Optimizing Diffusion-weighted MRI of Peripheral Nerves

HomeRadiologyVol. 302, No. 1 PreviousNext Reviews and CommentaryFree AccessEditorialOptimizing Diffusion-weighted MRI of Peripheral NervesHyungseok Jang, Jiang Du Hyungseok Jang, Jiang Du Author AffiliationsFrom the Department of Radiology, University of California, 9500 Gilman Dr, San Diego, CA 92093.Address correspondence to J.D. (e-mail: [email protected]).Hyungseok JangJiang Du Published Online:Oct 19 2021https://doi.org/10.1148/radiol.2021211907MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In See also the article by Foesleitner et al in this issue.Dr Jang is a scientist of radiology at the University of California San Diego. He is a junior fellow of the International Society for Magnetic Resonance in Medicine and a member of the Council of Early Career Investigators in Imaging and Academy for Radiology and Biomedical Imaging Research. His research focuses on the development of novel MRI techniques. Dr Jang is the principal investigator of one National Institutes of Health (NIH) grant.Download as PowerPointOpen in Image Viewer Dr Du is a professor of radiology at the University of California San Diego. He is a distinguished investigator of the Academy for Radiology and Biomedical Imaging Research and a fellow of the American Institute for Medical and Biological Engineering. His research focuses on the development of novel techniques for morphologic and quantitative MRI. Dr Du serves on NIH study sections and is the principal investigator of two NIH grants.Download as PowerPointOpen in Image Viewer With the recent advancement in MR hardware and imaging techniques, diffusion-weighted imaging (DWI) has emerged as one of the most important clinical diagnostic tools. DWI is based on sensitization of random molecular thermal motion (ie, Brownian motion), which can be used to characterize the connectivity and microstructure of a targeted tissue (1). Typically, the degree of motion is coded by a signal weighting, which is realized using a pair of motion-sensitizing gradients. The first gradient is meant to dephase the spins, and the second gradient, applied after a certain time delay (ie, mixing time), is meant to rephase the spins. Consequently, the spins with motion undergo residual dephasing and are not fully recovered by the time the second gradient is applied, resulting in signal decay.The degree of diffusion weighting (or signal decay) is determined by both the molecular motion constrained by tissue microstructure and the prescribed imaging parameters (eg, size of the diffusion-weighting gradient and the mixing time), which produce the b value (in the unit of sec/mm2), a fundamental imaging parameter in DWI. With use of multiple acquisitions of diffusion-weighted images with different b values, a diffusivity parameter (ie, apparent diffusion coefficient) can be fitted based on a known signal model. Under the assumption of Gaussian diffusion, the diffusion signal weighting over b values is modeled as a monoexponential decay. If there is a bulk motion or flow (ie, perfusion), the diffusion signal can be modeled as a biexponential decay (2). The non-Gaussian kurtosis model has also been proposed to better handle diffusion involving a more complex microstructure (3).Spin-echo and gradient-recalled echo-based echo-planar imaging have been commonly used to acquire a diffusion-weighted image. Typically, to estimate the diffusivity in one direction, at least one image with a nonzero b value is required, along with a T2-weighted image with a zero b value. To account for anisotropy in the tissue microstructure, DWI can be performed in three orthogonal directions, yielding both mean diffusivity (ie, trace) and fractional anisotropy. Diffusion tensor imaging (DTI) uses more advanced mathematical modeling to allow estimation of the magnitude of diffusivity (ie, eigenvalue) and its principal direction (ie, eigenvector). Given that diffusion tensor has six degrees of freedom (ie, the diffusion tensor is represented by a 3 × 3 symmetric positive definite matrix), at least six diffusion-weighted images with different gradient orientations are required, along with a T2-weighted image with a zero b value. For more complex modeling, such as the estimation of connectivity in crossing fibers, a greater number of diffusion images with more gradient orientations are commonly required. Thus, an optimal b value is a prerequisite to achieving reliable results in DWI and DTI. The b value should be high enough to allow for adequate diffusion sensitivity, but not too high to avoid non-Gaussian diffusion behavior (4). It is believed that an acceptable standard for the optimal b value for DWI of the central nervous system is around 1000 sec/mm2, as the signal deviates from the Gaussian diffusion model beyond this point (5). That said, to our knowledge, no consensus has been reached regarding the optimal b value for the peripheral nervous system.In this issue of Radiology, Foesleitner and colleagues (6) investigated non-Gaussian diffusion behavior to determine the optimal b value for DWI and DTI of the peripheral nervous system. A cross-sectional study of MR neurography of the sciatic nerve was performed in heathy individuals (n = 16) and participants with type 2 diabetes (n = 12). A wide range of b values (0–1500 sec/mm2) were tested on a 3-T clinical MR scanner. With use of the acquired DWI data, a monoexponential model, biexponential model, and kurtosis model were evaluated. Non-Gaussian diffusion behavior was observed beyond the b values of 600 sec/mm2 and 800 sec/mm2 in the axial and radial directions of the sciatic nerve, respectively, for both healthy individuals and participants with diabetes. Due to this phenomenon, the monoexponential model exhibited severe deviations in data fitting, while both the biexponential and kurtosis models showed much improved fitting. In addition, the potential diagnostic power of the kurtosis parameters (ie, Dk and K) was demonstrated by significant differences between the groups of healthy volunteers and participants with diabetes (P < .05). Based on their experimental results, the authors suggested an optimal b value of 700 sec/mm2 for the peripheral nervous system.It is interesting that the diffusion behavior in the peripheral nervous system is different from that in the central nervous system—namely, that a lower b value may be desirable to avoid non-Gaussian diffusion behavior in the peripheral nervous system. This is partly due to the more complex microstructure of peripheral nerves, which have multiple layers of connective tissue surrounding axons, including the endoneurium, which surrounds individual axons; the perineurium, which binds axons into fascicles; and the epineurium, which binds the fascicles into nerves. Peripheral nerves also include lipids, muscle, and blood vessels (7); therefore, the limited spatial resolution in clinical DWI is a major challenge given its potential impact on non-Gaussian behavior due to the partial volume effect. The double-echo steady-state sequence allows diffusion imaging with higher spatial resolution and may be an alternative that can reduce the partial volume effect (8). Another challenge is that the T2 value of a peripheral nerve is shorter than that of a nerve in the central nervous system. This inevitably affects the noise performance in DWI of the peripheral nervous system with high b values, resulting in a lower signal-to-noise ratio. This should be considered when determining the optimal b value. Ultrashort echo time–based DWI may be an alternative, given its ability to capture the signal at a much shorter echo time, allowing for a higher signal-to-noise ratio for the peripheral nerve with a relatively short T2 decay. More advanced techniques (eg, ultrashort echo time–based double-echo steady-state sequences) are expected to enable diffusion imaging of the peripheral nervous system with much higher spatial resolution and signal-to-noise ratio (9,10). Further study is required to investigate the aforementioned challenges in depth.In summary, Foesleitner and colleagues presented a new method to determine the optimal b value for DWI and DTI in MR neurography for the peripheral nervous system. Although it is too early to make a determination regarding the optimal b value for diffusion imaging of the peripheral nervous system, this study provides a meaningful reference for future studies.Disclosures of Conflicts of Interest: H.J. No relevant relationships. J.D. No relevant relationships.References1. Huisman TAGM. Diffusion-weighted and diffusion tensor imaging of the brain, made easy. Cancer Imaging 2010;10 Spec no A(1A):S163–S171. Crossref, Medline, Google Scholar2. 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. Link, Google Scholar3. Jensen JH, Helpern JA, Ramani A, Lu H, Kaczynski K. Diffusional kurtosis imaging: the quantification of non-gaussian water diffusion by means of magnetic resonance imaging. Magn Reson Med 2005;53(6):1432–1440. Crossref, Medline, Google Scholar4. Le Bihan D. Apparent diffusion coefficient and beyond: what diffusion MR imaging can tell us about tissue structure. Radiology 2013;268(2):318–322. Link, Google Scholar5. Jensen JH, Helpern JA. MRI quantification of non-Gaussian water diffusion by kurtosis analysis. NMR Biomed 2010;23(7):698–710. Crossref, Medline, Google Scholar6. Foesleitner O, Sulaj A, Sturm V, et al. Diffusion MRI in peripheral nerves: optimized b values and the role of non-Gaussian diffusion. Radiology 2021.https://doi.org/10.1148/radiol.2021204740. Published online October 19, 2021. Link, Google Scholar7. Chen Y, Haacke EM, Li J. Peripheral nerve magnetic resonance imaging. F1000 Res 2019;8:1803. Crossref, Google Scholar8. Raval SB, Britton CA, Zhao T, et al. Ultra-high field upper extremity peripheral nerve and non-contrast enhanced vascular imaging. PLoS One 2017;12(6):e0175629. Crossref, Medline, Google Scholar9. Chaudhari AS, Sveinsson B, Moran CJ, et al. Imaging and T2 relaxometry of short-T2 connective tissues in the knee using ultrashort echo-time double-echo steady-state (UTEDESS). Magn Reson Med 2017;78(6):2136–2148. Crossref, Medline, Google Scholar10. Jang H, Ma Y, Carl M, Jerban S, Chang EY, Du J. Ultrashort echo time Cones double echo steady state (UTE-Cones-DESS) for rapid morphological imaging of short T2 tissues. Magn Reson Med 2021;86(2):881–892. Crossref, Medline, Google ScholarArticle HistoryReceived: July 27 2021Revision requested: Aug 4 2021Revision received: Aug 6 2021Accepted: Aug 10 2021Published online: Oct 19 2021Published in print: Jan 2022 FiguresReferencesRelatedDetailsAccompanying This ArticleDiffusion MRI in Peripheral Nerves: Optimized b Values and the Role of Non-Gaussian DiffusionOct 19 2021RadiologyRecommended Articles Diffusion MRI in Peripheral Nerves: Optimized b Values and the Role of Non-Gaussian DiffusionRadiology2021Volume: 302Issue: 1pp. 153-161Imaging Review of Peripheral Nerve Injuries in Patients with COVID-19Radiology2020Volume: 298Issue: 3pp. E117-E130Functional MR Neurography in Evaluation of Peripheral Nerve Trauma and Postsurgical AssessmentRadioGraphics2019Volume: 39Issue: 2pp. 427-446Can Quantitative MRI Be Used to Differentiate Physiologic Changes Behind Muscle Weakness in Type 2 Diabetes Mellitus?Radiology2020Volume: 297Issue: 3pp. 620-621MR Neurography of Peripheral Nerve Injury in the Presence of Orthopedic Hardware: Technical ConsiderationsRadiology2021Volume: 300Issue: 2pp. 246-259See More RSNA Education Exhibits You’re Getting on My Nerves: A Review of Current and Future Applications of DTI Tractography of the Central and Peripheral Nervous SystemDigital Posters2019Besides Colorful Images, What Diffusion Tractography Of Peripheral Nerve Can Add?Digital Posters2021A Handbook of Non-EPI Diffusion Tensor Imaging Sequences: Physical Basis, Technical Adjustments, and Potential Clinical ApplicationsDigital Posters2019 RSNA Case Collection Intracranial TBRSNA Case Collection2022Leptomeningeal siderosisRSNA Case Collection2021Tuberculous Leptomeningitis with Vasculitic InfarctRSNA Case Collection2021 Vol. 302, No. 1 Metrics Altmetric Score PDF download