- Email: [email protected]
Magnetic resonance imaging with aneurysmal subarachnoid hemorrhage: comparison with computed tomography scan
Surg Neurol
71
1990;34:71-8
Magnetic Resonance Imaging with Aneurysmal Subarachnoid Hemorrhage: Comparison with Computed Tomography Scan Ken-ichi Matsumura, M.D., Masayuki Matsuda, M.D., Jyoji Handa, M.D., and Giro Todo, M.D. From the Department of Neurosurgery, Shiga University of Medical Science, Ohtsu, and the Department of Radiology, Nagahama Red Cross Hospital, Nagahama, Shiga-ken, Japan
Matsumura K, Matsuda M, Handa J, Todo G. Magnetic resonance imaging with aneurysmal subarachnoid hemorrhage: comparison with computed tomography scan. Surg Neurol 1990;34:71-8.
Magnetic resonance studies of 27 consecutive preoperative and 33 postoperative patients with cerebral aneurysm and subarachnoid hemorrhage were reviewed. Magnetic resonance imaging using a 0.5- or 0.22-Tesla unit was at least as accurate as computed tomography scan for detection of acute subarachnoid hemorrhage. Magnetic resonance imaging was superior to computed tomography scan for demonstrating blood in the ventricles or posterior fossa subarachnoid space, transependymal migration of the cerebrospinal fluid, and several kinds of iatrogenic intracranial pathologies in postoperative patients. Ischemic lesions, particularly fresh lesions caused by delayed cerebral vasospasm, were much better shown on magnetic resonance imaging than on computed tomography scan. Nonferromagnetic Sugita clips caused significant artifacts, but the area of artifacts was consistently smaller, and a reasonable evaluation of structures relatively closer to the clip was possible with magnetic resonance imaging rather than computed tomography scan. KEYWORDS: Cerebral aneurysm; Computed tomography; Intracranial hemorrhage; Magnetic resonance; Subarachnoid hemorrhage
Introduction Magnetic resonance imaging (MRI) has proven to be a noninvasive, highly sensitive, and accurate diagnostic method for evaluating a wide variety of intracranial diseases. In the evaluation of patients with suspected cerebrovascular diseases, noninvasive demonstration of blood flow without the use of iodinated contrast medium and
Address reprint requests to:Jyoji Handa, M.D., Department of Neurosurgery, Shiga University of Medical Science, Seta, Ohtsu, 520-21 Shiga-ken, Japan. Received December 27, 1989; accepted February 26, 1990. © 1990 by ElsevierSciencePublishingCo., Inc.
superb contrast sensitivity to abnormal signals associated with cerebral ischemia, as well as to previous hemorrhage, are apparent advantages of MRI over computed tomography (CT) scan or cerebral angiography. The unprecedented diagnostic value of MRI for cerebral vascular malformation, cerebral infarction, or intracerebral hemorrhage has been well established. On the other hand, little has been reported on the role of MRI for diagnosing aneurysmal subarachnoid hemorrhage (SAH). This is partly due to the fact that MRI is extremely sensitive to motion artifacts, and absolute immobilization of acutely ill patients is difficult to obtain. In addition, patients suffering from acute SAH are often connected to life-supporting equipment and monitoring devices, which could contain ferromagnetic materials and prohibit performance of MRI. The purpose of this paper is to report our initial experiences with MRI in patients with aneurysmal SAH, and compare the results with those of CT scan. Patients and Methods Magnetic resonance imaging was performed in 27 preoperative patients with aneurysmal SAH (group 1), and in 33 aneurysm patients after direct operation (group 2). Three aneurysms in group 1 were large or giant aneurysms (17, 21, and 25 mm in maximal diameter), and five patients with unruptured aneurysms were included in group 2. Of 27 patients in group 1 (aneurysmal SAH), 10 patients were male and 17 were female, and their ages ranged from 34 to 76, with an average of 58 years. Twenty patients were examined by MRI in the acute stage; namely 3-38 hours (mean = 21 hours) after SAH. Five patients were studied in the subacute stage, on day 5 through day 13 post-SAH (mean = day 7, the day of bleeding being day 0), and the remaining two patients were examined in the chronic stage, on day 21 and day 34, respectively. Neurological grades (World Federation of Neurological Surgery, WFNS) of patients at the time of MRI are shown in Table 1. 0090-3019/90/$3.50
72
Surg Neurol 1990;34:71-8
Matsumura et al
Table 1. Clinical Summary of 27 Patients with Aneurysmal Subarachnoid Hemorrhage (Group i) WFNS grade at MRI Stage Acute (day 0a--3) Subacute (day 4 - 1 4 ) Chronic (>day 15) Total
No. of patients
I
II
III
IV
V
20
11
5
2
2
0
5
2
1
2
0
0
2
2
0
0
0
0
27
15
6
4
2
0
Interval between onset and MR1 3 - 4 8 hours (mean = 21 hours) 5 - 1 3 days (mean = 7 days) 21 days and 34 days
Abbreviations: MR1, magnetic resonance imaging; WFNS, World Federation of Neurological Surgery. " Day 0 = day of hemorrhage. Three large or giant aneurysms are included, one in the acute, one in the subacute, and the remaining one in the chronic stage. Ten patients were male and 17 patients were female.
O f the 33 postoperative patients in group 2, 14 patients were male and 19 were female, and their ages ranged from 24 to 72, with an average of 54 years. In 31 patients the aneurysm was occluded with a nonferromagnetic Sugita clip. All these patients were operated upon at our institution, where all previous types of aneurysm clips had been replaced with the nonferromagnetic clips. N o patient in whom an aneurysm clip was placed at an outside institution was accepted for MRI, since we could not be absolutely sure o f the clip materials. The aneurysm was wrapped with a sheet o f cottonoid and reinforced with fibrin glue in the remaining two patients. Intervals between operation and MRI are shown in Table 2. Magnetic resonance imaging was performed using a superconducting Picker Vista MR unit operating at 0.5 Tesla, with the exception of 10 cases in whom the Toshiba MRT-22A (0.22 Tesla) was employed. Spin-echo pulse sequences were used, and T2-weighted images Table 2. ClinicalSummary of 33 PostoperativePatients (Group 2) No. of patients Gender
Ruptured or unruptured
Male Female
14 19
Ruptured
28 5
Unruptured
Operation
Neck clipping~ Wrapping
31 2
Interval between operation and MRI
--<1 week 1 week-lmonth 1 month-3 months 3 m o n t h s - 1 year > 1 year
6 11 9
Abbreviation: MRI, magnetic resonance imaging. Nonferromagnetic Sugita clip was used.
3 4
(T2WIs) with repetition times (TR) of 1 8 0 0 - 2 5 0 0 ms and echo times (TE) of 9 0 - 1 1 0 ms, Tl-weighted images (TIWIs) with T R of 5 0 0 - 7 0 0 ms and TE o f 2 5 - 3 0 ms, and proton density-weighted images (0WIs) with T R of 1800-2500 ms and TE of 30 ms were obtained in either the coronal, sagittal, or axial plane. Slice thickness was 10 mm in most cases, and images were displayed on 256 x 256 matrices. Computed tomography scan was performed using Siemens D R H , GE C T / T 8800, or YMS Quantex RX, and the slice thickness was either 5 or 10 mm. All o f our preoperative and postoperative patients underwent MRI and CT scan at least once, and none had just one of these two tests. Intervals between paired MRI and CT scan were less than 48 hours, and in most cases less than 24 hours, in the acute or subacute stage. Results Findings did not differ significantly between the two MRI units used, so that the results were analyzed together. The pertinent findings of MRI and CT scan in 27 consecutive patients with aneurysmal SAH (group 1) are compared in Table 3, and the representative images are illustrated in Figures 1-4. The T2WI was best suited for detection of blood in the basal cisterns and superficial subarachnoid spaces, showing a signal intensity higher than the cerebrospinal fluid (CSF) in 25 o f 27 patients, whereas CT scan confirmed SAH in 22 patients as the area of high density or isodensity (Figures 1 B, 1 C, and 2 A). Blood-fluid levels in the occipital horns, periventricular edema, cerebral infarction, and aneurysm per se were all better demonstrated on MRI than on CT scan (Figures 1 A, 1 B, 3, and 4). All infarcts found on MRIs were old ones, and no patient in the acute stage o f SAH showed abnormal signal intensities consistent with an area of edema or recent ischemia, possibly due to early vasospasm. N o n e of the postoperative patients (group 2) com-
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
Table 3. Comparisonof Magnetic Resonance Imaging and
Computed Tomography Scan in 27 Consecutive Patients with Aneurysmal Subarachnoid Hemorrhage No. of patients with positive findings Findings
Stage"
MRI
CT
Total no. of patients studied with both CT & MRI
Intraventricular blood-fluid level Blood in subarachnoid space Periventricular abnormal intensity/density Old infarction
Acute Subacute Chronic Acute Subacute Chronic Acute Subacute Chronic Acute Subacute Chronic
19 3 0 20 4 1 11 3 2 9 4 1 10
6 3 0 18 4 0 1 1 0 4 0 0 5
20 5 2 20 5 2 20 5 2 20 5 2 27
Identification of aneurysm
Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging. * Stage of patients after subarachnoid hemorrhage.
73
immediate vicinity of the valve, and it did not affect the diagnostic information of intracranial contents. Outside an area of artifacts, neither MRI nor CT scan showed any operative brain damage, such as contusional hemorrhage or focal edema due to retraction of the brain. The incidence of various postoperative pathologic changes seen on MRI and CT scan is compared in Table 4. Most of the extracerebral fluid collections were of minor clinical significance, such as a thin subdural fluid collection, minimal epidural blood collection, or hemorrhagic contusion of the temporal muscles, all without space-occupying effect (Figure 5 B, Table 4). Cerebral infarcts of various sizes and ages were much better seen on MRI than on CT scan. Both preoperative and postoperative MRIs were available for comparison in 11 of 33 patients in group 2 (Table 5, Figures 2 D and 5 B). In three of these 11 patients, no cerebral infarction was found on both preoperative and postoperative MRIs. In three patients, the number of infarcts remained unchanged, whereas in the remaining eight patients the number of foci of infarcts increased from 33 before operation to 52 after operation. Discussion
plained of any discomfort during MRI examinations, and no complication referable to metallic implant was encountered. Significant imaging artifacts were observed on both MRI and CT scan in all 31 postoperative patients in whom a metallic clip had been placed at operation (Figures 2 C and 5). Most patients had a single Sugita clip, but two patients had two and one had three clips. On CT scan, artifacts consisted of a center of complete attenuation of x-ray and radiating dense streaking artifacts, which were more pronounced along a long axis of the clip, and no diagnostic information was obtained in an area of such artifacts. On MRI, artifacts consisted of a central area of complete signal loss and a peripheral narrow rim of increased signal intensity. The size of the artifacts tended to be somewhat larger on T2WIs than on T1WIs. No diagnostic information was available in the region of artifacts. Close to the sharply demarcated rim of increased signal intensity, some distortion of the image was unavoidable, but it did not adversely affect diagnostic information to any great degree. An area of artifacts was consistently smaller on MRI than on CT scan (Figures 5 A and C, Table 4). In addition, a rapid transition to normal image quality outside an area of artifacts on MRI enabled a reasonable evaluation of structures relatively close to the clip. A Holter-Hausner shunt valve, that is also MRI compatible, caused metallic artifacts on both MRI and CT scan in four patients. The area of distortion of the MRI, however, was much smaller and restricted to the skin and the skull in the
Magnetic resonance imaging has been established as an excellent screening method for evaluation of a wide variety of diseases of the central nervous system, but its role in diagnosis of SAH has not been well documented. Several authors consider that acute and/or diffuse SAH is difficult to detect on MRI [2,5,8,29]. According to Zimmerman et al [29], higher oxygen tension of the CSF compared with that of the brain may lead to a low level of deoxyhemoglobin in the extravasated blood. In the absence of or with the very low level of deoxyhemoglobin, T2 shortening does not occur during an acute posthemorrhagic period. In addition, a pulse-shift effect of CSF pulsation may mask the effects of hemorrhage in the subarachnoid space and make detection of SAH difficult, particularly during the initial 3 days after hemorrhage [29]. Bradley et al [2] considered methemoglobin formation with T1 shortening as an important factor for increasing the signal intensity of the blood in the subarachnoid space over time, and they concluded that MRI might be useful in the diagnosis of SAH in the subacute stage but not in an acute stage. Measuring relaxation times of mixtures of normal human CSF and heparinized blood, Chakeres and Bryan [5] found a progressive shortening of T1 and T2 relaxation times with an increasing amount of blood. However, the concentrated subarachnoid blood in an acute stage of hemorrhage had relaxation times similar to the normal brain and was nearly isointense on MRIs. Partial volume effect of thin subarachnoid spaces also seemed to interfere with a definite delineation of acute SAH [5].
74
Surg Neurol 1990;34:71-8
Matsumura et al
Figure 1. (A,B) Ruptured aneurysm of the left internal carotid artery at the branching of the anterior choroidal artery. Magnetic resonance imaging on the day of S A H shows blood-fluid level (arrows) on o W l (A, T R / T E = 2100/30) and T 2 W I (B, T R / T E = 2100/90). Subarachnoid space shows intensities higher than the ventricular CSF on T2WI. (C) Computed tomography scan shows high densities in the cisterns but fails to show the intraventricular blood. (D) On follow-up M R I 220 days later, CSF in the subarachnoid space on T 2 W I is no longer higher in intensities than the ventricular CSF.
On the other hand, Jenkins et al [18] reported that they could identify the subarachnoid blood as often on MRI as on CT scan in 30 patients studied between 8 hours and 5 days after proven SAH. The blood-stained CSF showed a signal intensity higher than the brain and normal CSF on the axial T2-weighted spin-echo images [18]. Satoh and Kadoya [22] also found that in the acute stage, particularly within 24 hours after bleeding, the bloody CSF in the subarachnoid space showed high signal intensity relative to the brain on T2WI. Both Jenkins et al and Satoh and Kadoya used a Picker 0.15-Tesla resistive unit. Results of the present study are generally consonant with those of Jenkins et al [18] and Satoh and Kadoya [22], and substantiate the high sensitivity of MRI for detection of acute SAH (Table 3, Figure 1). Fine details of sulci and cisterns containing the bloody CSF, particularly those in the posterior fossa, were better shown on MRI than on CT scan. In addition, the blood in the
lateral ventricles was also seen more often on MRI than on CT scan. A high signal intensity of the blood-stained CSF on T2-weighted spin-echo images relative to the brain or the normal CSF is more or less difficult to explain, at least in the acute stage of SAH. In our study, T2 relaxation times of the bloody CSF in the lateral ventricles (144 + 9.7 ms) and subarachnoid cisterns (125 -+ 13.2 ms) were both shorter when compared with those of normal ventricular CSF (387 +- 68.8 ms). A high signal intensity seen on T2-weighted spin-echo images in the face of shortened T2 relaxation time is presumed to reflect a markedly shortened T1 relaxation time [18,22]. Even if this presumption is correct, the reason for such a marked shortening of T1 relaxation time in the bloody CSF in acute SAH remains to be elucidated. Production of methemoglobin is known to reduce T1 relaxation times [ 12]. During lysis of extravasated eryth-
r
w
Figure 2. (A,B) Ruptured aneurysm of the anterior communicating artery. Preoperative T2WIs (TR/TE = 2100/90) on day 1 post-SAH shows high intensities in the basal cisterns (arrowheads) (compare with the postoperative image in Figure 2 C). (C,D) Postoperative M R I shows an artifact due to the clip, but the structures in the immediate vicinity of the clip, such as the medial temporal lobe and the midbrain, are clearly seen. New patchy cerebral lesions are also seen on T 2 W I ( T R / T E = 2100/90).
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
75
2
Figure 3. (A-C) Ruptured aneurysm of the horizontal portion of the right anterior cerebral artery. The MRIs on the day of SAH show ventricular dilatation and periventricular hyperintensities (arrows) (A, TR/TE = 2100/30; B, TR/TE = 2100/60; C, T R / T E = 2100/90). The patient later developed symptomatic hydrocephalus. (D) Subarachnoid hemorrhage and the blood coating the choroid plexus were apparent also on CT scan; however, C T scan did not demonstrate periventricular abnormal densities.
rocytes, which begins within hours after SAH, oxyhemoglobin and smaller amounts of deoxyhemoglobin, iron, heme pigments, and methemoglobin are released into the CSF. Hayman et al [14] however, found that the highest CSF methemoglobin level reported so far was negligible at clays 0 to 7 after SAH, and they raised some doubts on the hypothesis that a marked T1 shortening was caused in the acute stage of SAH by the extremely low methemoglobin. As has been reported by various authors [9,23,25,26], MRI was more sensitive than CT scan for detecting cerebral infarctions both in the preoperative and postoperative studies (Tables 3 and 5). All the infarcts seen on MRIs taken in the acute or subacute stage were old lesions, and not one was
thought to have occurred in direct relation with the aneurysmal rupture. In the postoperative patients, however, the number of infarcts had considerably increased compared with the preoperative MRIs (Table 5). In five of 11 patients, the total number of ischemic foci increased from 33 before operation to 52 after operation. All five of these patients developed an angiographic vasospasm during the interval between the preoperative and postoperative MRI studies, and three of them developed clinical signs of focal cerebral ischemia (symptomatic vasospasm). New ischemic lesions on the postoperative MRIs were found in locations distant from the operative field, such as the corona radiata, centrum semiovale, or basal ganglia, and often bilaterally. We, therefore, assume that most
Figure 4. (A) Sagittal (TR/TE = 500/25) and (B,C)axial (B, TR/TE = I900/25; C, TR/TE = 1900/1 I0) MRIs on day 13 post-SAH, showing an aneurysm as an area of flow void (arrow) surrounded by circumscribed clot (arrowheads). Computed tomography scan without and with contrast enhancement fails to demonstrate aneurysm or subarachnoid clot. (D) Angiogram shows an aneurysm of the distal portion of the posterior inferior cerebellar artery (arrow).
76
Surg Neurol 1990;34:71-8
Matsumura et al
F i g u r e 5. Postoperative studies. Ruptured aneurysm of the left middle cerebral artery and unruptured aneurysm of the right middle cerebral artery were occluded using Sugita clips. An area of artifacts due to the clip is much smaller on MRI (A, T R / T E = 2100/100) than on C T scan (C). A thin layer of epidural clots and a spotty high-intensity lesion are evident on T 2 W I (B), but both of these intracranial pathologies are not apparent on C T scan (D).
of the new ischemic lesions can be ascribed to delayed vasospasm. Recently, patchy, deep cerebral lesions seen on MRIs have been described in patients with hydrocephalus, dementia, multiple sclerosis, or hypertension, and nonspecific white matter lesions also have been reported in clinically healthy elderly subjects [13,20,28]. Large Virchow-Robin spaces also have been reported to mimic lacunar infarcts [ 15]. Occasionally, it is difficult to differentiate lacunar infarcts from a large perivascular space. Lacunar infarcts are known to occur in the upper two thirds of the basal ganglia, and usually not of CSF intensity on all pulse sequences. By definition, a large Virchow-Robin space has to be isointense relative to the CSF on all pulse sequences and must conform to the path of the penetrating arteries. Such lesions must be differentiated from lacunar infarction. In this study, MRI was more sensitive than CT scan with and without contrast enhancement in the detection of aneurysms per se. The diameter of the aneurysm depicted on MRIs was 8 mm or larger. On CT scan, high-density clots surrounding the aneurysm tend to obscure the aneurysm in the acute stage of SAH, whereas rapidly flowing blood within the aneurysm usually produces a signal-void area on routine spin-echo MRIs and sharply contrasts with the high signal intensity of the bloody CSF or clot on T2-weighted spin-echo images (Figure 4) [1,18,22]. However, an aneurysmal sac may not always appear on MRIs as an area of signal void because of several factors, such as slice thickness, partial volume averaging effect, dimension of the aneurysmal sac, and others. In addition, a tortuous artery, CSF flow void, or an anatomic variant such as an asymmetrically pneumatized anterior clinoid process [1] may well mimic aneurysmal "signal void," and MRI diagnosis of an aneurysm has to be made with care. The use of ad-
junctive flow-sensitive imaging sequences such as gradient-recalled acquisition may be useful to confirm the suspicious area as an aneurysm [ 11,24,27]. T2-weighted spin-echo images demonstrated a periventricular zone of high signal intensity in 16 or 27 preoperative patients, whereas CT scan showed periventricular lucency in only two patients (Table 3). An area of periventricular hyperintensity on T2-weighted MRI has been observed in patients with cerebral infarction, white matter diseases such as multiple sclerosis, and hydrocephalus. A mild periventricular hyperintensity is also seen in elderly subjects with no evidence of intra-
Table 4. Comparison of Magnetic Resonance Imaging and Computed Tomography Findings in 33 Patients Who Underwent Direct Operation For Cerebral Aneurysm a No. of patients with positive findings Findings
MRI
CT
Infarction Subgaleal fluid collection Epidural hemorrhagic collection Subdural fluid collection Periventricular high intensity/ lucency Incidental tumor Clip artifact Size of artifact ~
26 3
14 3
13
6
13
13
10
7
1 31 3.1 (2.0-5.2)
1 31 11.3 (3.3-16.0)
Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging. Clipping of aneurysm with a Sugita clip was performed in 31 patients. Size of artifact in millimeters.
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
Table 5. Comparison of Location and Number of Cerebral Infarcts in 11 Patients Studied both Preoperatively and Postoperatively by Magnetic Resonance Tomography No. of foci of infarct Location of infarction
Preoperative MRI
Postoperative MRI
Internal capsule Corona radiata/ centrum semiovale Periventricular white matter Basal ganglia Cerebral cortex Pons Total
2 25
2 38
5
5
1 0 0 33
4 2 1 52
Abbreviation: MRI, magnetic resonance imaging. Preoperative examination was performed on day 0 (day of hemorrhage) through day 5, and postoperative study was carried out on day 21 through day 70.
cranial pathology [3,6,13,16,20,28]. Increased T2 signal intensity in the periventricular region is therefore not specific to one or another pathologic change: however, a zone of periventricular high intensity signal on T2WI in the present series was narrow (up to 2 mm) in width, smooth in contour, most prominent in the area of the anterior horn and trigone of the lateral ventricles, and bilaterally symmetrical. In addition, six of 14 patients who showed such a periventricular hyperintensity in the acute stage of SAH developed frank hydrocephalus days or weeks later. Therefore, we are of the opinion that in many patients, periventricular hyperintensities seen in the acute stage of SAH represent a transependymal migration of the CSF. In the postoperative cases, potential risks to patients with metallic surgical implants have to be seriously considered. Several authors carefully studied the martensitic content, magnetic field-induced displacement, heating effects, and other hazards of various kinds of aneurysm clips [7,10,19,21]. Clips inevitably caused significant artifacts on both CT scan and MRI, and interpretation was impossible in an area of artifacts. However, MRI proved to be far superior to CT scan in studying postoperative patients in that a size of artifacts was much smaller and a rapid change to the normal image quality immediately outside the central area of artifact enabled the evaluation of anatomic structures [4,17]. In conclusion, the present study confirms the important role of MRI in the management of patients with aneurysmal SAH.
References 1. Biondi A, Scilfa G, Scotti G. Intracranial aneurysms: MR imaging. Neuroradiology 1988;30:214-8.
77
2. Bradley WG Jr, Schmidt PG. Effect ofmethemoglobin formation on the MR appearance of subarachnoid hemorrhage. Radiology 1985;156:99-193. 3. Brant-Zawadzki M, Fein G, van Dyke C, Kiernan R, Davenport L, de Groot J. MR imaging of the aging brain: patchy whitematter lesions and dementia. AJNR 1985;6:675-82. 4. Brothers MF, Fox AJ, Lee DH, Pelz DM, DeveikisJP. MR imaging after surgery for vertebrobasilar aneurysm. AJNR 1990;11:149-61. 5. Chakeres DW, Bryan RN. Acute subarachnoid hemorrhage: in vitro comparison of magnetic resonance and computed tomography. AJNR 1986;7:223-8. 6. Chimowitz MI, Awad IA, Furlan AJ. Periventricular lesion on MRI. Facts and theories. Stroke 1989;20:963-7. 7. Davis PL, Crooks L, Arakawa M, McRee R, Kaufman L, Margulis AR. Potential hazards in NMR imaging: heating effects of changing magnetic fields and RF fields on small metallic implants. AJR 1981;137:857-60. 8. DeLaPaz RL, New PFJ, Buonanno FS, Kistler JP, Oot RF, Rosen BR, Taveras JM, Brady TJ. NMR imaging of intracranial hemorrhage. J Comput Assist Tomogr 1984;8:599-607. 9. Dewitt LD, Buonanno FS, Kistler JP, Brady TJ, Pykett IL, Goldman MR, Davis KR. Nuclear magnetic resonance imaging in evaluation of clinical stroke syndromes. Ann Neurol 1984;16:535-45. 10. Dujovny M, Kossovsky N, Kossowsky R, Valdivia R, Suk JS, Diaz FG, Berman SK, Cleary W. Aneurysm clip motion during magnetic resonance imaging: in vivo experimental study with metallurgical factor analysis. Neurosurgery 1985;17:543-8. 11. Gean AD, Pile-Spellman J, Heros RC. A pneumatized anterior clinoid mimicking an aneurysm on MR imaging--report of two cases. J Neurosurg 1989;71:128-32. 12. GomoriJM, Grossman RI, Goldberg HI, Zimmerman RA, Bilaniuk LT. Intracranial hematomas: imaging by high-field MR: Radiology 1985;157:87-93. 13. Gupta SR, Naheedy MH, Young JL, Ghobrial M, Rubino FA, Hindo W. Periventricular white matter changes and dementia. Clinical, neuropsychological, radiological, and pathological correlation. Arch Neurol 1988;45:637-41. 14. Hayman LA, PaganiJJ, KirkpatrickJB, Hinck VC. Pathophysiology of acute intracerebral and subarachnoid hemorrhage: applications to MR imaging. AJNR 1989;10:457-61. 15. Heier LA, Bauer CJ, Schwartz L, Zimmerman RD, Morgello S, Deck MDF. Large Virchow-Robin spaces: MR-clinical correlation. AJNR 1989;10:929-36. 16. Hendrie HC, Farlow MR, Austrom MG, Edwards MK, Williams MA. Foci of increased T2 signal intensity on brain MR scans of healthy elderly subjects. AJNR 1989;10:703-7. 17. Holt,s S, Olsson M, Romner B, Larsson E-M, S~iveland H, Brandt L. Comparison of MR imaging and CT in patients with intracranial aneurysm clips. AJNR 1988;9:891-7. 18. Jenkins A, Hadley DM, Teasdale GM, Condon B, Macpherson P, Patterson J. Magnetic resonance imaging of acute subarachnoid hemorrhage. J Neurosurg 1988;68:731-6. 19. Kean DM, Worthington BS, Firth JL, Hawkins RC. The effects of magnetic resonance imaging on different types of microsurgical clips. J Neurol Neurosurg Psychiatry 1985;48:286-7. 20. KirkpatrickJB, Hayman LA. White-matter lesions in MR imaging of clinically healthy brains of elderly subjects: possible pathologic basis. Radiology 1987;162:509-11. 21. New PFJ, Rosen BR, Brady TJ, Buonanno FS, Kistler JP, Burt CT, Hinshaw S, Newhouse JH, Pohost GM. Taveras JM. Potential hazards and artifacts of ferromagnetic and nonferromagnetic surgical and dental materials and devices in nuclear magnetic resonance imaging. Radiology 1983;147:139-48.
78
Surg N e u r o l 1990;34:71-8
22. Satoh S, Kadoya S. Magnetic resonance imaging of subarachnoid hemorrhage. Neuroradiology 1988;30:361-6. 23. Sipponen JT, Kaste M, Ketonen L, Sepponen RE, Katevuo K, Sivula A. Serial nuclear magnetic resonance (NMR) imaging in patients with cerebral infarction. J Comput Assist Tomogr 1983;7:585-9. 24. TsurudaJS, Halbach VV, Higashida RT, Mark AS, Hieshima GB, Norman D. MR evaluation of large intracranial aneurysm using cine low flip angle gradient-refocused imaging. AJNR 1988;9:415-24. 25. Unger EC, Gado MH, Fulling KF, Littlefield JL. Acute cerebral infarction in monkeys: an experimental study using MR imaging. Radiology 1987;162:789-95.
Matsumura et al
26. Virapongse C, Mancuso A, Quisling R. Human brain infarcts: Gd-DTPA-enhanced MR imaging. Radiology 1986;161: 785-94. 27. Wagte WA, Dumoulin CL, Souza SP, Cline HE. 3DFT MR angiography of carotid and basilar arteries. AJNR 1989;10: 911-9. 28. Zimmerman RD, Fleming CA, Lee BCP, Saint-Louis LA, Deck MDF. Periventricular hyperintensity as seen by magnetic resonance: prevalence and significance. AJNR 1986;7: 13-20. 29. Zimmerman RD, Heier LA, Snow RB, Liu DPC, Kelly AB, Deck MDF. Acute intracranial hemorrhage: intensity changes on sequential MR scans at 0.5 T. AJNR 1988;9:47-57.
71
1990;34:71-8
Magnetic Resonance Imaging with Aneurysmal Subarachnoid Hemorrhage: Comparison with Computed Tomography Scan Ken-ichi Matsumura, M.D., Masayuki Matsuda, M.D., Jyoji Handa, M.D., and Giro Todo, M.D. From the Department of Neurosurgery, Shiga University of Medical Science, Ohtsu, and the Department of Radiology, Nagahama Red Cross Hospital, Nagahama, Shiga-ken, Japan
Matsumura K, Matsuda M, Handa J, Todo G. Magnetic resonance imaging with aneurysmal subarachnoid hemorrhage: comparison with computed tomography scan. Surg Neurol 1990;34:71-8.
Magnetic resonance studies of 27 consecutive preoperative and 33 postoperative patients with cerebral aneurysm and subarachnoid hemorrhage were reviewed. Magnetic resonance imaging using a 0.5- or 0.22-Tesla unit was at least as accurate as computed tomography scan for detection of acute subarachnoid hemorrhage. Magnetic resonance imaging was superior to computed tomography scan for demonstrating blood in the ventricles or posterior fossa subarachnoid space, transependymal migration of the cerebrospinal fluid, and several kinds of iatrogenic intracranial pathologies in postoperative patients. Ischemic lesions, particularly fresh lesions caused by delayed cerebral vasospasm, were much better shown on magnetic resonance imaging than on computed tomography scan. Nonferromagnetic Sugita clips caused significant artifacts, but the area of artifacts was consistently smaller, and a reasonable evaluation of structures relatively closer to the clip was possible with magnetic resonance imaging rather than computed tomography scan. KEYWORDS: Cerebral aneurysm; Computed tomography; Intracranial hemorrhage; Magnetic resonance; Subarachnoid hemorrhage
Introduction Magnetic resonance imaging (MRI) has proven to be a noninvasive, highly sensitive, and accurate diagnostic method for evaluating a wide variety of intracranial diseases. In the evaluation of patients with suspected cerebrovascular diseases, noninvasive demonstration of blood flow without the use of iodinated contrast medium and
Address reprint requests to:Jyoji Handa, M.D., Department of Neurosurgery, Shiga University of Medical Science, Seta, Ohtsu, 520-21 Shiga-ken, Japan. Received December 27, 1989; accepted February 26, 1990. © 1990 by ElsevierSciencePublishingCo., Inc.
superb contrast sensitivity to abnormal signals associated with cerebral ischemia, as well as to previous hemorrhage, are apparent advantages of MRI over computed tomography (CT) scan or cerebral angiography. The unprecedented diagnostic value of MRI for cerebral vascular malformation, cerebral infarction, or intracerebral hemorrhage has been well established. On the other hand, little has been reported on the role of MRI for diagnosing aneurysmal subarachnoid hemorrhage (SAH). This is partly due to the fact that MRI is extremely sensitive to motion artifacts, and absolute immobilization of acutely ill patients is difficult to obtain. In addition, patients suffering from acute SAH are often connected to life-supporting equipment and monitoring devices, which could contain ferromagnetic materials and prohibit performance of MRI. The purpose of this paper is to report our initial experiences with MRI in patients with aneurysmal SAH, and compare the results with those of CT scan. Patients and Methods Magnetic resonance imaging was performed in 27 preoperative patients with aneurysmal SAH (group 1), and in 33 aneurysm patients after direct operation (group 2). Three aneurysms in group 1 were large or giant aneurysms (17, 21, and 25 mm in maximal diameter), and five patients with unruptured aneurysms were included in group 2. Of 27 patients in group 1 (aneurysmal SAH), 10 patients were male and 17 were female, and their ages ranged from 34 to 76, with an average of 58 years. Twenty patients were examined by MRI in the acute stage; namely 3-38 hours (mean = 21 hours) after SAH. Five patients were studied in the subacute stage, on day 5 through day 13 post-SAH (mean = day 7, the day of bleeding being day 0), and the remaining two patients were examined in the chronic stage, on day 21 and day 34, respectively. Neurological grades (World Federation of Neurological Surgery, WFNS) of patients at the time of MRI are shown in Table 1. 0090-3019/90/$3.50
72
Surg Neurol 1990;34:71-8
Matsumura et al
Table 1. Clinical Summary of 27 Patients with Aneurysmal Subarachnoid Hemorrhage (Group i) WFNS grade at MRI Stage Acute (day 0a--3) Subacute (day 4 - 1 4 ) Chronic (>day 15) Total
No. of patients
I
II
III
IV
V
20
11
5
2
2
0
5
2
1
2
0
0
2
2
0
0
0
0
27
15
6
4
2
0
Interval between onset and MR1 3 - 4 8 hours (mean = 21 hours) 5 - 1 3 days (mean = 7 days) 21 days and 34 days
Abbreviations: MR1, magnetic resonance imaging; WFNS, World Federation of Neurological Surgery. " Day 0 = day of hemorrhage. Three large or giant aneurysms are included, one in the acute, one in the subacute, and the remaining one in the chronic stage. Ten patients were male and 17 patients were female.
O f the 33 postoperative patients in group 2, 14 patients were male and 19 were female, and their ages ranged from 24 to 72, with an average of 54 years. In 31 patients the aneurysm was occluded with a nonferromagnetic Sugita clip. All these patients were operated upon at our institution, where all previous types of aneurysm clips had been replaced with the nonferromagnetic clips. N o patient in whom an aneurysm clip was placed at an outside institution was accepted for MRI, since we could not be absolutely sure o f the clip materials. The aneurysm was wrapped with a sheet o f cottonoid and reinforced with fibrin glue in the remaining two patients. Intervals between operation and MRI are shown in Table 2. Magnetic resonance imaging was performed using a superconducting Picker Vista MR unit operating at 0.5 Tesla, with the exception of 10 cases in whom the Toshiba MRT-22A (0.22 Tesla) was employed. Spin-echo pulse sequences were used, and T2-weighted images Table 2. ClinicalSummary of 33 PostoperativePatients (Group 2) No. of patients Gender
Ruptured or unruptured
Male Female
14 19
Ruptured
28 5
Unruptured
Operation
Neck clipping~ Wrapping
31 2
Interval between operation and MRI
--<1 week 1 week-lmonth 1 month-3 months 3 m o n t h s - 1 year > 1 year
6 11 9
Abbreviation: MRI, magnetic resonance imaging. Nonferromagnetic Sugita clip was used.
3 4
(T2WIs) with repetition times (TR) of 1 8 0 0 - 2 5 0 0 ms and echo times (TE) of 9 0 - 1 1 0 ms, Tl-weighted images (TIWIs) with T R of 5 0 0 - 7 0 0 ms and TE o f 2 5 - 3 0 ms, and proton density-weighted images (0WIs) with T R of 1800-2500 ms and TE of 30 ms were obtained in either the coronal, sagittal, or axial plane. Slice thickness was 10 mm in most cases, and images were displayed on 256 x 256 matrices. Computed tomography scan was performed using Siemens D R H , GE C T / T 8800, or YMS Quantex RX, and the slice thickness was either 5 or 10 mm. All o f our preoperative and postoperative patients underwent MRI and CT scan at least once, and none had just one of these two tests. Intervals between paired MRI and CT scan were less than 48 hours, and in most cases less than 24 hours, in the acute or subacute stage. Results Findings did not differ significantly between the two MRI units used, so that the results were analyzed together. The pertinent findings of MRI and CT scan in 27 consecutive patients with aneurysmal SAH (group 1) are compared in Table 3, and the representative images are illustrated in Figures 1-4. The T2WI was best suited for detection of blood in the basal cisterns and superficial subarachnoid spaces, showing a signal intensity higher than the cerebrospinal fluid (CSF) in 25 o f 27 patients, whereas CT scan confirmed SAH in 22 patients as the area of high density or isodensity (Figures 1 B, 1 C, and 2 A). Blood-fluid levels in the occipital horns, periventricular edema, cerebral infarction, and aneurysm per se were all better demonstrated on MRI than on CT scan (Figures 1 A, 1 B, 3, and 4). All infarcts found on MRIs were old ones, and no patient in the acute stage o f SAH showed abnormal signal intensities consistent with an area of edema or recent ischemia, possibly due to early vasospasm. N o n e of the postoperative patients (group 2) com-
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
Table 3. Comparisonof Magnetic Resonance Imaging and
Computed Tomography Scan in 27 Consecutive Patients with Aneurysmal Subarachnoid Hemorrhage No. of patients with positive findings Findings
Stage"
MRI
CT
Total no. of patients studied with both CT & MRI
Intraventricular blood-fluid level Blood in subarachnoid space Periventricular abnormal intensity/density Old infarction
Acute Subacute Chronic Acute Subacute Chronic Acute Subacute Chronic Acute Subacute Chronic
19 3 0 20 4 1 11 3 2 9 4 1 10
6 3 0 18 4 0 1 1 0 4 0 0 5
20 5 2 20 5 2 20 5 2 20 5 2 27
Identification of aneurysm
Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging. * Stage of patients after subarachnoid hemorrhage.
73
immediate vicinity of the valve, and it did not affect the diagnostic information of intracranial contents. Outside an area of artifacts, neither MRI nor CT scan showed any operative brain damage, such as contusional hemorrhage or focal edema due to retraction of the brain. The incidence of various postoperative pathologic changes seen on MRI and CT scan is compared in Table 4. Most of the extracerebral fluid collections were of minor clinical significance, such as a thin subdural fluid collection, minimal epidural blood collection, or hemorrhagic contusion of the temporal muscles, all without space-occupying effect (Figure 5 B, Table 4). Cerebral infarcts of various sizes and ages were much better seen on MRI than on CT scan. Both preoperative and postoperative MRIs were available for comparison in 11 of 33 patients in group 2 (Table 5, Figures 2 D and 5 B). In three of these 11 patients, no cerebral infarction was found on both preoperative and postoperative MRIs. In three patients, the number of infarcts remained unchanged, whereas in the remaining eight patients the number of foci of infarcts increased from 33 before operation to 52 after operation. Discussion
plained of any discomfort during MRI examinations, and no complication referable to metallic implant was encountered. Significant imaging artifacts were observed on both MRI and CT scan in all 31 postoperative patients in whom a metallic clip had been placed at operation (Figures 2 C and 5). Most patients had a single Sugita clip, but two patients had two and one had three clips. On CT scan, artifacts consisted of a center of complete attenuation of x-ray and radiating dense streaking artifacts, which were more pronounced along a long axis of the clip, and no diagnostic information was obtained in an area of such artifacts. On MRI, artifacts consisted of a central area of complete signal loss and a peripheral narrow rim of increased signal intensity. The size of the artifacts tended to be somewhat larger on T2WIs than on T1WIs. No diagnostic information was available in the region of artifacts. Close to the sharply demarcated rim of increased signal intensity, some distortion of the image was unavoidable, but it did not adversely affect diagnostic information to any great degree. An area of artifacts was consistently smaller on MRI than on CT scan (Figures 5 A and C, Table 4). In addition, a rapid transition to normal image quality outside an area of artifacts on MRI enabled a reasonable evaluation of structures relatively close to the clip. A Holter-Hausner shunt valve, that is also MRI compatible, caused metallic artifacts on both MRI and CT scan in four patients. The area of distortion of the MRI, however, was much smaller and restricted to the skin and the skull in the
Magnetic resonance imaging has been established as an excellent screening method for evaluation of a wide variety of diseases of the central nervous system, but its role in diagnosis of SAH has not been well documented. Several authors consider that acute and/or diffuse SAH is difficult to detect on MRI [2,5,8,29]. According to Zimmerman et al [29], higher oxygen tension of the CSF compared with that of the brain may lead to a low level of deoxyhemoglobin in the extravasated blood. In the absence of or with the very low level of deoxyhemoglobin, T2 shortening does not occur during an acute posthemorrhagic period. In addition, a pulse-shift effect of CSF pulsation may mask the effects of hemorrhage in the subarachnoid space and make detection of SAH difficult, particularly during the initial 3 days after hemorrhage [29]. Bradley et al [2] considered methemoglobin formation with T1 shortening as an important factor for increasing the signal intensity of the blood in the subarachnoid space over time, and they concluded that MRI might be useful in the diagnosis of SAH in the subacute stage but not in an acute stage. Measuring relaxation times of mixtures of normal human CSF and heparinized blood, Chakeres and Bryan [5] found a progressive shortening of T1 and T2 relaxation times with an increasing amount of blood. However, the concentrated subarachnoid blood in an acute stage of hemorrhage had relaxation times similar to the normal brain and was nearly isointense on MRIs. Partial volume effect of thin subarachnoid spaces also seemed to interfere with a definite delineation of acute SAH [5].
74
Surg Neurol 1990;34:71-8
Matsumura et al
Figure 1. (A,B) Ruptured aneurysm of the left internal carotid artery at the branching of the anterior choroidal artery. Magnetic resonance imaging on the day of S A H shows blood-fluid level (arrows) on o W l (A, T R / T E = 2100/30) and T 2 W I (B, T R / T E = 2100/90). Subarachnoid space shows intensities higher than the ventricular CSF on T2WI. (C) Computed tomography scan shows high densities in the cisterns but fails to show the intraventricular blood. (D) On follow-up M R I 220 days later, CSF in the subarachnoid space on T 2 W I is no longer higher in intensities than the ventricular CSF.
On the other hand, Jenkins et al [18] reported that they could identify the subarachnoid blood as often on MRI as on CT scan in 30 patients studied between 8 hours and 5 days after proven SAH. The blood-stained CSF showed a signal intensity higher than the brain and normal CSF on the axial T2-weighted spin-echo images [18]. Satoh and Kadoya [22] also found that in the acute stage, particularly within 24 hours after bleeding, the bloody CSF in the subarachnoid space showed high signal intensity relative to the brain on T2WI. Both Jenkins et al and Satoh and Kadoya used a Picker 0.15-Tesla resistive unit. Results of the present study are generally consonant with those of Jenkins et al [18] and Satoh and Kadoya [22], and substantiate the high sensitivity of MRI for detection of acute SAH (Table 3, Figure 1). Fine details of sulci and cisterns containing the bloody CSF, particularly those in the posterior fossa, were better shown on MRI than on CT scan. In addition, the blood in the
lateral ventricles was also seen more often on MRI than on CT scan. A high signal intensity of the blood-stained CSF on T2-weighted spin-echo images relative to the brain or the normal CSF is more or less difficult to explain, at least in the acute stage of SAH. In our study, T2 relaxation times of the bloody CSF in the lateral ventricles (144 + 9.7 ms) and subarachnoid cisterns (125 -+ 13.2 ms) were both shorter when compared with those of normal ventricular CSF (387 +- 68.8 ms). A high signal intensity seen on T2-weighted spin-echo images in the face of shortened T2 relaxation time is presumed to reflect a markedly shortened T1 relaxation time [18,22]. Even if this presumption is correct, the reason for such a marked shortening of T1 relaxation time in the bloody CSF in acute SAH remains to be elucidated. Production of methemoglobin is known to reduce T1 relaxation times [ 12]. During lysis of extravasated eryth-
r
w
Figure 2. (A,B) Ruptured aneurysm of the anterior communicating artery. Preoperative T2WIs (TR/TE = 2100/90) on day 1 post-SAH shows high intensities in the basal cisterns (arrowheads) (compare with the postoperative image in Figure 2 C). (C,D) Postoperative M R I shows an artifact due to the clip, but the structures in the immediate vicinity of the clip, such as the medial temporal lobe and the midbrain, are clearly seen. New patchy cerebral lesions are also seen on T 2 W I ( T R / T E = 2100/90).
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
75
2
Figure 3. (A-C) Ruptured aneurysm of the horizontal portion of the right anterior cerebral artery. The MRIs on the day of SAH show ventricular dilatation and periventricular hyperintensities (arrows) (A, TR/TE = 2100/30; B, TR/TE = 2100/60; C, T R / T E = 2100/90). The patient later developed symptomatic hydrocephalus. (D) Subarachnoid hemorrhage and the blood coating the choroid plexus were apparent also on CT scan; however, C T scan did not demonstrate periventricular abnormal densities.
rocytes, which begins within hours after SAH, oxyhemoglobin and smaller amounts of deoxyhemoglobin, iron, heme pigments, and methemoglobin are released into the CSF. Hayman et al [14] however, found that the highest CSF methemoglobin level reported so far was negligible at clays 0 to 7 after SAH, and they raised some doubts on the hypothesis that a marked T1 shortening was caused in the acute stage of SAH by the extremely low methemoglobin. As has been reported by various authors [9,23,25,26], MRI was more sensitive than CT scan for detecting cerebral infarctions both in the preoperative and postoperative studies (Tables 3 and 5). All the infarcts seen on MRIs taken in the acute or subacute stage were old lesions, and not one was
thought to have occurred in direct relation with the aneurysmal rupture. In the postoperative patients, however, the number of infarcts had considerably increased compared with the preoperative MRIs (Table 5). In five of 11 patients, the total number of ischemic foci increased from 33 before operation to 52 after operation. All five of these patients developed an angiographic vasospasm during the interval between the preoperative and postoperative MRI studies, and three of them developed clinical signs of focal cerebral ischemia (symptomatic vasospasm). New ischemic lesions on the postoperative MRIs were found in locations distant from the operative field, such as the corona radiata, centrum semiovale, or basal ganglia, and often bilaterally. We, therefore, assume that most
Figure 4. (A) Sagittal (TR/TE = 500/25) and (B,C)axial (B, TR/TE = I900/25; C, TR/TE = 1900/1 I0) MRIs on day 13 post-SAH, showing an aneurysm as an area of flow void (arrow) surrounded by circumscribed clot (arrowheads). Computed tomography scan without and with contrast enhancement fails to demonstrate aneurysm or subarachnoid clot. (D) Angiogram shows an aneurysm of the distal portion of the posterior inferior cerebellar artery (arrow).
76
Surg Neurol 1990;34:71-8
Matsumura et al
F i g u r e 5. Postoperative studies. Ruptured aneurysm of the left middle cerebral artery and unruptured aneurysm of the right middle cerebral artery were occluded using Sugita clips. An area of artifacts due to the clip is much smaller on MRI (A, T R / T E = 2100/100) than on C T scan (C). A thin layer of epidural clots and a spotty high-intensity lesion are evident on T 2 W I (B), but both of these intracranial pathologies are not apparent on C T scan (D).
of the new ischemic lesions can be ascribed to delayed vasospasm. Recently, patchy, deep cerebral lesions seen on MRIs have been described in patients with hydrocephalus, dementia, multiple sclerosis, or hypertension, and nonspecific white matter lesions also have been reported in clinically healthy elderly subjects [13,20,28]. Large Virchow-Robin spaces also have been reported to mimic lacunar infarcts [ 15]. Occasionally, it is difficult to differentiate lacunar infarcts from a large perivascular space. Lacunar infarcts are known to occur in the upper two thirds of the basal ganglia, and usually not of CSF intensity on all pulse sequences. By definition, a large Virchow-Robin space has to be isointense relative to the CSF on all pulse sequences and must conform to the path of the penetrating arteries. Such lesions must be differentiated from lacunar infarction. In this study, MRI was more sensitive than CT scan with and without contrast enhancement in the detection of aneurysms per se. The diameter of the aneurysm depicted on MRIs was 8 mm or larger. On CT scan, high-density clots surrounding the aneurysm tend to obscure the aneurysm in the acute stage of SAH, whereas rapidly flowing blood within the aneurysm usually produces a signal-void area on routine spin-echo MRIs and sharply contrasts with the high signal intensity of the bloody CSF or clot on T2-weighted spin-echo images (Figure 4) [1,18,22]. However, an aneurysmal sac may not always appear on MRIs as an area of signal void because of several factors, such as slice thickness, partial volume averaging effect, dimension of the aneurysmal sac, and others. In addition, a tortuous artery, CSF flow void, or an anatomic variant such as an asymmetrically pneumatized anterior clinoid process [1] may well mimic aneurysmal "signal void," and MRI diagnosis of an aneurysm has to be made with care. The use of ad-
junctive flow-sensitive imaging sequences such as gradient-recalled acquisition may be useful to confirm the suspicious area as an aneurysm [ 11,24,27]. T2-weighted spin-echo images demonstrated a periventricular zone of high signal intensity in 16 or 27 preoperative patients, whereas CT scan showed periventricular lucency in only two patients (Table 3). An area of periventricular hyperintensity on T2-weighted MRI has been observed in patients with cerebral infarction, white matter diseases such as multiple sclerosis, and hydrocephalus. A mild periventricular hyperintensity is also seen in elderly subjects with no evidence of intra-
Table 4. Comparison of Magnetic Resonance Imaging and Computed Tomography Findings in 33 Patients Who Underwent Direct Operation For Cerebral Aneurysm a No. of patients with positive findings Findings
MRI
CT
Infarction Subgaleal fluid collection Epidural hemorrhagic collection Subdural fluid collection Periventricular high intensity/ lucency Incidental tumor Clip artifact Size of artifact ~
26 3
14 3
13
6
13
13
10
7
1 31 3.1 (2.0-5.2)
1 31 11.3 (3.3-16.0)
Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging. Clipping of aneurysm with a Sugita clip was performed in 31 patients. Size of artifact in millimeters.
MRI with Aneurysmal SAH
Surg Neurol 1990;34:71-8
Table 5. Comparison of Location and Number of Cerebral Infarcts in 11 Patients Studied both Preoperatively and Postoperatively by Magnetic Resonance Tomography No. of foci of infarct Location of infarction
Preoperative MRI
Postoperative MRI
Internal capsule Corona radiata/ centrum semiovale Periventricular white matter Basal ganglia Cerebral cortex Pons Total
2 25
2 38
5
5
1 0 0 33
4 2 1 52
Abbreviation: MRI, magnetic resonance imaging. Preoperative examination was performed on day 0 (day of hemorrhage) through day 5, and postoperative study was carried out on day 21 through day 70.
cranial pathology [3,6,13,16,20,28]. Increased T2 signal intensity in the periventricular region is therefore not specific to one or another pathologic change: however, a zone of periventricular high intensity signal on T2WI in the present series was narrow (up to 2 mm) in width, smooth in contour, most prominent in the area of the anterior horn and trigone of the lateral ventricles, and bilaterally symmetrical. In addition, six of 14 patients who showed such a periventricular hyperintensity in the acute stage of SAH developed frank hydrocephalus days or weeks later. Therefore, we are of the opinion that in many patients, periventricular hyperintensities seen in the acute stage of SAH represent a transependymal migration of the CSF. In the postoperative cases, potential risks to patients with metallic surgical implants have to be seriously considered. Several authors carefully studied the martensitic content, magnetic field-induced displacement, heating effects, and other hazards of various kinds of aneurysm clips [7,10,19,21]. Clips inevitably caused significant artifacts on both CT scan and MRI, and interpretation was impossible in an area of artifacts. However, MRI proved to be far superior to CT scan in studying postoperative patients in that a size of artifacts was much smaller and a rapid change to the normal image quality immediately outside the central area of artifact enabled the evaluation of anatomic structures [4,17]. In conclusion, the present study confirms the important role of MRI in the management of patients with aneurysmal SAH.
References 1. Biondi A, Scilfa G, Scotti G. Intracranial aneurysms: MR imaging. Neuroradiology 1988;30:214-8.
77
2. Bradley WG Jr, Schmidt PG. Effect ofmethemoglobin formation on the MR appearance of subarachnoid hemorrhage. Radiology 1985;156:99-193. 3. Brant-Zawadzki M, Fein G, van Dyke C, Kiernan R, Davenport L, de Groot J. MR imaging of the aging brain: patchy whitematter lesions and dementia. AJNR 1985;6:675-82. 4. Brothers MF, Fox AJ, Lee DH, Pelz DM, DeveikisJP. MR imaging after surgery for vertebrobasilar aneurysm. AJNR 1990;11:149-61. 5. Chakeres DW, Bryan RN. Acute subarachnoid hemorrhage: in vitro comparison of magnetic resonance and computed tomography. AJNR 1986;7:223-8. 6. Chimowitz MI, Awad IA, Furlan AJ. Periventricular lesion on MRI. Facts and theories. Stroke 1989;20:963-7. 7. Davis PL, Crooks L, Arakawa M, McRee R, Kaufman L, Margulis AR. Potential hazards in NMR imaging: heating effects of changing magnetic fields and RF fields on small metallic implants. AJR 1981;137:857-60. 8. DeLaPaz RL, New PFJ, Buonanno FS, Kistler JP, Oot RF, Rosen BR, Taveras JM, Brady TJ. NMR imaging of intracranial hemorrhage. J Comput Assist Tomogr 1984;8:599-607. 9. Dewitt LD, Buonanno FS, Kistler JP, Brady TJ, Pykett IL, Goldman MR, Davis KR. Nuclear magnetic resonance imaging in evaluation of clinical stroke syndromes. Ann Neurol 1984;16:535-45. 10. Dujovny M, Kossovsky N, Kossowsky R, Valdivia R, Suk JS, Diaz FG, Berman SK, Cleary W. Aneurysm clip motion during magnetic resonance imaging: in vivo experimental study with metallurgical factor analysis. Neurosurgery 1985;17:543-8. 11. Gean AD, Pile-Spellman J, Heros RC. A pneumatized anterior clinoid mimicking an aneurysm on MR imaging--report of two cases. J Neurosurg 1989;71:128-32. 12. GomoriJM, Grossman RI, Goldberg HI, Zimmerman RA, Bilaniuk LT. Intracranial hematomas: imaging by high-field MR: Radiology 1985;157:87-93. 13. Gupta SR, Naheedy MH, Young JL, Ghobrial M, Rubino FA, Hindo W. Periventricular white matter changes and dementia. Clinical, neuropsychological, radiological, and pathological correlation. Arch Neurol 1988;45:637-41. 14. Hayman LA, PaganiJJ, KirkpatrickJB, Hinck VC. Pathophysiology of acute intracerebral and subarachnoid hemorrhage: applications to MR imaging. AJNR 1989;10:457-61. 15. Heier LA, Bauer CJ, Schwartz L, Zimmerman RD, Morgello S, Deck MDF. Large Virchow-Robin spaces: MR-clinical correlation. AJNR 1989;10:929-36. 16. Hendrie HC, Farlow MR, Austrom MG, Edwards MK, Williams MA. Foci of increased T2 signal intensity on brain MR scans of healthy elderly subjects. AJNR 1989;10:703-7. 17. Holt,s S, Olsson M, Romner B, Larsson E-M, S~iveland H, Brandt L. Comparison of MR imaging and CT in patients with intracranial aneurysm clips. AJNR 1988;9:891-7. 18. Jenkins A, Hadley DM, Teasdale GM, Condon B, Macpherson P, Patterson J. Magnetic resonance imaging of acute subarachnoid hemorrhage. J Neurosurg 1988;68:731-6. 19. Kean DM, Worthington BS, Firth JL, Hawkins RC. The effects of magnetic resonance imaging on different types of microsurgical clips. J Neurol Neurosurg Psychiatry 1985;48:286-7. 20. KirkpatrickJB, Hayman LA. White-matter lesions in MR imaging of clinically healthy brains of elderly subjects: possible pathologic basis. Radiology 1987;162:509-11. 21. New PFJ, Rosen BR, Brady TJ, Buonanno FS, Kistler JP, Burt CT, Hinshaw S, Newhouse JH, Pohost GM. Taveras JM. Potential hazards and artifacts of ferromagnetic and nonferromagnetic surgical and dental materials and devices in nuclear magnetic resonance imaging. Radiology 1983;147:139-48.
78
Surg N e u r o l 1990;34:71-8
22. Satoh S, Kadoya S. Magnetic resonance imaging of subarachnoid hemorrhage. Neuroradiology 1988;30:361-6. 23. Sipponen JT, Kaste M, Ketonen L, Sepponen RE, Katevuo K, Sivula A. Serial nuclear magnetic resonance (NMR) imaging in patients with cerebral infarction. J Comput Assist Tomogr 1983;7:585-9. 24. TsurudaJS, Halbach VV, Higashida RT, Mark AS, Hieshima GB, Norman D. MR evaluation of large intracranial aneurysm using cine low flip angle gradient-refocused imaging. AJNR 1988;9:415-24. 25. Unger EC, Gado MH, Fulling KF, Littlefield JL. Acute cerebral infarction in monkeys: an experimental study using MR imaging. Radiology 1987;162:789-95.
Matsumura et al
26. Virapongse C, Mancuso A, Quisling R. Human brain infarcts: Gd-DTPA-enhanced MR imaging. Radiology 1986;161: 785-94. 27. Wagte WA, Dumoulin CL, Souza SP, Cline HE. 3DFT MR angiography of carotid and basilar arteries. AJNR 1989;10: 911-9. 28. Zimmerman RD, Fleming CA, Lee BCP, Saint-Louis LA, Deck MDF. Periventricular hyperintensity as seen by magnetic resonance: prevalence and significance. AJNR 1986;7: 13-20. 29. Zimmerman RD, Heier LA, Snow RB, Liu DPC, Kelly AB, Deck MDF. Acute intracranial hemorrhage: intensity changes on sequential MR scans at 0.5 T. AJNR 1988;9:47-57.