Hemosiderin Detected by T2*-Weighted Magnetic Resonance Imaging in Patients with Unruptured Cerebral Aneurysms: Indication of Previous Bleeding?

Neurol Med Chir (Tokyo) 51, 275¿281, 2011

Hemosiderin Detected by T2*-Weighted Magnetic Resonance Imaging in Patients With Unruptured Cerebral Aneurysms: Indication of Previous Bleeding?

Shihomi TAKADA,1,2 Takashi INOUE,2 Kuniyasu NIIZUMA,1 Hiroaki SHIMIZU,2 and Teiji TOMINAGA1

1Department of Neurosurgery, Tohoku University Graduate School of Medicine, Sendai, Miyagi; 2Department of Neurosurgery, Kohnan Hospital, Sendai, Miyagi

Abstract Previous bleeding from a cerebral aneurysm indicates a higher risk of rupture. Hemosiderin may be detected during aneurysm surgery or by preoperative imaging sensitive to hemosiderin. The detection of

hemosiderin deposits by T2*-weighted magnetic resonance (MR) imaging was evaluated in 49 patients with unruptured cerebral aneurysms who underwent open surgery. MR imaging was performed using

3.0 tesla MR scanner. Two sequences of T2*-weighted imaging, and proton density images were obtained. Preliminary study in patients with old subarachnoid hemorrhage provided the definitions of

likely pathological findings during surgery and on T2*-weighted imaging due to previous hemorrhage. Hemosiderin deposits in the subarachnoid space were observed during surgery in 9 of the 49 patients, although no obvious rupture site was detected around the aneurysm wall. Size, presence of bleb, location, and number of aneurysms showed no significant difference between patients with and without

hemosiderin deposition. Hypointense areas on T2*-weighted imaging were recognized in four patients. The mean size of the aneurysms in these patients was 9.8 mm, significantly larger than those in other patients (p ＝ 0.029). Hemosiderin deposits were observed during surgery in sites close to the lesions on

T2*-weighted imaging in two of these four patients. Hemosiderin deposits are not rare in patients with unruptured aneurysms, and preoperative T2*-weighted imaging can detect such deposits.

Key words: 3 tesla, hemosiderin, T2*-weighted magnetic resonance imaging, unruptured aneurysm

Introduction minor leakage, with a mortality rate of 53%.9) The bleeding rate later than 4 weeks after such hemor- The cumulative rupture rate of unruptured cerebral rhage may still be higher than with truly unruptured aneurysmsinpatientswithnohistoryof aneurysms. Therefore, a method to detect indica- subarachnoid hemorrhage (SAH) is quite low, tions of previous bleeding may have a significant whereas the rupture rate of unruptured aneurysms clinical impact on decision making for the treatment is approximately 11 times higher in patients with a of cerebral aneurysms. history of rupture in other aneurysms.13) Minor leak- Surgery on patients with SAH at several weeks or age prior to major aneurysm rupture has been de- monthsaftertheictususuallyrevealsfibrinnetsin tected in 39% of patients with SAH, but unless SAH the major cisterns and spotty hemosiderin deposits is present the diagnosis may not be established.9) in the subarachnoid space, in amounts depending Patients with so-called unruptured aneurysms may on the initial amount of bleeding and time elapsed also have undetected minor leaks. Although the rup- before treatment. We have seen similar hemosiderin ture risk is not the same as just before minor leaks, deposits during surgery for asymptomatic ``unrup- patients with past minor leaks may have higher risk tured'' aneurysms. The clinical features of such of SAH than with truly unruptured aneurysms. hemosiderin deposits remain unclear. T2*-weighted Major rupture is reported to occur in 74% of patients magnetic resonance (MR) imaging is a promising within 24 hours to 4 weeks after findings suggesting method to detect intracerebral hemosiderin.2,7,8,14) Hemosiderin can be detected as a hypointense area Received October 21, 2010; Accepted January 6, 18 months after the onset of SAH.5) 2011 The present study investigated the correlation be-

275 276 S. Takada et al.

tween detection of hemosiderin by T2*-weighted MR Subjects and Methods imaging and the observation of hemosiderin deposits during surgery in patients who underwent Preliminary studies confirmed the intraoperative clipping of unruptured aneurysms. findings of hemosiderin deposition histologically in three patients who underwent surgery for aneurysm

Table 1 Patients with asymptomatic cerebral aneurysms

Aneurysms Presumed pathological hypointensity Case Age Sex Hemosiderin deposits No. (yrs) GRE T2*- EP T2*- in operative findings Number Size (mm) Location Bleb weighted image weighted image

147F 1 4 MCA ＋－－＋ 249F 1 4 MCA －－－－ 349F 1 5 MCA －－－－ 453F 1 3.8 MCA ＋－－－ 558F 1 5 MCA ＋－－＋ 663F 1 5 MCA －－－－ 764F 1 4 MCA ＋－－－ 868F 1 4,9 MCA ＋－－－ 968F 1 4 MCA －－－－ 10 71 F 1 9 MCA －－－－ 11 51 M 1 5 MCA －－－＋ 12 51 M 1 6 MCA －－－－ 13 55 M 1 6 MCA ＋－－－ 14 58 M 1 4 MCA －－－－ 15 58 M 1 8 MCA －－－－ 16 59 M 1 10 MCA －－－＋ 17 63 M 1 3 MCA －－－－ 18 68 M 1 4 MCA －－－－ 19 58 F 2 9, 2 MCA, MCA －－－－ 20 65 F 2 10, 9 MCA, MCA ＋－－－ 21 63 M 3 19.1, 3.1, 4.9 MCA, ICA, ACA ＋－－－ 22 56 F 2 12, 3 MCA, ACA －－－－ 23 65 F 2 5.5, 2.8 MCA, ACA －－－－ 24 61 F 3 4.6, 3.3, 3 MCA, ICA, BA －－－＋ 25 58 F 1 9 AcomA －－－－ 26 45 M 1 18.4 AcomA －－－－ 27 48 M 1 3 AcomA －－－－ 28 52 M 1 9.8 AcomA －－－＋ 29 53 M 1 6.7 AcomA －＋＋＋ 30 57 M 1 7 AcomA ＋－－－ 31 58 M 1 5.1 AcomA －－－－ 32 58 M 1 7.6 AcomA ＋＋＋－ 33 62 M 1 5 AcomA －－－－ 34 65 M 1 3.3 AcomA －－－－ 35 68 M 1 7.2 AcomA ＋－－－ 36 68 M 1 5 AcomA ＋－－－ 37 61 M 2 5, 2 AcomA, ICA ＋－－＋ 38 52 F 1 8.2 ICA －－－－ 39 55 F 1 5.6 ICA ＋－－－ 40 56 F 1 14.8 ICA －－＋＋ 41 56 F 1 4.9 ICA －－－－ 42 59 F 1 7.9 ICA －－－－ 43 64 F 1 9 ICA －－－－ 44 69 F 1 13.6 ICA －－－－ 45 69 F 1 8.4 ICA ＋－－－ 46 70 F 1 4.3 ICA －－－－ 47 71 F 1 7 ICA ＋－－－ 48 59 M 1 19 ICA －－－－ 49 72 F 1 10 ACA －－＋－

ACA: anterior cerebral artery, AcomA: anterior communicating artery, BA: basilar artery, EP: echo planar, F: female, GRE: gradient recalled echo, ICA: internal carotid artery, M: male, MCA: middle cerebral artery, ＋: positive, －:negative.

Neurol Med Chir (Tokyo) 51,April,2011 Hemosiderin on T2*-Weighted Imaging 277 clipping in the chronic stage (À11 months) after SAH. Previous studies of hemosiderin deposition in patients with SAH employed the gradient recalled 1,5,6) echo (GRE) T2*-weighted imaging sequence. The present study employed the gradient echo type echo planar (EP) imaging sequence in addition to the GRE sequence for T2*-weighted imaging, because we ex- pected a stronger T2* effect with the EP imaging sequence. Introduction of this new sequence needed preliminary evaluation in the normal brain and patients with known SAH. Two patients with fresh SAH were followed up using proton density (PD),

GRE T2*-weighted, and EP T2*-weighted imaging until 6 months after the onset. The present study included 49 non-consecutive patients (27 females and 22 males) aged 45–72 years (mean 60 years) with incidentally detected unruptured intracranial aneurysms without obvious past head trauma who underwent open surgery. Conven- tional MR imaging showed no obvious hemorrhagic lesion. Cerebral digital subtraction angiography or MR angiography was performed in all patients. Twenty-six aneurysms were located on the middle cerebral artery, 13 on the anterior communicating artery, 14 on the internal carotid artery, four on the anterior cerebral artery, and one on the basilar artery. Seven patients had multiple aneurysms. The clinical characteristics are shown in Table 1. The study protocol was approved by the local ethical committee. All patients gave written informed con- sent prior to the study. MR imaging was performed using a whole-body 3.0 tesla MR scanner (General Electric Medical Sys- tems,Milwaukee,Wisconsin,USA)andaparallel imaging head coil. All patients underwent GRE and

EP T2*-weighted imaging, and conventional spin echo PD imaging. Parameters of the PD imaging were repetition time (TR) 2800 msec, echo time (TE) 12.4 msec, matrix 512 × 320, field of view (FOV) 220 Fig. 1 Representative images showing hemosiderin deposits due to subarachnoid hemorrhage in a 65-year- mm, 6 mm slice thickness, and 2 mm slice interval. old female who underwent open surgery for unruptured Parameters for GRE T*-weighted imaging were TR 2 left internal carotid artery aneurysm 13 months after 500 msec, TE 10 msec, flip angle 309, matrix 320 × subarachnoid hemorrhage due to rupture of the basilar 192, FOV 220 mm, 6 mm slice thickness, and 2 mm tip aneurysm which was treated with coils. A: In- slice interval. Parameters for EP T2*-weighted imag- traoperative photograph through the left transsylvian ing were TR 800 msec, TE 17.5 msec, flip angle 309, approach showing a cluster of hemosiderin deposits (ar- matrix 256 × 128, FOV 220 mm, 6 mm slice thick- row) in the subarachnoid space on the insular cortex. ness, and 2 mm slice interval. B: Intraoperative photograph showing a similar but ThepresenceofhemosiderindepositsontheEP thicker cluster of hemosiderin deposits (arrow)onthe arachnoid membrane which was removed and examined T2*-weighted images was evaluated by three coauthors (ST, TI, HS) independently as described for pathological features. C: Photomicrograph of the in the preliminary studies. If individual judgments specimen of B confirming abnormal hemosiderin deposits in the arachnoid tissue. Hematoxylin and eo- differed, the final decision was made by consensus. sin stain, original magnification ×400. Two of the co-authors (ST, HS) viewed independently all operative video tapes which provided continu- ous records of all procedures from dissection of the

Neurol Med Chir (Tokyo) 51,April,2011 278 S. Takada et al. arachnoid membrane over the brain surface to com- pletion of clipping. Clusters of brown small spots on the brain surface or arachnoid membrane as shown in Fig. 1 were identified as hemosiderin deposits. Differences in characteristics listed in Table 1 between patients with and without hemosiderin deposition (intraoperative or on EP T2*-weighted images) were compared by the Mann-Whitney exact test. A probability of less than 5% was considered to indicate statistical significance.

Results

Thepreliminarystudydetectedbothfibrinnetsand sporadic clusters of hemosiderin deposits in the subarachnoid space and on the brain surface. Representative images from a patient who underwent open surgery 13 months after onset of SAH are showninFig.1.Anareaofhemosiderindeposits was observed over the insular cortex during surgery (Fig. 1A). A cluster of hemosiderin deposits in the arachnoid membrane was removed (Fig. 1B). Histo- logical examination of the specimen demonstrated hemosiderin deposition in the arachnoid tissue (Fig. 1C). Intraoperative observations of hemosiderin deposits were similar in the other two patients, and at least one part of the hemosiderin deposition was confirmed histologically. Therefore, we assumed that intraoperative findings similar to Fig. 1A or 1B were hemosiderin deposition without histological confirmation in the following patients with asymptomatic aneurysms.

Figure 2 shows the normal PD, GRE T2*-weighted, Fig. 2 Representative proton density (PD) (A), gradient recalled echo (GRE) T*-weighted (B), and gradient echo and EP T2*-weighted images. Normal subarachnoid 2 * space and sulci are depicted in principle as hyperin- type echo planar (EP) T2 -weighted images (C) of normal tense areas by EP T*-weighted imaging. Vascular subarachnoid space and sulci. Susceptibility artifacts 2 increase in the PD, GRE T*-weighted, and EP T*- structures can be distinguished by comparing the 2 2 weighted images in this order, and areas with the ar- PD, GRE T*-weighted, and EP T*-weighted images. 2 2 tifacts were not evaluated further (white arrows). Nor- A further preliminary study considered patients mal subarachnoid space and sulci are depicted as with known SAH. Two patients with fresh SAH hyperintense areas due to cerebrospinal fluid (dotted ar- were followed up using PD, GRE T2*-weighted, and rows). Cerebral arteries and large veins are best delineat- EP T2*-weighted imaging until 6 months after the on- ed on PD images as flow voids and show some widening set. Images from one of the patients are presented in on EP T2*-weighted images due to susceptibility artifacts Fig. 3 and provide the basis for the present defini- and lower spatial resolution (arrows). tion of hemosiderin deposition on EP T2*-weighted images. Hypointense areas on both GRE and EP T2*- tinguished based on computed tomography (CT) weighted images can be immediately assumed to scans obtained at the same time. represent hemosiderin deposition. In addition, com- Hemosiderin deposits were observed in the parison of early (Fig. 3B) and late (Fig. 3C) EP T2*- subarachnoid space during surgery in 9 of the 49 weighted images can identify hypointense areas on patients with unruptured aneurysms, although no

EP T2*-weighted images but not on GRE T2*-weighted obvious rupture site was detected around the images which can also be considered to show aneurysm wall. These nine patients included 5 of the hemosiderin deposition unless there are no respon- 24 patients with middle cerebral artery aneurysms, sible vascular components and artifacts. Hypoin- three of the 13 patients with anterior communicat- tense areas caused by calcification or air can be dis- ing artery aneurysms, three of the 14 patients with

Neurol Med Chir (Tokyo) 51,April,2011 Hemosiderin on T2*-Weighted Imaging 279

← Fig. 3 Representative images of old subarachnoid hemorrhage (SAH) or hemosiderin deposition in a patient who suffered from rupture of a left middle cerebral artery aneurysm which was clipped in the acute stage and recovered without neurological deficit, providing the basis for the present definition of hemosiderin deposits long after SAH. A: Computed tomography (CT) scans on the day of SAH onset. B:

Gradient echo type echo planar (EP) T2*-weighted images 1 month after the SAH onset. C–E: EP T2*-weighted images 6 months after the SAH onset (C) and corresponding

gradient recalled echo (GRE) T2*-weighted (D) and proton density (PD) images (E). Note clip artifact (white arrow) and slight slice differences between B and C.

Comparing EP T2*-weighted images at 6 month (C) and normal EP T2*-weighted images in Fig. 2C, hypointense areas within the subarachnoid spaces or sulci on EP T2*- weighted images other than vascular components are considered to be due to the hemorrhage. Hypointense

areas on all EP T2*-weighted, GRE T2*-weighted, and PD images (arrowheads) corresponded with the thick clot on initial CT scans and areas with moderate amount of clot

appeared hypointense on both EP and GRE T2*-weighted images but not on PD images (arrows). Remaining

hypointense areas that are clear on EP T2*-weighted images but not clear on GRE T2*-weighted and PD images are considered hemosiderin deposits (dotted arrows) only if no vascular components and artifacts are present. Note that these areas were more intense on EP

T2*-weighted images at 1 month (B), and small amounts of clot seen on CT scans at onset (A) indicate the presence of hemorrhage. Any obscure findings on EP

T2*-weighted images were not considered to show hemosiderin deposits. were outside the operative field and could not be confirmed during surgery in the other two patients.

Discussion internal carotid artery aneurysms, and none of the patients with anterior cerebral artery aneurysm. The present study observed hemosiderin deposits at There was no significant difference in aneurysm surgery in 9 of 49 patients with no obvious signs of characteristics, including size, presence of bleb, lo- past SAH. Such hemosiderin deposits are not neces- cation, and number of aneurysms, between patients sarily indicative of past SAH. Minor head trauma or with and without hemosiderin deposition. other causes of bleeding are other possible mechan-

Hemosiderin deposits were recognized on EP T2*- isms. However, clusters of iron-laden macrophages weighted images in four patients and on GRE T2*- canbeseeninthesubarachnoidspaceofpatients weighted images in two of these four patients. The withminorleakagepriortomajorruptureof mean size of these four aneurysms was 9.8 mm, sig- aneurysms,9) so some of the hemosiderin deposits nificantly larger than the aneurysms without observed in our patients are likely to be related to hemosiderin deposition (p ＝ 0.029). Other charac- previous minor leakage or rupture. No obvious rup- teristics of the aneurysms showed no significant ture site was detected around the aneurysms, which differences. Surgery on these four patients revealed may indicate long duration from previous leaks or hemosiderin deposits in two patients (Figs. 4 and 5) reconstruction of the aneurysm wall after rupture. close to the site indicated by the EP T2*-weighted im- Recently 3.0 T MR imaging has allowed high reso- aging findings were present, although stereotactic lution imaging in the clinical setting. One of the confirmation was not possible. The locations of the characteristic features is augmentation of the sus- hemosiderin deposits on EP T2*-weighted imaging ceptibility effects (shortening of T2* and T2), which

Neurol Med Chir (Tokyo) 51,April,2011 280 S. Takada et al.

Fig. 5 A 56-year-old female with a left internal carotid- posterior communicating artery aneurysm. A: In- traoperative photograph of hemosiderin deposits in the left sylvian fissure on the frontal operculum. B: Proton

density image. C: Gradient recalled echo T2*-weighted image. D: Gradient echo type echo planar T2*-weighted image. Hypointense area in D (arrow) corresponds to the intraoperative finding (circled area of A).

has both advantages and disadvantages. Disadvan- tages include signal intensity loss, susceptibility artifacts, and geometric distortion. Advantages include increased sensitivity to deoxyhemoglobin-in- duced contrast, which in turn improves the quality of functional MR imaging, and detection of hemor- Fig. 4 A 53-year-old male with an anterior com- rhage or hemosiderin.10) municating artery aneurysm treated through a right 1.5 T T2*-weighted imaging demonstrated old SAH pterional approach. A: Intraoperative photographs of more sensitively compared with other MR se- hemosiderin deposits in the right frontal operculum quences in patients several weeks or months after (left) and superior temporal gyrus (right). Note the tem- SAH.5) Acute SAH can also be reliably detected with porary clip on the left A and the aneurysm (arrows). B: 1 12) 3.0 T T2*-weighted imaging. The present study ap- Proton density images. C: Gradient recalled echo T2*- weighted images. D: Gradient echo type echo planar plied 3.0 T T2*-weighted imaging to patients with un- * (EP) T2*-weighted images. Hypointense areas (arrows)in ruptured aneurysms. We used EP T2 -weighted imag- the right sylvian fissure and the sulcus in the right fron- ing to maximize the susceptibility effect and tal lobe cannot be accounted for by vascular structures detectability of hemosiderin. Hypointense areas on or artifacts, indicating likely hemosiderin deposits. T2*-weighted imaging are not specific for hemor- Hypointense area (dotted arrow) is not distinguishable rhage. Other causes include calcification, physiolog- from sphenoidal bone artifacts and is not considered in- ical ferritin, melanoma, air, and some paramagnetic dicative of hemosiderin, although the circled area of A contrast agents.1) Therefore, we carefully excluded left may be included. Area with hemosiderin deposits causes other than hemosiderin by referring to the shown in circled area of A right was also obscure on EP CT scans, and conventional PD and GRE T*-weight- T*-weighted images due to skull base artifacts. 2 2 ed images. This study has several limitations. First, histologi-

Neurol Med Chir (Tokyo) 51,April,2011 Hemosiderin on T2*-Weighted Imaging 281 cal samples could not be obtained in all patients with 1.5 T. Comparison with spin-echo imaging and clini- hemosiderin deposits during surgery because most cal applications. Radiology 168: 803–807, 1988 deposits were on the brain surface, so histological 2) Gomori JM, Grossman RI, Hackney DB, Goldberg HI, Zimmerman RA, Bilaniuk LT: Variable appearances samples were difficult to take without damaging the of subacute intracranial hematomas on high-field brain. However, we consider that the similarities be- spin-echo MR. AJR Am J Roentgenol 150: 171–178, tween Fig. 1A and Figs. 4A, 4B, and 5A support our 1988 hypothesis that these findings are due to hemoside- 3) Haacke EM, Xu Y, Cheng YC, Reichenbach JR: Sus- rin deposits. Second, the findings on EP T*-weight- ceptibility weighted imaging (SWI). Magn Reson Med 2 52: 612–618, 2004 ed images of the four patients were not individually 4) Idbaih A, Boukobza M, Crassard I, Porcher R, Bouss- confirmed as hemosiderin deposits during surgery. er MG, Chabriat H: MRI of clot in cerebral venous Hemosiderin deposits did approximately cor- thrombosis: high diagnostic value of susceptibility- respond to the MR imaging findings in two patients, weighted images. Stroke 37: 991–995, 2006 but no stereotactic method such as a neuronaviga- 5) Imaizumi T, Chiba M, Honma T, Niwa J: Detection of hemosiderindepositionbyT2*-weighted MRI after tion system was employed to confirm the coinci- subarachnoid hemorrhage. Stroke 34: 1693–1698, dence. The apparent hemosiderin deposits were out- 2003 side the surgical field in the other two patients. 6) Imaizumi T, Horita Y, Chiba M, Hashimoto Y, Hon- Third, surgery demonstrated hemosiderin deposits ma T, Niwa J: Dot-like hemosiderin spots on gradient in nine patients, but seven showed no abnormal echo T2*-weighted magnetic resonance imaging are associated with past history of small vessel disease in * finding on EP T2 -weighted imaging. We can suggest patients with intracerebral hemorrhage. JNeuro- two possible explanations, other than the amount of imaging 14: 251–257, 2004 hemosiderin: low spatial resolution and small sus- 7) Kinoshita T, Okudera T, Tamura H, Ogawa T, Hatazawa J: Assessment of lacunar hemorrhage asso- ceptibility effect of the EP T2*-weighted imaging to detect tiny hemosiderin deposits. The findings of ciated with hypertensive stroke by echo-planar gradient-echo T2*-weighted MRI. Stroke 31: 1646– * T2 -weighted imaging in this study does not directly 1650, 2000 evidence previous rupture of aneurysm or 8) Kwa VI, Franke CL, Verbeeten B Jr, Stam J: Silent in- hemosiderin deposits. The relationship between tracerebral microhemorrhages in patients with ischemic stroke. Amsterdam Vascular Medicine aneurysm size and hemosiderin deposits or T2*- weighted imaging findings was different in this Group. Ann Neurol 44: 372–377, 1998 9) Leblanc R: The minor leak preceding subarachnoid * study, suggesting that T2 -weighted imaging findings hemorrhage. JNeurosurg66: 35–39, 1987 need to be interpreted with caution. 10) Lin W, An H, Chen Y, Nicholas P, Zhai G, Gerig G, Susceptibility-weighted imaging is a novel method Gilmore J, Bullitt E: Practical consideration for 3T to demonstrate differences in susceptibility.3) Sus- imaging. Magn Reson Imaging Clin N Am 11: 615– ceptibility-weighted imaging is a high spatial resolu- 639, vi, 2003 11) Sehgal V, Delproposto Z, Haddar D, Haacke EM, tion, three-dimensional, and fully velocity-com- SloanAE,ZamoranoLJ,BargerG,HuJ,XuY,Prab- pensated gradient-echo sequence that provides in- hakaran KP, Elangovan IR, Neelavalli J, Reichenbach formation about blood products, venous vascula- JR: Susceptibility-weighted imaging to visualize ture, and thrombus.3,4,11) Although not presently blood products and improve tumor contrast in the available, it may be worth investigating in a further study of brain masses. J Magn Reson Imaging 24: 41–51, 2006 study. 12) Sohn CH, Baik SK, Lee HJ, Lee SM, Kim IM, Yim The present study of hemosiderin deposition in MB, Hwang JS, Lauzon ML, Sevick RJ: MR imaging patients with unruptured aneurysms showed that of hyperacute subarachnoid and intraventricular hemosiderin deposits were not rare, and that hemorrhage at 3T: a preliminary report of gradient * preoperative EP T*-weighted imaging has the poten- echo T2 -weighted sequences. AJNR Am J Neuroradiol 2 26: 662–665, 2005 tial to detect such deposits. Although a relatively 13) Unruptured intracranial aneurysms—risk of rupture small number of cases was included in this study, and risks of surgical intervention. International previous rupture, hemosiderin deposits, and low in- Study of Unruptured Intracranial Aneurysms Inves- tigators. NEnglJMed339: 1725–1733, 1998 tensity signals on T2*-weighted imaging may be related. Further investigation with a larger number of 14) Wardlaw JM, Statham PF: How often is haemoside- rin not visible on routine MRI following traumatic cases and more sensitive imaging methods such as intracerebral haemorrhage? Neuroradiology 42: 81– susceptibility-weighted imaging are needed. 84, 2000 References Address reprint requests to: Takashi Inoue, MD, Depart- ment of Neurosurgery, Kohnan Hospital, 4–20–1 Nagamachi–minami, Taihaku–ku, Sendai, Miyagi 1) Atlas SW, Mark AS, Grossman RI, Gomori JM: In- 982–8523, Japan. tracranial hemorrhage: gradient-echo MR imaging at e-mail:tainoue＠kohnan-sendai.or.jp

Neurol Med Chir (Tokyo) 51,April,2011