- Email: [email protected]
Surface rendered three-dimensional MR imaging for the evaluation of trigeminal neuralgia and hemifacial spasm
Journal of Clinical Neuroscience (2004) 11(8), 840–844 0967-5868/$ - see front matter ª 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jocn.2003.06.010
Clinical study
Surface rendered three-dimensional MR imaging for the evaluation of trigeminal neuralgia and hemifacial spasm Masakazu Ogiwara MD, Tsuneo Shimizu MD Department of Neurosurgery, Kanto Neurosurgical Hospital, Saitama, Japan
Summary Surface rendered three-dimensional (SR3D) magnetic resonance (MR) imaging was used to visualize the cerebellopontine angle nerves, blood vessels and brain stem and compared with the operative view in patients undergoing microvascular decompression. Thirty-one patients with 32 affected nerves manifesting as trigeminal neuralgia (TN, 16 cases), hemifacial spasm (HFS, 14 cases) and glossopharyngeal neuralgia (2 cases) underwent microvascular decompression at the Kanto Neurosurgical Hospital between February 2001 and July 2002. Preoperative MR imaging and MR angiography were performed in all patients. The SR3D MR image and the operative view correlated in 30 of the 31 patients. A very thin artery was not detected in one patient with TN. SR3D MR imaging can clearly demonstrate the three-dimensional relationship between the nerves and the offending vessel, so is very effective for the preoperative planning of microvascular decompression in cases of TN and HFS. ª 2004 Elsevier Ltd. All rights reserved. Keywords: microvascular decompression, magnetic resonance imaging, magnetic resonance angiography, three-dimensional
INTRODUCTION Trigeminal neuralgia (TN), hemifacial spasm (HFS) and glossopharyngeal neuralgia (GN) can be caused by vascular compression by the superior cerebellar artery (SCA), anterior inferior cerebellar artery (AICA), posterior inferior cerebellar artery (PICA) or a vein of the root entry/exit zone of the trigeminal nerve, the facial nerve and the glossopharyngeal nerve, respectively.1,2 Microvascular decompression (MVD) was first described in 19623,4 and since has become an established treatment strategy.5–8 Identification of the cause of TN and HFS previously depended on computed tomography and angiography to detect tumors and arteriovenous malformations, respectively.9 Microvascular compression caused by the SCA, AICA and PICA could be detected by displacement and elongation of the offending vessel. Preoperative magnetic resonance (MR) imaging and MR angiography now provide a noninvasive method for the visualization of the affected cranial nerve and offending vessel.10–16 However, the three-dimensional structure is difficult to assess, even if identification of the offending vessel is possible, so the operative view is difficult to simulate. Cerebral three-dimensional images were first obtained from MR images in 1990.17 Three-dimensional images reconstructed using this method from spoiled gradient-recalledecho (SPGR) MR imaging and MR angiography data were very useful for operation planning in cases of TN and HFS.18 Moreover, although virtual MR imaging based on constructive interference in steady state three-dimensional Fourier transformation (CISS 3DFT) MR imaging is useful, the boundary of the nerve and offending vessel may not be clearly discriminated.19 The present study digitally reconstructed the three-dimensional appearance of the nerve, blood vessel, brain stem and cerebellum by surface rendered three-dimensional (SR3D) imaging techniques using MR imaging and MR angiography data in a series Received 20 March 2003 Accepted 10 June 2003 Correspondence to: Masakazu Ogiwara MD, Department of Neurosurgery, Kanto Neurosurgical Hospital, 1120 Dai, Kumagaya, Saitama 360-0804, Japan. Tel.: +81 48 521 3133; Fax: +81 48 524 6190; E-mail: [email protected]
840
of patients with TN and HFS and compared them with the intraoperative view. CLINICAL MATERIAL AND METHODS Thirty-one patients underwent MVD for TN, HFS and GN at the Kanto Neurosurgical Hospital between February 2001 and July 2002 (Table 1). Fifteen patients had TN (5 males and 10 females, aged 57–77 years). Thirteen patients had HFS (2 males and 11 females, aged 40–70 years). One patient had both TN and HFS (female aged 64 years). Two patients had GN (females, 53 and 75 years). All patients underwent preoperative MR imaging and MR angiography. MR imaging was performed using a 1.5-Tesla MR imaging system (Gyroscan Intera; Philips) with the three-dimensional-turbo spin echo (3D-TSE) sequence (repetition time/echo time [TR/ TE] 1500 ms/250 ms; flip angle 90; field of view 150 mm; matrix 256 · 512; slice thickness 0.7 (1.4); 40 slices). MR angiography used the T1-first field echo (T1-FFE) sequence (TR/TE 26 ms/ 3.0 ms; flip angle 22; field of view 150 mm; matrix 256 · 512; slice thickness 0.55; 86 slices). The acquired data from MR imaging and MR angiography were transferred to a workstation (Easy Vision R.4.3; Sunmicrosystems) for SR3D MR image reconstruction. The brain stem, cerebellum, trigeminal nerve and facial nerve were separated based on the 3D-TSE data using the segmentation tool. The vertebrobasilar arterial system, SCA, AICA, PICA and petrosal vein were separated based on the 3D-TSE and T1-FFE data using the segmentation tool. Surface rendering processing was then performed. The brain stem, cerebellum, nerves, arteries and veins were rendered in different colors. Translucent processing of the cerebellum and vessels was performed when required. The operative views were compared with the SR3D MR images in all cases. RESULTS The brain stem, trigeminal nerve and vessels were easily separated on the SR3D MR images, and the offending vessel and point of contact rapidly identified. Addition of the petrosal vein showed
SR3D MR imaging for the evaluation of TN and HFS 841
Table 1 Characteristics of the patients Case no.
Sex
Age
Side
Type of neurovascular compression
Offending vessels by MR imaging
Offending vessels at operation SCA AICA AICA VA AICA PICA AICA VA SCA SCA AICA-PICA PICA AICA-PICA AICA-PICA SCA SCA Vein SCA SCA AICA SCA PICA SCA AICA SCA AICA AICA SCA SCA Vein PICA SCA VA AICA PICA AICA PICA AICA-PICA AICA AICA AICA VA BA AICA
1
Female
62
Right
TN
SCA
2
Female
73
Right
TN
3 4 5
Female Female Female
49 43 61
Left Left Right
HFS HFS TN
6 7 8 9 10 11 12
Male Female Female Male Female Female Male
74 64 75 51 63 77 66
Right Left Left Left Left Left Right
TN HFS GN HFS HFS TN TN
13 14
Female Male
66 60
Right Right
TN TN
15 16 17
Male Female Female
66 62 71
Right Left Right
TN HFS TN
18
Female
61
Right
TN
19 20 21
Female Female Male
69 76 57
Left Left Left
HFS TN TN
22 23 24
Female Female Female
53 72 60
Left Left Right
GN TN HFS
25
Female
63
Left
HFS
26 27 28 29 30
Female Male Female Female Female
51 40 76 70 54
Right Left Right Right Left
HFS HFS TN HFS HFS
31
Female
64
Left
TN HFS
AICA VA AICA PICA AICA VA SCA SCA AICA-PICA PICA AICA-PICA AICA-PICA SCA SCA Vein SCA SCA AICA SCA PICA SCA AICA SCA AICA AICA SCA SCA Vein PICA SCA VA AICA PICA AICA PICA AICA-PICA AICA AICA AICA VA BA AICA
TN, trigeminal neuralgia; HFS, hemifacial spasm; GN, glossopharyngeal neuralgia; BA, basilar artery; VA, vertebral artery; SCA, superior cerebellar artery; AICA, anterior inferior cerebellar artery; PICA, posterior inferior cerebellar artery.
the relationship to the trigeminal nerve for preoperative planning. The nerve and non-offending vessels were clearly dissociated. The SR3D MR image and the operative view coincided in 15 of the 16 cases of TN. The SR3D MR image did not demonstrate the involvement of the AICA in one patient with TN. Subsequent repeated reconstruction of the SR3D MR image confirmed the involvement of the AICA. Postoperative SR3D MR imaging showed that the offending vessels were moved from the root entry zone of the trigeminal nerve. The brain stem, nerve, and the offending and nonoffending vessels were similarly identified on the SR3D MR images in all 14 cases of HFS. However, the acoustic nerve and the facial nerve could not be distinguished in almost all cases. The SR3D MR image and the operative view coincided in all 14 patients. The SR3D MR image can demonstrate the SCA, AICA and PICA with a thickness of at least 1 mm, but depiction of thinner blood vessels is less reliable, so the presence of the offending vessels around the nerve must be confirmed during the operation. ª 2004 Elsevier Ltd. All rights reserved.
REPRESENTATIVE CASE REPORTS Case 12 A 66-year-old male presented with TN of the right third division, which had begun one year previous. The offending vessels were identified as the SCA and vein on the SR3D MR image (Fig. 1). Left MVD was performed. Intraoperative inspection confirmed that the offending vessels were the SCA and vein. The trigeminal nerve was located deep to the SCA and superficial to the petrosal vein.
Case 16 A 62-year-old female presented with left clonic type HFS, which had begun 5 years before. MR angiography suggested the involvement of the PICA and AICA. However, the SR3D MR image showed that the offending vessel was the PICA (Fig. 2). The AICA was in contact with the distal part of the facial nerve. Journal of Clinical Neuroscience (2004) 11(8), 840–844
842 Ogiwara and Shimizu
Fig. 1 Case 12. A patient with right TN. Upper left: 3D-TSE MR image showing the offending artery (arrow). Upper right: operative photograph showing that the offending vessels were the SCA and petrosal vein (arrows). Lower left: upper anteroposterior SR3D MR image. Lower right: SR3D MR image of the operative view showing that the offending vessels were the SCA and petrosal vein (arrows).
MVD was performed. Intraoperative inspection confirmed that the offending vessel was the PICA. DISCUSSION The present method of SR3D MR imaging used different colors to model the nerves, blood vessels, brain stem and cerebellum for the preoperative planning of MVD. The three-dimensional images simulated the operative view, and were confirmed to be accurate at operation in 30 of 31 cases. SR3D MR imaging is non-invasive method for visualizing the trigeminal and facial nerves and the offending vessels. The three-dimensional color-coded representation of the anatomical structures clearly demonstrates the contact point between the affected nerve and the offending vessel. Processing of the data on a workstation provides views at various angles, including the operative approach, which allows detailed preoperative planning of the displacement of the offending vessels based on the anatomical relationships around the point of vascular compression. For example, the operator’s view of the nerves and the offending vessels could easily be simulated. The possibility of the petrosal vein obscuring the operative view, and the best approach to optimize the view could be predicted. Problems with transposition in the presence of elongated blood vessels, basilar artery or vestibular artery could be modeled to assess the best method of decompression. This reduces the likelihood of complications. In the presented series of 31 cases, no complications, such as facial paralysis or auditory disturbance occurred in any patient. Journal of Clinical Neuroscience (2004) 11(8), 840–844
SR3D MR imaging may omit or exaggerate the size of a thin nerve or blood vessel of about 1 mm or less in diameter as a result of the surface rendering processing. Therefore, some subjective use of the segmentation tool to separate the nerve and microvascular vessels is required. The method requires high resolution MR imaging equipment, a high speed workstation, and about 30–60 min is required to create the SR3D MR images in each case. However, SR3D MR imaging can also be used for screening by creating isolated images of the blood vessels and nerves, requiring only 10–20 min. Detailed knowledge of the anatomy around the brain stem is essential to correctly interpret the MR images for surface processing. In particular, thin nerves, microvascular vessels, and vessels with slower blood flow are difficult to visualize by MR imaging and so are difficult to dissociate. However, the high resolution of 3D-TSE MR imaging can demonstrate even thin blood vessels for use in SR3D MR imaging. The SR3D MR image and the operative view did not coincide in one of our cases examined early in the study, when the technique had not yet been fully established. The offending vessel was a thin AICA that coursed horizontally and was hard to visualize by MR imaging, so was not separated by the threshold value of the segmentation tool. Repeated SR3D MR imaging after the operation with manual separation of the AICA coincided with the operative view. SR3D MR imaging of the trigeminal nerve was satisfactory with clear depiction of the modification and displacement by the offending vessels. However, the facial nerve was not clearly ª 2004 Elsevier Ltd. All rights reserved.
SR3D MR imaging for the evaluation of TN and HFS 843
Fig. 2 Case 16. A patient with HFS. Upper: operative photograph showing that the offending artery was the PICA (arrow). Lower left: lower anteroposterior SR3D MR image. Lower right: SR3D MR image of the operative view showing that the offending artery was the PICA (arrow).
discriminated from the acoustic nerve, even by processing based on 3D-TSE MR imaging, so improvements are still needed in this respect. In other reports oblique sagittal MR imaging could identify the offending vessel in 75% of cases of HFS.20 CISS 3DFT MR imaging showed the offending vessel in 62.5% of cases of TN (100% when including equivocal cases) and 100% of cases of HFS.21 SPGR MR imaging and three-dimensional time-of-flight MR angiography detected the offending artery in 67% of cases of TN and 87% of cases of HFS.22 The offending vessel was demonstrated in 30 of our 31 cases (96.8%) based on 3D-TSE and T1-FFE MR imaging. Therefore, SR3D MR imaging provides comparable accuracy in diagnosis, but has the advantages of high resolution and definition, and color rendering of the nerves, blood vessels brain stem and cerebellum, thus it is very easy to understand and interpret. Furthermore, 3D-TSE MR imaging can visualize the petrosal vein, which is not detected by MR angiography, for processing on the SR3D MR image. Translucent processing is also useful to clarify locations where nerves and blood vessels overlap. ‘‘Real” 3D CISS imaging is also effective for cases of tumor and aneurysms of the cerebellopontine angle.23 SR3D MR imaging is also able to evaluate the relationships between aneurysms and lower cranial nerves, cerebellopontine angle tumor and nerves, the relationships between pituitary adenomas ª 2004 Elsevier Ltd. All rights reserved.
and the optic nerve, and the anterior communicating artery and internal carotid artery aneurysms and the optic nerve. CONCLUSIONS SR3D MR imaging can clearly visualize the microsurgical anatomical relationships of the nerves, blood vessels, veins, brain stem and cerebellum, enabling each structure to be displayed in a different color. Since the resolution and quality of the images is high, the three-dimensional structure can easily be understood, which is very useful for pre-operative discussions with the patient. Furthermore, the SR3D MR image can display the same view as the anticipated operative approach, helping reduce the likelihood of complications after MVD. The clinical history is very important in the diagnosis of TN and HFS, but SR3D MR imaging is also very effective in the preoperative evaluation of TN and HFS. REFERENCES 1. Dandy WE. The treatment of trigeminal neuralgia by the cerebellar route. Ann Surg 1932; 96: 787–795. 2. Dandy WE. Concerning the cause of trigeminal neuralgia. Am J Surg 1934; 24: 447–455. 3. Gardner WJ, Sava GA. Hemifacial spasm-a reversible pathophysiologic state. J Neurosurg 1962; 19: 240–247.
Journal of Clinical Neuroscience (2004) 11(8), 840–844
844 Ogiwara and Shimizu
4. Gardner WJ. Concerning the mechanism of trigeminal neuralgia and hemifacial spasm. J Neurosurg 1962; 19: 947–958. 5. Jannetta PJ. Arterial compression of the trigeminal nerve at the pons in patients with trigeminal neuralgia. J Neurosurg 1967; 26: 159–162. 6. Jannetta PJ. Microsurgical approach to the trigeminal nerve for tic douloureux. Prog Neurol Surg 1976; 7: 180–200. 7. Jannetta PJ, Abbasy M, Maroon JC, Ramos FM, Albin MS. Etiology and definitive microsurgical treatment of hemifacial spasm. Operative techniques and results in 47 patients. J Neurosurg 1977; 47: 321–328. 8. Jannetta PJ. Neurovascular compression in cranial nerve and systemic disease. Ann Surg 1980; 192: 518–525. 9. Digre KB, Corbett JJ, Smoker WR, McKusker S. CT and hemifacial spasm. Neurology 1988; 38: 1111–1113. 10. Arbab AS, Nishiyama Y, Aoki S, et al. Simultaneous display of MRA and MPR in detecting vascular compression for trigeminal neuralgia or hemifacial spasm: comparison with oblique sagittal views of MRI. Eur Radiol 2000; 10: 1056–1060. 11. Barker II FG, Jannetta PJ, Bissonette DJ, Shields PT, Larkins MV, Jho HD. Microvascular decompression for hemifacial spasm. J Neurosurg 1995; 82: 201–210. 12. Casselman JW, Kuhweide R, Deimling M, Ampe W, Dehaene I, Meeus L. Constructive interference in steady state-3DFT MR imaging of the inner ear and cerebellopontine angle. AJNR Am J Neuroradiol 1993; 14: 47–57. 13. Girard N, Poncet M, Caces F, et al. Three-dimensional MRI of hemifacial spasm with surgical correlation. Neuroradiology 1997; 39: 46–51. 14. Tash RR, Kier EL, Chyatte D. Hemifacial spasm caused by a tortuous vertebral artery: MR demonstration. J Comput Assist Tomogr 1988; 12: 492–494.
Journal of Clinical Neuroscience (2004) 11(8), 840–844
15. Tash RR, Sze G, Leslie DR. Trigeminal neuralgia: MR imaging features. Radiology 1989; 172: 767–770. 16. Wong BY, Steinberg GK, Rosen L. Magnetic resonance imaging of vascular compression in trigeminal neuralgia. Case report. J Neurosurg 1989; 70: 132–134. 17. Hu XP, Tan KK, Levin DN, et al. Three-dimensional magnetic resonance images of the brain: application to neurosurgical planning. J Neurosurg 1990; 72: 433–440. 18. Kumon Y, Sakaki S, Kohno K, Ohta S, Ohue S, Miki H. Three-dimensional imaging for presentation of the causative vessels in patients with hemifacial spasm and trigeminal neuralgia. Surg Neurol 1997; 47: 178–184. 19. Shigematsu Y, Korogi Y, Hirai T, et al. Virtual MRI endoscopy of the intracranial cerebrospinal fluid spaces. Neuroradiology 1998; 40: 644–650. 20. Nagaseki Y, Horikoshi T, Omata T, et al. Oblique sagittal magnetic resonance imaging visualizing vascular compression of the trigeminal or facial nerve. J Neurosurg 1992; 77: 379–386. 21. Yamakami I, Kobayashi E, Hirai S, Yamaura A. Preoperative assessment of trigeminal neuralgia and hemifacial spasm using constructive interference in steady state-three-dimensional Fourier transformation magnetic resonance imaging. Neurol Med Chir (Tokyo) 2000; 40: 545–556. 22. Niwa Y, Shiotani M, Karasawa H, Ohseto K, Naganuma Y. [Identification of offending vessels in trigeminal neuralgia and hemifacial spasm using SPGR-MRI and 3D-TOF-MRA]. Rinsho Shinkeigaku (Jpn) 1996; 36: 544–550. 23. Kakizawa Y, Hongo K, Takasawa H, et al. ‘‘Real” three-dimensional constructive interference in steady-state imaging to discern microneurosurgical anatomy. Technical note. J Neurosurg 2003; 98: 625–630.
ª 2004 Elsevier Ltd. All rights reserved.
Clinical study
Surface rendered three-dimensional MR imaging for the evaluation of trigeminal neuralgia and hemifacial spasm Masakazu Ogiwara MD, Tsuneo Shimizu MD Department of Neurosurgery, Kanto Neurosurgical Hospital, Saitama, Japan
Summary Surface rendered three-dimensional (SR3D) magnetic resonance (MR) imaging was used to visualize the cerebellopontine angle nerves, blood vessels and brain stem and compared with the operative view in patients undergoing microvascular decompression. Thirty-one patients with 32 affected nerves manifesting as trigeminal neuralgia (TN, 16 cases), hemifacial spasm (HFS, 14 cases) and glossopharyngeal neuralgia (2 cases) underwent microvascular decompression at the Kanto Neurosurgical Hospital between February 2001 and July 2002. Preoperative MR imaging and MR angiography were performed in all patients. The SR3D MR image and the operative view correlated in 30 of the 31 patients. A very thin artery was not detected in one patient with TN. SR3D MR imaging can clearly demonstrate the three-dimensional relationship between the nerves and the offending vessel, so is very effective for the preoperative planning of microvascular decompression in cases of TN and HFS. ª 2004 Elsevier Ltd. All rights reserved. Keywords: microvascular decompression, magnetic resonance imaging, magnetic resonance angiography, three-dimensional
INTRODUCTION Trigeminal neuralgia (TN), hemifacial spasm (HFS) and glossopharyngeal neuralgia (GN) can be caused by vascular compression by the superior cerebellar artery (SCA), anterior inferior cerebellar artery (AICA), posterior inferior cerebellar artery (PICA) or a vein of the root entry/exit zone of the trigeminal nerve, the facial nerve and the glossopharyngeal nerve, respectively.1,2 Microvascular decompression (MVD) was first described in 19623,4 and since has become an established treatment strategy.5–8 Identification of the cause of TN and HFS previously depended on computed tomography and angiography to detect tumors and arteriovenous malformations, respectively.9 Microvascular compression caused by the SCA, AICA and PICA could be detected by displacement and elongation of the offending vessel. Preoperative magnetic resonance (MR) imaging and MR angiography now provide a noninvasive method for the visualization of the affected cranial nerve and offending vessel.10–16 However, the three-dimensional structure is difficult to assess, even if identification of the offending vessel is possible, so the operative view is difficult to simulate. Cerebral three-dimensional images were first obtained from MR images in 1990.17 Three-dimensional images reconstructed using this method from spoiled gradient-recalledecho (SPGR) MR imaging and MR angiography data were very useful for operation planning in cases of TN and HFS.18 Moreover, although virtual MR imaging based on constructive interference in steady state three-dimensional Fourier transformation (CISS 3DFT) MR imaging is useful, the boundary of the nerve and offending vessel may not be clearly discriminated.19 The present study digitally reconstructed the three-dimensional appearance of the nerve, blood vessel, brain stem and cerebellum by surface rendered three-dimensional (SR3D) imaging techniques using MR imaging and MR angiography data in a series Received 20 March 2003 Accepted 10 June 2003 Correspondence to: Masakazu Ogiwara MD, Department of Neurosurgery, Kanto Neurosurgical Hospital, 1120 Dai, Kumagaya, Saitama 360-0804, Japan. Tel.: +81 48 521 3133; Fax: +81 48 524 6190; E-mail: [email protected]
840
of patients with TN and HFS and compared them with the intraoperative view. CLINICAL MATERIAL AND METHODS Thirty-one patients underwent MVD for TN, HFS and GN at the Kanto Neurosurgical Hospital between February 2001 and July 2002 (Table 1). Fifteen patients had TN (5 males and 10 females, aged 57–77 years). Thirteen patients had HFS (2 males and 11 females, aged 40–70 years). One patient had both TN and HFS (female aged 64 years). Two patients had GN (females, 53 and 75 years). All patients underwent preoperative MR imaging and MR angiography. MR imaging was performed using a 1.5-Tesla MR imaging system (Gyroscan Intera; Philips) with the three-dimensional-turbo spin echo (3D-TSE) sequence (repetition time/echo time [TR/ TE] 1500 ms/250 ms; flip angle 90; field of view 150 mm; matrix 256 · 512; slice thickness 0.7 (1.4); 40 slices). MR angiography used the T1-first field echo (T1-FFE) sequence (TR/TE 26 ms/ 3.0 ms; flip angle 22; field of view 150 mm; matrix 256 · 512; slice thickness 0.55; 86 slices). The acquired data from MR imaging and MR angiography were transferred to a workstation (Easy Vision R.4.3; Sunmicrosystems) for SR3D MR image reconstruction. The brain stem, cerebellum, trigeminal nerve and facial nerve were separated based on the 3D-TSE data using the segmentation tool. The vertebrobasilar arterial system, SCA, AICA, PICA and petrosal vein were separated based on the 3D-TSE and T1-FFE data using the segmentation tool. Surface rendering processing was then performed. The brain stem, cerebellum, nerves, arteries and veins were rendered in different colors. Translucent processing of the cerebellum and vessels was performed when required. The operative views were compared with the SR3D MR images in all cases. RESULTS The brain stem, trigeminal nerve and vessels were easily separated on the SR3D MR images, and the offending vessel and point of contact rapidly identified. Addition of the petrosal vein showed
SR3D MR imaging for the evaluation of TN and HFS 841
Table 1 Characteristics of the patients Case no.
Sex
Age
Side
Type of neurovascular compression
Offending vessels by MR imaging
Offending vessels at operation SCA AICA AICA VA AICA PICA AICA VA SCA SCA AICA-PICA PICA AICA-PICA AICA-PICA SCA SCA Vein SCA SCA AICA SCA PICA SCA AICA SCA AICA AICA SCA SCA Vein PICA SCA VA AICA PICA AICA PICA AICA-PICA AICA AICA AICA VA BA AICA
1
Female
62
Right
TN
SCA
2
Female
73
Right
TN
3 4 5
Female Female Female
49 43 61
Left Left Right
HFS HFS TN
6 7 8 9 10 11 12
Male Female Female Male Female Female Male
74 64 75 51 63 77 66
Right Left Left Left Left Left Right
TN HFS GN HFS HFS TN TN
13 14
Female Male
66 60
Right Right
TN TN
15 16 17
Male Female Female
66 62 71
Right Left Right
TN HFS TN
18
Female
61
Right
TN
19 20 21
Female Female Male
69 76 57
Left Left Left
HFS TN TN
22 23 24
Female Female Female
53 72 60
Left Left Right
GN TN HFS
25
Female
63
Left
HFS
26 27 28 29 30
Female Male Female Female Female
51 40 76 70 54
Right Left Right Right Left
HFS HFS TN HFS HFS
31
Female
64
Left
TN HFS
AICA VA AICA PICA AICA VA SCA SCA AICA-PICA PICA AICA-PICA AICA-PICA SCA SCA Vein SCA SCA AICA SCA PICA SCA AICA SCA AICA AICA SCA SCA Vein PICA SCA VA AICA PICA AICA PICA AICA-PICA AICA AICA AICA VA BA AICA
TN, trigeminal neuralgia; HFS, hemifacial spasm; GN, glossopharyngeal neuralgia; BA, basilar artery; VA, vertebral artery; SCA, superior cerebellar artery; AICA, anterior inferior cerebellar artery; PICA, posterior inferior cerebellar artery.
the relationship to the trigeminal nerve for preoperative planning. The nerve and non-offending vessels were clearly dissociated. The SR3D MR image and the operative view coincided in 15 of the 16 cases of TN. The SR3D MR image did not demonstrate the involvement of the AICA in one patient with TN. Subsequent repeated reconstruction of the SR3D MR image confirmed the involvement of the AICA. Postoperative SR3D MR imaging showed that the offending vessels were moved from the root entry zone of the trigeminal nerve. The brain stem, nerve, and the offending and nonoffending vessels were similarly identified on the SR3D MR images in all 14 cases of HFS. However, the acoustic nerve and the facial nerve could not be distinguished in almost all cases. The SR3D MR image and the operative view coincided in all 14 patients. The SR3D MR image can demonstrate the SCA, AICA and PICA with a thickness of at least 1 mm, but depiction of thinner blood vessels is less reliable, so the presence of the offending vessels around the nerve must be confirmed during the operation. ª 2004 Elsevier Ltd. All rights reserved.
REPRESENTATIVE CASE REPORTS Case 12 A 66-year-old male presented with TN of the right third division, which had begun one year previous. The offending vessels were identified as the SCA and vein on the SR3D MR image (Fig. 1). Left MVD was performed. Intraoperative inspection confirmed that the offending vessels were the SCA and vein. The trigeminal nerve was located deep to the SCA and superficial to the petrosal vein.
Case 16 A 62-year-old female presented with left clonic type HFS, which had begun 5 years before. MR angiography suggested the involvement of the PICA and AICA. However, the SR3D MR image showed that the offending vessel was the PICA (Fig. 2). The AICA was in contact with the distal part of the facial nerve. Journal of Clinical Neuroscience (2004) 11(8), 840–844
842 Ogiwara and Shimizu
Fig. 1 Case 12. A patient with right TN. Upper left: 3D-TSE MR image showing the offending artery (arrow). Upper right: operative photograph showing that the offending vessels were the SCA and petrosal vein (arrows). Lower left: upper anteroposterior SR3D MR image. Lower right: SR3D MR image of the operative view showing that the offending vessels were the SCA and petrosal vein (arrows).
MVD was performed. Intraoperative inspection confirmed that the offending vessel was the PICA. DISCUSSION The present method of SR3D MR imaging used different colors to model the nerves, blood vessels, brain stem and cerebellum for the preoperative planning of MVD. The three-dimensional images simulated the operative view, and were confirmed to be accurate at operation in 30 of 31 cases. SR3D MR imaging is non-invasive method for visualizing the trigeminal and facial nerves and the offending vessels. The three-dimensional color-coded representation of the anatomical structures clearly demonstrates the contact point between the affected nerve and the offending vessel. Processing of the data on a workstation provides views at various angles, including the operative approach, which allows detailed preoperative planning of the displacement of the offending vessels based on the anatomical relationships around the point of vascular compression. For example, the operator’s view of the nerves and the offending vessels could easily be simulated. The possibility of the petrosal vein obscuring the operative view, and the best approach to optimize the view could be predicted. Problems with transposition in the presence of elongated blood vessels, basilar artery or vestibular artery could be modeled to assess the best method of decompression. This reduces the likelihood of complications. In the presented series of 31 cases, no complications, such as facial paralysis or auditory disturbance occurred in any patient. Journal of Clinical Neuroscience (2004) 11(8), 840–844
SR3D MR imaging may omit or exaggerate the size of a thin nerve or blood vessel of about 1 mm or less in diameter as a result of the surface rendering processing. Therefore, some subjective use of the segmentation tool to separate the nerve and microvascular vessels is required. The method requires high resolution MR imaging equipment, a high speed workstation, and about 30–60 min is required to create the SR3D MR images in each case. However, SR3D MR imaging can also be used for screening by creating isolated images of the blood vessels and nerves, requiring only 10–20 min. Detailed knowledge of the anatomy around the brain stem is essential to correctly interpret the MR images for surface processing. In particular, thin nerves, microvascular vessels, and vessels with slower blood flow are difficult to visualize by MR imaging and so are difficult to dissociate. However, the high resolution of 3D-TSE MR imaging can demonstrate even thin blood vessels for use in SR3D MR imaging. The SR3D MR image and the operative view did not coincide in one of our cases examined early in the study, when the technique had not yet been fully established. The offending vessel was a thin AICA that coursed horizontally and was hard to visualize by MR imaging, so was not separated by the threshold value of the segmentation tool. Repeated SR3D MR imaging after the operation with manual separation of the AICA coincided with the operative view. SR3D MR imaging of the trigeminal nerve was satisfactory with clear depiction of the modification and displacement by the offending vessels. However, the facial nerve was not clearly ª 2004 Elsevier Ltd. All rights reserved.
SR3D MR imaging for the evaluation of TN and HFS 843
Fig. 2 Case 16. A patient with HFS. Upper: operative photograph showing that the offending artery was the PICA (arrow). Lower left: lower anteroposterior SR3D MR image. Lower right: SR3D MR image of the operative view showing that the offending artery was the PICA (arrow).
discriminated from the acoustic nerve, even by processing based on 3D-TSE MR imaging, so improvements are still needed in this respect. In other reports oblique sagittal MR imaging could identify the offending vessel in 75% of cases of HFS.20 CISS 3DFT MR imaging showed the offending vessel in 62.5% of cases of TN (100% when including equivocal cases) and 100% of cases of HFS.21 SPGR MR imaging and three-dimensional time-of-flight MR angiography detected the offending artery in 67% of cases of TN and 87% of cases of HFS.22 The offending vessel was demonstrated in 30 of our 31 cases (96.8%) based on 3D-TSE and T1-FFE MR imaging. Therefore, SR3D MR imaging provides comparable accuracy in diagnosis, but has the advantages of high resolution and definition, and color rendering of the nerves, blood vessels brain stem and cerebellum, thus it is very easy to understand and interpret. Furthermore, 3D-TSE MR imaging can visualize the petrosal vein, which is not detected by MR angiography, for processing on the SR3D MR image. Translucent processing is also useful to clarify locations where nerves and blood vessels overlap. ‘‘Real” 3D CISS imaging is also effective for cases of tumor and aneurysms of the cerebellopontine angle.23 SR3D MR imaging is also able to evaluate the relationships between aneurysms and lower cranial nerves, cerebellopontine angle tumor and nerves, the relationships between pituitary adenomas ª 2004 Elsevier Ltd. All rights reserved.
and the optic nerve, and the anterior communicating artery and internal carotid artery aneurysms and the optic nerve. CONCLUSIONS SR3D MR imaging can clearly visualize the microsurgical anatomical relationships of the nerves, blood vessels, veins, brain stem and cerebellum, enabling each structure to be displayed in a different color. Since the resolution and quality of the images is high, the three-dimensional structure can easily be understood, which is very useful for pre-operative discussions with the patient. Furthermore, the SR3D MR image can display the same view as the anticipated operative approach, helping reduce the likelihood of complications after MVD. The clinical history is very important in the diagnosis of TN and HFS, but SR3D MR imaging is also very effective in the preoperative evaluation of TN and HFS. REFERENCES 1. Dandy WE. The treatment of trigeminal neuralgia by the cerebellar route. Ann Surg 1932; 96: 787–795. 2. Dandy WE. Concerning the cause of trigeminal neuralgia. Am J Surg 1934; 24: 447–455. 3. Gardner WJ, Sava GA. Hemifacial spasm-a reversible pathophysiologic state. J Neurosurg 1962; 19: 240–247.
Journal of Clinical Neuroscience (2004) 11(8), 840–844
844 Ogiwara and Shimizu
4. Gardner WJ. Concerning the mechanism of trigeminal neuralgia and hemifacial spasm. J Neurosurg 1962; 19: 947–958. 5. Jannetta PJ. Arterial compression of the trigeminal nerve at the pons in patients with trigeminal neuralgia. J Neurosurg 1967; 26: 159–162. 6. Jannetta PJ. Microsurgical approach to the trigeminal nerve for tic douloureux. Prog Neurol Surg 1976; 7: 180–200. 7. Jannetta PJ, Abbasy M, Maroon JC, Ramos FM, Albin MS. Etiology and definitive microsurgical treatment of hemifacial spasm. Operative techniques and results in 47 patients. J Neurosurg 1977; 47: 321–328. 8. Jannetta PJ. Neurovascular compression in cranial nerve and systemic disease. Ann Surg 1980; 192: 518–525. 9. Digre KB, Corbett JJ, Smoker WR, McKusker S. CT and hemifacial spasm. Neurology 1988; 38: 1111–1113. 10. Arbab AS, Nishiyama Y, Aoki S, et al. Simultaneous display of MRA and MPR in detecting vascular compression for trigeminal neuralgia or hemifacial spasm: comparison with oblique sagittal views of MRI. Eur Radiol 2000; 10: 1056–1060. 11. Barker II FG, Jannetta PJ, Bissonette DJ, Shields PT, Larkins MV, Jho HD. Microvascular decompression for hemifacial spasm. J Neurosurg 1995; 82: 201–210. 12. Casselman JW, Kuhweide R, Deimling M, Ampe W, Dehaene I, Meeus L. Constructive interference in steady state-3DFT MR imaging of the inner ear and cerebellopontine angle. AJNR Am J Neuroradiol 1993; 14: 47–57. 13. Girard N, Poncet M, Caces F, et al. Three-dimensional MRI of hemifacial spasm with surgical correlation. Neuroradiology 1997; 39: 46–51. 14. Tash RR, Kier EL, Chyatte D. Hemifacial spasm caused by a tortuous vertebral artery: MR demonstration. J Comput Assist Tomogr 1988; 12: 492–494.
Journal of Clinical Neuroscience (2004) 11(8), 840–844
15. Tash RR, Sze G, Leslie DR. Trigeminal neuralgia: MR imaging features. Radiology 1989; 172: 767–770. 16. Wong BY, Steinberg GK, Rosen L. Magnetic resonance imaging of vascular compression in trigeminal neuralgia. Case report. J Neurosurg 1989; 70: 132–134. 17. Hu XP, Tan KK, Levin DN, et al. Three-dimensional magnetic resonance images of the brain: application to neurosurgical planning. J Neurosurg 1990; 72: 433–440. 18. Kumon Y, Sakaki S, Kohno K, Ohta S, Ohue S, Miki H. Three-dimensional imaging for presentation of the causative vessels in patients with hemifacial spasm and trigeminal neuralgia. Surg Neurol 1997; 47: 178–184. 19. Shigematsu Y, Korogi Y, Hirai T, et al. Virtual MRI endoscopy of the intracranial cerebrospinal fluid spaces. Neuroradiology 1998; 40: 644–650. 20. Nagaseki Y, Horikoshi T, Omata T, et al. Oblique sagittal magnetic resonance imaging visualizing vascular compression of the trigeminal or facial nerve. J Neurosurg 1992; 77: 379–386. 21. Yamakami I, Kobayashi E, Hirai S, Yamaura A. Preoperative assessment of trigeminal neuralgia and hemifacial spasm using constructive interference in steady state-three-dimensional Fourier transformation magnetic resonance imaging. Neurol Med Chir (Tokyo) 2000; 40: 545–556. 22. Niwa Y, Shiotani M, Karasawa H, Ohseto K, Naganuma Y. [Identification of offending vessels in trigeminal neuralgia and hemifacial spasm using SPGR-MRI and 3D-TOF-MRA]. Rinsho Shinkeigaku (Jpn) 1996; 36: 544–550. 23. Kakizawa Y, Hongo K, Takasawa H, et al. ‘‘Real” three-dimensional constructive interference in steady-state imaging to discern microneurosurgical anatomy. Technical note. J Neurosurg 2003; 98: 625–630.
ª 2004 Elsevier Ltd. All rights reserved.