1774

Clinically Feasible Optic Nerve Diffusion Basis Spectrum Imaging at 3T
Joo-won Kim1,2,3, Peng Sun4, Sheng-Kwei Song4, Samantha Lancia5, Courtney Dula5, Robert T Naismith5, and Junqian Xu1,2,3,6

1Translational and Molecular Imaging Institute, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 2Department of Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 3Graduate School of Biomedical Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 4Department of Radiology, Washington University, Saint Louis, MO, United States, 5Department of Neurology, Washington University, Saint Louis, MO, United States, 6Department of Neuroscience, Icahn School of Medicine at Mount Sinai, New York, United States

Synopsis

Optic nerve MRI is susceptible to eyeball movement. The relatively long acquisition time of advanced diffusion MRI (dMRI) methods exacerbates the motion sensitivity in optic nerve dMRI and limits the clinical implementation of these methods. In this work, we evaluate a short (less than 2.5 min per eye) single slice coronal optic nerve dMRI acquisition protocol at 3T and propose a 2D optic nerve center searching algorithm customized for such dMRI data. We demonstrate improved optic nerve center contrast after image alignment and the expected benefits of reduced partial volume effects from diffusion basis spectrum imaging (DBSI) analysis.

Purpose

Advanced diffusion MRI (dMRI) acquisition and signal modeling, such as diffusion basis spectrum imaging (DBSI)1, hold the promise to offer microstructural information specific to axonal degeneration that is not accessible with conventional dMRI methods, such as diffusion tensor imaging (DTI). However, these advanced dMRI methods typically require relatively dense q-space sampling. Consequently, the longer dMRI acquisition time limits integration into routine clinical neuroradiology protocol. The challenge of shortening the dMRI acquisition time is especially salient for clinical optic nerve MRI protocol, because of the unavoidable eyeball movement-induced non-linear optic nerve movement2 (Fig. 1) during the dMRI acquisition and the competing requirement of high resolution to resolve the small optic nerve. Here, we present a short clinically-feasible (less than 2.5 min for each eye) optic nerve dMRI acquisition protocol at 3T and the associated optic nerve center alignment procedure, and demonstrate sufficient data quality for DBSI modeling.

Methods

MRI: Optic nerve dMRI data were acquired in five healthy adult subjects on a 3T scanner (Trio, Siemens) with a 32-ch head coil (12 anterior receive elements only). An inner-volume-imaging spin echo EPI diffusion sequence modified from our previous implementation3 (Fig. 2) was used to acquire reduced field-of-view (FOV) optic nerve images at 1.3 mm × 1.3 mm in-plane resolution, FOV=166 mm x 41.5 mm, matrix=128 x 32, single 3 mm thick slice (coronal slice positioned perpendicular to and at the center of the intraorbital optic nerve, Fig. 3), phase-encoding (PE) direction=foot->head, 6/8 partial Fourier, echo spacing=0.87 ms, ETL=24 (20.9 ms), bandwidth=1346 Hz/Px, TR/TE=2500/56 ms, monopolar diffusion encoding, optimized 25 multi-bval (linearly spaced) multi-bvec diffusion scheme with bmax=1000 s/mm2 and two b0, two effective averages with alternating diffusion gradient polarity, Tacq=2 min 20 sec per eye. The two sequence modifications included (i) implementation of hyperbolic secant (HS) adiabatic inversion/refocusing pulses for improved slab profile in the PE direction and (ii) placement of the second inversion/refocusing pulse immediately before the excitation pulse to reduce the time interval between the two inversion/refocusing pulses.

Optic nerve center alignment: A manual mask for all dMRI frames was drawn including the optic nerve but excluding extraocular muscles for image alignment (Fig. 4, cyan color). On each frame, the optic nerve center was defined on 0.1 voxel unit space (spline interpolation) within the mask region by finding a square 5×5 (voxel) kernel whose sum of image intensities was maximal (Fig. 4, red color). The center of the 5×5 kernel was assigned as the optic nerve center and the image intensity was resampled to the original voxel resolution. The resulting 5×5 kernel from each dMRI frame were concatenated for diffusion signal modeling.

Diffusion analysis: DBSI/DTI maps were generated using an in-house MATLAB program. The region-of-interest (ROI) of the optic nerve was the center voxel of the 5×5 kernel.

Results

Sufficient optic nerve contrast to the surrounding suppressed fat signal allows definition of the optic nerve (including CSF and dura) for every dMRI frame (Fig. 4). The optic nerve center is hypointense compared to the surrounding CSF on the mean b0 frames of the registered optic nerve (Fig. 4, b0 mean), while the optic nerve center is hyperintense on the mean of diffusion-weighted (DW) frames (Fig. 4, DW mean). The DBSI maps also showed the expected contrast of optic nerve center (e.g., Fig. 4, fiber ratio). The DBSI results at the optic nerve center were comparable to our previous studies in mouse optic nerve4 and human brain white matter5 and demonstrated the expected robustness to partial volume effects as compared to DTI results (Table 1).

Discussion

Partial volume effects from the surrounding CSF is a significant confound in optic nerve dMRI (even at 1.3 mm isotropic resolution as we have previously shown), especially for atrophic optic nerves due to chronic neurodegeneration6,7. Advanced dMRI methods capable of modeling multiple tissue compartments, such as DBSI, can alleviate CSF contamination to offer more specific and less biased microstructure parameters. However, clinical applicability of optic nerve DBSI has been hampered by higher q-space sampling and data quality requirements. We addressed this limitation by implementing a short single slice coronal optic nerve dMRI protocol with specific optimization in dMRI acquisition and optic nerve center alignment. Compared to previous coronal optic nerve dMRI studies8,9, our individualized optic nerve acquisition allows slice positioning perpendicular to each nerve for better optic nerve center definition using our proposed searching algorithm.

Conclusion

High resolution optic nerve DBSI with sufficient data quality can be achieved with a short (<2.5 min) single slice coronal optic nerve dMRI protocol at 3T for clinical neuroradiology integration.

Acknowledgements

NIH/NINDS R21NS090910 and NIH/NEI U01EY025500

References

1. Wang, Y. et al, Quantification of increased cellularity during inflammatory demyelination, Brain, 2011; 134(12) 3587-3598.

2. Kim, J.-W. et al, Non-linear Distortion Correction in Human Optic Nerve Diffusion Imaging, Proceedings of 24th International Society for Magnetic Resonance in Medicine. 2016; poster #2054.

3. Xu, J. et al, Assessing optic nerve pathology with diffusion MRI: from mouse to human, NMR Biomed. 2008; 21(9):928-40.

4. Wang, X. et al, Diffusion basis spectrum imaging detects and distinguishes coexisting subclinical inflammation, demyelination and axonal injury in experimental autoimmune encephalomyelitis mice, NMR Biomed. 2014; 27(7) 843-852.

5. Wang, Y. et al, Differentiation and quantification of inflammation, demyelination and axon injury or loss in multiple sclerosis, Brain. 2015; 138(Pt 5):1223-1238.

6. Naismith, R. T. et al, Radial diffusivity in remote optic neuritis discriminates visual outcomes, Neurology. 2010; 74(21):1702-1710.

7. Chang, S. T. et al, Optic nerve diffusion tensor imaging parameters and their correlation with optic disc topography and disease severity in adult glaucoma patients and controls, J Glaucoma. 2014; 23(8):513-520.

8. Wheeler-Kingshott, C. A. et al, ADC mapping of the human optic nerve: increased resolution, coverage, and reliability with CSF-suppressed ZOOM-EPI, Magn Reson Med. 2002; 47(1):24-31.

9. Hickman, S. J. et al, Optic nerve diffusion measurement from diffusion-weighted imaging in optic neuritis, AJNR Am J Neuroradiol. 2005; 26(4):951-956.

Figures

Figure 1. Representative coronal (A, B) and axial (C, D) acquisitions of optic nerve dMRI frames to show the effects of eyeball movement on the optic nerve dMRI image, with an illustration of the optic nerve (E) describing eyeball movement and the resulting non-linear optic nerve movement (red arrows). Red- and yellow- colored regions (C, D) are optic nerve segmentation from C and D, respectively.

Figure 2. Sequence diagram of our previously published inner-volume-imaging optic nerve dMRI acquisition with modification to incorporate adiabatic RF pulses.

Figure 3. Axial structural MRI localizer for dMRI slice positioning (A) and single slice optic nerve dMRI b0 image from one of the eyes (B). Red lines (A) indicate the optic nerves and blue lines (A) indicate the dMRI slice location.

Figure 4. Illustration of the customized 2D optic nerve center registration (left panel) and the results after image registration (right panel). Three representative dMRI frames (first column) and the 5x5 optic nerve region (second column, same as the red squares in the first column). Cyan-colored region (first column) is the manual mask over all frames. The top-left plot in the right panel illustrates the optic nerve center (center voxel, white), partial-volumed optic nerve with more (dark blue) or less (light blue) CSF, and saturated fat (black). The nerve regions were outlined (blue) in the DBSI and DTI maps.

Table 1. Optic nerve diffusion basis spectrum imaging (DBSI) and diffusion tensor imaging (DTI) measurements of five healthy subjects. For each subject, measurements from the left and right eyes were averaged. FA: fractional anisotropy, RD: radial diffusivity, AD: axial diffusivity, MD: mean diffusivity.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)
1774