Towards in-vivo quantification of the optic nerves' microstructure with diffusion MRI
Poster No:
2187
Submission Type:
Abstract Submission
Authors:
Patryk Filipiak1,2, Kamri Clarke1,2, Zakia Gironda Ben Youss1,2, Mary Bruno1,2, David Tran3, Timothy Shepherd1,2, Dimitris Placantonakis4, Steven Baete1,2,5
Institutions:
1Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health, New York, NY, 2Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health, New York, NY, 3Hansjörg Wyss Department of Plastic Surgery, NYU Langone Health, New York, NY, 4Department of Neurosurgery, Kimmel Center for Stem Cell Biology, NYU Grossman School of Medicine, New York, NY, 5Vilcek Institute of Graduate Biomedical Sciences, NYU Grossman School of Medicine, New York, NY
First Author:
Patryk Filipiak
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Co-Author(s):
Kamri Clarke
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Zakia Gironda Ben Youss
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Mary Bruno
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Timothy Shepherd
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health
New York, NY|New York, NY
Dimitris Placantonakis
Department of Neurosurgery, Kimmel Center for Stem Cell Biology, NYU Grossman School of Medicine
New York, NY
Department of Neurosurgery, Kimmel Center for Stem Cell Biology, NYU Grossman School of Medicine
New York, NY
Steven Baete
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health|Vilcek Institute of Graduate Biomedical Sciences, NYU Grossman School of Medicine
New York, NY|New York, NY|New York, NY
Bernard and Irene Schwartz Center for Biomedical Imaging, NYU Langone Health|Center for Advanced Imaging Innovation and Research (CAI2R), NYU Langone Health|Vilcek Institute of Graduate Biomedical Sciences, NYU Grossman School of Medicine
New York, NY|New York, NY|New York, NY
Introduction:
Diffusion MRI (dMRI) enables in-vivo quantification of the central nervous system microstructure; however, the accuracy of this approach varies in different anatomical locations [1]. Probing the microstructure of optic nerves (ONs) is particularly challenging due to involuntary eye movements during dMRI acquisition [2] and Echo Planar Imaging (EPI) artifacts at the air-tissue interface near sinuses [3].
The goal of this preliminary study was to estimate the intra-axonal volume fraction, diffusivity coefficients, and the diameter of axons inside the ONs from Radial Diffusion Spectrum Imaging (RDSI) [4]. To assess the reproducibility of our results, we scanned each subject twice. During the second scan, we asked the participants to watch a video aiming to minimize the involuntary eye movements.
Our results show that in-vivo quantification of ONs can be feasible and reproducible. This approach could be applied in tracking progression of ON tissue degeneration in injuries, glaucoma, or acquired blindness [5].
The goal of this preliminary study was to estimate the intra-axonal volume fraction, diffusivity coefficients, and the diameter of axons inside the ONs from Radial Diffusion Spectrum Imaging (RDSI) [4]. To assess the reproducibility of our results, we scanned each subject twice. During the second scan, we asked the participants to watch a video aiming to minimize the involuntary eye movements.
Our results show that in-vivo quantification of ONs can be feasible and reproducible. This approach could be applied in tracking progression of ON tissue degeneration in injuries, glaucoma, or acquired blindness [5].
Methods:
We acquired dMRI of 4 healthy subjects (1 female, 32±5 y.o.) at 2mm isotropic resolution, single-shot EPI, TE/TR=74/5800ms, 60 diffusion encoding directions sampled at RDSI radial lines [4] with b=250,1000,2250,4000s/mm², interleaved with 17 images at b=0. Each scanning session comprised two identical parts, i.e., test and retest, separated with a 15-min. break. During the retest, the participants were watching a video of their choice.
Our postprocessing in MRtrix3 [6] included denoising, Gibbs ringing removal, correction of B1 field inhomogeneity and eddy currents. Next, we virtually dissected the optic pathways in DSI-Studio [7] (Fig. 1) and fitted the dMRI signal along the streamlines to the Ball&Cylinder [8] microstructure model in dmipy [9]. For the latter, we employed the Van Gelderen axonal model [10] due to its realistic assumption of Gaussian diffusion during the dMRI gradient pulse.
We assessed the plausibility of our estimated microstructure parameters through comparison with earlier studies in the brain white matter. Also, we computed the two-sided statistical t-test with the significance level α=0.05 between the aggregated test and retest measurements to verify the reproducibility of our procedure and potential motion reduction due to video watching during acquisition.
Our postprocessing in MRtrix3 [6] included denoising, Gibbs ringing removal, correction of B1 field inhomogeneity and eddy currents. Next, we virtually dissected the optic pathways in DSI-Studio [7] (Fig. 1) and fitted the dMRI signal along the streamlines to the Ball&Cylinder [8] microstructure model in dmipy [9]. For the latter, we employed the Van Gelderen axonal model [10] due to its realistic assumption of Gaussian diffusion during the dMRI gradient pulse.
We assessed the plausibility of our estimated microstructure parameters through comparison with earlier studies in the brain white matter. Also, we computed the two-sided statistical t-test with the significance level α=0.05 between the aggregated test and retest measurements to verify the reproducibility of our procedure and potential motion reduction due to video watching during acquisition.
Results:
The intra-axonal volume fraction p1 oscillated around 0.27-0.29, dropping below 0.20 on both ends of the ONs – in proximity to the globe and near the optic chiasm, respectively (Fig. 1&2). Consequently, the extra-axonal volume fraction piso demonstrated the exact opposite behavior. The intra-axonal diffusivity Da remained at the stable yet relatively low average level 0.52–0.65 ×10-9m²/s, whereas the extra-axonal diffusivity Diso≈2.6×10-9m²/s was close to the values reported for the brain white matter [11]. The estimated axon diameter of approx. 10-20µm was an order of magnitude higher than the values reported in histology [12], which conforms with the systemic overestimation observed in dMRI-based methods [13,14]. Finally, the t-test showed no significant differences between the test and retest measurements regardless of the video watching during acquisition (p≫0.05, Fig. 2).
·Fig. 1: 3D maps of the intra-axonal volume fraction (a) and the axon diameter (b) estimated along the optic pathways.
·Fig. 2. Means and standard deviations of the microstructure parameters estimated along the optic nerves.
Conclusions:
Our RDSI acquisition protocol and the assumed Ball&Cylinder diffusion model were sufficient for reproducible quantification of the human ON microstructure parameters in vivo. The observed variability of the intra-axonal fraction was likely attributed to the partial volume effect at the extremities of the ONs. Another limitation of our study was the simplicity of the diffusion model which led to underestimation of the intra-axonal diffusivity. Nonetheless, the numerical stability of our estimates suggests that dMRI-based microstructure quantification may be considered for probing the integrity of ONs. Future work should include disease controls and potential extensions of the diffusion model.
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 2
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
White Matter Anatomy, Fiber Pathways and Connectivity 1
Novel Imaging Acquisition Methods:
Diffusion MRI
Keywords:
Acquisition
Modeling
MRI
Nerves
OPTICAL
STRUCTURAL MRI
Tractography
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - optic nerves, reproducibility
1|2Indicates the priority used for review
Provide references using author date format
[2] Hoch et al. (2017), 'Advanced MRI of the optic nerve', Neuro-Ophthalmology, 37(2), 187-196.
[3] Neufeld et al. (2005), 'Susceptibility-matched envelope for the correction of EPI artifacts', Magnetic Resonance Imaging, 23(9), 947-951.
[4] Baete et al. (2016), 'Radial q‐space sampling for DSI', Magnetic resonance in medicine, 76(3), 769-780.
[5] Mendoza et al. (2022), 'Advanced Diffusion MRI of the Visual System in Glaucoma: From Experimental Animal Models to Humans', Biology, 11(3), 454.
[6] Tournier et al. (2019), 'MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation', NeuroImage, vol. 202, pp. 116137
[7] Yeh et al. (2013), 'Deterministic diffusion fiber tracking improved by quantitative anisotropy', PLoS ONE, vol. 8(11), pp. e80713
[8] Assaf et al. (2004), 'New modeling and experimental framework to characterize hindered and restricted water diffusion in brain white matter', Magnetic Resonance in Medicine, 52(5), 965-978.
[9] Fick et al. (2019), 'The dmipy toolbox: Diffusion MRI multi-compartment modeling and microstructure recovery made easy', Frontiers in Neuroinformatics, 13, 64.
[10] Van Gelderen et al. (1994). 'Evaluation of restricted diffusion in cylinders. Phosphocreatine in rabbit leg muscle', Magnetic Resonance, Series B, 103(3), 255-260.
[11] Dhital et al. (2019), 'Intra-axonal diffusivity in brain white matter', NeuroImage, 189, 543-550.
[12] Mikelberg et al. (1989), 'The normal human optic nerve: axon count and axon diameter distribution', Ophthalmology, 96(9), 1325-1328.
[13] Alexander et al. (2010), 'Orientationally invariant indices of axon diameter and density from diffusion MRI', NeuroImage, 52(4), 1374-1389.
[14] Veraart et al. (2020), 'Noninvasive quantification of axon radii using diffusion MRI', eLife, 9, e49855.