardial micronecrosis. Am J Physiol Heart Circ Physiol. 2011;300(3):H754-H761. doi:10.1152/ajpheart.00486.2010.

17. Kassab GS, Fung YC. Topology and dimensions of pig coronary capillary network. Am J Physiol. 1994;267(1 Pt 2):H319-25. doi:10.1152/ajpheart.1994.267.1.H319.

18. Karampinos DC, King KF, Sutton BP, Georgiadis JG. Intravoxel partially coherent motion technique: Characterization of the anisotropy of skeletal muscle microvasculature. J Magn Reson Imaging. 2010;31(4):942-953. doi:10.1002/jmri.22100.

19. Fibich G, Lanir Y, Liron N. Mathematical model of blood flow in a coronary capillary. Am J Physiol Hear Circ Physiol. 1993;265(5):H1829-1840. doi:10.1021/ic0616434.

20. Poole DC, Mathieu-Costello O. Analysis of capillary geometry in rat subepicardium and subendocardium. Am J Physiol - Hear Circ Physiol. 1990;259(1):H204-H210.

21. Mathieu O, Cruz‐Orive L ‐M, Hoppeler H, Weibel ER. Estimating length density and quantifying anisotropy in skeletal muscle capillaries. J Microsc. 1983;131(2):131-146. doi:10.1111/j.1365-2818.1983.tb04240.x.

22. Buecker A, Adam G, Neuerburg JM, Vaals JJ Van, Guenther RW. MR-guided biopsy using a T2- weighted single-shot zoom imaging sequence (local look technique ). J Magn Reson Imaging. 1998;8:955-959.

23. Jones DK, Horsfield M a, Simmons a. Optimal strategies for measuring diffusion in anisotropic systems by magnetic resonance imaging. Magn Reson Med. 1999;42(3):515-525.

Figures

Figure 1: Illustration of myocardial perfusion simulation. Capillary segments (red) follow myocytes (purple). The start positions (black circles, velocity vectors) are uniformly distributed on the initial segment, the segment length follows a Weibull distribution and the segment angulation is parameterized using the azimuth and zenith angles: the former assumes a uniform distribution while the latter follows a von Mises distribution. The endpoints on the M-th segment are denoted as black dots. The principal myocyte direction is indicated as dark green arrow (eigenvector corresponding to the largest diffusion tensor eigenvalue), the diffusion encoding gradient direction is represented by the light green arrow.

Figure 2: Normalized phase distributions, magnitude attenuation and phases. Three gradient (zenith) angles are simulated: parallel to the main capillary direction, 45° offset and orthogonal to it. Top row: M0-SE and STEAM sequences exhibit non-vanishing net phases, Gaussian modelling approximates the distributions for STEAM and M0-SE for orthogonal gradients, distributions created by M1-SE and M2-SE sequences approach Laplace distributions. Middle row: dashed lines are magnitude attenuations calculated from a Gaussian Phase Approximation (GPA), which hold only for STEAM. Bottom row: M0-SE/STEAM produce a max. phase of approx. 3/10 π respectively.

Figure 3: Top row: ROI-averaged magnitudes; the dotted lines help guide the eye to visualize differences. Middle row: parameter maps on top of the DWIs for STEAM. Bottom row: the phase derived from subtracting the anti-parallel from the parallel direction is shown as ROI-averaged values vs. b-values and as maps for b=1000 s/mm² with arrows indicating the LCX branch vessels. A phase-contrast flow-sensitive scan is shown (values smaller 0.5 cm/s are excluded for noise suppression) as reference with an arrow indicating the LCX vessels at the same location in the heart.

Figure 4: Ex-vivo IVIM parameters as mean±standard deviation. The asterisk * indicates p<0.05 between the baseline and perfusion scans (Wilcoxon rank sum test).

Figure 5: Top row: DTI and IVIM parameters exhibit locally reduced apparent diffusion and perfusion parameters (white arrows) if all data or only a subset is used. The corresponding histograms show increased counts of reduced parameters (black arrows). Bottom row: the phase of STEAM and the flow-sensitive phase contrast scan show no presence of perfused LCX vessels, but perfusion in interventricular vessels (green & red arrows). Velocities below 0.5 cm/s are not shown for noise suppression. The corresponding magnitude image also reveals perfusion in both the anterior and posterior interventricular vessels, indicated by signal elevation (white arrows).

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)

1231