Trigeminal neuralgia (TN) is characterized by recurrent episodes of intense lancinating pain localized to the sensory supply areas of the trigeminal n

Transcription

1 Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at ORIGINAL RESEARCH n NEURORADIOLOGY Trigeminal Neuralgia due to Neurovascular Compression: High-Spatial-Resolution Diffusion- Tensor Imaging Reveals Microstructural Neural Changes 1 Juergen Lutz, MD Jennifer Linn, MD Jan H. Mehrkens, MD Niklas Thon, MD Robert Stahl, MD Klaus Seelos, MD Hartmut Brückmann, MD, PhD Markus Holtmannspötter, MD Purpose: Materials and Methods: Results: Conclusion: To preoperatively detect, by using diffusion-tensor imaging coregistered with anatomic magnetic resonance (MR) imaging, suspected microstructural tissue changes of the trigeminal nerves in patients with trigeminal neuralgia (TN) resulting from neurovascular compression. The study was approved by the institutional review board, and written informed consent was obtained from all patients. Twenty patients (mean age, 51.3 years) with TN and evidence of neurovascular contact were examined with use of a 3.0-T MR unit combined with an eight-channel head coil before undergoing surgical decompression. A single-shot diffusion-tensor echo-planar sequence was used along 15 different diffusion directions, with a b value of 1000 sec/mm 2 and a section thickness of 2 mm. For anatomic correlation, 0.6-mm isotropic three-dimensional fast imaging employing steady-state images were acquired for coregistration with the functional diffusion-tensor maps. After region of interest placement, mean fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values were calculated for each nerve by using the paired-sample two-tailed t test (with P,.005 indicating significance) and compared with surgical findings. FA was significantly lower ( P =.004) on the trigeminal neuralgia affected side (mean FA, 0.203) than on the contralateral side (mean FA, 0.239). ADCs were nearly identical between the normal and TN-affected nerve tissues. These findings suggest that diffusion-tensor imaging enables the identification and quantification of anisotropic changes between normal nerve tissue and TN-affected trigeminal nerves. Coregistration of anatomic three-dimensional fast imaging employing steady-state imaging and diffusion-tensor imaging facilitates excellent delineation of the cisternal segments of the trigeminal nerves. 1 From the Departments of Neuroradiology (J. Lutz, J. Linn, K.S., H.B., M.H.), Neurosurgery (J.H.M., N.T.), and Radiology (R.S.), Klinikum Grosshadern, Ludwig-Maximilians University, Marchioninistrasse 5, Munich, Germany. Received March 12, 2010; revision requested April 27; revision received July 2; accepted July 22; fi nal version accepted August 17. Address correspondence to J. Lutz ( Juergen.Lutz@med.uni-muenchen.de ). q RSNA, 2010 q RSNA, radiology.rsna.org n Radiology: Volume 258: Number 2 February 2011

2 Trigeminal neuralgia (TN) is characterized by recurrent episodes of intense lancinating pain localized to the sensory supply areas of the trigeminal nerve (CNV). The most frequent cause of TN is a mechanical irritation of the nerve caused by neurovascular contact the neurovascular compression syndrome (NVCS) ( 1 ). It is widely believed that compression at vulnerable sites only those in the so-called root entry or exit zone of the nerve causes NVCS. Most investigators define the root entry or exit zone as the region extending from the nerve s point of entry into or exit from the brainstem to the point of transition from the central myelin (derived from the oligodendroglia) to the peripheral myelin (derived from Schwann cells) ( 1 ). The superior cerebellar artery is responsible for most (60% 90%) cases of NVCS; the anteroinferior cerebellar artery and basilar artery follow ( 1 3 ). The referencestandard treatment for refractive TN caused by NVCS is microvascular decompression. Preoperatively, high-spatialresolution magnetic resonance (MR) imaging is performed to detect the neurovascular contact and exclude other causes of TN. While interpreting the imaging data, one has to take into account that a neurovascular contact can also be present in the form of an anatomic variant in healthy subjects or on the unaffected side in patients with TN. Advances in Knowledge n The coregistration of high-spatial- resolution three-dimensional anatomic magnetic resonance (MR) imaging with diffusion-tensor imaging, as compared with the use of diffusion-tensor imaging alone, has the potential to enhance the detection of small anatomic structures such as the trigeminal nerve (CNV). n Alterations in a diffusivity index, the fractional anisotropy (FA) of the CNV, are present in patients with trigeminal neuralgia (TN) and can be identified by using anatomic MR imaging coregistered with diffusion-tensor imaging. Histopathologic studies have revealed focal axonal degeneration and demyelination in the postoperative specimens collected from patients with TN due to NVCS ( 4 7 ). Some study investigators have proposed the use of an imaging protocol involving conventional MR sequences to image the posterior fossa in patients with TN ( 8 ). Diffusion-tensor imaging enables the analysis of white matter integrity by means of in vivo measurement of molecular diffusivity ( 9 ). There are numeric indexes to describe the anisotropic diffusion and underlying tissue integrity. The restriction in water diffusivity can be expressed as fractional anisotropy (FA) ( 10 ). Measurement of FA represents a robust method of assessing the degree of directionality of diffusion (ie, anisotropic diffusion) that occurs in a particular region. Diffusivity, expressed as the apparent diffusion coefficient (ADC), is a quantitative measure of water motility (independent of orientation) in a certain voxel ( 11 ). Degeneration or damage of white matter tracts is expected to result in reduced FA owing to the loss of directionality of diffusion and the increase in ADC that are due to diffusivity being averaged in all spatial directions as a result of the loss of myelin and axonal membranes ( 12,13 ). Thus, the aim of this study was to preoperatively evaluate possible microstructural tissue changes of the CNVs in patients with TN due to neurovascular compression by using a high-spatialresolution diffusion-tensor imaging sequence coregistered with a threedimensional fast imaging employing Implications for Patient Care n Coregistration of anatomic MR imaging with diffusion-tensor imaging could change the diagnostic work-up of patients with TN by yielding direct evidence of microstructural nerve alterations. n FA alterations may play a role in the differentiation of CNV abnormalities such as neurovascular compression induced versus encephalitis disseminata induced TN. steady-state acquisition (FIESTA; GE Healthcare, Milwaukee, Wis) sequence. Materials and Methods Subjects The study was approved by our institutional review board, and written informed consent was obtained from all patients. In February and September of 2009, we prospectively examined 20 consecutive patients (mean age, 51.3 years; age range, years) with long-standing (. 2 years) TN: 11 patients were women (mean age, 52 years 6 9 [standard deviation]; age range, years), and nine were men (mean age, 64 years 6 4; age range, years). All patients were examined by an experienced neurosurgeon (J.H.M., 12 years experience in functional neurosurgery) and met the criteria of the International Headache Society for TN ( 1 ). The routine preoperative work-up included high-spatialresolution computed tomography (CT) of the skull base, acoustic evoked potentials testing, and cranial MR imaging. Imaging Protocol Imaging was performed by using a 3.0-T MR imaging unit (Signa HDx 3T; GE Healthcare, Milwaukee, Wis) with an eight-channel head coil and the following Published online before print /radiol Radiology 2011; 258: Abbreviations: ADC = apparent diffusion coeffi cient CNV = trigeminal nerve FA = fractional anisotropy NVCS = neurovascular compression syndrome TN = trigeminal neuralgia Author contributions: Guarantors of integrity of entire study, J. Lutz, J. Linn; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript fi nal version approval, all authors; literature research, J. Lutz, J. Linn, J.H.M., K.S., H.B., M.H.; clinical studies, J. Lutz, J. Linn, J.H.M., N.T., R.S., K.S.; statistical analysis, J. Lutz, J. Linn, N.T., R.S., K.S.; and manuscript editing, J. Lutz, J. Linn, J.H.M., K.S., H.B., M.H. Authors stated no fi nancial relationship to disclose. Radiology: Volume 258: Number 2 February 2011 n radiology.rsna.org 525

3 Figure 1 Figure 1: Axial (a) FA, (b) ADC, and (c) color-coded diffusion-tensor index maps with orientation encoding obtained in 63-year-old man. In a and b, arrow delineates left CNV. In c, arrows delineate CNV in anteroposterior direction. routine sequences, which have been recommended for imaging the posterior fossa in patients with TN ( 8 ): (a) an axial fluid-attenuated inversionrecovery sequence (8500/120/2250 [repetition time msec/echo time msec/inversion time msec], mm field of view, matrix, 5-mm section thickness) to encompass the whole brain, (b) an axial T2-weighted periodically rotated overlapping parallel lines with enhanced reconstruction sequence (5100/75 [repetition time msec/echo time msec], mm field of view, matrix, 3-mm section thickness) to encompass the brainstem and posterior fossa, and (c) whole-brain threedimensional T1-weighted fast spoiled gradient-recalled pulse sequences (3.2/7.9, mm field of view, matrix, 1.4-mm section thickness) performed before and after the intravenous administration of 0.1 mmol of gadobenate dimeglumine (MultiHance; Bracco Imaging, Milan, Italy) per kilogram of body weight to rule out other possible causes of TN, such as schwannoma and multiple sclerosis ( 1 ). In addition, we performed the following high-spatial-resolution sequences: (a) a three-dimensional fast imaging employing steady-state sequence (FIESTA) (4.5/1.8, 50 flip angle, matrix, mm field of view, 0.6-mm section thickness, spatial resolution of mm, acquisition time of 7 minutes 45 seconds), (b) a contrast material enhanced gradient-echo MR an giographic sequence performed after the injection of 0.1 mmol/kg gadobenate dimeglumine (6.6/2.3, 20 flip angle, matrix, mm field of view, 0.8-mm section thickness, spatial resolution of mm, acquisition time of 7 minutes 15 seconds), and (c) a spin-echo single-shot echo-planar diffusion-tensor imaging sequence (10.5/6.5, matrix, mm field of view, 2-mm section thickness, voxel size of mm). Diffusion gradients were applied in 15 spatial directions, as described by Pierpaoli and Basser ( 10 ). The maximal b value was 1000 sec/mm 2, and 32 axial sections were acquired. Therefore, 10 measurements were obtained and averaged to result in an acquisition time of 11 minutes 33 seconds. The three-dimensional fast imaging employing steady-state images and diffusion-tensor images were angulated exactly parallel to the cisternal segment of the CNV, as visualized on the T1- weighted images. Subsequent to imaging, all patients underwent microvascular decompression. The intraoperative findings were correlated with the presurgical MR imaging results. Data Analysis The three-dimensional fast imaging employing steady-state images and contrast- enhanced MR angiograms were analyzed collaboratively by two experienced neuroradiologists (J. Lutz, J. Linn, 4 and 5 years experience, respectively, in diffusion-tensor and cranial nerve imaging) who were not aware of the clinical data or the affected side. Multiplanar reconstructions were used to assess the right and left CNVs separately for the presence of a neurovascular conflict, and, if applicable, the respective vessel was identified on the contrast-enhanced MR angiogram. The diffusion-tensor imaging data were analyzed (J. Lutz) by using Func- Tool software, release AW4.3_07 (GE Health care, Buc, France) at a commercially available workstation (AW 4.2; GE Healthcare, Buc, France). First, diffusiontensor trace images were generated by averaging all 15 diffusion-weighted images. Then, a motion correction algorithm was applied to correct for the head motion and image distortion due to eddy current artifacts. From these images, ADC and FA values in each voxel and color-coded maps (orientation encoding) were automatically calculated and displayed ( Fig 1 ). Data Fusion and Region of Interest Analysis With use of the FuncTool software, the calculated FA and ADC maps were matched with the fast imaging employing steady-state MR images and visualized 526 radiology.rsna.org n Radiology: Volume 258: Number 2 February 2011

4 Demographic Data on Patients with TN Patient No./ Sex/Age (y) Symptoms Duration (y) Affected Side/CNV Branch(es) * 1/M/68 1 L/V2 2/M/68 2 L/V2 3/M/72 4 L/V2 4/F/40 4 R/V2 + V3 5/F/64 4 L/V3 6/F/51 2 L/V2 + V3 7/F/59 2 R/V3 8/F/74 8 R/V3 9/F/39 4 R/V2 + V3 10/M/77 9 L/V1 + V2 + V3 11/F/68 8 R/V2 + V3 12/F/46 5 R/V1 + V2 + V3 13/F/49 10 R/V2 14/F/43 4 R/V2 15/M/58 4 R/V2 + V3 16/F/49 8 R/V2 17/M/47 6 L/V2 + V3 18/M/68 3 L/V2 19/M/63 7 R/V2 20/M/59 8 R/V2 Note. In most patients, the left CNV was involved. In all patients except subjects 7 and 10, the superior cerebellar artery was responsible for the NVCS. In patients 7 and 10, the anteroinferior cerebellar artery was responsible for the NVCS. The MR and surgical fi ndings matched in all 20 patients. * V1 = fi rst branch of CNV, V2 = second branch of CNV, V3 = third branch of CNV. both alone and as fused images. For quantitative analysis, a standard approximately 6-mm 2 (range, 4 6 mm 2 ) region of interest was positioned in the root entry or exit zone of each CNV on the fused images. All regions of interest were marked by the same two neuroradiologists (J. Lutz, J. Linn) in consensus. ADC and FA values were measured in the regions of interest. The mean values and standard deviations of these anisotropy indexes were calculated and compared with those for the contralateral side. Statistical Analyses ADC and FA data were tested for normal distribution by using the Kolmogorov-Smirnov test. As the next step, the clinical and intraoperative findings were studied to determine the affected side in the patients. Then, the mean values were compared between the affected and nonaffected sides in the same patient by using a paired-sample two-tailed t test because all of the data were normally distributed. P,.005 was considered to indicate a significant difference. Data are presented as means 6 standard deviations. All statistical calculations were performed with SPSS V.14 software (SPSS, Chicago, Ill). Results Clinical and Surgical Findings All patients with TN had long-standing (. 2 years) disease. In eight patients, the left side was affected, and in 12 patients, the right side was involved. Twelve (60%) of the 20 patients were found at presentation to have isolated involvement of one branch of the CNV: the second branch of the CNV in nine cases and the third branch in three. Six (30%) patients had combined involvement of the second and third branches of the CNV, and two (10%) had combined involvement of all three branches ( Table ). During surgery, direct neurovascular contact was detected in all cases. Anatomic Imaging Findings Fast imaging employing steady-state imaging and contrast-enhanced MR angiography revealed neurovascular contact at the root entry or exit zone of the CNV on the affected side in all 20 patients. In 18 (90%) patients, contact with the superior cerebellar artery was identified ( Figs 2 and 3 ), while in two (10%) patients, contact with the anteroinferior cerebellar artery was found. No neurovascular contact was detected on the unaffected contralateral side in any patient. The intraoperative findings were in accordance with the MR imaging findings in all cases. Diffusion-Tensor Imaging All diffusion-tensor imaging data sets were amenable to postprocessing with the FuncTool fusion software ( Fig 4 ). Motion correction was successful in all cases. In all 20 patients, the CNVs could be delineated on both sides on the fused images. FA was significantly lower on the affected side (mean FA, ) than on the contralateral unaffected side (mean FA, ) ( P =.004). ADCs were nearly identical between the affected (mean ADC, mm 2 / sec ) and uninvolved contralateral (mean ADC, mm 2 /sec ) sides ( P =.865). Discussion We found significantly lower FA in the affected nerve compared with the FA in the normal nerve in the patients with surgically proved TN due to NVCS. Conversely, ADCs were similar between the affected and nonaffected nerves. These findings indicate that diffusion-tensor imaging FA measurement enables in vivo visualization of the microstructural changes of the CNV in these patients. TN is a facial pain syndrome characterized by sudden and intense pain in one of the branches of the CNV. Demyelination of the sensory fibers of the CNV due to neurovascular compression at the root entry or exit zone is widely regarded as the underlying pathomechanism, accounting for 90% of TN cases ( 6 ). To exclude other causes of TN (eg, posterior fossa tumors and multiple sclerosis lesions) ( 6,14 ), MR imaging is essential. High-spatial-resolution MR imaging performed with heavily T2-weighted steady-state sequences and MR angiography constitutes an excellent and established method of imaging cranial nerve abnormalities ( 15,16 ). One limitation of this purely anatomic imaging method is that neurovascular contacts can also be found in a considerable portion of asymptomatic healthy subjects ( 2 ). Therefore, the sole finding of nerve vessel contact on MR images is not very specific for NVCS. Histopathologically, large demyelinated axons ( 5,6 ), loss of axon quantity ( 7 ), and abnormal remyelinization ( 1 ) are found in the nerve specimens collected from patients with TN. Agreeing with these histopathologic findings, neurosurgeons describe the affected CNV as flattened and edematous in situ. Although investigators in recent structural MR imaging studies have tried to estimate Radiology: Volume 258: Number 2 February 2011 n radiology.rsna.org 527

5 Figure 2 Figure 2: Axial (a) fast imaging employing steady-state image and (b) contrast-enhanced MR angiographic maximum intensity projection of prepontine fossa in 58-year-old man show long-standing TN involving second and third branches of CNV. Images show both cisternal segments of CNVs exiting the pons (short arrows). On right side, a vessel loop, the superior cerebellar artery, impresses and deviates right CNV posteriorly and laterally. Long arrow points to neurovascular confl ict. Findings were confi rmed with surgery. Basilar artery ( ) and bilateral cavernous segments of internal carotid arteries (arrowheads) also are seen. Figure 3 Figure 3: Axial (a) fast imaging employing steady-state image and (b) contrast-enhanced MR angiographic maximum intensity projection of prepontine fossa in 43-year-old woman show TN involving second branch of CNV. Another, less obvious example of neurovascular confl ict is shown in right CNV (short arrow). Arterial loop of superior cerebellar artery (long arrow) crosses right CNV in middle of cisternal course. Neurovascular confl ict was confi rmed with surgery. Basilar artery ( ) is seen in a. the degree of compression ( 17 ), this approach yields no information about the integrity of the nerve tissue. On the contrary, diffusion-tensor imaging enables assessment of the axonal architecture of white matter tracts in vivo ( 9 ). FA and ADC represent the two diffusiontensor imaging indexes that are most widely used to investigate white matter changes because these measurements reliably indicate abnormalities (or the absence thereof) in white matter tracts independent of their cause ( 12,13 ). Degeneration of white matter tracts results in a reduction in FA due to a loss of the directionality of diffusion and an increase in ADC that are due to diffusivity being averaged in all spatial directions as a result of the loss of myelin and axonal membranes ( 12 ). More recently, diffusion-tensor imaging has been used to assess the integrity of different cranial nerves in vivo. In patients with glaucoma, for example, microstructural changes in the white matter of the structures of the visual pathway could be demonstrated ( 18 ). Herweh et al ( 19 ) applied diffusiontensor imaging in five patients with TN and visible vessel nerve contact and 528 radiology.rsna.org n Radiology: Volume 258: Number 2 February 2011

6 Figure 4 Figure 4: (a) Axial high-spatial-resolution fast imaging employing steady-state image in 40-year-old woman shows clearly visible cisternal aspects of left CNV (arrow) and right CNV. (b) Axial fast imaging employing steady-state image fused with FA image shows region of interest drawn on left CNV (arrow). (c) Axial fast imaging employing steady-state image fused with FA index map shows same region of interest on left CNV (arrow). found reduced FA in three of them. On the basis of these promising initial results, we initiated a prospective study to systematically evaluate the potential of diffusion-tensor imaging in this clinical setting. The use of a higher field strength (3 T vs 1.5 T), a diffusion-tensor imaging sequence with 15 instead of six noncolinear diffusion directions, and diffusion-tensor images fused with highspatial-resolution fast imaging employing steady-state images enabled us to identify the CNVs and measure diffusiontensor imaging parameters in all cases. All patients underwent decompression surgery so that the intraoperative findings could be used to confirm our diagnosis. The neural and vessel structures of the prepontine fossa range between 0.5 and 3.0 mm in diameter. This small diameter has to be considered in diffusiontensor imaging studies in terms of the limited spatial resolution of the sequences used ( 15 ). To overcome this limitation, we used the information generated by using the high-spatial-resolution threedimensional fast imaging employing steady-state sequence. The fusion of the diffusion-tensor index maps and the fast imaging employing steady-state images served as a template for placement of the regions of interest on the diffusiontensor images. Despite the difference in section thickness between the fast imaging employing steady-state (0.6 mm) and diffusion-tensor (2.0 mm) images, we had no interference on the superimposed images. For manual region of interest placement ( 20,21 ), most diffusiontensor imaging studies have involved the use of images with a b value of 0 sec/mm 2 for exact delineation of the CNV; this is a widely used and acceptable procedure according to studies in the literature ( 19,21 ). We used fast imaging employing steady-state for detailed depiction of the cisternal course of the CNV and used contrast-enhanced MR angiography to delineate even smaller arteries and veins ( 16,22 ). The detailed anatomic information generated by using this sequence enabled detailed region of interest placement within each CNV. In terms of a potential limitation of our study, we did not include an agematched healthy control group of subjects with and those without neurovascular conflict; rather, we compared the affected nerve intraindividually with the nerve on the clinically healthy contralateral side. We used this as a straightforward method for routine clinical use. In conclusion, diffusion-tensor imaging combined with structural MR imaging enables the visualization of microstructural changes based on objective parameters independent of clinical symptoms in the affected nerve in patients with TN due to NVCS. We suggest that diffusion-tensor imaging is a valuable tool for delineating symptomatic compression in patients with facial pain syndromes ( 23 ) and equivocal clinical or imaging findings. References 1. Love S, Coakham HB. Trigeminal neuralgia: pathology and pathogenesis. Brain 2001 ; 124 ( pt 12 ): Adamczyk M, Bulski T, Sowińska J, Furmanek A, Bekiesińska-Figatowska M. Trigeminal nerve: artery contact in people without trigeminal neuralgia MR study. Med Sci Monit 2007 ; 13 ( suppl 1 ): Headache Classification Committee of the International Headache Society. Classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988 ; 8 ( suppl 7 ): Sindou M, Howeidy T, Acevedo G. Anatomical observations during microvascular decompression for idiopathic trigeminal neuralgia (with correlations between topography of pain and site of the neurovascular conflict): prospective study in a series of 579 patients. Acta Neurochir (Wien) 2002 ; 144 ( 1 ): 1 12; discussion Hilton DA, Love S, Gradidge T, Coakham HB. Pathological findings associated with trigeminal neuralgia caused by vascular compression. Neurosurgery 1994 ; 35 ( 2 ): ; discussion Love S, Hilton DA, Coakham HB. Central demyelination of the Vth nerve root in trigeminal neuralgia associated with vascular compression. Brain Pathol 1998 ; 8 ( 1 ): 1 11; discussion Devor M, Govrin-Lippmann R, Rappaport ZH. Mechanism of trigeminal neuralgia: an Radiology: Volume 258: Number 2 February 2011 n radiology.rsna.org 529

7 ultrastructural analysis of trigeminal root specimens obtained during microvascular decompression surgery. J Neurosurg 2002 ; 96 ( 3 ): Yoshino N, Akimoto H, Yamada I, et al. Trigeminal neuralgia: evaluation of neuralgic manifestation and site of neurovascular compression with 3D CISS MR imaging and MR angiography. Radiology 2003 ; 228 ( 2 ): Le Bihan D. Molecular diffusion, tissue microdynamics and microstructure. NMR Biomed 1995 ; 8 ( 7-8 ): Pierpaoli C, Basser PJ. Toward a quantitative assessment of diffusion anisotropy. Magn Reson Med 1996 ; 36 ( 6 ): Pierpaoli C, Jezzard P, Basser PJ, Barnett A, Di Chiro G. Diffusion tensor MR imaging of the human brain. Radiology 1996 ; 201 ( 3 ): Basser PJ, Pierpaoli C. Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 1996 ; 111 ( 3 ): Moseley ME, Cohen Y, Kucharczyk J, et al. Diffusion-weighted MR imaging of anisotropic water diffusion in cat central nervous system. Radiology 1990 ; 176 ( 2 ): Nurmikko TJ, Eldridge PR. Trigeminal neuralgia: pathophysiology, diagnosis and current treatment. Br J Anaesth 2001 ; 87 ( 1 ): Yousry I, Camelio S, Schmid UD, et al. Visualization of cranial nerves I-XII: value of 3D CISS and T2-weighted FSE sequences. Eur Radiol 2000 ; 10 ( 7 ): Linn J, Peters F, Moriggl B, Naidich TP, Brückmann H, Yousry I. The jugular foramen: imaging strategy and detailed anatomy at 3T. AJNR Am J Neuroradiol 2009 ; 30 ( 1 ): Leal PR, Hermier M, Froment JC, Souza MA, Cristino-Filho G, Sindou M. Preoperative demonstration of the neurovascular compression characteristics with special emphasis on the degree of compression, using high-resolution magnetic resonance imaging: a prospective study, with comparison to surgical findings, in 100 consecutive patients who underwent microvascular decompression for trigeminal neuralgia. Acta Neurochir (Wien) 2010 ; 152 ( 5 ): Garaci FG, Bolacchi F, Cerulli A, et al. Optic nerve and optic radiation neurodegeneration in patients with glaucoma: in vivo analysis with 3-T diffusion-tensor MR imaging. Radiology 2009 ; 252 ( 2 ): Herweh C, Kress B, Rasche D, et al. Loss of anisotropy in trigeminal neuralgia revealed by diffusion tensor imaging. Neurology 2007 ; 68 ( 10 ): Khong PL, Zhou LJ, Ooi GC, Chung BH, Cheung RT, Wong VC. The evaluation of Wallerian degeneration in chronic paediatric middle cerebral artery infarction using diffusion tensor MR imaging. Cerebrovasc Dis 2004 ; 18 ( 3 ): Sasaki M, Li B, Lankford KL, Radtke C, Kocsis JD. Remyelination of the injured spinal cord. Prog Brain Res 2007 ; 161 : Yousry I, Moriggl B, Dieterich M, Naidich TP, Schmid UD, Yousry TA. MR anatomy of the proximal cisternal segment of the trochlear nerve: neurovascular relationships and landmarks. Radiology 2002 ; 223 ( 1 ): Kress B, Schindler M, Rasche D, Hähnel S, Tronnier V, Sartor K. Trigeminal neuralgia: how often are trigeminal nerve-vessel contacts found by MRI in normal volunteers. Rofo 2006 ; 178 ( 3 ): radiology.rsna.org n Radiology: Volume 258: Number 2 February 2011