METHOD OF DIAGNOSING DEGREE OF NEUROPATHY OF A NERVE BY DIFFUSION TENSOR MAPPING

Abstract

The invention is related to a method of diagnosing degree of neuropathy of a nerve by diffusion tensor mapping, said method comprising the following steps: a) acquiring diffusion tensor imaging, DTI, digital images of the nerve based on magnetic resonance, MR, data of the part of the body comprising said nerve, b) for each voxel in said DTI digital images calculating a diffusion tensor D

[00001]$\stackrel{?}{D}=\left[\begin{array}{ccc}{D}_{\mathrm{xx}}& D_{xy}& D_{xz}\\ D_{yx}& D_{yy}& D_{yz}\\ D_{zx}& D_{zy}& D_{\mathrm{zz}}\end{array}\right].$

and finding its eigenvalues L.sub.1>L.sub.2>L.sub.3:

[00002]$\stackrel{?}{L}=\left[\begin{array}{ccc}{L}_1& 0& 0\\ 0& L_2& 0\\ 0& 0& L_3\end{array}\right]$ c) selecting a region of interest in at least one of said DTI digital images, said region of interest comprising pixels, d) calculating the threshold value ?.sub.th of a cutoff parameter ?, ranging from 0 to L.sub.1, as the minimal value of ? for which the following cutoff condition

[00003]$L_1?RD+?$

is false for all pixels within the region of interest,
wherein RD is a radial diffusivity defined as

[00004]$RD=\left(L_2+L_3\right)/2.$

Claims

1. A method of diagnosing degree of neuropathy of a nerve by diffusion tensor mapping, said method comprising the following steps: a) acquiring diffusion tensor imaging, DTI, digital images of the nerve based on magnetic resonance, MR, data of the part of the body comprising said nerve, b) for each voxel in said DTI digital images calculating a diffusion tensor D $\hat{D}=\left[\begin{array}{ccc}{D}_{\mathrm{xx}}& D_{\mathrm{xy}}& D_{\mathrm{xz}}\\ D_{\mathrm{yx}}& D_{\mathrm{yy}}& D_{\mathrm{yz}}\\ D_{\mathrm{zx}}& D_{\mathrm{zy}}& D_{\mathrm{zz}}\end{array}\right].$ and finding its eigenvalues L.sub.1>L.sub.2>L.sub.3: $\hat{L}=\left[\begin{array}{ccc}{L}_1& 0& 0\\ 0& L_2& 0\\ 0& 0& L_3\end{array}\right]$ c) selecting a region of interest in at least one of said DTI digital images, said region of interest comprising pixels, d) calculating the threshold value ?.sub.th of a cutoff parameter ?, ranging from 0 to L.sub.1, as the minimal value of ? for which the following cutoff condition $L_1?\mathrm{RD}+?$ is false for all pixels within the region of interest, wherein RD is a radial diffusivity defined as $\mathrm{RD}=\left(L_2+L_3\right)/2.$

2. The method according to claim 1, wherein said nerve is a human or an animal nerve.

3. The method according to claim 1, wherein said nerve is an optic, an auditory nerve, an oculomotor nerve, a trigeminal nerve, a vestibulocochlear nerve, white matter of a brain or a spinal cord.

4. The method according to claim 1, wherein the acquisition of said images in step a) is conducted for no more than 15 minutes.

5. The method according to claim 1, wherein said nerve is an optic nerve and which comprises the following additional steps, performed after step b) and before step c): b1) for each voxel in said DTI digital images calculating fractional anisotropy FA, mean diffusivity MD, and radial diffusivity RD, wherein: $\mathrm{FA}=\frac{{\left(L_1-L_2\right)}^2+{\left(L_2-L_3\right)}^2+{\left(L_1-L_3\right)}^2}{2\left(L_1^1+L_2^2+L_3^2\right)},\mathrm{MD}=\left(L_1+L_2+L_3\right)/3,\mathrm{RD}=\left(L_2+L_3\right)/2.$ b2) associatingwith the first eigenvector in each voxela line segment representing the main direction of the diffusion in this voxel, said line segment having a length, an opacity and an RGB color assignment, b3) for each line segmentadjusting its length to a value given by: ${?\left(L_1-\mathrm{RD}\right)}^{?},$ wherein ? is a scaling parameter ranging from 1 to 1000 and ? is a scaling parameter ranging from 1 to 3, b4) for each line segmentadjusting its opacity to a value given by: $?.Math.\mathrm{MD}.Math.\mathrm{FA},$ wherein ? is a scaling parameter ranging from 1 to 1000, b5) for each line segmentadjusting its RGB color assignment as follows: $\mathrm{Red}=?L_1,\mathrm{Green}=1-L_1,\mathrm{Blue}=0$ wherein ? is a scaling parameter for setting the intensity of red color ranging from 0 to 3, thereby obtaining an adjusted three-dimensional shape of the nerve b6) projecting said adjusted three-dimensional shape of the nerve onto a transverse plane by superposing two or more layers of the adjusted three-dimensional shape on said transverse plane, thereby obtaining a planar superposed image, to be further used as an DTI digital image for steps c) and d).

6. The method according to claim 5, wherein ? is 200.

7. The method according to claim 5, wherein ? is 2.2.

8. The method according to claim 5, wherein ? is 500.

9. The method according to claim 5, wherein ? is 2.

10. The method according to claim 5, wherein the layers mentioned in step b6) are selected such that they are essentially parallel to the axis of the optical nerve.

11. The method according to claim 5, wherein neuropathy of the optical nerve is diagnosed for ?.sub.th lower than 0.27.

12. The method according to claim 1, wherein the digital images acquired in step a) are echo planar imaging diffusion tensor imaging, EPI DTI, digital images.

13. The method according to claim 1 which is computer-implemented.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Preferred embodiments of the present invention are presented in a more detailed way with reference to the attached drawing, in which:

[0051] FIG. 1a-f presents the visual effect of the algorithm in six steps for a healthy subject, wherein the subsequent steps are: aprincipal diffusion directions shown as line segments; blength coding; copacity coding; dcolor coding; eseven layers projection; fseven layers projection and data thresholding; FIGS. 1a-c are presented in radiological orientation convention.

[0052] FIG. 2 shows visual pathway shape on different plane projections, wherein white and blue arrows indicate right and left optic nerve, respectively, and wherein real voxel proportions are preserved.

[0053] FIG. 3 presents visual pathways of a male subject with right optic nerve atrophy, shown by orange arrows (FIG. 3a) and of healthy male subject (FIG. 3b); visual pathways are visualized with the DTI algorithm superimposed on the same T1 image, whereby optic nerves are also presented on diffusion scalar maps FA and MD, and on morphological T1-weighted images, and whereby the orientation convention used is radiological.

[0054] FIG. 4 presents images for male subject with right optic nerve neuropathy, in radiological orientation convention, generated for ? values equal to A) 0, B) 0.15, C) 0.25, and D) 0.35; FIG. 4 presents also density plots showing all voxels in the datasets used for visualization (black dots) and voxels omitted according to condition (9) (grey dots).

DETAILED DESCRIPTION

[0055] Preferred embodiments of the invention are described in details below. The examples serve only as an illustration and do not limit the scope of the present invention.

Brief Description of the Study

[0056] To develop a method according to the invention, it was important to perform a case study, which is described below. The used set-up, including all parameters of the scan, as well as software and hardware used, are exemplary and should not be treated as limiting the scope of the present invention.

[0057] Images from a healthy male subject (age: 32) and a male subject with penetrating right eyeball injury (age: 46) were used. Both subjects underwent a full ophthalmological examination, which included visual acuity, color vision, applanation measurement of intraocular pressure, evaluation of the anterior chamber and the fundus of the eye using a slit lamp and the Volk lens, optical coherence tomography examination of the macula and optic nerve disc (OCT), neurological field vision, and visual evoked potential (VEP) testing. In the tests conducted, no deviations from the normal condition were found in the subjects, while the trauma patient had endotamponade of the vitreous chamber of the right eye with silicone oil, lack of sense of light, secondary glaucoma, and atrophy of the right eye optic nerve, and correct ophthalmological examination of the left eye. In his medical history, the patient reported that prior to the penetrating injury (metal shard from a masonry hammer driven into the eyeball), the patient did not experience any deterioration in vision compared to the left eye. During active scanning, the patients tried not to blink and did not move their heads, while gazing at the red-locking LED outside the camera reflected in the mirror of the coil. Imaging acquisition required about 10 minutes of gaze fixation.

[0058] In diffusion tensor imaging, the subject uncontrolled blinking, eyeball movements, and spatial and temporal variations of the static magnetic field caused by susceptibility effects and time-varying eddy currents result in severe distortions, blurring, and misregistration artifacts. Due to the limitations of acquisition time and attempts to shorten it to a minimum for the patient's comfort, only the part of the head containing the optic nerve and the tract was scanned.

[0059] MRI data were acquired using a Biograph mMR 3-Tesla scanner (Siemens, Erlangen, Germany) with a 16-channel head-neck coil. Acquisition parameters were set up based on the article (Gerlach, D. A.; Marshall-Goebel, K.; Hasan, K. M.; Kramer, L. A.; Alperin, N.; Rittweger, J. MRI-Derived Diffusion Parameters in the Human Optic Nerve and Its Surrounding Sheath during Head-down Tilt. npj Microgravity 2017) and technician experience. The total duration of the EPI DTI examination was 10:34 min (8:05 and 2:29 for A>>P and P>>A phase encoding directions respectively). MRI acquisitions were obtained using a scanner with a maximum gradient amplitude of 45 mT/m and a maximum slew rate of 200 T/m/s. Parameters for used for echo planar (EPI) acquisition method were: TR 7000 ms, TE 113 ms, echo spacing 0.97 ms, bandwidth 1184 Hz/Px, EPI factor 128, voxel size 1.3?1.3?2.0 mm, 25 slices (transverse orientation), distance factor 0% for transverse plane and 50% for other planes, base resolution 128 and prescan normalization filtering. Sixty-four non-collinear diffusion directions for a b value of 1000 s/mm2 with one average were acquired for the A>>P encoding direction and six averages without diffusion-weighted acquisition for b value of 0 s/mm2 were acquired for both the A>>P and the P>>A encoding directions. Parallel imaging with an acceleration factor of 2 was enabled using the GRAPPA algorithm. Additionally, an anatomical T1-weighted (MP RAGE), scan time 5:21 min, voxel size 1.0?1.0?1.2 mm, was performed.

[0060] In addition to coronal magnetization-prepared rapid acquisition gradient-echo reconstruction, oblique transverse sections through both optic nerves were done. The exact positions and angulation of slices were further graphically specified and adjusted in the central sections of the left and right sagittal optic nerves. This procedure allowed for an exact adjustment of the EPI and T1 slices through both optic nerves. In general, positioning was focused on the part of the optic nerve closest to the eye to allow for optimal comparability in the case of a curved optic nerve.

[0061] The orientation of layers was chosen such that they were parallel to the optic nerve axis to a large extent, which is different than standard brain scanning, and which is an important aspect of the present invention since it may affect the results significantly.

[0062] Diffusion images were converted from the DICOM format to the NIfTI format using the open source dom2niix software, and the data were post-processed using the FSL software library (Jenkinson M.; Beckmann C. F.; Behrens T. E.; Woolrich M. W.; Smith S. M. FSL; 2011). The ITK-SNAP software was used to create a binary mask of the brain, eyeballs, and optic nerves. FSL's tools were used for susceptibility-induced and eddy current-induced distortion correction, and movement artifacts correction (what is known in the art). The correction procedure assumes that the eddy current-induced field can be modelled as a combination of linear and quadratic terms. Calculations were parallelized with CUDA v9.1 on Nvidia Tesla K80 GPU. Eddy current and motion-corrected diffusion-weighted images were used for voxel-by-voxel-based tensor calculations. Consequently, eigenvectors (V.sub.1, V.sub.2 and V.sub.3) and eigenvalues (L.sub.1?L.sub.2?L.sub.3) were obtained. Throughout the text, the largest eigenvalue L; will be called axial diffusivity. For all other calculations, the Wolfram Mathematica software was used and NIfTI images were imported using the QMRITools toolbox developed for the Wolfram language (Froeling M. QMRTools: A Mathematica Toolbox for Quantitative MRI Analysis. J Open Source Softw 2019, 4, 1204). Fractional anisotropy (FA), mean diffusivity (MD) and radial diffusivity (RD) were defined as in: Minati L.; Weglarz W. P. Physical Foundations, Models, and Methods of Diffusion Magnetic Resonance Imaging of the Brain: A Review. Concepts Magn Reson Part A 2007, 30, 278-307; Mukherjee P.; Chung S. W.; Berman J. I.; Hess C. P.; Henry R. G. Diffusion Tensor MR Imaging and Fiber Tractography: Technical Considerations. PHYSICS REVIEW AJNR Am J Neuroradiol 2008, 29, 843-852; Gerlach D. A.; Marshall-Goebel K.; Hasan K. M.; Kramer L. A.; Alperin N.; Rittweger J. MRI-Derived Diffusion Parameters in the Human Optic Nerve and Its Surrounding Sheath during Head-down Tilt. NPJ Microgravity 2017, 3, 18, i.e.:

[00013]$\begin{array}{cc}\mathrm{FA}=\frac{{\left(L_1-L_2\right)}^2+{\left(L_2-L_3\right)}^2+{\left(L_1-L_3\right)}^2}{2\left(L_1^1+L_2^2+L_3^3\right)}& \left(1\right)\end{array}$$\begin{array}{cc}\mathrm{MD}=\frac{\left(L_1+L_2+L_3\right)}{3}& \left(2\right)\end{array}$$\begin{array}{cc}\mathrm{RD}=\frac{\left(L_2+L_3\right)}{2}& \left(3\right)\end{array}$

Algorithm

[0063] The physical property of the diffusion tensor is that all eigenvalues are non-negative. However, correction algorithms and low-resolution data can lead to negative eigenvalues that are clearly non-physical and are usually the results of numerical algorithms imperfection. Therefore, in the present procedure all negative, nonphysical eigenvalues are set to zero before normalization. Replacing them with zero is preferred over replacing them with a modulus, which also is possible, however much less justified numerically.

[0064] The first step of a method of diagnosing degree of neuropathy of a nerve by diffusion tensor mapping is acquiring data to estimate diffusion tensor imaging (DTI) metrics, especially with echo planar imaging sequences (EPI DTI). The data are digital images of the nerve based on magnetic resonance, MRof the part of the body comprising said nerve. Digital images acquired in this step are especially obtained by echo planar imaging diffusion tensor imaging (e.g. SE EPI, GRE EPI), but using other sequences is also possible. Preferably, while using EPI, the acquisition of said images is conducted for no more than 15 minutes.

[0065] Further, for each voxel in said DTI digital images a diffusion tensor D is calculated:

[00014]$\begin{array}{cc}\hat{D}=\left[\begin{array}{ccc}{D}_{\mathrm{xx}}& D_{\mathrm{xy}}& D_{\mathrm{xz}}\\ D_{\mathrm{yx}}& D_{\mathrm{yy}}& D_{\mathrm{yz}}\\ D_{\mathrm{zx}}& D_{\mathrm{zy}}& D_{\mathrm{zz}}\end{array}\right].& \left(4\right)\end{array}$

Along each of the x, y and z principal directions of diffusion, the diffusion coefficient is L.sub.1, L.sub.2, L.sub.3, respectively. The unit vectors (1,0,0).sup.T, (0,1,0).sup.T and (0,0,1).sup.T, directed along x, y, and z, are called eigenvectors and are referred to as e.sub.1, e.sub.2, e.sub.3. Each eigenvector corresponds to one eigenvalue. Eigenvectors and eigenvalues are related by the fact that when a tensor is multiplied by an eigenvector, the result is the same eigenvector multiplied by the eigenvalue:

[00015]$\begin{array}{cc}\begin{array}{cc}\hat{D}{e}_i=L_1{e}_i& i=\left\{1,2,3\right\},\end{array}& \left(5\right)\end{array}$

[0066] Further, eigenvalues L.sub.1>L.sub.2>L.sub.3 of the a diffusion tensor {circumflex over (D)} are found.

[0067] For every subject eigenvalues are normalized by dividing by maximum value of the L.sub.1 among all voxels. An alternative method is dividing by median value of the L.sub.1. Changing the normalization method would result in the necessity of re-calculating ? values (? is referred to below).

[0068] For each voxel in said DTI digital images, fractional anisotropy FA, mean diffusivity MD, and radial diffusivity RD are calculated, according to equations (1), (2) and (3) above.

[0069] The data visualization algorithm itself will be described below, and consists of six steps each having a significant impact on the displayed image. The visual effect accomplished by each step is presented in FIG. 1a-f.

[0070] The first step (FIG. 1a) is associatingwith the first eigenvectora line segment representing the main direction of the diffusion in this voxel, said line segment having a length, an opacity and an RGB color assignment. The first eigenvector in each voxel is represented as a line segment representing the main direction of diffusion in that voxel, and the line segments are given some initial length, which may be the same length. Because the morphological structure of the optic nerve favors the presence of one particularly distinguished direction of diffusion, especially voxels in which diffusion is strongly directional i.e., L.sub.1>>L.sub.2 should be taken into account. A picture resulting from the first step does not have to be actually displayed.

[0071] Hence, in the second step (FIG. 1b), the length of the line segment in each voxel is adjusted to a value given by

[00016]$\begin{array}{cc}{?\left(L_1-\mathrm{RD}\right)}^{?}& \left(6\right)\end{array}$

where ? is a scaling parameter. This means that the more L; deviates from RD, the longer is the line segment for that voxel. The scaling parameter ? ranges from 1 to 1000. In the present embodiment, parameter ? is set to 200. ? is also a scaling parameter, ranging from 1 to 3. For example, scaling parameter ?, in this embodiment, is equal to 2.2.

[0072] Another important piece of information is the strength of diffusion in each voxel. Thus, in the third step (FIG. 1c), the opacity of the line segment in each voxel is adjusted, i.e. is set proportional to the product of the mean diffusivity and fractional anisotropy, to a value given by

[00017]$\begin{array}{cc}?*\mathrm{MD}*\mathrm{FA}& \left(7\right)\end{array}$

where ? is a scaling parameter, ranging from 1 to 1000. This means that the stronger the diffusion, the opaquer the segment for that voxel. In the present embodiment, parameter B is set to 500.

[0073] The fourth step in the algorithm (FIG. 1d) is an RGB color assignment that improves image clarity based on L.sub.1 values as follows

[00018]$\begin{array}{cc}\mathrm{Red}=?L_1\mathrm{Green}=1-L_1\mathrm{Blue}=0& \left(8\right)\end{array}$

where ? is a scaling parameter for setting the intensity of red color, ? ranging from 0 to 3. In the present embodiment, parameter ? is set to 2. The white arrow in FIG. 1d shows oculomotor nerve. The goal of choosing a different color scaling than in the standard tractography is to assume that the L; values should be similar along the entire length of the optic pathway, and therefore, it should be presented in a similar color. Furthermore, standard tractography color coding appears to be ineffective for the optic pathway because this structure does not form a straight-line segment in three-dimensional projections and such coding would unnecessarily introduce different colors on different sections of the optic nerve and tract. For this reason, for optic pathway visualization, more than one layer must be used.

[0074] In step five (FIG. 1e), the three-dimensional shape of the visual pathways is projected onto the transverse plane, by superposing two or more layers of the adjusted three-dimensional shape on said transverse plane. In the present embodiment, seven layers were superimposed. The layers are preferably selected such that they are essentially parallel to the axis of the optical nerve. Thereby, a planar superimposed image is obtained, to be further used as an DTI digital image for the next steps. As a consequence, although the information about shape in the direction of the projection is lost, a much-improved visibility in the two other directions is gained (i.e., the plane perpendicular to the projection direction). On the transverse plane, the visual pathway is best visualized, but to show the complete location of the nerve, the segmented optic pathway is projected manually onto the sagittal and coronal planes (FIG. 2). In other words, after the projection of a few layers onto one plane, a superposition of line segments in each voxel is obtained, producing enhanced visibility.

[0075] Then, a region of interest (ROI) in at least one of said DTI digital images is selected, said region of interest comprising pixels.

[0076] Segment superposition can affect image clarity and for this reason a sixth step is implemented (FIG. 1f), in which short line segments are discarded based on the following cutoff condition

[00019]$\begin{array}{cc}{L}_1>\mathrm{RD}+?,& \left(9\right)\end{array}$

where ? is a threshold value (FIG. 4). In other words, the threshold value ?.sub.th of a cutoff parameter ? is calculated, ranging from 0 to L.sub.1, as the minimal value of ? for which the cutoff condition given by equation (9) is false for all pixels within the region of interest, wherein RD is radial diffusivity defined as in equation (3). Filtering of data used for imaging by applying a threshold value (abbreviated by ?.sub.th) found by analysis of diffusion tensor components of the image, and abandonment data in all pixels for which radial component of diffusivity (abbreviated by RD) increased by the threshold value (?.sub.th) exceeds the first main component (abbreviated by L.sub.1).
For the sake of clarity: the threshold value of the ? estimated by equation (9) is the maximal value of the ? for which condition (9) is true for all pixels within the region of interest (as well as the minimal value of ? for which the cutoff condition given by equation (9) is false for all pixels within the region of interest). In another words, the threshold value of the ? is equal to the largest difference between axial and radial diffusivity for the pixels inside the region of interest. A boundary of the region of interest includes an optic nerve ranging from eyeball to the optic canal.
The result is that line segments are displayed only if the axial diffusivity exceeds the radial diffusivity by more than a chosen value for ?. This will be described in more detail further in the text.
Thus, the ? parameter can be used as a quantitative measure in the diagnosis of nerve neuropathy (atrophy). Preferably, and in the present embodiment, neuropathy of the nerve, here optical nerve, is diagnosed for ?.sub.th lower than 0.27.

[0077] The presented algorithm and projections in three directions are particularly suited for the imaging of fiber structures. The optic nerves are clearly visualized in FIG. 2 (arrows). The right-hand side nerve is visualized as a single fiber, while the left-hand side nerve is split on coronal and sagittal projections because it is located just between the two planes used in the imaging protocol. Parameters ?, ?, ? and ? were selected once to ensure the best visibility of the visual pathways. These parameters may require modifications for other nerves.

[0078] Diffusion tensor estimation can produce indefinite diffusion tensors that arise in locations where the diffusion-weighted signal is disturbed by noise and the degree of anisotropy is high. In biological tissue, all eigenvalues of the diffusion tensor are assumed to be positive but due to noise or signal drop, negative eigenvalues may be generated. The fraction of voxels for which negative eigenvalues have been generated relative to the total number of all voxels, expressed as a percentage, can be used as a measure of image quality. For a healthy subject, those values were 0.61%, 6.5% and 9.0% for L.sub.1, L.sub.2 and L.sub.3 maps, respectively. In the case of a subject with a damaged eyeball, these values reached the values of 0.99%, 2.3%, and 9.5%, respectively.

[0079] Datasets from nine healthy subjects (the Young Adult Diffusion Dataset release of healthy adults between the ages of 20 and 59) from the HCP database (https://www. humanconnectome.org) were used for experimental evaluation. The HCP data scanning protocol was approved by the local Institutional Review Board (IRB) at Washington University. The HCP database provides diffusion data that was acquired with a high-quality image acquisition protocol using a customized Connectome Siemens Skyra scanner and processed using a well-designed processing pipeline including motion correction, eddy current correction and distortion correction. The diffusion acquisition parameters in HCP were: TR=5.520 ms, TE=89.5 ms, FA=78?, voxel size=1.25?1.25?1.25 mm, and FOV=210?180 mm. A total of 288 images were acquired in each diffusion dataset, including 18 baseline images with a low diffusion weighting b=5 s/mm.sup.2 and 270 diffusion weighted images evenly distributed at three shells of b=1,000/2,000/3,000 s/mm.sup.2. In this study only single-shell b=1,000 s/mm.sup.2 data, consisting of 90 diffusion weighted images and 18 baseline images, were used for calculating with the present algorithm. Single-shell b=1,000 s/mm.sup.2 data were used because it is similar to our clinical acquisition protocol. Furthermore, single-shell b=1,000 s/mm.sup.2 data has been shown to be more effective for identification of cranial nerves than higher b values. A visual check of the diffusion data for each of the subject was performed and any subjects whose diffusion data had incomplete optic nerve coverage were excluded. An FSL Nudge Tool was used to manually adjust the affine for every image to cover the layers along the optic nerve and then the diffusion tensor was calculated. Finally, nine subjects were chosen, and thresholds of ? value were calculated. These nine subjects from HCP database were not underwent a full ophthalmological examination, and if there are any visual diseases, this may be the reason of the ? values fluctuations. No visual fixation and eyeballs movement during the examination made it difficult to distinguish the optic nerve from the periocular muscles.

[0080] The presented algorithm was used to generate images of the visual pathway in a healthy male subject (FIG. 3b) and a male subject with the right optic nerve atrophy caused by eyeball injury (FIG. 3a). Blue arrows in FIG. 3b show optic canal and locations of visible discontinuities. Values of the ?, ?, ?, ? and ? used in visualization were 200, 2.0, 500, 2.0 and 0.2, respectively. Actual values of FA and MD measured in a ROI containing the optic nerve are presented in FIG. 3 in green. For progressive optic neuropathy in the second subject an increased MD and decreased FA in the optic nerve have been found, compared to the healthy optic nerve of the same subject, and compared to the healthy optic nerve of the first subject (Table 1). In T1 image a constriction of the atrophic nerve compared to the left optic nerve have been found (FIG. 3). The DTI image represented as line segments generated by the said algorithm showed significant difference between the atrophic and fully functional optic nerves with no segmentation required. The most important for unidirectional diffusion data were extracted by introduced threshold parameter ?.sub.th. A set of images generated for different values of ? show that this parameter can be used as an indicator of the degree of neuropathy (FIG. 4). Both optic nerves from a healthy subject and left optic nerve from the subject with right eyeball injury are removed from the image when ? exceeds almost the same thresholding level (Table 1). In contrast, the atrophic nerve is not visible even for ? equal to zero (FIG. 4a). An example of the region of interest used for determining the ?.sub.th value is shown in green in FIG. 3.

TABLE-US-00001 TABLE 1 Fractional anisotropy (FA), mean diffusivity (MD) in the optic nerve and ? threshold value for which the last pixel inside the optic nerve is no longer displayed. subject with penetrating right eyeball injury healthy subject right eye left eye right eye left eye FA 0.40(19) 0.52(12).sup. 0.53(20).sup. 0.53(20).sup. MD [mm.sup.2/s] 2810(830) 728(280) 897(280) 900(360) ?.sub.th 0 0.35 0.30 0.36

[0081] Threshold values of ? for which the last pixel inside the optic nerve is no longer displayed, calculated for 21 non-atrophic optic nerve (every nerve from 9 HCP subjects, two nerves from healthy subject, and one healthy nerve from subject with right eyeball injury) were similar (Table 2). The data presented in Table 2 show that standard deviation of the ?.sub.th parameter is over ten times smaller than the average value confirming reasonably good reproducibility. It was also checked that the particular choice of parameters ?, ?, ? and ? (Equations (6), (7), (8) and FIG. 3) that resulted in improvement of nerve visualization does not affect the threshold value of ? at which the last pixel inside the optic nerve is no longer displayed (FIG. 4). FIG. 4 presents images for male subject with right optic nerve neuropathy generated for ?.sub.th values equal to A) 0, B) 0.15, C) 0.25, and D) 0.35. FIG. 4 presents also density plots showing all voxels in the datasets used for visualization (black dots) and voxels omitted according to condition (9) (grey dots). The dots are obtained from all voxels in the selected layers. The red dashed lines correspond to the condition given by equation

[00020]$\begin{array}{cc}{L}_1=\mathrm{RD}+?.& \left(10\right)\end{array}$

The left optic nerve, shown by white arrow, is visible on FIG. 4A,B,C until a certain value of ? is exceeded (FIG. 4D).

[0082] It is important to note that the presented algorithm allows for visualization of the oculomotor nerve (indicated by the white arrow in FIG. 1d), consistent with the anatomical atlases. The bone canal through which the optic nerve passes (optic canal) may be the reason for the observed optic nerve discontinuities at this point (marked with blue arrows in FIG. 3).

TABLE-US-00002 TABLE 2 Distribution of ?.sub.th value for 21 non-atrophic optic nerves. parameter value median 0.33 mean 0.33 standard deviation 0.03 min 0.27 max 0.39

[0083] By finding the ? threshold value for which the last pixel inside the optic nerve is no longer displayed, a quantitative measure of optic nerve neuropathy is obtained. Other numerical voxel-based measurements are possible by optic nerve segmentation from images in which voxel brightness intensities are set based on the product of segment length and opacity. In such a monochrome image, it is easy to draw a region of interest's mask, for example, on the part of the optic nerve starting behind the eyeball and ending on the skull optic canal, where the line continuity is lost. The ? parameter can also be helpful in drawing ROIs for measurements of diffusion tensor scalar values. In other words, founded masks can be automatically corrected by ? parameter, which may be a threshold value used for semi-automatic segmentation, or active contour segmentation. Both methods can be alternatives to different tractography segmentations and manual methods.

[0084] Knowing the sigma (?) value at which the optic nerve disappears from the image allows to obtain a quantitative measure of optic neuropathy without the need for manual segmentation, which can be difficult due to the small size of the optic nerve.

[0085] Thus, as said above, the important conclusion is that the ? parameter can be used as a quantitative measure in the diagnosis of nerve neuropathy.

[0086] The selection of the ?, ?, ? and ? parameters does not affect the value of the 0th parameter, for which the optic nerve disappears, because the result of the ? parameter is presented in the image in an objective way, and the change in the parameters ?, ?, ? and n affect only subjectively the visibility of the optic nerves. The measurement of the threshold value of the ? parameter, calculated using normalized eigenvalues, is absolute, which makes it possible to compare optic nerves obtained on different scanners and with different acquisition parameters.

[0087] It is also expected that the nerve in the image will always disappear at the same ? value, regardless of the ?, ?, ? and ? parameters set, although these parameters must be set so that before increasing the ? value, the optic nerve is visible.

[0088] The present invention also shows that the projection of a few layers on a single plane is a preferential method to view fibrous tissues. Even though some 3D information is lost, the visibility enhancement of the entire structure is significant. Since multilayer projections cannot be merged with single layer morphological images due to possible mismatch, in FIG. 2 only a part of the T1 image in the background is shown.

[0089] The proposed new approach of diffusion tensor visualization of the optic nerve and tract can be a complementary to the standard tractography methods, particularly in cases when the standard results that do not fit well with the known anatomy.

[0090] The presented algorithm allows for a direct visualization of data for fibrous structures, such as the optic nerve, by diffusion tensor mapping, without relying on probabilistic imaging tools. Appropriate matching and combining scalar values of the diffusion tensor with the eigenvector data can result in the generation of an image that shows a significant advantage in nerve neuropathy assessment. As shown on the example of optic nerve, the atrophic and healthy nerves can be differentiated with the present algorithm without any segmentation and can be correlated with diffusion voxel-based metrics such as FA or MD. The visibility of the optic nerve, chiasm and tract by an analysis of eigenvectors and eigenvalues and an optimized assignment of visual components that include segments length, opacity and color depth which together enhance the visibility of the fiber structures. As shown, the images generated by the presented algorithm clearly show significant differences between the atrophic and healthy optic nerves from the eyeball to the optic chiasm and the ? parameter in the algorithm provides a quantitative measure of nerve atrophy. The estimation of the parameter on 21 non-atrophic optic nerve show reasonably good reproducibility since the standard deviation is ten times smaller than its average value.

[0091] Decreased FA and increased MD found in the optic nerve with progressive optic neuropathy coincide with the values reported for amblyopathy, glaucoma, optic neuritis, retinitis pigmentosa and multiple sclerosis, what has been shown by many studies, and the proposed method may help to diagnose such diseases.

[0092] The described approach may be a helpful complement to existing tractography methods not only of the optic nerve, but also other nerves having substantially fibrous shape, for example an auditory nerve, an oculomotor nerve, a trigeminal nerve, a vestibulocochlear nerve, white matter of a brain or a spinal cord.