J Neurosurg 106:263–269, 2007 Ischemic complications associated with resection of opercular glioma TOSHIHIRO KUMABE, M.D.,1 SHUICHI HIGANO, M.D.,2 SHOKI TAKAHASHI, M.D.,2 AND TEIJI TOMINAGA, M.D.1 Departments of 1Neurosurgery and 2Diagnostic Radiology, Tohoku University Graduate School of Medicine, Sendai, Japan Object. Opercular glioma inferolateral to the hand/digit sensorimotor area can be resected safely using a neuronavigation system and functional brain mapping techniques. However, the surgery can still sometimes cause postoperative ischemic complications, the character of which remains unclear. The authors of this study investigated the occurrence of infarction associated with resection of opercular glioma and the arterial supply to this region. Methods. The study involved 11 consecutive patients with gliomas located in the opercular region around the orofacial primary motor and somatosensory cortices but not involving either the hand/digit area or the insula, who had been treated in their department after 1997. Both pre- and postoperative diffusion-weighted magnetic resonance (MR) imaging was performed in the nine consecutive patients after 1998 to detect ischemic complications. All patients underwent open surgery for maximum tumor resection. Postoperative MR imaging identified infarction beneath the resection cavity in all patients. Permanent motor deficits associated with infarction involving the descending motor pathway developed in two patients. Cadaveric angiography showed that the distributing arteries to the corona radiata were the long insular arteries and/or medullary arteries from the opercular and cortical segments of the middle cerebral artery. Conclusions. Subcortical resection around the upper limiting sulcus of the posterior region of the insula and wide resection in the anteroposterior and cephalocaudal directions of the opercular region were considered to be risk factors of the critical infarction. Surgeons should be aware that resection of opercular glioma can disrupt the blood supply of the corona radiata, and carries the risk of permanent motor deficits. KEY WORDS • infarction • complication • glioma • descending motor pathway • operculum • diffusion-weighted imaging P RECISE localization of a glioma in the frontoparietal opercular region inferolateral to the hand/digit sensorimotor area is now possible using various methods including functional brain mapping techniques, neuronavigation systems, intraoperative MR imaging, and photodynamic diagnosis using various photosensitizers. Therefore, gliomas in this location, even in the dominant hemisphere, can be totally resected without causing permanent neurological deficits.5,9,11,12 During such procedures, surgical techniques for opercular glioma have concentrated on the identification and preservation of the cortical and subcortical functions.5,11,12 However, little is known about the ischemic complications that can occur after the resection of an opercular glioma. Diffusion-weighted MR imaging, which reflects the degree of water diffusion in vivo, is an invaluable tool for the diagnosis of acute stroke and other types of brain injury.14 Abbreviations used in this paper: DW = diffusion-weighted; GBM = glioblastoma multiforme; MCA = middle cerebral artery; MR = magnetic resonance. J. Neurosurg. / Volume 106 / February, 2007 Several potential applications of DW MR imaging in patients with gliomas have been recently investigated, mainly for the evaluation of tumor cellularity.7,8,16,21 Recently, postoperative DW MR imaging has been proposed as a routine study to identify ischemic complications after resection of the glioma.15 Diffusion-weighted MR imaging detected abnormalities after resection in approximately two thirds of newly diagnosed gliomas. At our institution, postoperative MR imaging including DW imaging has been performed for nearly 10 years as one of the examinations used to determine the postoperative state of patients after tumor removal and has disclosed evidence of postoperative ischemic complications. In the present study we investigated the postoperative ischemic complications and DW MR imaging findings in 11 patients with pure opercular gliomas surgically treated during the past 9 years. Microangiography studies of cadavers were also analyzed to identify the blood supply for the corona radiata. Finally, referring to the results of the microangiographic analysis, we tried to determine the risk factors for critical infarction at the corona radiata that were likely to result in permanent motor deficits. 263 T. Kumabe et al. TABLE 1 Summary of characteristics in 11 cases of opercular glioma* Case No. Age (yrs), Sex Diagnosis WHO Grade Tumor Location Dominant Hemisphere Presenting Symptoms Date of Surgery State of Anesthesia 1 26, M A II rt face motor area no 1/23/97 awake 2 31, M AA III yes 9/11/97 awake 3 68, M AA III no lt facial seizure 1/27/98 awake 4 3, M DNET I lt face/tongue motor area rt tongue sensorimotor area rt face motor area lt facial seizure followed by generalized convulsion generalized convulsion no lt facial seizure 12/16/99 general 5 38, M AA III lt face motor area yes lt facial seizure 2/28/00 general 6 69, F GBM IV no general 49, M OA II lt facial seizure followed by lt hemiconvulsion incidental 1/24/02 7 2/5/02 general 8 21, M A II 9 38, M AA III rt face/tongue sensorimotor area rt tongue sensorimotor area lt tongue sensorimotor area rt face motor area 10 34, F O II 11 23, M AA III lt face/tongue motor area lt face motor area no no 4/11/02 general no rt facial seizure followed by generalized convulsion lt facial seizure 7/22/02 general yes rt facial seizure 4/28/03 awake yes loss of consciousness 6/9/05 awake * A = astrocytoma; AA = anaplastic astrocytoma; DCS = direct cortical stimulation; DNET = dysembryoplastic neuroepithelial tumor; KPS = Karnofsky Performance Scale; O = oligodendroglioma; OA = oligoastrocytoma; SCS = subcortical stimulation; US = ultrasonography; WHO = World Health Organization. Clinical Material and Methods Patient Population and Tumor Characteristics This study included 11 consecutive patients, nine males and two females with ages from 3 to 69 years (mean 36.4 6 19.8 years). Each patient harbored a newly diagnosed glioma located at the opercular region around the orofacial primary motor and somatosensory cortices but not involving either the hand/digit area or the insula and treated in our department after 1997. The tumors included one dysembryoplastic neuroepithelial tumor, two astrocytomas, one oligoastrocytoma, one oligodendroglioma, five anaplastic astrocytomas, and one GBM. Patient characteristics are summarized in Table 1. Informed consent for this study was obtained from all the patients, and institutional review board approval was waived because of the retrospective nature of the study. Surgical Procedure and Intraoperative Neurophysiological Monitoring All patients underwent open surgery for maximum tumor resection. Cortical mapping was performed in eight patients, using electric stimuli of 3 to 12 mA to identify the sensorimotor and language cortices, according to a method described previously.1–4 Five patients were treated while in an awake condition. Primary speech cortex was identified based on speech arrest or hesitation due to stimulation during counting or object naming by the patient. A frameless stereotactic navigation device (ISG Viewing Wand, ISG Technologies; ViewScope, Elekta IGS; or Vector Vision, BrainLAB) was used in all nine cases after 1998; ultrasonography was used in the initial two cases. 264 Tumor resection was performed using an Ultrasonic Surgical Aspirator (Sonopet, Miwatec Co., Ltd.). All of the opercular arteries were carefully dissected and preserved. If the tumor infiltrated down to the sylvian fissure, thorough dissection of the affected sylvian fissure was initially performed to identify and preserve the insular and opercular segments of the MCAs. Only the minimum amount of Surgicel (Ethicon, Inc.) was routinely used for hemostasis. Dexamethasone (4 mg) was given every 6 hours with varied tapering schedules, and the administration of anticonvulsants (typically phenytoin and/or zonisamide) and antibiotics was begun or continued in patients during the immediate postoperative period. Neuroimaging Studies All patients underwent preoperative, postoperative, and subsequent follow-up MR imaging at our department. Postoperative imaging was performed within 72 hours of surgery. Data from pre- and postoperative DW MR imaging were available for the nine consecutive patients treated after 1998. For MR imaging, we used a 1.5-tesla system (Signa Horizon LX CV/i, GE Medical Systems) with a conventional quadrature head coil. We obtained T1-weighted images before and after Gd administration, T2-weighted images, and DW images during the same imaging session without repositioning the patient’s head. Axial DW images were obtained using fat-suppressed spin echo–echo planar imaging (TR 5000 msec, TE 72 msec, number of excitations 2, slice thickness 6 mm, gap 2 mm, matrix 128 3 128, and field of view 23 3 23 cm) with three orthogonal directional motionprobing gradients (b = 1000 seconds/mm2), followed by automatic generation of isotropic DW images. To evaluate the J. Neurosurg. / Volume 106 / February, 2007 Ischemic complications and resection of opercular glioma TABLE 1 (continued) Stimulation Navigation System Extent of Resection DW Image Immediate Postop Outcome Long-Term Deficit Follow Up (3/26/06) DCS US total not examined lt facial palsy none no recurrence, KPS 100 DCS US total not examined slight dysarthria no recurrence, KPS 90 DCS viewing wand total ViewScope total lt fine movement disorder none dead 5/14/99 (dissemination) none no recurrence, KPS 100 DCS & SCS ViewScope total DCS & SCS ViewScope subtotal none ViewScope total none ViewScope total DCS & SCS ViewScope total DCS & SCS ViewScope subtotal DCS & SCS Vector Vision subtotal infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity infarction beneath resection cavity rt facial palsy, dysarthria lt hemiparesis spatial relationship between the postoperative ischemic lesion and the pyramidal tract, coronal DW images were also obtained using similar conditions, but the motion-probing gradient was applied in only the anteroposterior direction. The pyramidal tracts, which run in the cephalocaudal direction, were delineated as slightly hyperintense to adjacent brain parenchyma in these coronal images. The extent of the resection was evaluated according to MR images obtained within 72 hours of surgery. If the tumor had been enhanced on the preoperative MR images, its gross-total resection was defined as no residual enhanced tumor, its subtotal resection as more than 75% removal, and its partial resection as less than 75% removal. If the tumor had not been enhanced on the preoperative MR images, resection was evaluated based on the presence of residual high-intensity lesion on the T2-weighted MR images. Postoperative Neurological Outcomes The postoperative neurological outcome was recorded and confirmed by retrospective review of all hospital records and physician notes. Immediate postoperative neurological function was determined during the first 7 days after surgery, and long-term function was determined between 3 and 6 months after surgery. Microangiographic Analysis of Vascular Supply to the Corona Radiata Coronal and axial microangiograms of five cadaveric brains without gross brain pathological features, which were part of a microangiographic study on the distribution of the basal perforating arteries that had been conducted by one of the authors (S.T.) in 1985,17,18 were reanalyzed to examine J. Neurosurg. / Volume 106 / February, 2007 lt hemiparesis rt facial palsy, dysarthria lt hemiparesis slight dysarthria no recurrence, KPS 90 lt hemiparesis lt facial palsy none dead 4/28/04 (local recurrence & dissemination) no recurrence, KPS 100 none none no recurrence, KPS 100 none none no recurrence, KPS 100 rt facial palsy, dysarthria dysarthria rt fine movement none disorder, dysarthria no recurrence, KPS 90 no recurrence, KPS 100 pial cortical arteries in and around the insuloopercular region and to identify the blood supply to the corona radiata. Results Extent of Resection Table 1 outlines the extent of resection in each case, as determined by quantitative volumetric analysis using postoperative MR imaging. Gross-total resection of the lesions was accomplished in eight patients (Cases 1–5 and 7–9). Subtotal resection was achieved in three patients (Cases 6, 10, and 11). Postoperative MR Imaging and Neurological Outcomes Postoperative DW MR images showed markedly hyperintense areas representing restricted diffusion in all nine patients treated after 1998. These lesions were all contiguous with the resection cavity. The size of the lesions beneath the resection cavity varied from case to case. All lesions appeared as high-intensity areas on T2-weighted MR images. Similar high-intensity lesions were also depicted on postoperative T2-weighted images in the two initial patients without DW imaging data. Details on postoperative neurological deficits are shown in Table 1. Eight patients, all of whom had undergone relatively small areas of resection mostly located in the face motor area, did not suffer impairment of long-tract function after surgery; three patients (Cases 3, 4, and 6) did have impaired long-tract function immediately after surgery. The lesions with restricted diffusion involved the descending motor pathway in these three patients. Tumor was located in the orofacial sensorimotor area in two of these patients 265 T. Kumabe et al. Microangiographic Analysis of Vascular Supply to the Corona Radiata Coronal microangiography of the cadavers showed that the corona radiata is constantly supplied by the lateral striate arteries, the long insular arteries originating from insular portions of the MCA, and the medullary arteries from the opercular and cortical portions of the MCA (Fig. 5). Surgical removal for opercular glioma, even if not involving the insula, could compromise the latter two fine arteries, resulting in cerebral infarction at the corona radiata. Illustrative Cases Case 3 FIG. 1. Case 3. Images obtained in a 68-year-old man with an anaplastic astrocytoma in the right tongue sensorimotor area. Preoperative axial T2-weighted (A) and coronal fluid-attenuated inversion-recovery (B) MR images revealing a high-intensity mass in the right frontoparietal opercular region. Postoperative axial T2-weighted (C) and coronal DW (D) MR images showing total removal of the tumor and a new high-intensity lesion (arrows) beneath the resection cavity. The lesion extends to the region of the corona radiata, probably involving the corticospinal tract. This tract appears as a bandlike area of slightly high intensity on the coronal DW image with the motion-probing gradient applied only in the anteroposterior direction. (Cases 3 and 6). A relatively small resection in these two patients was performed around the white matter above the upper limiting sulcus of the posterior region of the insula, and wide resection was undertaken in the anteroposterior and cephalocaudal directions of the opercular region (Figs. 1–4). History and Examination. This 68-year-old man presented with an anaplastic astrocytoma manifesting as left facial seizures. Results of T2-weighted MR imaging demonstrated a hyperintense lesion in the right opercular portions of the inferior frontal, precentral, and postcentral gyri inferolateral to the precentral knob, not involving the insula (Fig. 1A). Administration of a contrast medium caused no enhancement. Neurological and neuropsychological examination revealed no abnormality. Operation. A right frontoparietotemporal craniotomy was performed with the patient in an awake condition. Direct cortical stimulation identified the face motor area and the primary sensory sites of the tongue and face. The sylvian fissure was thoroughly dissected toward the distal end, and the insular surface was exposed under the operating microscope. The precentral and central arteries were separated from the tumor and preserved. The lesion was totally removed up to the face motor area and toward the deepest portion using the upper limiting sulcus as the anatomical landmark and the ISG Viewing Wand (Fig. 2). Postoperative Course. Almost complete left hemiparesis was observed postoperatively. Magnetic resonance images showed that the entire lesion had been resected but revealed an ischemic area beneath the resection cavity involving the descending motor pathway (Fig. 1B). As adjuvant therapy, the patient received 72 Gy hyperfractionated radiation to the extended local field. His left hemiparesis resolved except FIG. 2. Case 3. Intraoperative photographs obtained before (left) and after (right) tumor resection. Left: Note the results of functional brain mapping: 1, face motor; 2, tongue sensory; and 3, face sensory. The outline indicates the location of the tumor; the arrows represent the central sulcus. Right: Note preservation of the precentral (4) and central (5) arteries. 266 J. Neurosurg. / Volume 106 / February, 2007 Ischemic complications and resection of opercular glioma FIG. 3. Case 6. Images obtained in a 69-year-old woman with a right frontoparietal GBM. A: Preoperative axial Gdenhanced T1-weighted MR image demonstrating an irregularly enhanced mass lesion in the opercular region around the central sulcus (arrow). B: Postoperative axial T2-weighted MR image depicting total removal of the tumor and a new lesion of high intensity (arrow) beneath the resection cavity. C: Postoperative axial (left) and coronal (right) DW MR images revealing the new lesion (arrows) beneath the resection cavity with reduced diffusion probably involving the corticospinal tract. for impaired fine movement of the left finger. He was discharged home and able to ambulate 2 months postsurgery. Case 6 History and Examination. This 69-year-old woman presented with a GBM manifesting as left facial seizures followed by left hemiconvulsion. Preoperative T1-weighted MR images with contrast medium exhibited an enhanced mass in the right opercular portions of the inferior frontal, precentral, and postcentral gyri inferolateral to the precentral knob, not involving the insula (Fig. 3A). Neurological and neuropsychological examination revealed no abnormality. Operation. A right frontoparietotemporal craniotomy was performed with the patient in a state of general anesthesia. Direct cortical stimulation identified the hand/digit motor area. The sylvian fissure was thoroughly dissected toward the distal end, and the insular surface was exposed under the operating microscope. The precentral, central, and anterior parietal arteries were separated from the tumor and preserved. The lower portion of the lesion was resected toward the deepest portion by using the upper limiting sulcus as the anatomical landmark and the ViewScope. The tumor was removed in a stepwise manner while monitoring the muscle J. Neurosurg. / Volume 106 / February, 2007 contraction by direct cortical stimulation to the hand/digit motor area (Fig. 4). The positive cortical response gradually became duller at the end of the surgery. Postoperative Course. Almost complete left hemiparesis was observed postoperatively. Magnetic resonance images obtained after surgery showed that most of the enhanced lesion had been resected, but an ischemic area was found beneath the resection cavity involving the descending motor pathway (Fig. 3B and C). The patient’s left hemiparesis did not resolve. Discussion Surgical removal of glioma in the opercular region presents many challenges. In 1991, LeRoux et al.11 first reported that gliomas involving the nondominant face motor cortex can be safely removed using brain mapping techniques to localize the rolandic cortex and avoid resection of the hand motor cortex and descending subcortical motor pathways. However, resection of the face motor cortex in the dominant hemisphere was not recommended because language localization in the cortical zones is contiguous with this region. In 1995, Ebeling and Kothbauer5 supposed that radical tumor resection of a purely opercular glioma, not including the insula, in the dominant hemisphere can be 267 T. Kumabe et al. FIG. 4. Case 6. Intraoperative photographs obtained before (left) and after (right) tumor resection. Left: Note the results of functional brain mapping: 1 and 2, hand/digit motor. The outline indicates the tumor location; the arrows represent the central sulcus. Right: Note preservation of the precentral (3), central (4), and anterior parietal (5) arteries. achieved without significant lasting morbidity. However, only biopsy was recommended for large dominant insular or opercular–insular tumors, because the lenticulostriate arteries hinder total resection and no clear border toward the internal capsule can be found. In 2004, Peraud et al.12 reported the surgical results of 14 cases of opercular gliomas, stressing the importance of intraoperative neuromonitoring as an aid to surgery in the dominant opercular region. The severity and duration of postoperative deficits was well correlated with the distance from the resection margin to the next positive stimulation point(s), and a distance of more than 5 mm was found to avoid major impairments. Clearly, intraoperative functional brain mapping techniques can help preserve cortical and subcortical functions. However, vascular damage during resection of opercular glioma remains less well understood. Recently, restricted diffusion abnormalities were found adjacent to the resection cavity on immediately postopera- tive images in 64% of cases.15 Both cortical and subcortical lesions with restricted diffusion were observed after surgery. In the present study, postoperative MR images including DW images disclosed infarcted lesions in all patients, and these lesions unexpectedly extended to the descending motor pathway in the corona radiata in three patients, which could have resulted in the impairment of long-tract function. These infarcted lesions were probably caused by disruption of the blood supply during the surgical procedures. Microangiographic analysis in this study revealed that the corona radiata is supplied by the lateral striate arteries, long insular arteries, and medullary arteries from the opercular and cortical segments of the MCA passing over the frontoparietal operculum. According to the study data collected by Ture et al.,20 approximately 85 to 90% of insular arteries are short and supply the insular cortex and extreme capsule, 10% are medium length and supply the claustrum and external capsule, and 3 to 5% are long and extend as far FIG. 5. Coronal (left) and axial (right) microangiograms of a cadaveric brain slice through the interventricular foramina. Both long insular arteries (arrows) arising from the insular portions of the MCA and the medullary arteries (arrowhead) from the opercular and cortical portions of the MCA course toward the ventricular wall and supply the region of the corona radiata. 268 J. Neurosurg. / Volume 106 / February, 2007 Ischemic complications and resection of opercular glioma as the corona radiata. Interruption of blood flow to these long insular arteries during the resection of intrinsic insular tumor may result in hemiparesis; thus, these arteries should be preserved to prevent infarction of the corona radiata.6,10,19,20 Data in the present study demonstrated that, in addition to the long insular arteries, the long medullary arteries from the opercular and cortical segments of the MCA passing over the frontoparietal operculum contribute to the arterial supply to the corona radiata. Although the vascular supply can show individual variations, these long medullary arteries can be impaired during the resection of a pure opercular glioma. Adequate collateral blood supply would not be expected13 because intraparenchymal arterioles such as the lateral striate arteries, long insular arteries, and long medullary arteries are all end arteries without substantial anastomoses with other arteries except in pathological cases like moyamoya disease. If the impaired arteries supply most of the descending motor pathway, resection of opercular glioma is likely to result in hemiparesis. The long insular arteries are mostly located in the posterior region of the insula,20 most commonly on the posterior half of the central insular sulcus and on the long gyri.19 Thus, subcortical resection around the upper limiting sulcus of the posterior region of the insula carries a higher risk of sacrifice of the long insular arteries, which may lead to extensive corona radiata infarction, and ultimately critical damage to the descending motor pathway. Similarly, a wide resection in the anteroposterior and cephalocaudal directions of the opercular region could damage a large number of medullary arteries from the opercular and cortical segments of the MCA over the frontoparietal operculum. In our experience, these two maneuvers appeared to be risk factors for critical infarction in the corona radiata after resection of an opercular glioma. However, reliable methods for avoiding damage to the long insular and medullary arteries are not available. Limited resection of the operculum as well as sparing of the posterior region of the insula may be the only measures presently available to avoid injury to a large number of long insular arteries and long medullary arteries. The development of new surgical devices to remove an opercular glioma with preservation of thin blood vessels like these arteries is to be expected in the future. Conclusions In the present study we found that ischemic complications occurring beneath the resection cavity including the pyramidal tract within the corona radiata are caused by damage to the distributing arteries—in particular, the long insular arteries and/or medullary arteries from the opercular and cortical segments of the MCA passing over the frontoparietal operculum—after resection of glioma in the frontoparietal opercular region inferolateral to the hand/digit sensorimotor area. Surgeons should be aware of the risk of ischemic complications during resection of opercular glioma and the possibility of permanent motor deficits. References 1. Berger MS: Malignant astrocytomas: surgical aspects. Semin Oncol 21:172–185, 1994 2. Berger MS, Deliganis AV, Dobbins J, Keles GE: The effect of J. Neurosurg. / Volume 106 / February, 2007 extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas. Cancer 74:1784–1791, 1994 3. Berger MS, Kincaid J, Ojemann GA, Lettich E: Brain mapping techniques to maximize resection, safety, and seizure control in children with brain tumors. Neurosurgery 25:786–792, 1989 4. Berger MS, Ojemann GA, Lettich E: Neurophysiological monitoring during astrocytoma surgery. Neurosurg Clin N Am 1: 65–80, 1990 5. Ebeling U, Kothbauer K: Circumscribed low grade astrocytomas in the dominant opercular and insular region: a pilot study. Acta Neurochir 132:66–74, 1995 6. Hentschel SJ, Lang FF: Surgical resection of intrinsic insular tumors. Neurosurgery 57 (1 Suppl):176–183, 2005 7. Higano S, Yun X, Kumabe T, Watanabe M, Mugikura S, Umetsu A, et al: Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis. Radiology 241:839–846, 2006 8. Kitis O, Altay H, Calli C, Yunten N, Akalin T, Yurtseven T: Minimum apparent diffusion coefficients in the evaluation of brain tumors. Eur J Radiol 55:393–400, 2005 9. Kumabe T, Nakasato N, Suzuki K, Sato K, Sonoda Y, Kawagishi J, et al: Two-staged resection of a left frontal astrocytoma involving the operculum and insula using intraoperative neurophysiological monitoring—case report. Neurol Med Chir 38:503–507, 1998 10. Lang FF, Olansen NE, DeMonte F, Gokaslan ZL, Holland EC, Kalhorn C, et al: Surgical resection of intrinsic insular tumors: complication avoidance. J Neurosurg 95:638–650, 2001 11. LeRoux PD, Berger MS, Haglund MM, Pilcher WH, Ojemann GA: Resection of intrinsic tumors from nondominant face motor cortex using stimulation mapping: report of two cases. Surg Neurol 36:44–48, 1991 12. Peraud A, Ilmberger J, Reulen HJ: Surgical resection of gliomas WHO grade II and III located in the opercular region. Acta Neurochir 146:9–18, 2004 13. Phan TG, Donnan GA, Wright PM, Reutens DC: A digital map of middle cerebral artery infarcts associated with middle cerebral artery trunk and branch occlusion. Stroke 36:986–991, 2005 14. Schaefer PW, Grant PE, Gonzalez RG: Diffusion-weighted MR imaging of the brain. Radiology 217:331–345, 2000 15. Smith JS, Cha S, Mayo MC, McDermott MW, Parsa AT, Chang SM, et al: Serial diffusion-weighted magnetic resonance imaging in cases of glioma: distinguishing tumor recurrence from postresection injury. J Neurosurg 103:428–438, 2005 16. Sugahara T, Korogi Y, Kochi M, Ikushima I, Shigematu Y, Hirai T, et al: Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 9:53–60, 1999 17. Takahashi S, Goto K, Fukasawa H, Kawata Y, Uemura K, Suzuki K: Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part I: Striate arterial group. Radiology 155:107–118, 1985 18. Takahashi S, Goto K, Fukasawa H, Kawata Y, Uemura K, Yaguchi K: Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part II: Thalamic arterial group. Radiology 155:119–130, 1985 19. Tanriover N, Rhoton AL Jr, Kawashima M, Ulm AJ, Yasuda A: Microsurgical anatomy of the insula and the sylvian fissure. J Neurosurg 100:891–922, 2004 20. Ture U, Yaşargil MG, Al-Mefty O, Yaşargil DC: Arteries of the insula. J Neurosurg 92:676–687, 2000 21. Zimmerman RD: Is there a role for diffusion-weighted imaging in patients with brain tumors or is the “bloom off the rose”? AJNR 22:1013–1014, 2001 Manuscript submitted March 26, 2006. Accepted August 7, 2006. Address reprint requests to: Toshihiro Kumabe, M.D., Department of Neurosurgery, Tohoku University Graduate School of Medicine, 1–1 Seiryo-machi, Aoba-ku, Sendai 980–8574, Japan. email: kuma@nsg.med.tohoku.ac.jp. 269