Diffusion tensor imaging and beyond

The diffusion of water molecules inside organic tissues is often anisotropic (1). Namely, if there are aligned structures in the tissue, the apparent diffusion coefficient (ADC) of water may vary depending on the orientation along which the diffusion-weighted (DW) measurements are taken. In the late 1980s, diffusion-weighted imaging (DWI) became possible by combining MR diffusion measurements with imaging, enabling the mapping of both diffusion constants and diffusion anisotropy inside the brain and revealing valuable information about axonal architectures (2-14). In the beginning of the 1990s, the diffusion tensor model was introduced to describe the degree of anisotropy and the structural orientation information quantitatively (15, 16). This diffusion tensor imaging (DTI) approach provided a simple and elegant way to model this complex neuroanatomical information using only six parameters. Since then, we have witnessed a tremendous amount of growth in this research field, including more sophisticated nontensor models to describe diffusion properties and to extract finer anatomical information from each voxel. Three-dimensional (3D) reconstruction technologies for white matter tracts are also developing beyond the initial deterministic line-propagation models (17-20). As these new reconstruction methods are an area of very active research, it is important to remember that the theory cannot be dissociated from practical aspects of the technology. Importantly, DWI is inherently a noise-sensitive and artifact-prone technique (Fig. 1). Thus, we cannot overemphasize the importance of image quality assurance and robust image analysis techniques. Last but not least, data acquisition technologies have also been steadfastly evolving. In this article, we review the recent advances in these areas since 2000. Examples of typical artifacts: (i) signal/slice dropouts, (ii) eddy-current induced geometric distortions, (iii) systematic vibration artifacts, and (iv) ghosting (insufficient/incorrect fat-suppression). DWI is inherently a low-resolution and low-SNR technique. Image quality problems are further exacerbated by its high sensitivity to physiological motion. DWI is sensitized to translational motion of water molecules, which is of the order of 5–15 μm assuming typical measurement times. A small amount of subject motion, even cardiac pulsation, can lead to a significant amount of signal phase shift or signal loss, which can severely affect image quality (21-25). To reduce motion sensitivity, single-shot echo-planar imaging (EPI) is commonly used. However, these single-shot data acquisitions usually suffer from artifacts and other limitations. The images are distorted due to B0 susceptibility effects and are prone to eddy current-induced distortions (Figs. 2 and 3). In addition, T signal decay during the lengthy echo train leads to severe imaging blurring and limits the spatial resolution. Three examples of DW images with shears, stretches, and translations induced by eddy currents in the frequency-encoded left-right (LR) direction (a), the phase-encoded anterior–posterior (AP) direction (b), and the slice-select encoded inferior–superior (IS) direction (c), respectively. For each example, the undistorted B0 image (i) is shown with lines overlaid in red indicating brain edges and boundaries of the lateral ventricles. The mismatches of these prominent contours when overlaid on the distorted DW images, shown in (ii) and (iii), now become obvious (see also the enlarged regions corresponding with the arrowheads). Notice the difference in polarity of the eddy current induced gradient between (ii) and (iii) for each example. The images in (c) are shown in a sagittal view to highlight the linearly varying image translation as a function of slice position. Note that this distortion, induced by eddy currents along the IS orientation, may be considered as a shear in the AP-IS plane. Static geometric deformations in the phase-encoded direction by B0 field inhomogeneities (or susceptibility-induced off-resonance fields). The deformations are clearly visible when fused rigidly with a structural T1 weighted image (the enlarged region shows the misalignment of the genu of the corpus between the color-encoded FA image and the T1 map). To reduce these EPI artifacts, techniques are required to shorten the echo train length and reduce the echo spacing. Parallel imaging and segmented k-space sampling are two widely used such methods. Parallel imaging was introduced in the late 1990s (26) and is ideally suited to DWI, as it allows a substantial shortening of both the echo train length and echo spacing, while retaining the robustness to motion of single-shot EPI. The resulting reduction of B0-susceptibility artifacts is substantial. With parallel imaging capability now standard on modern magnetic resonance imaging (MRI) scanners, most DWI studies currently use parallel imaging as part of a routine protocol. However, the acceleration factor (parallel imaging factor) is practically limited to 2–4 depending on the number of receiver channels and coil geometry. If one requires further reduction of susceptibility-induced distortions and/or higher image resolution, a multishot segmented scanning scheme needs to be used. In terms of pulse programming, segmented k-space sampling is straightforward. However, the extreme motion sensitivity of DWI poses a unique challenge. For each shot, translational motion of water leads to signal loss (incoherent motion) or phase shifts (bulk motion). When single-shot imaging is used, phase shifts are irrelevant because the phase information is discarded by the magnitude calculation. However, if the k-space is acquired over multiple shots, phase coherence between shots has to be preserved, which cannot be guaranteed if motion-induced phase shifts occurs. Phase monitoring and postprocessing correction are, therefore, imperative to avoid severe artifacts in multishot DWI (27-34). In the past, phase navigation techniques have been used for segmented EPI (35), spiral EPI (36), and FSE type scans. More sophisticated approaches have been proposed including self-navigated blade-type scans such as PROPELLER (37, 38), and more recently vertically segmented EPI (39-41). One drawback of all segmented scans is their reduced SNR per scan time compared with single-shot imaging. Even with single-shot scans and with a relatively low spatial resolution (2–2.5 mm isotropic voxels), DTI typically requires an acquisition time of ∼5 min to obtain appropriate data. It remains to be seen whether the improved B0 susceptibility distortion and spatial resolution provided by segmented imaging techniques can provide enough SNR and robustness to motion within a clinically feasible scan time. If so, this would of course open up new and exciting opportunities for research studies in which scan time and success rate are less of an issue. Another image quality issue related to the usage of the single-shot EPI readout is that of image distortion due to eddy currents introduced by the rapid switching of the large diffusion gradient pulses. While this was an important issue in the ’90s due to limitations in the gradient systems, this type of distortion is much less problematic on moderns MRI scanners due to advancements in hardware quality and eddy-current compensation schemes. Moreover, eddy current effects caused by the DW gradient pulses can be further suppressed using appropriate pulse sequences, including the use of bipolar gradients (42, 43). Several studies have investigated the impact of cardiac pulsation in detail, which causes nonlinear motion and local deformations of the brain parenchyma (44, 45), and can corrupt the measured diffusion signal (9, 21, 22, 46, 47). The resulting signal drop-outs and residual misalignments between the DW images will lead to erroneous estimates of the diffusion tensor (or any other model) and, consequently, of any subsequently derived diffusion measure of interest. To avoid these pulsatile artifacts, one can trigger the acquisition sequence to the cardiac cycle, which introduces a dependence of the effective pulse repetition time on the heart rate of the subject. Triggering can for example be performed based on the signal obtained from a pulse-oximeter placed on the subject’s forefinger. The downside of cardiac gating is that the acquisition typically requires a longer and unpredictable scan time. Diffusion-weighting by a pair of strong gradient pulses introduces a number of imaging parameters unique to diffusion imaging. This includes the magnitude (b-value) and orientations as well as the number of the least DW images (so-called b = 0 images). The majority of DTI studies nowadays use b-values in the range of 700–1000 s/mm2, leading to 30–50% signal reduction assuming the mean diffusivity of normal white matter is around 0.8 to 1.0 × 10−3 mm2/s. The determination of the optimum b-value (48, 49) is complicated by the involvement of many factors (50), including: SNR (the higher the SNR, the more accurately signal attenuation can be measured with higher b-values), echo time (the smaller the b-value, the shorter the achievable echo time), and other factors that are more difficult to assess such as eddy current and motion artifacts (in general, smaller b-values produce less artifacts). Determination of the optimal number and distribution of gradient directions is also not straightforward. Simulation studies have shown improvements in fractional anisotropy (FA) estimation (a reduced dependency of the accuracy on the fiber orientation) by increasing the number of orientations up to 30 orientations, suggesting that as many DW gradient orientations should be used as time allows (51-53). On modern MR scanners, assuming 50–65 axial slices (2.0–2.5 mm thickness) and a 5 min scan time, there should be enough time to acquire about 30–40 imaging volumes. If 5 b = 0 images are acquired, the rest of the time can be spent sampling as many orientations as possible (25–35 orientations). A related question is whether for example a 36 direction scheme is better than 3 repeats of a 12 direction scheme (which have equivalent scan times), and if so, by how much? This question is compounded by various other potentially more dominant sources of inaccuracy such as physiological noise (motion-induced intensity fluctuations or misregistration) and instrument imperfection (eddy-currents, stability, etc). In a human study comparing 6, 10, 15, and 30 orientation schemes, the differences in test–retest reproducibility were found to be relatively minor for studies with scan times shorter than 10 min, provided the orientations used are well distributed in space (54, 55). From a practical point of view, multiple repeat scans (e.g., 3 × 12 orientations rather than 36 orientations) may provide a better way to judge the quality of the postprocessing (e.g., coregistration) or existence of motion-corrupted voxels, by comparing corresponding images across the multiple repeats. It is, therefore, difficult to conclude that multiple repeated sets of a smaller number of orientations should not be used. Please note that these discussions are applicable only for tensor-based approaches, where only six parameters need to be estimated, and the system is, therefore, overdetermined. For nontensor or multiple-tensor based analyses, the number of parameters that need to be estimated may be much larger, and the correspondingly larger minimum number of directions may preclude the acquisition of multiple repeats within a feasible scan time. These issues will be discussed in more detail in Sections “Data Processing and Quality Control” and “Beyond the tensor model.” Although general guidelines exist for optimizing a DTI acquisition protocol in terms of SNR, b-value, voxel size, diffusion gradient directions, cardiac gating, etc. (56), there are still large variations in data quality across imaging centers due to, for instance, differences in scanner hardware, pulse sequences, and available scan times. Despite recent efforts to ensure high-data quality (e.g., by considering specialized diffusion phantoms to further improve acquisition settings and optimize postprocessing methods) (57-61), there is currently no consensus over which approach is preferred for DTI quality assessment. Detailed slice-by-slice inspection of the DW MRI data to detect potential artifacts can be extremely time-consuming given the vast amount of acquired data.1 However, by simply looping through the DW images at a relatively high frame rate (∼10 fps), large signal dropouts and geometric distortions can be spotted instantly in a short amount of time. Subtle system drifts, which may result in apparent translations (i.e., the “levitation” artifact), may also be detected by quickly toggling between the views of the first and last acquired DW image. Perhaps considered trivial, but often overlooked, is to inspect the images in different “orthogonal” views and not only to look at the image plane that the data were acquired in. In doing so, interslice and intravolume instabilities (i.e., the “zebra pattern” or “zipper” artifact), such as relative offsets in slice location or differences in signal intensity, can be easily observed (see Fig. 4). A DW image shown in three orthogonal views. The interslice instabilities (encircled) in this axially interleaved acquisition might not be seen on the axial slices, but are very prominent on the sagittal and coronal through-plane views. Calculating the standard deviation across the different DW images (SDWI) for each voxel provides an efficient way to investigate image misalignment artifacts (62, 63). In Fig. 5, SDWI is shown for (Fig. 5a) the raw uncorrected DW images and (Fig. 5b) the same DW images, but now corrected for subject motion and eddy-current induced geometric distortions (64). The size and brightness of the rims at brain edges and tissue interfaces in these SDWI maps reflect the degree of misalignment between the different DW images. If multiple b = 0 images are available, taking the standard deviation across these images can be a sensitive approach to locate cardiac pulsation artifacts (see Fig. 6). To assess subject motion, or more generally, misalignment between the DW images, computing the standard deviation across the DW images SDWI is an efficient and more quantitative approach than inspection of the DW images on a slice-by-slice basis. The bright rim in the SDWI map shown in (a) is not present after correction for subject motion and eddy current induced geometric distortions (b) (see arrowheads). In the b = 0 images (i.e., the non-DWIs), unreliable regions in terms of signal variability (most likely due to pulsation artifacts) can be visualized readily by taking the standard deviation across the b = 0 images (Sb = 0). For the example shown here, six b = 0 images were acquired. Notice the high variability near the medial parts of the brainstem, cerebellum, and the lateral ventricles. By definition, FA is bound between 0 and 1. However, FA values larger than 1 can be obtained if the diffusion tensor contains one (or more) negative eigenvalues. These negative eigenvalues are typically encountered when there is a non-DW signal that is smaller than (any of) the DW signals. A binary map that flags such physically implausible signals proves to be a powerful method to spot additional artifacts (e.g., Gibbs ringing in the b = 0 images) (Fig. 7). Examples of “physically implausible signals” (i.e., voxels where the B0 intensity is lower than the DW intensities) that contaminate the fractional anisotropy (FA) values. When comparing the locations of the corrupted voxels (in red) between the middle and the left image, these FA values are typically overestimated. A more detailed investigation revealed the presence of negative eigenvalues, which were caused by ill-conditioned diffusion tensor estimations. The corresponding B0 images on the right contain the artifacts (Gibbs-ringing) that formed the basis of these error accumulations: parallel to the interface between the CSF and the surrounding white matter, artificially low intensity rims can be observed (see arrowheads), which is typically seen in images with a relatively small acquisition matrix. Three examples that showcase the high sensitivity of diffusion tensor residual maps (R) to detect artifacts. a: The experienced DTI user will immediately spot the phantom commissural pathways that connect left and right occipital lobes on the directionally color-encoded fractional anisotropy (FA) maps (encircled). If one is unfamiliar with the rich (colorful) information contained in these images, however, or when pathology is involved, the corresponding R map is a useful tool to differentiate between low and high quality regions. In this example, the artifact is clearly visible on the R map as well. By contrast, in (b) and (c), the artifacts, i.e., ghosting due to insufficient fat suppression and RF interference (i)–(iii)/slice dropout (iv), respectively, are not visible on the FA maps and can hardly be seen on the individual DW images themselves. a: Diffusion tensor residuals calculated for each DW image (averaged across all brain voxels—see Eq. 3, with the error bars representing the inter-quartile range). In the example shown, five diffusion volumes were “heavily” corrupted as indicated by the higher residuals (encircled). b: A more quantitative feel of the significance of high residual values is obtained by calculating the “statistical” outliers of these tensor residuals. The percentage of outliers per DW image may then serve as a marker to identify artifacts. c: To increase the specificity of detecting artifacts, the same procedure can be applied to each slice separately and along the different (coronal, axial, and sagittal) image views. In this way, a summary statistic of data quality can be shown for each slice and for all the DW gradient directions simultaneously in a single matrix. Retrospective identification of “problematic” slices is then facilitated by the “hot (see enlarged When diffusion MRI strong magnetic field gradient pulses are applied to the diffusion along a The rapid in the magnetic field with the and of such large gradients will in of the MRI scanner (e.g., the or the gradient These eddy currents will in produce magnetic field gradients that will to the applied decay of the eddy currents over the signal acquisition the gradient will the in k-space and, as a geometric distortions will be introduced (Fig. In addition, not a difference between the and the b-value will for each gradient as it on gradient field properties that are by the eddy have been to for eddy-current induced geometric distortions based additional information obtained during postprocessing such as image and a of the two (42, (Fig. a: color-encoded fractional anisotropy (FA) maps (i) and after (ii) for subject motion and eddy current-induced geometric The bright rim (see enlarged clearly visible in is practically in the corrected image In this example, geometric distortions and subject motion in the DW images are corrected for As a the orientation of the diffusion gradients should be to potential In (b), the difference in orientation of the estimated first between the uncorrected (i) and the corrected (ii) gradient directions is shown in a region of the To the of this the of the first are shown in (iii) on the which with the in (i) and Although the shown in (iii) small and, therefore, the in (c) clearly the deviation in the fiber pathways when subject motion and eddy current-induced geometric distortions are not The model of water diffusion would use a single diffusion which would that the system has isotropic However, it is now that the diffusion of water within brain white matter tissue is The 3 × 3 tensor model was proposed as the and most elegant way to such a only six parameters to be the tensor model may the (see “Beyond the tensor for more Thus, it is important to derived from the tensor model with only six parameters to the DW information within a is a significant reduction in the DW information reduction It is also important to that and which usually requires the tensor to be further to a further information reduction reduction values for the degree of anisotropy (e.g., and diffusivity and of these maps will often further information reduction reduction and For example, across are often performed by regions of a vast amount of the not much information but spatial between images is usually not and the SNR is often low to detect For these a substantial amount of spatial is often the amount of The use of spatial is also a subject of as it may not the of structural misalignment between images, and as the are on the amount and type of the In addition, to be it has been shown that of approach and can further affect the of the The reduction of the anatomical information to a tensor and then to a that when or differences are found in one of the it is often difficult to any about the at the While this can be considered as a drawback of the systematic information reduction can also be as an with more than an number of and of of the human brain is a complex the of which is currently beyond If one to its anatomical and it between different a quantitative method that can reduce the anatomical information a size is which is DTI can and within a short shows a of different anatomical factors that affect diffusion These factors can be and From the point of view, there are various that diffusion anisotropy such as and which are structures in the order of μm These structures are often and axonal These then become of white matter structures on a much larger While the factors are for water to diffusion are not because the factors can fiber orientations within a voxel can lead to isotropic diffusion If a (or in diffusion anisotropy is it is difficult to immediately conclude which of these multiple factors for the anatomical factors that the diffusion anisotropy measurement by DTI and their To more information about the of the the information reduction and needs to be For example, when FA is there are at least three potential in terms of the relative the of the diffusion parallel is the shorter and are or both of these By the 3 × 3 tensor information to a single FA these three become By each more information can potentially be to the of the studies have shown that loss is with an increase in axonal loss is related more to a in the parallel diffusivity However, there is the that such are to the model used and may not For instance, studies not the (e.g., axonal loss may lead to a in the parallel but such a may not mean axonal Note that this only on potential be when such in regions to contain fiber orientations Although in its the diffusion tensor model has been shown to be in the many regions of the brain that contain (62, two or more fiber are within the same voxel. The is as it includes any where multiple fiber orientations to the signal measured for the same imaging voxel. this also to that may not have been of as for example, fiber that each other within the same imaging or even or (Fig. are to DWI, due to its resolution to compared with the white matter structures of the tracts are only in regions recent studies have shown that a significant of the white matter contain with the most recent that multiple fiber orientations can be detected in over of white matter voxels (Fig. of complex fiber at the length of a single voxel Note that any different from a single fiber is typically to as including (i) (e.g., and (ii) (e.g., fiber as shown in (iii), might in the region of the where the lateral of the corpus with the By contrast, the shown in (iv) fiber such as the and the of the corpus of the within a single voxel. The number of fiber orientations detected within each voxel overlaid on the corresponding anatomical image. voxel within the is to the number of orientations detected single two more than two orientations). These were from data obtained from a of repeats of 30 DW directions, acquired at b = s/mm2, using within a Image of These effects have an obvious impact on the diffusion tensor and any derived from it As the mean is anisotropy such as FA are sensitive to the presence of as are the axial and (Fig. This has important for their as are commonly as of white matter the and impact of fiber such should only be with extreme are even more problematic for tensor-based methods (62, if one corrupt orientation is the may course an white matter leading to both and Moreover, the is than might be any given