Eur Radiol (2008) 18: 2283–2291 DOI 10.1007/s00330-008-0971-4

C. Khalil C. Hancart V. Le Thuc C. Chantelot D. Chechin A. Cotten

Received: 5 August 2007 Revised: 5 March 2008 Accepted: 13 March 2008 Published online: 17 April 2008 # European Society of Radiology 2008

C. Khalil (*) . C. Hancart . V. Le Thuc . A. Cotten Service de Radiologie Ostéoarticulaire, Hôpital Roger Salengro, CHRU de Lille, 59037, France e-mail: [email protected] Tel.: +33-3-20446103 Fax: +33-3-20446135 C. Chantelot Clinique d’ Orthopédie, Hôpital Roger Salengro, CHRU de Lille, France D. Chechin Philips Medical Systems, Suresnes, France

MUSCULO SKELETAL

Diffusion tensor imaging and tractography of the median nerve in carpal tunnel syndrome: preliminary results

Abstract The purpose was to demonstrate the feasibility of in vivo diffusion tensor imaging (DTI) and tractography of the human median nerve with a 1.5-T MR scanner and to assess potential differences in diffusion between healthy volunteers and patients suffering from carpal tunnel syndrome. The median nerve was examined in 13 patients and 13 healthy volunteers with MR DTI and tractography using a 1.5-T MRI scanner with a dedicated wrist coil. T1-weighted images were performed for anatomical correlation. Mean fractional anisotropy (FA) and mean apparent diffusion coefficient (ADC) values were quantified in the median nerve on tractography images. In all subjects, the nerve orientation and course could be detected with

Introduction Diffusion tensor imaging (DTI) based on magnetic resonance (MR) can provide valuable information about tissue microstructure and architectural organization by monitoring in vivo random microscopic motion of water protons, and measuring anisotropy (i.e., diffusion inhomogeneities of water protons in space) [1–5]. Furthermore, tissue orientation and course can be visualized with recently developed algorithms processing DTI data. This technique is referred to as MR tractography or fiber tracking [6–11]. DTI with fiber tracking has found clinical applications in the evaluation of the central nervous system [5, 12–19]. Indeed, it is most interesting in tissues with an organized microstructure, such as white matter tracts of the human brain [5]. In the latter, diffusion is slightly higher along the

tractography. Mean FA values were significantly lower in patients (p=0.03). However, no statistically significant differences were found for mean ADC values. In vivo assessment of the median nerve in the carpal tunnel using DTI with tractography on a 1.5-T MRI scanner is possible. Microstructural parameters can be easily obtained from tractography images. A significant decrease of mean FA values was found in patients suffering from chronic compression of the median nerve. Further investigations are necessary to determine if mean FA values may be correlated with the severity of nerve entrapment. Keywords MRI . DTI . Tractography . Carpal tunnel syndrome . Median nerve

axons than perpendicular to them, resulting in anisotropic diffusion. Even though fiber tracking has been extensively used to image white matter tracts, its application in the assessment of peripheral nerves has been more limited. DTI with fiber tracking of a proximal peripheral nerve (i.e., sciatic nerve) has been reported in a study performed on three healthy volunteers, using a 1.5-T MR scanner [20]. To the best of our knowledge, distal peripheral nerve assessment has only been reported in three studies performed on 3-T MR scanners, but never on a 1.5-T magnet [21–23]. Furthermore, and as suggested by Hiltunen et al., we believe that DTI with fiber tracking may provide clinically relevant information and show abnormalities that are beyond the resolution of conventional MR techniques for the diagnosis and follow-up of nerve entrapments [22]. We decided to test this hypothesis

2284

in patients with carpal tunnel syndrome, as this syndrome is the most common nerve compression disorder in the upper extremity, and its positive diagnosis is usually established with confidence, relying upon characteristic clinical features and nerve conduction analysis [24]. The aims of this preliminary study were to (1) demonstrate the feasibility of in vivo DTI and fiber tracking of the human median nerve with a 1.5-T MR scanner, (2) provide a reliable way of obtaining microstructural parameters [mean fractional anisotropy (FA) and mean apparent diffusion coefficient (ADC)] of the median nerve on tractography images and (3) assess potential differences in diffusion within the median nerve between healthy volunteers and patients suffering from carpal tunnel syndrome.

Materials and methods –

Population

Our study had received prior approval from our Institutional Review Board and Ethics Committee. Thirteen consecutive patients (11 females, 2 males; mean age: 54 years; range: 45 to 71 years) suffering from carpal tunnel syndrome (CTS) and referred to our institution were prospectively included in this study. They were matched to 13 asymptomatic volunteers according to age and sex (11 females, 2 males; mean age: 55 years; range: 47 to 69 years). Each volunteer or patient provided informed consent. Exclusion criteria were a previous history of trauma or surgery of the studied wrist and/or the existence of contraindications to MRI. All volunteers were asymptomatic, did not suffer from a disease and had not received any drug that could alter nerve conduction. A normal electromyogram (EMG) was mandatory for each volunteer. All patients had a clinical history consistent with CTS, which had been supported by electrodiagnostic investigation. The diagnosis required a median sensory nerve conduction velocity within the index finger-to-wrist segment of less than 45 m/s [25], a difference of >0.3 ms between the median and ulnar nerve palm-to-wrist latencies [26] or a median distal motor latency >4.4 ms [26–28]. All patients had normal ulnar nerve conduction studies. In addition, patients with a clinical history or electrodiagnostic findings suggesting a concomitant peripheral neuropathy were excluded. –

MRI examination

(1) Acquisition parameters MRI scans were performed on a 1.5-T full-body scanner (Achieva, Philips Medical Systems, Best, The Nether-

lands). Volunteers and patients were scanned in supine position, arms alongside the body, palm against thigh, feet first. A dedicated four-channel phased array wrist coil (INVIVO corp, Orlando, FL) was used to scan the symptomatic wrist of each patient or, arbitrarily, the dominant wrist of each volunteer. The examined wrist was immobilized with cushions and bandages. No exogenous contrast agent was required. Prior to DTI, high-order shimming with 10-cm field of view (FOV) was applied to reduce the inhomogeneities of the main magnetic field in the imaging area. Diffusion tensor imaging was performed in the axial plane with a spin-echo single-shot diffusion-weighted echo-planar imaging sequence, a b-value of 400 s/mm2 in 32 different diffusion gradient orientations in addition to baseline T2weighted EPI images (b=0 images). Δ was 43.8 ms; δ was 26.4 ms. The diffusion images acquisition parameters were as follows: TR=3,198 ms, TE=89 ms, FOV 100 mm, parallel imaging technique was used with a sense factor of 2, partial Fourier acquisition (half scan factor: 0.681), matrix scan=80×80, reconstruction matrix=128×128 using zero-filling interpolation, the in-plane resolution was 0.78×0.77 mm. A frequency selective fat saturation was used (SPIR). Twenty contiguous slices were acquired with a slice thickness of 2 mm, placed in order to cover the entire carpal tunnel with the most distal slice located at the level of the carpo-metacarpal joint. In order to prevent any error in slice positioning, the latter was under the supervision of the same radiologist (observer 1) for each examination. T1-weighted turbo spin echo sequence was performed in the axial plane to match the anatomical location of the DW images with the following parameters: TR=500 ms, TE=16 ms, image matrix=512×512. All the other parameters (field of view, slice thickness, numbers and overlap) were the same as for DTI imaging. Total acquisition time was 6 min 14 s (1 min 55 s for DTI and 4 min 19 s for T1weighted images). (2) Diffusion registration and post-processing Eddy current and motion-related misalignment of the diffusion tensor MR images was corrected off-line using Automated Image Registration software (Philips Pride Diffusion Registration and IDL, ITT, Boulder, CO). All diffusion-weighted images were reoriented to match the b0 images. This preliminary coregistration procedure helped to minimize distortions and artifacts, and achieve proper alignment prior to image analysis [29, 30]. (3) Data analysis The MR imaging analysis was performed independently by two radiologists (with 3 years and 13 years of experience in MR imaging) blinded to the clinical status and electrophysiological findings. The first reader performed a second blinded analysis after a 1-month period.

2285

a

Fractional anisotropy maps

FA maps were calculated using an IDL-based (Research Systems INC., Boulder, CO) PRIDE (Philips Research Imaging Development Environment) fiber-tracking program (software version 6.2) provided by the manufacturer. The diffusion tensors at each voxel were diagonalized so as to obtain eigenvalues and eigenvectors for each voxel. The eigenvector associated with the largest eigenvalue was used to represent the local fiber direction. FA maps were calculated from the eigenvalues on the basis of standard formulas [4, 31]. Wrist color maps were created on the basis of the three vector elements for each voxel [31, 32]. The absolute values of the principal eigenvector were assigned to different colors: red (left-right component), green (volar-dorsal component) and blue (proximal-distal component). If the main eigenvector was aligned along the proximal-distal axis, pure blue was assigned to the corresponding voxel, whereas if the eigenvector was 45° between the left-right and proximal-distal axes, purple (red plus blue) was assigned to the voxel. Color intensity in each voxel was correlated to the degree of FA. The resulting image was weighted by values of FA image to exclude the tissues with isotropic diffusion. This color information was then combined with the results of further fiber tracking. b

Median nerve fiber tracking

Fiber tracking was performed using a line propagation technique from the Philips PRIDE fiber tracking software. Tracking was launched from a seed region of interest (ROI) from which a line was propagated in both the retrograde and anterograde directions according to the main eigenvector at each voxel [9]. The median nerve was first identified on the anatomic TSE T1-weighted images at the level of the pisiform bone. Then, a freehand ROI was positioned on the cross-sectional area of the median nerve on the FA color-coded map. The following parameters were used for fiber tracking: manual ROI position, FA threshold=0.3, direction thresholds=±7 degrees, step size=0.9. Figs. 1 and 2 show tractography results in respectively a volunteer and a patient, illustrating the median nerve at different levels of the carpal tunnel. c

Assessed parameters

Once fiber tracking was performed, with a single right click with the mouse on the bundle of fibers, the software gave the following measured quantitative parameters: mean FA value, mean ADC value (Fig. 3). d

Reproducibility of DTI data in time

Fig. 1 Medial view of the median nerve imaged with DTI in a volunteer showing the 3D course of the nerve through consecutive levels of the carpal tunnel: a at the level of the hook of the hamate, b at the level of the carpometacarpal joint, c distally at the level of the metacarpal bases. Due to its orientation (proximal-distal direction), the median nerve appears predominantly coded in blue

mean ADC were calculated in order to assess reproducibility of the diffusion data in time. (4) Signal to noise measurements

A volunteer underwent three additional MRI examinations (once a day, every 2 days) out of which mean FA and

Since signal-to-noise ratio (SNR) is a critical point in fiber tracking, it was calculated according to the method

2286

median nerve, carefully positioned in order not to include any obvious artifacts, on either one of the original unsubstracted images. Signal is the mean value of the pixel intensity in that ROI is minus any offset. The noise is defined as the standard deviation (SD) derived from using the same ROI on the subtracted image. The calculated SNR is as follows: SNR=S√2 / SD [33]. –

Statistical analysis

Because of the small sample size, statistical analysis was performed with non-parametric tests. Spearman coefficient was used in order to calculate inter- and intra-observer agreement rates: a κ0.80 excellent agreement. Comparisons between groups were performed with the Wilcoxon (or Mann Whitney) test. P values inferior to 0.05 were considered statistically significant. Fig. 2 Tractography image demonstrating the median nerve, coded in blue, with an excellent correlation, with the reference T1weighted image in a patient suffering from carpal tunnel syndrome

described by Price et al. as follows: two consecutive scans with identical scan parameters were acquired on the same subject (registration was also performed for each scan). Both scans were subsequently subtracted at b=0, hence creating a third pixel-by-pixel difference image. The signal was measured using a ROI drawn in the center of the Fig. 3 Example of mean ADC and FA values obtained on the median nerve with the tractography software

Results 1

Direction encoded color maps (DEC maps)

FA maps appeared as color-coded axial images in which nerves and muscles were color-coded depending on their fibers orientation. Median nerve could be identified on all examinations as a round or oval shaped structure isointense

2287

to muscles on T1-weighted images and blue coded on FA maps. There was no aberration in the color encoding scheme. Distal branching was not assessed. Susceptibility effect related distorsions or signal dropout at tissue-bone interfaces were minimal, mostly on the most proximal and most distal slices, but never for the assessment of the median nerve. A careful visual comparison made by both observers between the DEC maps and the T1-weighted images at each slice position insured that the nerve was followed through the entire volume and that therefore the fiber tracking procedure did not include any of the low anisotropy surrounding tissues. The maps were of great value for ROI positioning for tractography. 2

Fiber tracking

In all healthy and pathological wrists, fiber tracking based on the diffusion tensor images identified a bundle of fibers in the median nerve. The location of these fibers matched the location of the median nerve on reference T1 images (Fig. 4). As expected, the median nerve, because of its course, was predominantly coded in blue (proximal-distal direction) (Figs. 1 and 2). There was no visual modification of the tractography images between patients and volunteers. No focal nerve flattening was identified, especially in patients.

3

Mean ADC and mean FA values

No significant differences in the mean ADC values were observed between healthy volunteers and patients suffering from carpal tunnel syndrome (p=0.2). Results of mean ADC values are reported in Table 1. Mean FA values were statistically lower (p=0.03) in patients suffering from carpal tunnel syndrome. Results of mean FA values are reported in Table 2. Interobserver and intraobserver agreement were excellent for both mean ADC values (respectively κ=0.94, and κ=0.90) and mean FA values (respectively κ=0.89, κ=0.95) (Fig. 5). A mild overlap between FA values in patients and volunteers was observed (Fig. 6). 4

Reproducibility of DTI data in time

In the volunteer who underwent three additional examinations, mean FA over time was 0.536 (standard deviation over time was 0.021), mean ADC over time was 1.862 mm2/s (standard deviation over time was 0.133 mm2/s). 5

SNR measurements

The measured SNR, at b=0, in the median nerve was 17.61 (mean of the ROI in the nerve: 19.43, standard deviation of the noise: 1.56).

Discussion

Fig. 4 Cross-sectional color-coded FA map at the level of the carpal tunnel superimposed with the corresponding reference T1-weighted image showing that the tracked fibers matched the location of the median nerve. White arrow: median nerve

Our study demonstrates the feasibility of in vivo DTI and fiber tracking of the human median nerve with a 1.5-T MR scanner. This imaging is possible both in healthy volunteers and in patients suffering form carpal tunnel syndrome. The use of T1-weighted MRI sequence as an anatomical reference confirmed that we were dealing with the median nerve. To the best of our knowledge, peripheral nerve assessment with DTI tractography on a 1.5-T scan has been reported in only one study for a large proximal nerve (i.e., sciatic nerve) of three healthy volunteers [20]. Visualizing smaller peripheral nerves using DTI tractography has only been reported in three studies performed with a high-field 3-T MRI scan (respectively in 6 healthy volunteers, in 3 healthy volunteers and 1 patient, and in 20 volunteers and 2 patients) [21–23]. Our study is the first to directly compare two populations of patients and volunteers. We used tractography as a tool for the analysis of the median nerve microstructure, since nerve fibers were extracted from a single ROI. This prevented us from going through the time-consuming process of crosssectional area measurements of mean FA and mean ADC values by drawing a ROI at each level. Another advantage of this technique is its reproducibility as interobserver and intraobserver agreements were excellent for both mean

2288

Table 1 Mean ADC values: there were no significant differences in mean ADC values between patients and volunteers Participants

Observer 1

Observer 1

Observer 2

Volunteers Median Standard deviation Patients Median Standard deviation

Observer 1, first analysis 1.733 0.250 Observer 1, first analysis 1.693 0.211

Observer 1, second analysis 1.832 0.383 Observer 1, second analysis 1.696 0.196

Observer 2 1.735 0.277 Observer 2 1.703 0.240

ADC values and mean FA values, and as DTI data results in a volunteer were reproducible in time. Our study demonstrated a significant decrease of mean FA in patients suffering from carpal tunnel syndrome, which may suggest an alteration of water proton diffusion along the median nerve axis. The baseline SNR measured in our study seemed good enough to ensure the quantified FA studies. Indeed, as reported by Papadakis et al. in their simulation, whenever baseline SNR (b=0) is superior to 15, the value of FA bias never exceeds 2% of the theoretical value [34]. In our study, the baseline SNR, calculated in vivo, was superior to that threshold. We were also conforted by the low standard deviation of mean FA values measured in both patients and volunteers. The only two other cases of mean FA variations in a pathological median nerve were reported by Meek et al. in a patient who had suffered accidental transsection [21] and by Kabakci et al. in two patients suffering from CTS [23]. In the former study, the patient underwent two DTI examinations, respectively 1 and 2 months after fascicular nerve repair. In the first scan, the median nerve could only be tracked up to the site of the lesion. It could be tracked more distally in the second scan, even though no nerve function had clinically improved yet, demonstrating that nerve regeneration could be followed with DTI and tractography. In the latter study, Kabakci et al. observed a decrease of mean FA values in two patients, but in their study, there was no direct comparison of patients and volunteers. The significant mean FA decrease observed in our patients might be related to histological changes occurring after chronic compression. These abnormalities consist in varying combinations of segmental demyelination, Wallerian degeneration and eventually axonal damage

[35]. Chronic pressure also affects the connective tissue components of nerves. Post-mortem studies of the median nerve in severe carpal tunnel syndrome have reported a marked increase of the endoneurial and perineurial connective tissue under and proximal to the flexor retinaculum [36, 37]. Intrafascicular edema due to local venous compression may also be associated. These changes lead to an increase of the distance between axons and between axons fascicles. One can therefore assume that these changes might alter the diffusion of water protons along the nerve fibers, because of a slight increase of diffusion vectors in a perpendicular direction. The diffusion ellipsoid would therefore tend towards a more spherical shape consistent with the decrease of mean FA values calculated in patients suffering from carpal tunnel syndrome. In contrast, no statistically significant difference in mean ADC values was found between volunteers and patients. One might have expected an increase in ADC values in patients suffering from carpal tunnel syndrome, similarly to what has been reported in Wallerian degeneration of the pyramidal tract after ischemic stroke [38, 39]. Nevertheless, previous DTI studies have also reported a relatively smaller increase of ADC values compared with the decrease of FA in Wallerian degeneration [39, 40]. Indeed, in Wallerian degeneration, there is neither significant water accumulation in the interstitial space nor formation of cysts, both of which could lead to a marked increase in ADC. Furthermore, as Pierpaoli et al. state, the reduction in diffusion anisotropy accompanied by decreased diffusivity parallel to the fibers increased diffusivity perpendicular to them and a relatively small change in ADC, which taken together suggest that there is an increase of isotropic tissue

Table 2 Mean FA values: mean FA values were significantly lower in patients for both observers Participants

Observer 1

Observer 1

Observer 2

Volunteers Median Standard deviation Patients Median Standard deviation

Observer 1, first analysis 0.588 0.058 Observer 1, first analysis 0.528 0.063

Observer 1, second analysis 0.601 0.053 Observer 1, second analysis 0.520 0.063

Observer 2 0.599 0.052 Observer 2 0.514 0.067

2289

Inter-observer agreement (ADC)

Fig. 5 Correlation plots showing excellent interobserver and intraobserver agreement

Inter-observer agreement (FA)

2,5

0,8

0,7

Reader 2

Reader 2

2

1,5

0,6

0,5

0,4

1 1

1,5

2

0,3

2,5

0,3

0,4

Reader 1

Intra-observer agreement (ADC)

0,5 0,6 Reader 1

0,7

0,8

Intra-observer agreement (FA)

2,5

0,8

2

Second analysis

Second analysis

0,7

1,5

0,6

0,5

0,4

1

0,3 1

1,5

2

First analysis

0.8

Fa Values

0.7

0.6

0.5

0.4

Volunteers

Patients

Fig. 6 Box plots showing a mild overlap between FA values in patients and volunteers

2,5

0,3

0,4

0,5

0,6

0,7

0,8

First analysis

structures in the regions where Wallerian degeneration has occurred, which is consistent with the presence of gliosis and the possible increase in extracellular matrix found histologically in WD [40]. Similar histological changes occur in case of nerve entrapment (thickening of the interstitial compartment: i.e., epineurium and endoneurium), and since only a slight decrease of mean FA values was found, we hypothesize that an even slighter increase of mean ADC might be observed. The small sample size of our study, and hence its limited power of statistical analysis, might have prevented us from detecting such a mild modification. Currently, one of the major drawbacks of DTI is its poor spatial resolution, making it still impossible to assess very small nerves (the minimal caliber remaining unknown). Therefore, it should be stressed that the feasibility of in vivo MR tractography in human median nerve using a 1.5T magnet requires several parameters to be combined in order to improve spatial resolution. We chose a b value of 400 s/mm2 (lower to the one that is usually applied in CNS studies) following the parameters of Skorpil et al. in their

2290

assessment of the sciatic nerve in healthy volunteers [20]. To minimize distortions and artifacts, image coregistration was performed. Furthermore, diffusion weighting gradients were applied in 32 directions (instead of the minimal 6), improving the accuracy of fiber tracking. Finally, the use of a dedicated wrist coil allowed us to achieve higher resolution images. None of the three studies assessing the median nerve with DTI and tractography on a 3-T MRI scan combined these technical modifications. Two of them were performed applying diffusion weighting gradients in 32 encoding directions, but were limited by the use of standard head coils (respectively 6- and 8-channel), whereas the other, using a dedicated wrist coil, was performed with only 13 different diffusion gradient orientations [21–23]. However, in order not to lengthen the acquisition time, we were forced to cut down the number of acquisitions for each encoding direction to one. We acknowledge that our study has several limitations. The first limitation is that only a relatively small number of subjects were investigated. Second, we did not correlate our results with the morphological abnormalities reported in CTS on conventional MRI sequences, as our purpose was to assess the alteration of DTI parameters and tractography in a nerve entrapment syndrome. Third, because of the mild overlap we observed between mean FA values in patients and volunteers’ median nerves, we were not able to determine a cut-off point under which mean FA would be considered as pathological (Fig. 6). Further studies are required to determine such a threshold and eventually to correlate the decrease of mean FA values with the severity of nerve entrapment on electrophysiological studies. Another limitation may be that positioning could have been optimized to assess the median nerve: a prone position with the subject’s arm above the head would have allowed the wrist to be evaluated in the center of the

tube, even though it would have added some discomfort to the examination. Finally signal-to-noise ratio could have been improved with the use of multiple acquisitions for each encoding gradient direction, at the cost of a longer examination time [23, 29]. Further studies may demonstrate a better compromise between the quality of tractography and the acquisition time. The improvement of this technique in the evaluation of CTS prior to surgery, a condition for which electrophysiological studies show characteristic patterns, might help in the assessment of symptom recurrence after surgery, which remains a diagnostic challenge. DTI with tractography might also represent an interesting tool in cases of discrepancies between clinical and electrophysiological findings.

Conclusion In conclusion, the results of this preliminary study demonstrate the feasibility of providing in vivo 3D visualization of the median nerve using tractography on a 1.5-T magnet and microstructural diffusion parameters (mean FA and mean ADC). We found significant changes of the diffusion properties of the median nerve in patients suffering from carpal tunnel syndrome. Indeed, patients demonstrated a significant decrease in mean FA values when compared with healthy volunteers. The non-invasive assessment of distal peripheral nerves with MR DTI appears to provide complementary information to the usual sequences. Further studies should be undertaken to confirm this potential in CTS as well as in other nerve compression disorders. Acknowledgments The authors would like to thank the department of Dr. J.F. Hurtevent, MD, PhD, for his help in the electrophysiological assessment of all subjects in this study.

References 1. Basser PJ, Jones DK (2002) Diffusiontensor MRI: theory, experimental design and data analysis-a technical review. NMR Biomed 15:456–467 2. Beaulieu C, Allen PS (1994) Determinants of anisotropic water diffusion in nerves. Magn Reson Med 31:394–400 3. Beaulieu C, Does MD, Snyder RE et al (1996) Changes in water diffusion due to Wallerian degeneration in peripheral nerve. Magn Reson Med 36:627–631 4. Basser PJ, Pierpaoli C (1996) Microstructural and physiological features of tissues elucidated by quantitativediffusion-tensor MRI. J Magn Reson B 111:209–219

5. Le Bihan D, Mangin JF, Poupon C et al (2001) Diffusion tensor imaging: concepts and applications. J Magn Reson Imaging 13:534–546 6. Bammer R (2003) Basic principles of diffusion-weighted imaging. Eur J Radiol 45:169–184 7. Bammer R, Acar B, Moseley ME (2003) In vivo MR tractography using diffusion imaging. Eur J Radiol 45:223–234 8. Mori S, van Zijl PC (2002) Fiber tracking: principles and strategies-a technical review. NMR Biomed 15:468–480 9. Mori S, Crain BJ, Chacko VP et al (1999) Three-dimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann Neurol 45:265–269

10. Basser PJ, Pajevic S, Pierpaoli C et al (2000) In vivo fiber tractography using DT-MRI data. Magn Reson Med 44:625–632 11. Budzik JF, Le Thuc V, Demondion X et al (2007) In vivo MR tractography of thigh muscles using diffusion imaging: initial results. Eur Radiol 17:3079– 3085 12. Rovaris M, Filippi M (2007) Diffusion tensor MRI in multiple sclerosis. J Neuroimaging 17(Suppl 1):27S–30S 13. Sugiyama K, Kondo T, Higano S et al (2007) Diffusion tensor imaging fiber tractography for evaluating diffuse axonal injury. Brain Inj 21:413–419

2291

14. Oppenheim C, Ducreux D, Rodrigo S et al (2007) Diffusion tensor imaging and tractography of the brain and spinal cord. J Radiol 88:510–520 15. Firbank MJ, Blamire AM, Krishnan MS et al (2007) Diffusion tensor imaging in dementia with Lewy bodies and Alzheimer’s disease. Psychiatry Res 155:135–145 16. Nguyen TH, Yoshida M, Stievenart JL et al (2005) MR tractography with diffusion tensor imaging in clinical routine. Neuroradiology 47:334–343 17. Mukherjee P (2005) Diffusion tensor imaging and fiber tractography in acute stroke. Neuroimaging Clin N Am 15:655–665, xii 18. Ohgiya Y, Oka M, Hiwatashi A et al (2007) Diffusion tensor MR imaging of the cervical spinal cord in patients with multiple sclerosis. Eur Radiol 17:2499– 2504 19. Bammer R, Fazekas F (2003) Diffusion imaging of the human spinal cord and the vertebral column. Top Magn Reson Imaging 14:461–476 20. Skorpil M, Karlsson M, Nordell A (2004) Peripheral nerve diffusion tensor imaging. Magn Reson Imaging 22:743–745 21. Meek MF, Stenekes MW, Hoogduin HM et al (2006) In vivo three-dimensional reconstruction of human median nerves by diffusion tensor imaging. Exp Neurol 198:479–482 22. Hiltunen J, Suortti T, Arvela S et al (2005) Diffusion tensor imaging and tractography of distal peripheral nerves at 3 T. Clin Neurophysiol 116:2315– 2323

23. Kabakci N, Gurses B, Firat Z et al (2007) Diffusion tensor imaging and tractography of median nerve: normative diffusion values. AJR Am J Roentgenol 189:923–927 24. Seror P (1993) Sensitivity of various electrophysiological studies for the diagnosis of carpal tunnel syndrome (150 cases). Muscle Nerve 16:1418–1419 25. Hansson S (1994) Does forearm mixed nerve conduction velocity reflect retrograde changes in carpal tunnel syndrome? Muscle Nerve 17:725–729 26. Stevens JC (1997) AAEM minimonograph #26: the electrodiagnosis of carpal tunnel syndrome. American Association of Electrodiagnostic Medicine. Muscle Nerve 20:1477–1486 27. Melvin JL, Schuchmann JA, Lanese RR (1973) Diagnostic specificity of motor and sensory nerve conduction variables in the carpal tunnel syndrome. Arch Phys Med Rehabil 54:69–74 28. Kimura J (2001) Electrodiagnosis in diseases of nerve and muscle: Principles and practice. Oxford University Press, New York 29. Mangin JF, Poupon C, Clark C et al (2002) Distortion correction and robust tensor estimation for MR diffusion imaging. Med Image Anal 6:191–198 30. Nielsen JF, Ghugre NR, Panigrahy A (2004) Affine and polynomial mutual information coregistration for artifact elimination in diffusion tensor imaging of newborns. Magn Reson Imaging 22:1319–1323 31. Jellison B, Field A, Medow J et al (2004) Diffusion tensor imaging of cerebral white matter:a pictorial rewiew of physics, fiber tracts anatomy and tumor imaging pattern. AJNR 25:356– 369

32. Douek P, Turner R, Pekar J et al (1991) MR color mapping of myelin fiber orientation. J Comput Assist Tomogr 15:923–929 33. Price RR, Axel L, Morgan T et al (1990) Quality assurance methods and phantoms for magnetic resonance imaging: report of AAPM nuclear magnetic resonance Task Group No. 1. Med Phys 17:287–295 34. Papadakis NG, Murrills CD, Hall LD et al (2000) Minimal gradient encoding for robust estimation of diffusion anisotropy. Magn Reson Imaging 18:671–679 35. Stewart JD (1993) Focal peripheral neuropathies. Lippincott Williams & Wilkins, Philadelphia 36. Thomas P-K, Fullerton P-M (1969) Nerve fiber size in the carpal tunnel syndrome. J Neurol Neurosurg Psychiatry 26:520–527 37. Marie P, Foix C (1913) Atrophie isolée de l’éminence thénar d’origine névritique: rôle du ligament annulaire antérieur dans la localisation de la lésion. Rev Neurol 26:647–649 38. Thomalla G, Glauche V, Koch MA et al (2004) Diffusion tensor imaging detects early Wallerian degeneration of the pyramidal tract after ischemic stroke. Neuroimage 22:1767–1774 39. Thomalla G, Glauche V, Weiller C et al (2005) Time course of wallerian degeneration after ischaemic stroke revealed by diffusion tensor imaging. J Neurol Neurosurg Psychiatry 76:266–268 40. Pierpaoli C, Barnett A, Pajevic S et al (2001) Water diffusion changes in Wallerian degeneration and their dependence on white matter architecture. Neuroimage 13:1174–1185