Computer Methods and Programs in Biomedicine 66 (2001) 99 – 103 www.elsevier.com/locate/cmpb
Virtual MR microscopy for unruptured aneurysm Naoki Kobayashi *, Takaaki Hosoya, Michito Adachi, Tamami Haku, Koichi Yamaguchi Department of Radiology, Yamagata Uni6ersity School of Medicine, 2 -2 -2 Iida-nishi, Yamagata 990 -9585, Japan
Abstract Purpose: This study was performed to evaluate the usefulness of virtual magnetic resonance microscopy (VMRM) for the diagnosis of cerebral aneurysms. Materials and methods: We reviewed 11 patients with unruptured aneurysms confirmed by angiography or surgical therapy. We evaluated the ability of VMRM to represent aneurysms and the findings of the form and neck of the aneurysms. Results: VMRM revealed 17 aneurysms, one of which was not detected by MR angiography (MRA). One suspected aneurysm by MRA was denied by VMRM. Although VMRM did not clearly demonstrate either one giant aneurysm or distal middle cerebral artery, two aneurysms in the cavernous sinus were clearly visualized by VMRM. VMRM seems to be almost equivalent to computed tomography angiography when detecting aneurysms without additional radiation exposure. Conclusion: VMRM is found valuable not only to plan the microscopic surgical therapy, but to visualize aneurysms. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Virtual MR microscopy; Unruptured aneurysm; Perspective volume rendering; MR angiography; CT angiography; Virtual CT microscopy
1. Introduction Virtual microscopy is said to be a perspective three-dimensional (3D) imaging technique that is processed on the software developed for computed tomography angiography (CTA). The term ‘microscopy’ is not derived from histopathological microscopy, but from microscopic images resembling microscopic surgical views. We applied this technique to magnetic resonance angiography * Corresponding author. Tel.: + 81-23-6285386; fax: +8123-6285389. E-mail address:
[email protected] (N. Kobayashi).
(MRA), which is called virtual MR microscopy (VMRM). We evaluated here the clinical usefulness of VMRM in patients with unruptured aneurysms compared with common techniques such as MRA, CTA, and virtual CT microscopy (VCTM) that was processed with the volume data of CTA.
2. Materials and methods We reviewed 11 patients with unruptured aneurysms confirmed by angiography or surgical therapy that ranged in age from 45 to 84 years (mean, 63.5 years). All 11 patients underwent examinations of MRA, CTA and angiography with digital
0169-2607/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 9 - 2 6 0 7 ( 0 1 ) 0 0 1 4 2 - 0
100
N. Kobayashi et al. / Computer Methods and Programs in Biomedicine 66 (2001) 99–103
subtraction angiography and biplane stereo magnification radiography. 3D volume data of MRA were acquired on a 1.5-Tesla MRI system (SIGNA Horizon; General Electric Medical System, Milwaukee, WI), using 3D spoiled gradient-recalled acquisition in the steady state (SPGR). The 3D SPGR sequence was performed at TR=21 ms, TE=3.4 ms, NEX = 1, and 25° flip angle with ramped RF and flow compensation pulses, without fat saturation or multi-slab techniques. The slab thickness was 40 mm, resulting in a section thickness of 0.5 mm. The field of view was 210× 150 mm2 with a 512 ×256×80 matrix. The total scanning time was 5 min and 42 s. MRA images were processed by a technique of maximum intensity display (MIP) on an independent console (General Electric Medical System, Milwaukee, WI). Scanning for CTA was performed with a helical CT system (HiSpeed Advantage; General Electric Medical System, Milwaukee, WI), using parameters of 1-mm slice thickness, 1 mm/s table speed (pitch of 1) and 18 cm in field of view. The slab thickness was 40 mm. All patients received an intravenous injection of 150 ml non-ionic contrast medium (Omnipaque, iohexol of 300 mgI/ml), which was set at a rate of 2.5– 3.5 ml/s and 25 s scan delay. The total scanning time was 40 s. After scanning, 3D volume data were obtained by reconstruction with 0.5-mm intervals, which gave 12 – 15 cm of deformed field of view and were transferred to the workstation (Advantage Windows; General Electric Medical System, Milwaukee, WI). CTA images were reconstructed with a surface-rendering algorithm on the workstation. Virtual microscopy was processed with the 3D volume data of MRA and CTA on the same workstation (Advantage Windows) using a perspective volume rendering algorithm (Navigator), where we chose ‘black in white’ paradigm at 60° of camera field of view. The threshold value was set at 140–170 of signal intensity in VMRM and 150– 200 HU in VCTM. Virtual microscopic images in at least six directions including anterior, posterior, left, right, superior and inferior were obtained by 1–2 mm step, respectively. From 20 to 30 min was required for image processing.
VMRM images were evaluated as the ability to represent aneurysms that were confirmed by surgery or angiography, comparing with MRA. We also evaluated the findings of the form and neck of the aneurysms by VMRM in comparison with CTA and VCTM.
3. Results Seventeen aneurysms, one of which was not detected by MRA, were confirmed by surgery and/or angiography in 11 patients. In a patient who had one suspected aneurysm, which was diagnosed as an infundibular dilatation, VMRM and CTA showed a new aneurysm at the contralateral side, which was clipped by surgery (Fig. 1). Thus, 16 of 17 aneurysms, which were suspected by MRA, were confirmed by surgery or angiography. In a patient with two aneurysms, one aneurysm was diagnosed by surgery and the other was confirmed by angiography. Finally, five aneurysms of five patients were confirmed by surgery, 12 aneurysms in seven patients were confirmed by angiography. Seven of 11 patients had solitary lesions, two patients had two aneurysms, and two had three aneurysms. Six aneurysms were located in the internal carotid arteries, two in the anterior cerebral arteries, five in the middle cerebral arteries, and four in the basilar arteries. The size of aneurysms ranged from 1.5 to 30 mm in diameter. The mean size was 7.4 mm. Three of 17 aneurysms were larger than 10 mm in diameter. VMRM images showed all 17 aneurysms that were confirmed by surgery and/or angiography. Two aneurysms, which had complicated forms, were more precisely demonstrated by VMRM than by MRA (Fig. 2). Two aneurysms in the cavernous sinus were more clearly demonstrated by VMRM in comparison with CTA and VCTM. VMRM revealed two hypoplastic arteries (an anterior cerebral artery and a posterior communicating artery) near the aneurysms more clearly than that VCTM did. In five of 11 patients, VMRM provided more useful information than that CTA and VCTM did. On the other hand, VMRM did not clearly demonstrate a giant aneu-
N. Kobayashi et al. / Computer Methods and Programs in Biomedicine 66 (2001) 99–103
rysm in one case or distal branches to the middle cerebral artery aneurysm in another case compared with CTA and VCTM. In two patients, CTA and VCTM were superior to VMRM.
4. Discussion Perspective volume rendering [1], which has been developed as image processing software to
101
process the volume data of helical CT, is beginning to be applied to clinical use such as virtual endoscopy [1–3] and virtual microscopy. We applied it to MRA with 3D volume data. MRA is available for screening unruptured aneurysms [4], although this method is ineffective to confirm them in some cases [5,6]. In this study, VMRM revealed an aneurysm that was not detected by MRA, and confirmed an infundibular dilatation that was suspected to be an aneurysm
Fig. 1. An aneurysm of the left internal carotid artery and an infundibular dilatation of the right posterior-communicating artery. By MRA ((a) superior view, (b) right ICA, lateral view, (c) left ICA, lateral view), we suspected an aneurysm of the right internal carotid artery (ICA) at the branching of the posterior-communicating artery. CTA ((d) superior view) and VMRM ((e) superior view) show it was not an aneurysm, but an infundibular dilatation of the posterior-communicating artery. In addition, they demonstrated an aneurysm at the contralateral internal carotid artery.
102
N. Kobayashi et al. / Computer Methods and Programs in Biomedicine 66 (2001) 99–103
Fig. 2. An aneurysm of the left internal carotid artery at the branching of the posterior-communicating artery. Although a simple berry aneurysm (white arrow) was depicted by CTA ((a) left lateral view), VMRM ((b) left lateral view) and VCTM ((c) left lateral view) show an aneurysm consisting of two dome-like dilated portions (black arrow). VMRM demonstrated more clearly the relationship between the aneurysm and ICA than VCTM.
by MRA. VMRM decreased the false-positive and false-negative findings by MRA, and the sensitivity and specificity of VMRM were 100%. It was suggested that VMRM was almost equivalent to CTA in diagnosing the existence of an aneurysm. VMRM images can be obtained by postprocessing techniques using 3D data of MRA alone. We consider VMRM to be a very useful technique, because patients require no other examinations or additional radiation exposure. Clinically, CTA does not accurately demonstrate aneurysms in the cavernous sinus. VMRM more clearly revealed two aneurysms in the cavernous sinus than CTA and VCTM, because the venous flow or bones did not affect VMRM as much as MRA. VMRM was considered to be
useful not only to detect aneurysms in the cavernous sinus, but also to plan the microscopic surgical therapy. Although VMRM did not clearly show a giant aneurysm in one case or distal branches to a middle cerebral artery aneurysm in another case, VMRM tend to reveal hypoplastic arteries more precisely in comparison with VCTM. It may be because spatial resolution of MRA is superior that of CTA but the blood speed has not been sufficient for MRA in some case [7,8]. Due to its eddy diffusion, VMRM fails to present the whole body of the giant aneurysm. In such cases, VMRM is not adequate to demonstrate the aneurysm. In perspective volume rendering images, the closer the viewpoint, the higher the resolution.
N. Kobayashi et al. / Computer Methods and Programs in Biomedicine 66 (2001) 99–103
Such perspective images are considered useful to plan microscopic surgical therapy. Although VMRM was very useful to diagnose aneurysms in the cavernous sinus, it is necessary to perform CTA and VCTM to plan the surgical therapy because it is very important to recognize the spatial relationship between the aneurysm and vein, and between the aneurysm and bone [9,10]. Regarding preoperative information, we consider that CTA, VCTM and VMRM are complementary techniques each other.
5. Conclusion VMRM is found capable of demonstrating the details of the aneurysms, especially those in the cavernous sinuses that were not demonstrated by CTA. VMRM is valuable not only to plan the microscopic surgical therapy, but also to visualize aneurysms.
References [1] G.D. Rubin, C.F. Beaulieu, V. Argiro, et al., Perspective volume rendering of CT and MR images: applications for endoscopic imaging, Radiology 199 (1996) 321 –330.
.
103
[2] Y. Shigematsu, Y. Korogi, T. Hirai, et al., New developments: 2. Virtual MR endscopy in the central nervous system. ISMR-M (1998) 289 – 296. [3] N. Bontozoglou, H. Spanos, P. Lasjaunias, G. Zarifis, Intracranial aneurysms: endovascular evaluation with three-dimensional-display MR angiography, Radiology 197 (1995) 876 – 879. [4] N. Bontozoglou, H. Spanos, P. Lasjaunias, G. Zarifis, Three-dimentional display of the orifice of intracranial aneurysms, Neuroradiology 36 (1994) 346 – 349. [5] K.W. Stock, E.W. Radue, A.L. Jacob, et al., Intracranial arteries: prospective blinded comparative study of MR angiography, Radiology 195 (1995) 451 – 456. [6] S.W. Atlas, L. Sheppard, H.I. Goldberg, Intracranial aneurysms: detection and characterization with MR angiography with use of an advanced postprocessing technique in a blinded-reader study, Radiology 203 (1997) 807 – 814. [7] P. Brugieres, J. Blustajn, C.L. Guerinel, et al., Magnetic resonance angiography of giant intracranial aneurysms, Neuroradiology 40 (1998) 96 – 102. [8] N. Kurihara, S. Takahashi, S. Higano, et al., Evaluation of large intracranial aneurysm with three-dimensional MRI, J. Comput. Assisted Tomography 19 (1995) 707 – 712. [9] M.J. Harrison, B.A. Johnson, G.M. Gardner, et al., Preliminary results on the management of unruptured intracranial aneurysms with magnetic resonance angiography and computed tomographic angiography, Neurosurgery 40 (1997) 947 – 957. [10] D. Tampieri, R. Leblanc, J. Oleszek, et al., Three-dimensional computed tomographic angiography of cerebral aneurysms, Neurosurgery 36 (4) (1995) 749 – 755.