4584

Correlation analysis between quantitative-T2 and diffusion-MRI measurements in human white matter: a study across different field strengths.
Elisa Marchetto1,2, Sebastian Flassback1, Patricia Johnson1, Jakob Assländer1, and Jelle Veraart1
1Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States, 2Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States

Synopsis

Keywords: Quantitative Imaging, Quantitative Imaging

Motivation: Quantitative MRI is increasingly translated to low (<=0.55T) and high (>=7T) field strengths. Yet, differences in the underlying biophysics across field strengths remain poorly understood.

Goal(s): To identify and understand deviations in the correlation of T2 and diffusion metrics at extreme field strengths.

Approach: Quantitative T2 and diffusion MRI maps were computed from a comprehensive MRI protocol at 0.55T, 1.5T, 3T, and 7T.

Results: We observe decreased correlation between T2 relaxation and diffusivity at high field strengths, while the results at 0.55T align well with clinical field strengths, which shows the ground for optimization of experimental design at low- and ultra-low fields.

Impact: T2 and diffusion-MRI commonly complement each other in clinical imaging. Our results show a similar correlation of related metrics, suggesting shared information on tissue properties, which can be used to optimize experimental design at low- and ultra-low fields.

Introduction

Quantitative MRI is important for developing biomarkers for the early detection and monitoring of neurological diseases and for evaluating drug treatments. Quantitative MRI models are most widely used on MRI scanners with clinical field strengths (1.5T–3T), but their interpretation remains under-studied at lower (0.55T) or higher (7T) field strength1. In this study, we analyze the relation between T2 relaxation and diffusion across different field strengths and show the premises and advantages of performing quantitative MRI on a wide range of field strengths.

Methods

A comprehensive MRI protocol was designed for 0.55T, 1.5T, 3T, and 7T systems, and tested on one healthy subject. The protocol includes:
  1. MP-RAGE2.
  2. Diffusion-weighted imaging (DWI) using a 2D-EPI sequence with 2 mm isotropic resolution. The repetition times ranged between 5.5 and 6.8 seconds across field strengths, the echo time was 73 ms, and the b-values were 10x0, 60x500, and 60x1000 s/mm².
  3. T2 mapping, employing a 2D multi-echo spin-echo3,4 sequence with 2 mm isotropic resolution. 10 ms echo-spacing was adopted consistently across all field strengths.
The diffusion data was pre-processed with the DESIGNER pipeline5. Denoising6 was applied only to the multi-SE data acquired at 0.55T. Image registration was performed to align the data with the 3T MP-RAGE data using ANTs7 with rigid transformation.
Data analysis involved:
  1. T2 mapping using EMC dictionary matching8.
  2. Diffusion tensor imaging and extraction of Fractional Anisotropy (FA) and Mean Diffusivity (MD) maps.
  3. Region of interest (ROI) analysis using the Johns Hopkins University (JHU) atlas9,10,11 segmentation with diffeomorphic registration through ANTs12. The fiber orientation within each voxel was calculated using peak extraction following constrained spherical deconvolution13,14. ROI-average fiber orientation w.r.t B0 (𝜃) was computed from the resulting dyadic tensor15, 16.

Results

Example quantitative maps are shown in Fig. 1 for all field strengths. A white matter ROI analysis suggests increased variability of R2 at higher B0 fields (Fig. 2A). We observed a weaker correlation between R2 and the mean diffusivity at 7T. In contrast, the correlation coefficient does not further increase when moving from 1.5 or 3T to 0.55T. Regardless of the field strength, we did not find a significant correlation between R2 and the fiber orientation 𝜃, i.e., the angle between the fibers and B0 (Fig. 2B). In Fig. 3, we show the linear regression of R2, with and without regressing out the mean diffusivity and 𝜃. The results confirm the above-described findings: the data suggests that the mean diffusivity is a strong predictor of R2, while the fiber orientation is not (Fig. 3C).

Discussion

The effect of B0 on the correlation between T2 relaxation and diffusion, shown in Figs. 2–4, is in line with previous literature17,18. The slight reduction of diffusivity and R2 at 0.55T (Fig. 2 and 4) are potentially caused by a Rician bias that will require a tailored correction19. We hypothesize that the reduced MD and increased FA at 7T compared to 1.5T and 3T (Fig. 4) result from the presence of multiple tissue compartments in the white matter, including intra- and extra-axonal spaces20. However, we found no significant dependence of R2 to the fiber orientation 𝜃 (Fig. 2B), which suggests that susceptibility effects, e.g. due to iron concentrations, could be the major factor in our R2 differences across B017. The strong correlation between R2 and MD at lower-field strength (0.55T), shows that quantitative T2 and diffusion-MRI encode shared information on white matter properties. Gaining more insight into the relation between the two metrics will be important to optimize the experimental design of low-field and ultra-low field MRI (<0.1T) experiments, thereby advancing reproducible quantitative measurements in a feasible scan time across field strengths. While we are collecting additional data, we here show the premise of using cross-scanner evaluations of quantitative T2 and diffusion-MRI metrics to improve understanding of quantitative models.

Acknowledgements

This work was performed under the rubric of the Center for Advanced Imaging Innovation and Research (CAI2R, www.cai2r.net), an NIBIB National Center for Biomedical Imaging and Bioengineering (NIH P41 EB017183).

References

  1. Hori M, Hagiwara A, Goto M, Wada A, Aoki S. Low-Field Magnetic Resonance Imaging: Its History and Renaissance. Invest Radiol. 2021 Nov 1;56(11):669-679.
  2. Brant-Zawadzki M, Gillan GD, Nitz WR. MP RAGE: a three-dimensional, T1-weighted, gradient-echo sequence--initial experience in the brain. Radiology. 1992 Mar;182(3):769-75.
  3. Carr H. Y. and Purcell E. M. Effects of Diffusion on Free Precession in Nuclear Magnetic Resonance Experiments. Phys. Rev. 94, 630.
  4. Meiboom S, Gill D. Modified Spin-Echo Method for Measuring Nuclear Relaxation Times. Rev Sci Instrum. 1958;29(8):688–691.
  5. Ades-Aron B, Veraart J, Kochunov P, et al. Evaluation of the accuracy and precision of the diffusion parameter EStImation with Gibbs and NoisE removal pipeline. Neuroimage. 2018;183:532-543.
  6. Veraart J, Novikov DS, Christiaens D, et al. Denoising of diffusion MRI using random matrix theory. Neuroimage. 2016;142:394-406.
  7. Avants BB, Tustison NJ, Stauffer M, Song G, Wu B, Gee JC. The Insight ToolKit image registration framework. Front Neuroinform. 2014;8:44.
  8. Ben-Eliezer N, Sodickson DK, Block KT. Rapid and accurate T2 mapping from multi-spin-echo data using Bloch-simulation-based reconstruction. Magn Reson Med. 2015;73(2):809-17.
  9. Mori et al., MRI Atlas of Human White Matter. Elsevier, Amsterdam, The Netherlands 2005.
  10. Wakana et al., Reproducibility of quantitative tractography methods applied to cerebral white matter. NeuroImage. 2007; 36:630-644.
  11. Hua et al., Tract probability maps in stereotaxic spaces: analysis of white matter anatomy and tract-specific quantification. NeuroImage. 2008; 39(1):336-347.
  12. Avants BB, Epstein CL, Grossman M, Gee JC. Symmetric diffeomorphic image registration with cross-correlation: evaluating automated labeling of elderly and neurodegenerative brain. Med Image Anal. 2008;12(1):26-41.
  13. C.M.W. Tax, B. Jeurissen, S.B.Vos, M.A. Viergever, et al. Recursive calibration of the fiber response function for spherical deconvolution of diffusion MRI data. NeuroImage.2014; 86; 67–80.
  14. J.-D. Tournier, F. Calamante, and A. Connelly. Robust determination of the fibre orientation distribution in diffusion MRI: non-negativity constrained super-resolved spherical deconvolution. Neuroimage. 2007; 35;1459–72.
  15. Basser PJ, Pajevic S. Statistical artifacts in diffusion tensor MRI (DT-MRI) caused by background noise. Magn Reson Med. 2000 Jul;44(1):41-50.
  16. Jones DK. Determining and visualizing uncertainty in estimates of fiber orientation from diffusion tensor MRI. Magn Reson Med. 2003 Jan;49(1):7-12.
  17. Gossuin, Y., Roch, A., Muller, R.N. et al. Relaxation induced by ferritin and ferritin-like magnetic particles: The role of proton exchange. Magn. Reson. Med. 2000; 43: 237-243.
  18. Zhan L, Mueller BA, Jahanshad N, et al. Magnetic resonance field strength effects on diffusion measures and brain connectivity networks. Brain Connect. 2013;3(1):72-86.
  19. Koay CG, Ozarslan E, Basser PJ. A signal transformational framework for breaking the noise floor and its applications in MRI. J Magn Reson. 2009;197(2):108-19.
  20. Kleban E, Tax CMW, Rudrapatna US, et al. Strong diffusion gradients allow the separation of intra- and extra-axonal gradient-echo signals in the human brain. Neuroimage. 2020. 15;217:116793.

Figures

Fig. 1: Quantitative maps across the four different field strengths for one subject. All maps were coregistered to a common reference (3T MP-RAGE) using ANTs rigid registration. The apparent residual misalignment of the T2 map at 7T is due to gaps between the acquired slices.

Fig. 2: (a) Scatter plot between mean diffusivity and R2 in different white matter ROIs and across field strengths. We observe a decreasing correlation at 7T compared to the other field strength. At 0.55T, we observe a correlation coefficient that is comparable to clinical field strengths (1.5T, 3T). (b) R2 is only weakly correlated with the orientation 𝜃, i.e., the angle between fibers and B0.

Fig. 3: (a) R2 as a function of MD and the orientation angle 𝜃. (b) Comparison after regressing the orientation dependency out, which shows only minor differences. (c) In contrast, R2 shows a strong dependency on the field strength after regressing out the fiber orientation.

Fig. 4: Quantitative diffusion metrics as a function of the field strengths. The MD is reduced at the extreme field strengths 0.55T and 7T. The reduction at 0.55T is likely due to a Rician noise bias, while the reduction at 7T is likely due to susceptibility effects.

Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)
4584
DOI: https://doi.org/10.58530/2024/4584