Temporal lobe arteriovenous malformations: Surgical management and outcome

ELSEVIER

Featured Subject: Arteriovenous Malformations

TEMPORAL LOBE ARTERIOVENOUS MALFORMATIONS: SURGICAL MANAGEMENT AND OUTCOME Ghaus M. Malik, M.D.,* Donald M. Seyfried, M.D.,* and Jay K. Morgan, M.D.-t *Department of Neurosurgery, Henry Ford Hospital, Detroit, Michigan, and TSierra Neurosurgery Group, Reno, Nevada

Malik GH, Seyfried DM, Morgan JK. Temporal lobe arteriovenous malformations: surgical management and outcome. Surg Neurol 1996;46:106-15.

KEY

WORDS

Arteriovenous malformations, intracranial hemorrhage, management outcome, microneurosurgery, temporal lobe.

BACKGROUND

Temporal lobe arteriovenous malformations (AVMs) represent a subgroup of intracranial AVMs with particular characteristics and management issues. METHODS

We performed a retrospective analysis of 24 consecutive patients with temporal lobe AVMs treated with surgical excision. Factors such as location, size, arterial feeders, venous drainage, and clinical follow-up were recorded for each. Results were compared with those of 132 patients with nontemporal lobe AVMs surgically treated over the same time period. RESULTS

Sixteen of the temporal AVMs were located in the convexity, six in the mesotemporal region, and two were predominantly intraventricular. The mode of presentation was seizure in 11 patients, hemorrhage in 7, headache in 4, and 2 were asymptomatic. Patients with convexity AVMs more commonly presented with seizures, whereas patients with mesotemporal or intraventricular AVMs were more likely to present with hemorrhage. One patient with subarachnoid hemorrhage from a basilar artery aneurysm died. Postoperatively, 2 patients (8.3%) had a new hemiparesis and dysphasia, 1 (4%) had a new dysphasia and hemianopsia, and 3 others (13%) were left with an isolated superior quadrant field deficit. Lasting surgical morbidity other than isolated field deficit was 13% for patients with temporal AVMs and 15% for those with nontemporal AVMs. CONCLUSIONS

Temporal lobe AVMs may be successfully resected using a direct microsurgical approach with limited morbidity and excellent prognosis for recovery. Most of the deficits relating to AVM hemorrhage and those of the immediate postoperative period improved significantly over the subsequent few months.

Address reprint requests to: Editorial Office, Department of Neurosurgery, Henry Ford Hospital, 2799 West Grand Blvd., Detroit, Michigan 48202. Received May 19, 1994; accepted January 2, 1996. 0090-3019/96/$15.00 PI1 SOO90-3019(96)000845

T

emporal lobe arteriovenous malformations (AVMs) are relatively uncommon. Their surgical treatment has been considered in various se ries, often described in the context of AVMs of the limbic system, sylvian fissure, tentorial ring, anterior choroidal artery, and medial temporal lobe. Described surgical approaches have been based on the specific location of the AVM within the temporal lobe. This engenders the subclassification of the lesions and overlap with AVMs outside the temporal lobe proper. Few series are devoted to manage ment morbidity of temporal lobe AVMs as a group. We therefore present the clinical course and surgical management of our series of patients with AVMs confined to the temporal lobe and compare the results with those of surgically treated nontemporal AVMs.

METHODS From 1978 to 1991, 24 consecutive patients underwent craniotomy for surgical excision of a temporal lobe AVM to prevent future risk of hemorrhage. Seven patients had preoperative embolization attempts. Medical records and angiograms were reviewed retrospectively. Factors such as location, size (volume estimate = [length X width X height]/ 2), arterial feeders, venous drainage, and immediate and long-term clinical follow-up were recorded for each. Surgical excision was verified by postoperative angiogram, except in those patients requiring treatment of other malformations, those who refused the study, or those who died. Those AVMs 655 Avenue

0 1996 by Elsevier Science Inc. of the Americas, New York, NY 10010

Surg Neurol 1996;46:106-15

Temporal Lobe Arteriovenous Malformations

with a sylvian nidus were not included in the temporal AVM group. During the same time period (1978-1991) 132 patients with nontemporal AVMs were surgically treated and had necessary follow-up and records available to be included in this study for purposes of comparison. Convexity temporal lobe AVMs were approached directly, using a corticotomy when necessary. Attention was paid to staying within the gliotic margin and avoiding ventricular hemorrhage. The intraventricular lesions were approached by a transcortical route. Anterior mesotemporal AVMs were reached by opening of the sylvian fissure and limited temporal lobe resection. Those of the posterior temporal region were approached subtemporally, with or without corticotomy.

PATIENTS Ages of patients with temporal lobe AVMs at the time of surgery ranged from 11 to 67 years (mean, 33 years). There were 11 males and 13 females. Lesions were right-sided in 11 patients, and on the left in 13. Eleven patients presented with seizures, which were generalized in 10. Seven patients had intracranial hemorrhage from the AVM, with associated neurologic deficit ranging from hemianopsia alone to combined paresis and dysphasia. Four patients presented with headache alone, and 2 others were found to have AVMs incidentally. One patient (#14, Table 1) had a subarachnoid hemorrhage from a ruptured basilar artery aneurysm, with a posterior temporal lobe AVM discovered on angiography.

RESULTS The average length of follow-up for the patients with temporal lobe AVMs was 27 months, and the results are shown in Table 1. Sixteen patients had their AVM located on the convexity, in 6 it was mesotemporal, and in 2 it was primarily intraventricular. Of the patients with convexity AVMs, only 2 presented with hemorrhage; AVMs of the other 14 patients were found either after seizure, headache, or incidentally. Of the 8 patients with mesotemporal or intraventricular AVMs, 5 had hemorrhages and only 2 of the 8 were neurologically intact on admission. The volumes of the temporal AVMs ranged from 1 to 53 cm3 (mean, 17 cm3). AVMs of the convexity were predominantly supplied by middle cerebral artery (MCA) branches and had contributions from the posterior cerebral artery (PCA) when located in

107

the posterior temporal lobe. Convexity lesions were drained by the superficial venous system, and 19% (3/16) also had drainage into the deep venous system. Two of those with combined drainage had a mean volume of 44 cm3, and the other one was a l-cm3 AVM that presented with hemorrhage. Mesotemporal (N = 6) or intraventricular @ = 2) lesions were more likely to be fed by the choroidal and PCA vessels, and all had a major component of deep venous drainage. Arterial supply by AVM location is shown in Table 2. Of the seven preoperative embo lization attempts in cases with temporal AVMs, two were unsuccessful and two others (#4 and #17, Table 1) resulted in neurologic deficit. Embolization results for AVMs in all locations are presented in Table 3. One patient had a postoperative intracerebra1 hemorrhage, and another had a postoperative epidural. One patient (#24, Table 1) had a delayed hemorrhage from a residual AVM; he underwent a second craniotomy with a postoperative angiogram showing complete AVM excision. One patient (#14, Table 1) died following brain stem infarct after clipping of a ruptured basilar artery aneurysm and resection of a temporal lobe AVM. Two patients had hydrocephalus requiring ventricular drainage. There were three wound infections. Of the group of 132 patients with surgically treated nontemporal brain AVMs, there was a 3.8% mortality and a 15% morbidity (neurologic deficit other than an isolated field loss lasting at least 3 months) (Table 4). Infection rate was 1.5%, giving an overall infection rate of 3.2% for all AVM patients.

DISCUSSION Temporal lobe AVMs historically account for 12% to 16% of intracranial AVMs and comprised 15% (24/ 156) of all brain AVMs in our series [2,3,14]. They present a challenge to the neurosurgeon because of potentially associated morbidity and their unique anatomic relationships. Most of the recent clinical reviews group temporal lobe AVMs with those of other brain regions. For example, temporal lobe AVMs comprise a variable percentage of malformations of the sylvian fissure [12], tentorial ring [9], limbic system [lo], anterior choroidal artery [5,13], PCA [5], and the medial temporal lobe [4]. Intracranial AVMs generally present with hemorrhage and less commonly with seizures [ 1,3,6,15]. In our group of nontemporal AVMs, those in deep basal ganglia and posterior fossa locations presented with hemorrhage in 82% and 89% of instances, respectively. In our series of temporal lobe AVMs, seizures were a more common presenting

0

Malik et al

Surg Neural 1996;46:106-15

108

Temporal lobe AVMs: characteristics and outcome.

PATENT

A-1 SEX

PRE%NTATION

16/F 20/F

Intracerebral hemorrhage Seizure

22/M

Seizure

65/M

Headache

28/M

Seizure

41/M

Seizure

36/F

Seizure

52/F

Seizure

9

67/F

Headache

10

19/F

Seizure

11

28/F

Incidental

12

29/M

Seizure

13

25/F

14

42/M

Intracerebral hemorrhage Incidental

15

35/M

Headache

16

44/F

Seizure

17

51/F

Headache

18

31/M

19

23/F

20

41/M

lntraventricular hemorrhage Intracerebral hemorrhage Seizure

21

32/F

Seizure

22

1 l/F

23

19/M

24

27/M

Intracerebral hemorrhage Intracerebral hemorrhage Intracerebral hemorrhage

Abbreviations: PCA = posterior

GREATEST DIAMETER 63)

ARTERIAL FFXDERS

3; 10; Left posterior 6; 53; Right posterior 4; 34; Right anterior 2; 4; Left posterior 4; 19; Right posterior 2; 3; Right anterior 2; 1; Left posterior 6; 39; Left convexity

MCA; PCA

6; 24; Right posterior 2.5; 2; Right middle 3; 8; Right superior 5; 28; Left posterior 2; 1; Left posterior 3; 8; Left posterior 4; 22; Left posterior 3; 9; Right convexity 6; 16; Left mesotemporal 5; 32; Right mesotemporal 2; 3; Left posterior 6; 23; Left anterior 3; 9; Right posterior 4; 13; Right posterior 4; 20; Left temporal horn 5; 24; Left temporal horn

ACh; WA;

PCA

MCA PCA

ACh = anterior choroidal artery; cerebral artery; PCOM = posterior

Intact

Right superior quadrantanopsia Left superior quadrantanopsia Intact

Superficial and deep Superficial and deep Superficial

Intact

Improved hemiparesis and dysphasia Intact Intact

Superficial

MCA

Superficial

Intact

MCA

Superficial

Peroneal

MCA; PCA

Superficial

Intact

MCA; PCA

Superficial

Intact

Right hemiparesis; moderate dysphasia Intact

MCA

Superficial

Intact

Intact

MCA

Superficial

Intact

Intact

ACh; MCA; PCA

Superficial

MCA

Superficial

Right central VII palsy Expressive dysphasia

Right central VII palsy Intact

PCA

Superficial

Intact

Coma and death

MCA; PCA; ECA

Superficial

Intact

MCA

Superficial

ACh; MCA; PCA

Superficial and deep Superficial and deep Deep

Mild left arm motor and sensory deficit Right homonomous hemianopsia Left homonomous hemianopsia Right central VII palsy Right hemiparesis

Right superior quadrantanopsia Intact

MCA PCA ACh; MCA; PCA; PCOM MCA: PCA Ach; PCA Thalamoperforators; ACh ACh; PCh PCh = posterior choroidal communicating artery

MCA

PCA

ACH

14/16 416 o/2

9116 516 012

2116 316 32

PCA = posterior

cere

Abbreviations: MCA = middle cerebral artery; bra1 artery: ACh = anterior choroidal artery.

Dysphasia

MCA; PCA

Arterial feeders to the AVM.

Convexity Mesotemporal Intraventricular

Superficial

LAST FOLLOW-UP

Dysphasia and right hemiparesis Intact

symptom, occurring in 46% of patients, compared with the incidence of 24% in nontemporal AVM patients. Mahalick et al reported a similar experience [7]. This may relate to the larger number of AVMs

q

PRF.oPFRAlwE EXAM

VENOUS DRAINAGE

Superficial and deep Superficial and deep Superficial and deep Deep Deep artery;

atrophy

Peroneal

atrophy

Improved dysphasia: hemianospia; paresis Left superior quadrantanopsia Intact

Intact

Dysphasia: paresis; hemianopsia Intact

Left homonomous hemianopsia Dysphasia; right hemianopsia Dysphasia; right hemianopsia; paresis

Left homonomous hemianopsia Minimal right field constriction Improved hemianopsia; dysphasia: paresis

ECA = external

carotid

artery;

McA = middle cerebral

artery;

being on the convexity and involving the cortex, rather than deep within the temporal lobe. We noted a greater percentage of the convexity AVMs presenting with seizure, whereas the deeper mesotemporal lesions had a propensity to present with hemorrhage, perhaps behaving as ventricular AVMs, with choroidal artery supply and deep venous drainage. These latter findings are consistent with previous reports [4,5]. The mesotemporal AVMs were also more likely to present with neurologic deficit. Extensive supply by the PCA and location along the tentorium have also been associated with bleeding [9,13]. Of the seven patients who presented with hemorrhage and neurologic deficit,

Temporal Lobe Arteriovenous

EIIEmbolization

Malformations

and morbidity:

Surg Neurol 1996;46:106-15

temporal versus nontemporal

Temporal Nontemporal

&JCCESSW

L~cA~~N (NUMB@ Temporal (24) FrontotemDoral (76) Occipital (26) Deep/basal ganglia (111 Po&kior fossa (19)

EIWKXIZATION

MORBID~

7 26

No MORBIDITY

2 (29%) 9 (35%)

five had marked improvement or resolution of the deficit at the most recent follow-up. Thus, it ap pears that the outcome may be good to excellent, despite the initial presentation with hemorrhage, even with medially located AVMs. Arterial supply to AVMs of the convexity predominantly involved the MCA but also included vessels from the PCAwhen the lesion was large or extended into the posterior temporal lobe. Deep feeders from the anterior choroidal artery and PCA must be looked for on the preoperative angiogram. Convexity lesions drained superficially in all instances, and there was dual drainage (superficial and deep) in a minority of patients (19%). Mesotemporal and ventricular AVMs often had supply from the anterior choroidal artery and were more likely than convexity lesions to have deep perforators from the Ml or PCA. Deep temporal AVMs (mesotemporal and intraventricular) always had a major component of venous drainage into the deep venous system via the basal vein of Rosenthal or the vein of Galen. As Stein has noted, when these AVMs have a dual venous drainage, one or the other may be sacrificed during the exposure to assist in the resection [lo]. The vascular patterns help determine the feasibility and need for preoperative embolization. Hodes et al described the endovascular techniques with six AVMs of the temporal lobe embolized via the anterior choroidal artery [5]. One patient had neurologic deficit following the procedure, and sev-

L3 ’ AVM presentation

AVMs.

Ehl~ouit.kTIoN ATTEMPTS

AVMLOCATION

and surgical morbidity

109

3 (43%) 13 (50%)

era1 patients required multiple embolizations. Luessenhop and Rosa have advocated intravascular techniques, either as an adjunct to surgery in patients with large AVMs (i.e., those at risk for circulatory breakthrough) or as a primary treatment in cases that are inoperable because of the patient’s age or the size or location of the AVM [ 61. Pasqualin et al correlated preoperative embolization with decreased intraoperative bleeding and noted a trend towards reduction of hyperemic changes [8]. From our experience, we believe embolization is helpful for the deep feeders from the anterior choroidal artery and large temporal branches of the MCA and PCA, since these are arising medially and may be difficult to control when approaching the AVM from the lateral side. When embolizing the anterior choroidal artery, the catheter must be beyond the plexal point to avoid occluding the perforators to the optic radiations or the internal capsule. Complications from embolization attempts seem to relate to AVMs with numerous perforators arising from the PCA, therefore, extra caution may be warranted when considering embolization of these particular lesions. One such case in our series resulted from overdistension of the vessel wall with dubsequent rupture. Availability of newer catheters and additional experience in their use may reduce the complications from endovascular procedures. If necessary, a staged approach may be more desir-

by location.

PRJTSENTATION

SURGICAL MORTALITY

ALL SURGICAL MORBIDITY LA!+TING 23 MONTHS

SURGICAL MORRIDITY EXCLUDING ISOLATED FIELD DEFIQTS 13% 16%

HF,MORRHACE

SEIZURE

OTHER

29% 41%

46% 34%

25% 25%

4% 4%

25% 17%

46% 82%

19% 9%

35% 9%

0% 0%

50% 36%

89%

0%

11%

11%

5%

5%

1 10

Surg Neurol 1996;46:106-15

able for AVMs of large dimensions incorporating multiple PCA feeders [6]. In our series, convexity AVMs (Figure 1) were the most frequent of the temporal lobe malformations (16/24 [67%]). We employed a direct approach for dissection of the majority of the convexity AVMs, staying within pial planes and the gliotic margins. Large convexity lesions were exposed through a gyrus incision to avoid the arterial vessels within the sulci. When making a cortical incision, an attempt was made to remain parallel to fiber tracts and dissect along the gliotic, nonfunctioning tissue at the margin of the AVM. Dissection proceeds more easily when remaining perpendicular to the AVM, rather than tangential to it [ll]. When the malformation involves exposure through the dominant superior temporal gyrus, dissection through the lower half may spare the language fibers or result in only a transient dysphasia. While dissecting medially along a convexity AVM, a cautionary approach is necessary with the arterial feeders from the posterior communicating and anterior choroidal arteries. If the temporal horn is entered, intraventricular hemorrhage should be minimized by covering the exposed ependymal surface with a cottonoid, while obtaining meticulous hemostasis of the choroidal feeders. Surgical excision of the anterior temporal AVMs and those adjacent to the sylvian fissure requires partial splitting of the fissure. With this maneuver, avoidance of the major cortical draining veins is important during the early part of the operation. Three of our 16 patients with convexity AVMs were left with a superior quadrantanopsia, and 1 had a permanent mild hemiparesis and expressive dysphasia related to the surgery. We use the term “mesotemporal” to refer to lesions located deep within the temporal lobe (Figure 2). Other series refer to these AVMs as hippocampal [3], tentorial [9], anterior choroidal [ 131, medial temporal [4], and limbic [lo]. Drake [3] has resected most through a transcortical incision, with this approach being especially desirable for larger lesions [lo]. Heros used a transcortical incision through the inferior temporal gyri to excise three AVMs of the medial temporal lobe [4]. We operated on six patients with AVMs of the mesotemporal area, employing a transcortical route, with complete excision in all cases. When located anteriorly, these lesions were approached through the sylvian fissure. At times, a limited resection of the distal 3 cm of the temporal lobe was performed. AVMs of the midtemporal region were dissected with transcortical incisions to provide the most direct route to the malformation, while minimizing temporal lobe retraction. When the lesion was more pos-

Malik et al

terior, a subtemporal approach was helpful, with resection of the inferior temporal gyrus when necessary. Although resection of mesotemporal AVMs is more difficult, five of our six patients returned to work or school, while the remaining patient had a persistent spastic hemiparesis. Two of our patients had intraventricular, temporal horn AVMs, presenting with intracerebral and intraventricular hemorrhage with impending herniation (Figure 3). One patient was treated by an emergent clot evacuation and stabilized with a dysphasia and hemiparesis. The other patient responded to ventricular drainage. Both underwent delayed resection of their malformation via a transcortical approach. Care was used to obtain meticulous hemostasis of the choroidal feeders and to excise the veins draining into the basal vein of Rosenthal. One 19-year-old patient had only a mild residual field cut and returned to school the following year. The second patient was ambulatory with a 4/5 hemiparesis and an expressive dysphasia. Of those 12 patients who had neurologic deficit secondary to the AVM or embolization preoperatively, 8 had improved neurologic status at the most recent follow-up. Two patients were the same neurologically, and 2 had increased neurologic deficit. Postoperatively, 8 of the 24 patients showed transient worsening on neurologic examination, usually dysphasia or a field cut. Most deficits, whether caused by the initial hemorrhage or the surgery, tended to improve significantly or resolved over a 3-month period. This may result from the displacement, rather than destruction, of parenchymal tissue by the AVM, hemorrhage, or the surgical dissection. Permanent visual field abnormalities without other neurologic deficit were the most fre quent operative morbidity. Three such patients in our series had their field defect restricted to a superior quadrant. All of these patients had posterior temporal convexity lesions. Postoperatively, 3 (13%) of the 24 patients experienced a new deficit, other than a field cut, which has persisted to the most recent follow-up. All of these patients were ambulatory, however. These results are comparable with other series having a subset of patients with temporal lobe AVMs [3,9,10,14]. We found a similarity of surgical results between temporal and nontemporal AVMs (Table 4): 13% of temporal AVM patients had a lasting nonvisual deficit, while 15% of patients with nontemporal AVMs had similar deficits. Mortality rates also were similar between the two groups: 4% for patients with temporal AVMs and 3.8% for all others. The incidence of surgical morbidity in our series may be reflective of the large percentage of patients being

of patient (#9, Table 1) with a tempo bra1 IItive A-PAngiogram convexity AVM; A-P (a), lateral (b), and postope :ra(c) views.

Angiogram (a-d) and magnetic resonance image (e) qAngiogram of patient ~317, Table 1) with a mesotemporal AVM. shows filling from the internal carotid artery (a and b) and vertebral artery (c and d). Magnetic resonance image demonstrates the AVM in the axial plane (e).

Temporal Lobe Arteriovenous

Malformations

Surg Neurol 1996;46:106-15

1 13

Angiogram of patient (#23, Table 1) with an AVM of El the temporal horn; A-P and lateral views (a and b) and a postoperative A-P view (c).

1 14

Surg Neurol

Malik et al

1996;46:106-15 treated with no prior hemorrhage (71% of the temporal AVM patients and 48% of the nontemporal AVM patients). The indication for AVM resection has been to provide immediate and complete prevention of future hemorrhage. The incidence of residual AVM after surgery (without the use of intraoperative angiography), which was then subsequently resected without rehemorrhage, was 4.5% for nontemporal AVMs. Residual AVM causing hemorrhage was 4% in the temporal group and 3% in the nontemporal group. This underscores the value of intraoperative angiography to avoid this potential cause of morbidity. Temporal lobe AVMs represent a unique and challenging group of intracranial malformations. Those of the mesotemporal region often present with hemorrhage and neurologic deficit, whereas those of the convexity appear to present more often with seizures. Visual-field defects are the most common morbidity following surgery, and neurologic deficits from the hemorrhage or the surgery tend to improve dramatically during the first 3 months. Preoperative embolization, while helpful in decreasing the blood supply to the medial aspect of the lesions, does have potential risks that need to be balanced against benefits of reducing flow and providing ease of surgical excision. An effective surgical approach to these AVMs is through a transcortical route, although the mesotemporal lesions remain the most difficult to excise. REFERENCES 1. Batjer HH, Devous MD, Seibert GB, Purdy PD, Bonte

2.

3. 4. 5.

6.

7. 8.

FJ. Intracranial arteriovenous malformations: relationship between clinical factors and surgical complications. Neurosurgery 1989;24:75-9. Brown RD, Weibers DO, Forbes G, O’Fallon WM, Piepgras DG, Marsh WR, Maciunas RJ. The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 1988;68:352-7. Drake CG. Cerebral arteriovenous malformations: considerations for and experience with surgical treatment in 166 cases. Clin Neurosurg 1979;26:145-208. Heros RC. Arteriovenous malformations of the medial temporal lobe: surgical approach and neuroradiological characterization. J Neurosurg 1982;56:44-52. Hodes JE, Aymard A, Casassco A, Rufenacht D, Reizine D, Merland JJ. Embolization of arteriovenous malformations of the temporal lobe via the anterior choroidal artery. AJNR 1991;12:775-80. Luessenhop AJ, Rosa L. Cerebral arteriovenous malformations: indications for and results of surgery, and the role of intravascular techniques. J Neurosurg 1984;60:14-22. Mahalick DM, Ruff RM, U HS. Neuropsychological se quelae of arteriovenous malformations. Neurosurgery 1991;29:351-7. Pasqualin A, Scienza R, Cioffi F, Barone G, Benati A, Beltramello A, Pian RD. Treatment of cerebral arte-

9. 10. 11.

12. 13.

14. 15.

riovenous malformations with a combination of preoperative embolization and surgery. Neurosurgery 1991;29:358-68. Solomon RA, Stein BM. Surgical management of arteriovenous malformations that follow the tentorial ring. Neurosurgery 1986;18:708-15. Stein BM. Arteriovenous malformations of the medial cerebral hemisphere and the limbic system. J Neurosurg 1984;60:23-31. Stein BM. Arteriovenous malformations of the cerebral convexities. In: Wilson CB, Stein BM, eds. Intracranial arteriovenous malformations. Baltimore: Williams and Wilkins, 1984:156-83. Sugita K, Takemae T, Kobayashi S. Sylvian fissure arteriovenous malformations. Neurosurgery 1987;21: 7-14. Wilson CB, Martin NA. Deep supratentorial arterio venous malformations. In: Wilson CB, Stein BM, eds. Intracranial arteriovenous malformations. Baltimore: Williams and Wilkins, 1984:184-208. Yasargil MC. Microneurosurgery, Volume IIIB. New York: Georg Thieme Verlag, 1988: 112-5 1. Yeh HS, Kashiwagi S, Tew JM, Berger TS. Surgical management of epilepsy associated with cerebral arteriovenous malformations. J Neurosurg 1990;72:21623.

COMMENTARY

Dr. Malik et al have reported a series of 24 patients who underwent surgical treatment of AVMs of the temporal lobe over a 13-year period. The locations of the malformations are divided into three groups: convexity, mesotemporal, and intraventricular. The clinical presentation, size, vascular supply and drainage, preoperative neurologic exam, and last follow-up exam are reported. The surgical morbidity is compared to 132 patients who underwent treatment of nontemporal lobe AVMs over the same time period. The authors briefly discuss the use of preoperative embolization and the associated morbidity, as well as the choice of operative approach for the three groups of temporal lobe lesions. Not surprisingly, most of the patients in this series presented with either seizures (46%) or hemorrhage (29%) while the rest were discovered either incidentally (8%) or as a result of headaches (17%). The operative morbidity is roughly similar to the morbidity associated with nontemporal AVMs in the cohort series offered for comparison. Embolization in the temporal lobe AVM patients was used in 7 patients and carried a 30% morbidity. This figure is high and may be related to embolization of the anterior choroidal artery and deep perforate feeders from the PCA. As a rule, we do not embolize large feeding branches from the anterior or middle temporal branches of the MCA, if they can be accessed surgically by opening the sylvian fissure, and usually do not attempt catheterization or em-