microsurgical and angiographic anatomy of middle cerebral artery aneurysms

ONS344 | VOLUME 62 | OPERATIVE NEUROSURGERY 2 | MAY 2008 www.neurosurgery-online.com

Surgical Anatomy and Technique

MICROSURGICAL AND ANGIOGRAPHIC ANATOMY OFMIDDLE CEREBRAL ARTERY ANEURYSMS: PREVALENCEAND SIGNIFICANCE OF EARLY BRANCH ANEURYSMS

OBJECTIVE: To determine the prevalence of early branch aneurysms, characterize theselesions angiographically and anatomically, and determine their clinical significance.METHODS: The authors conducted a retrospective review of 125 consecutive patientswith a diagnosis of middle cerebral artery (MCA) aneurysm. Eighty-four patients har-boring 100 MCA aneurysms were studied; 41 patients were excluded for lack of ade-quate imaging or for fusiform morphology of the aneurysm. Demographic characteris-tics including age, side, sex, subarachnoid hemorrhage, intracerebral hematoma, multipleaneurysms, and type of treatment were obtained.RESULTS: The average patient age was 57.3 years (range, 29–79 yr); 69 were women and15 were men. Fifty-eight were right MCA aneurysms and 42 were left aneurysms. Fourteenpatients had multiple MCA aneurysms. Thirty-nine of 100 aneurysms were associatedwith subarachnoid hemorrhage. Twelve of 100 aneurysms were associated with an intrac-erebral hematoma. The average aneurysm sizes were 9.1 mm overall (range, 2.0–27.0mm), 12.3 mm for ruptured aneurysms, and 7.5 mm for unruptured. There were 36 M1bifurcation aneurysms, 39 early frontal branch aneurysms, 18 early temporal branchaneurysms, four lenticulostriate artery aneurysms, and three trifurcation aneurysms.CONCLUSION: In our retrospective review, the majority of MCA aneurysms arosealong the M1 segment proximal to the M1 bifurcation. Early frontal branch aneurysmswere more common than typical M1 segment bifurcation aneurysms. M1 segmentaneurysms arising from early frontal and early temporal branches have distinct anatomicfeatures that impact surgical management and outcome. Understanding the relationshipbetween the recurrent lenticulostriate arteries arising from the proximal segments ofthese early branches and the aneurysm neck should allow surgeons to avoid many post-operative ischemic complications when dealing with these challenging lesions.

KEY WORDS: Anatomy, Aneurysm, Angiography, Microsurgery, Middle cerebral artery

Neurosurgery 62[ONS Suppl 2]:ONS344–ONS353, 2008 DOI: 10.1227/01.NEU.0000310700.14628.AE

VASCULAR

Arthur J. Ulm, M.D.Georgia Neurosurgical Institute,Mercer University School of Medicine,Macon, Georgia, andDepartment of Neurological Surgery,Vanderbilt University Medical Center,Nashville, Tennessee

Gregory L. Fautheree, M.D.Department of Neurological Surgery,University of Florida,Gainesville, Florida

Necmettin Tanriover, M.D.Department of Neurological Surgery,University of Florida,Gainesville, Florida

Antonino Russo, M.D.Georgia Neurosurgical Institute,Mercer University School of Medicine,Macon, Georgia

Erminia Albanese, M.D.Georgia Neurosurgical Institute,Mercer University School of Medicine,Macon, Georgia

Albert L. Rhoton Jr., M.D.Department of Neurological Surgery,University of Florida,Gainesville, Florida

Robert A. Mericle, M.D.Department of Neurological Surgery,Vanderbilt University Medical Center,Nashville, Tennessee

Stephen B. Lewis, M.D.Department of Neurological Surgery,University of Florida,Gainesville, Florida

Reprint requests:Arthur J. Ulm, M.D.,Mercer University School of Medicine,Georgia Neurosurgical Institute,840 Pine Street,Suite 880,Macon, GA 31201.Email: [email protected]

Received, June 20, 2007.

Accepted, January 15, 2008.

Aneurysms arising from the middle cere-bral artery (MCA) account for one-fifthof the total number of intracranial aneu-

rysms and are one of the most common sites ofaneurysm rupture (2, 11). The MCA has themost complex branching pattern of any of themajor intracranial arteries (1, 3, 4, 6, 8, 9, 15,17–19, 22, 23). Its complex anatomy, coupledwith its superficial location, has made surgicalclipping the preferred method of treatment forthese lesions (2, 5, 7, 10–14, 20, 21).

MCA aneurysms have classically beendescribed as occurring at one of six locations:1) the bifurcation of the M1 segment, 2) the M1

trifurcation, 3) the takeoff of the lenticulostriatearteries, 4) the branch site of M1 segment tem-poral arteries, 5) the M1 frontal branch arteries,and 6) distal sites on the M2, M3, and M4 seg-ments (11). In multiple studies, the MCA bifur-cation has been reported as the most commonsite of aneurysm occurrence with percentagesas high as 80 to 90% (11, 13, 16).

Our group has previously detailed the anat-omy of the M1 segment in relation to the pres-ence and character of M1 segment frontal andtemporal arteries, termed early temporal andearly frontal branches. In the 159 specimensexamined anatomically or angiographically,

NEUROSURGERY VOLUME 62 | OPERATIVE NEUROSURGERY 2 | MAY 2008 | ONS345

ANATOMY OF MIDDLE CEREBRAL ARTERY ANEURYSMS

early temporal branches were identified in more than 90% ofcases, and early frontal branches were found in more than 30%(22). Many of these M1 segment arteries were large and associ-ated with recurrent lenticulostriate arteries.

Angiographically, many of these proximal early temporaland early frontal branches resembled M1 postbifurcation infe-rior or superior trunks. There have been previous studies andreports detailing M1 segment aneurysms (13, 14, 24). Nonethe-less, a detailed description of these aneurysms is still lacking.The goal of the current study is to determine the prevalence ofearly branch aneurysms, characterize these lesions angiograph-ically, and determine the clinical significance of theseaneurysms in relation to MCA aneurysms as a whole.

MATERIALS AND METHODS

Retrospective ReviewThe anatomic data discussed in the present article are from a pre-

viously published article detailing the microanatomy of early corti-cal branches of the M1 segment of the MCA (22). The MCA wasexamined using 3� to 40� magnification in 50 cerebral hemispheresobtained from 25 adult cadavers. The arteries and veins were per-fused with colored latex to facilitate dissection, aid in defining thesmall-caliber perforating arteries, and improve photographic detail.

The cadaveric heads were placed in the position used for the stan-dard pterional approach, and the sylvian fissure was opened widelyby using microsurgical techniques to expose the proximal portion ofthe MCA and its early branches. The early branches were character-ized according to their sites and patterns of origin, diameters, rela-tive proximity to the internal carotid artery bifurcation, and numbersof lenticulostriate branches. The course and branching patterns ofcortical arteries arising from the early branches were also character-ized (Fig. 1).

Between 2001 and 2004, we retrospectively reviewed a total of 125consecutive patients with a diagnosis of MCA aneurysm. Patientswere excluded if imaging studies could not be obtained, or if theiraneurysms were fusiform in morphology. All remaining catheterangiogram or computed tomographic angiogram images were indi-vidually reviewed. Demographic characteristics, including age, sideon which the aneurysm was located, sex, presence of subarachnoidhemorrhage, location of intracerebral hematoma, presence of multipleaneurysms, and type of treatment, were retrospectively obtained froma chart review.

Based on the previously reported classification scheme (22), the spe-cific location of each MCA aneurysm was determined on each angio-graphic image. Each aneurysm was classified by location as 1) M1 seg-ment bifurcation, 2) M1 segment trifurcation, 3) lenticulostriate artery,4) early temporal branch, 5) early frontal branch, or 6) distal M2, M3,or M4 segments. Aneurysm size in millimeters, as well as aneurysmmorphology, was recorded.

FIGURE 1. Images showing anatomic dissection of cadaveric brain. A, coronaldissection demonstrating the course of the middle cerebral artery (MCA) withinthe sphenoidal and insular compartments of the sylvian fissure. The M1 pre- andpostbifurcation trunks run within the sphenoidal compartment. At the limeninsula, the artery turns to run posteriorly within the insular compartment. Thegenu marks the division between the M1 and M2 segments of the artery. TheM2 arteries give off branches to the lateral cortex, which course over the frontal,parietal, and temporal opercula. The opercular portions of the MCA correspondto the M3 segments. B, left pterional exposure demonstrating the internalcarotid artery (ICA) bifurcation into the anterior cerebral artery (ACA) andMCA. The M1 prebifurcation segment of the MCA extends from the ICA bifur-cation to the MCA bifurcation and runs in the sphenoidal compartment of thesylvian fissure. The M1 segment frequently gives off cortical branches before itsbifurcation. These branches are early frontal and early temporal arteries.Lentriculostriate arteries (LSAs) often arise from the proximal segment of theseearly branches. The M1 segment continues a variable distance as M1 postbifur-cation trunks before the genu at the level of the limen insula. At the limeninsula, the postbifurcation trunks turn to run along the surface of the insula asthe M2 segment. The M2 segments run within the insular compartment of thesylvian fissure. The inset shows the patient orientation. CNIII, oculomotornerve; PCA, posterior cerebral artery; M1, M1 segment of MCA; M2, M2 seg-ment of MCA; M3, M3 segment of MCA; Genu, genu of MCA; Int.Cer.V.,internal cerebral vein; Optic Tr., optic tract; A1, A1 segment of ACA; Chiasm,optic chiasm; Limen Ins., limen insula; Tent., tentorium; Post.clin., posteriorclinoid; Early temp. Br., early temporal branch; M.C.A bif., MCA bifurcation;Inf. Tr., inferior trunk; Sup. Tr., superior trunk; Olf.Tr., olfactory tract;Sphenoidal, sphenoidal compartment of sylvian fissure; Insular, insular com-partment of sylvian fissure; CN II, optic nerve. (A, from, Tanriover N,Kawashima M, Rhoton AL, Ulm AJ, Mericle RA: Microsurgical anatomy of theearly branches of the middle cerebral artery: Morphometric analysis and classi-fication with angiographic correlation. J Neurosurg 98:1277–1290, 2003 [22].)

A

B


ULM ET AL.

The size was measured directly using image processing software onthe hospitalwide Picture Archiving and Communication System, whenavailable, or estimated based on the size of the internal carotid artery(ICA) as it enters the petrous carotid canal (5 mm) when viewed onhard-copy angiograms. The distance between the ICA bifurcation andthe aneurysm origin was measured in millimeters. The direction of theaneurysm dome was recorded in two planes; on anteroposteriorangiography, the dome direction was denoted as being superior(toward the frontal lobe), inferior (toward the temporal lobe), or neu-tral (in line with the sylvian fissure); on lateral angiography, the domedirection was classified as anteriorly projecting (in front of the line ofthe M1 segment), posteriorly projecting (behind the line of the M1 seg-ment), or neutral (not projecting either anterior or posterior to the lineof the M1 segment).

For aneurysms associated with an early M1 branch (lenticulostriateartery, early frontal branch, early temporal branch), additional angio-graphic characteristics were recorded. The location of the aneurysmalong the M1 segment between the ICA bifurcation and the genu of theMCA was quantified as being along either the proximal or the distal halfof the segment. The size of the early branch was recorded as being lessthan 50% or 50% or greater of the caliber of the M1 artery on anteropos-terior angiography.

RESULTS

DemographicsThe average age in our patient population was 57.3 years

(range, 29–79 yr). There were 69 women and 15 men. Theaneurysm was on the right side in 58 cases and on the left sidein 42 cases. Twelve women and one man had two MCAaneurysms, and one woman had three. Thirty-eight of 125patients (30%) had multiple aneurysms (in any intracerebrallocation). Subarachnoid hemorrhage was associated with 39 of100 included MCA aneurysms (39.0%). Intracerebral hematomawas associated with 12 of 39 ruptured aneurysms (30.8%).

Angiographic FeaturesThe average aneurysm size was 9.1 mm (range, 2.0–27.0

mm). The average size of ruptured aneurysms was 12.3 mm,and of unruptured aneurysms 7.5 mm. Of the 100 includedMCA aneurysms, we found 36 M1 bifurcation aneurysms (Fig.2), 39 early frontal branch aneurysms (Fig. 3), 18 early tempo-ral branch aneurysms (Fig. 4), four lenticulostriate arteryaneurysms (Fig. 5), and three M1 trifurcation aneurysms.

Of the 100 MCA aneurysms, the direction of the aneurysmdome in the anteroposterior plane for all aneurysms wasfrontal in 45 (45%), temporal in 39 (39%), and neutral in 15(15%), and was obscured by clip artifact in one case (1%). Thedirection of the dome in the lateral plane was anterior in 33(33%), posterior in seven (7%), and neutral in 59 (59%), and onewas obscured by clip artifact (1%).

For the 36 aneurysms occurring at the true M1 bifurcation,the aneurysm dome in the anteroposterior plane projectedfrontally in six cases (16.7%). Twenty were temporal in projec-tion (55.6%), and nine aneurysms projected down the line ofthe sylvian fissure (25.0%). In one case, clip artifact obscuredthe direction of the dome (2.8%). In the lateral plane, 11 pro-

jected anteriorly (30.6%), one projected posteriorly (2.8%), 23were neutral (63.9%), and one was obscured by clip artifact(2.8%) (Fig. 6).

FIGURE 2. A, left internal carotid angiogram demon-strating an MCA bifurcation aneurysm. The aneurysmoccurs at the point of division of the M1 segment intoterminal superior and inferior trunks. The majority ofbifurcation aneurysms occur in close proximity to thelimen insula and genu of the MCA. B, right internalcarotid angiogram showing another example of a typi-cal MCA bifurcation aneurysm. The aneurysm occursat the division of the short M1 segment into a terminalsuperior and inferior trunk. Note an early temporalbranch originating off of the proximal M1 trunk withinthe sphenoidal compartment of the sylvian fissure. M1,M1 segment of MCA; A1, A1 segment of anterior cere-bral artery; Genu, genu of MCA; Bif., bifurcation ofMCA; An., aneurysm; M1 pre-bif, prebifurcation M1segment; M1 post-bif, postbifurcation M1 segment.

A

B

Most of the 39 early frontal branch aneurysms projectedsuperiorly toward the frontal lobe in the anteroposterior angio-graphic plane. Thirty-three of 39 aneurysms projected frontally(84.6%), six were in line with the sylvian fissure (15.4%), and noearly frontal branch aneurysms projected inferiorly toward thetemporal lobe. In the lateral angiographic plane, 11 early frontalbranch aneurysms projected anteriorly (28.2%), four projectedposteriorly (10.3%), and 24 were neutral (61.5%) (Fig. 3).

In contrast to early frontal branch aneurysms, the majority ofearly temporal branch aneurysms projected toward the tempo-ral lobe. In the anteroposterior plane, 16 of 18 early temporalbranch aneurysms projected temporally (88.9%), and twoaneurysm domes projected toward the frontal lobe (11.1%). Inthe lateral plane, eight projected anteriorly (44.4%), one projectedposteriorly (5.6%), and nine projected neutrally (50.0%) (Fig. 4).

Among the four lenticulostriate artery early branchaneurysms, in the anteroposterior plane, three projectedfrontally (75.0%) and one projected temporally (25.0%). In thelateral plane, two were anterior (50.0%), one was posterior(25.0%), and one projected neutrally (25.0%) (Fig. 5).

Of the three trifurcation aneurysms, in the anteroposteriorplane, the aneurysm dome projected toward the frontal lobe inone case (33.3%) and toward the temporal lobe in two cases(66.7%). In the lateral plane, one projected anteriorly (33.3%)and two were neutral in projection in relation to the M1 line(66.7%) (Fig. 6; Table 1).

The average distance from the ICA bifurcation to the proximalneck for all aneurysms measured was 15.5 mm. The average dis-tance between the ICA bifurcation and the proximal neck wassignificantly less for all types of M1 segment early branchaneurysms than the average distance in typical bifurcationaneurysms. For typical M1 bifurcation aneurysms, the averagedistance was 20.2 mm; for early frontal branch aneurysms, 13.1mm; and for early temporal branch aneurysms, 12.7 mm.

With regard to location along the M1 segment, early frontalbranch aneurysms were located along the proximal half of theM1 segment in 19 of 31 aneurysms (48.7%) and along the distalhalf in the remaining 20 cases (51.3%). Early temporal branchaneurysms were located along the proximal half in 11 aneurysms(61.1%) and the distal half in seven (38.9%). Lenticulostriate

FIGURE 3. A, anatomic dissection of a left cadaveric hemisphere (pterionalexposure) demonstrating a complex M1 segment branching pattern with anearly frontal branch originating from the proximal M1 segment, an early tem-poral branch arising from the distal M1 segment, and the M1 bifurcationoccurring at the level of the genu and limen insula. Note the lenticulostriatearteries arising from the proximal portion of the early frontal branch. B, leftICA angiogram demonstrating an early frontal branch aneurysm arising fromthe distal one-half of the M1 segment. The aneurysm is located within thesphenoidal compartment of the sylvian fissure, proximal to the MCA bifurca-tion, genu, and limen insula. The dome of this aneurysm projects superiorlytoward the frontal lobe. At surgery, the dome of the aneurysm was partiallyburied beneath the surface of the limen insula. The patient imaged in B awokewith a partial right-sided hemiparesis. C, postoperative computed tomographicscan showing an ischemic event within the left internal capsule. The injurywas thought to be secondary to the incorporation of an early frontal branchLSA into the clip during reconstruction. D, right ICA angiogram demonstrat-ing a proximal early frontal branch aneurysm. The caliber of the early frontalbranch is similar to the parent M1 segment. This aneurysm has been previ-ously classified as an MCA bifurcation aneurysm associated with a short M1prebifurcation segment. However, note the bifurcation of the M1 segment intoa superior and inferior trunk at the typical location near the genu and limen

insula, distal to the aneurysm. Thisaneurysm projects superiorlytoward the frontal lobe, is locatedwithin the sphenoidal compartmentof the sylvian fissure, and is proxi-mal to the genu and limen insula.ETB, early temporal branch; M2,M2 segment of the MCA; limen,limen insula; Bif., MCA bifurca-tion; M1, M1 segment of MCA;CN II, optic nerve; EFB, earlyfrontal branch; Fr.lobe, frontal lobe;Genu, genu of the MCA; An.,aneurysm; (sup.tr.) M2, M2 seg-ment, superior trunk; (inf.tr.) M2,M2 segment, inferior trunk; A1, A1segment of anterior cerebral artery; A2, A2 segment of anterior cerebral artery.(A, from, Tanriover N, Kawashima M, Rhoton AL, Ulm AJ, Mericle RA:Microsurgical anatomy of the early branches of the middle cerebral artery:Morphometric analysis and classification with angiographic correlation.J Neurosurg 98:1277–1290, 2003 [22].)

A B C

D




ULM ET AL.

artery aneurysms were located along the proximal half in twocases (50.0%) and the distal half in two cases (50.0%).

The caliber of the early frontal and early temporal branchesharboring aneurysms was compared with the caliber of the M1

segment. Among the early frontal branch aneurysms, 21 weregreater than 50%, whereas 18 were less than 50% the caliber oftheir M1 segments. Twelve of 18 early temporal branches weregreater than 50% and six were less than 50% of the M1 caliber.

FIGURE 4. A, anatomic dissection of a right MCA. A large early tempo-ral branch can be seen originating from the proximal M1 segment. Thebifurcation is at the level of the limen insula and genu of the MCA. SeveralLSAs can be seen coming off the proximal portion of the early temporalbranch. B, right ICA angiogram demonstrating an aneurysm at the originof a large early temporal branch. The aneurysm is within the sphenoidalcompartment of the sylvian fissure and is located proximal to the genu andMCA bifurcation. C, left ICA angiogram demonstrating a proximal M1segment early temporal branch MCA aneurysm. The aneurysm dome is

clearly delineated from the parent vasculature on this anteroposterior pro-jection. CN II, optic nerve; M1, M1 segment of the MCA; Genu, genu ofthe MCA; ETB, early temporal branch; Temp. Lobe, temporal lobe; Bif.,MCA bifurcation; An., aneurysm; A1, A1 segment of the anterior cerebralartery; A2, A2 segment of the anterior cerebral artery. (A, from, TanrioverN, Kawashima M, Rhoton AL, Ulm AJ, Mericle RA: Microsurgicalanatomy of the early branches of the middle cerebral artery: Morphometricanalysis and classification with angiographic correlation. J Neurosurg98:1277–1290, 2003 [22].)

FIGURE 5. A, anatomic dissection demonstrating an M1 segment thatdivides into three terminal trunks. The trifurcation occurs just proximal tothe genu at the level of the limen insula. An early temporal branch can be seenoriginating along the proximal M1 segment. Ten to 15% of anatomic speci-mens have an M1 segment that terminates into a trifurcation. B, left ICAangiogram in a patient with an LSA aneurysm. The dome of the aneurysm isprojecting superiorly. The aneurysm is located within the sphenoidal compart-ment of the sylvian fissure and arises from the proximal portion of the M1

segment. The bifurcation and genu of the MCA are distal to the aneurysm. C,left ICA angiogram showing another example of an LSA aneurysm project-ing superiorly. Fr. Lobe, frontal lobe; MCA tri., MCA trifurcation; ETB,early temporal branch; M1, M1 segment of MCA; M2, M2 segment of MCA;A1, A1 segment of the anterior cerebral artery; A2, A2 segment of the ante-rior cerebral artery; CNII, optic nerve; Genu, genu of the MCA; Bif., MCAbifurcation; An., aneurysm.

A B C

A B C

Intracerebral HematomasThere were a total of 12 intracerebral hematomas associated

with 39 ruptured MCA aneurysms, for a hematoma rate of30.8%. Seven of 36 bifurcation aneurysms were associated withan intracerebral hematoma (19.4%). Three of the hemorrhageswere localized in the frontal lobe and four were temporal lobeclots. One of 39 early frontal branch aneurysms had an associ-ated intracerebral hematoma (2.6%), and it was frontal in loca-tion. Three early temporal branch aneurysms had an associatedintracerebral hematoma (16.7%), and all were located in thetemporal lobe. One lenticulostriate artery aneurysm (25.0%)was associated with an intracerebral hematoma, and the clotwas localized to the temporal lobe.

DISCUSSION

MCA aneurysms are a frequent cause of aneurysmal sub-arachnoid hemorrhage and are among the most commonlyclipped aneurysms (2, 11). Despite the importance of MCAaneurysms to neurosurgeons, the literature detailing themicroanatomy of these lesions is sparse. The current nomencla-ture of MCA aneurysms is confusing, often overlapping, andlacking in descriptive anatomic detail. The literature is largelydevoid of references to early frontal branch aneurysms as adistinct entity.

We identified five distinct types of saccular MCA aneurysmswith characteristic angiographic and anatomic features (Fig. 6;Table 1). A sixth type of saccular MCA aneurysm, distal to thebifurcation, has been reported. However, in our retrospectivereview, all the aneurysms located distal to the bifurcation wereof a fusiform morphology and were excluded.

Early Frontal Branch AneurysmsSurprisingly, the most common type of MCA aneurysm

encountered in our review arose from the M1 trunk and wasassociated with an early frontal branch (Fig. 3). Yasargil (24)noted that lenticulostriate arteries (LSAs) usually arise fromthe prebifurcation trunk of the M1 segment and consideredthose large cortical branches arising proximal to the most lat-eral LSA to be early branches, referring to their origin as a“false early bifurcation.” The location of the LSAs is importantin defining the site of the main bifurcation, because the mainbifurcation is usually located distal to the origin of the LSAs.

Early frontal branches occur anatomically in only 32% ofspecimens, and yet they accounted for the majority of MCAaneurysms in our retrospective review (22). Earlier reports havesuggested that all together, M1 segment aneurysms make upless than 10% of the total number of MCA aneurysms (2, 11, 13).Early frontal branch MCA aneurysms have been poorlydescribed in the literature and have escaped distinct classifica-tion in many cases. We believe that many of these aneurysmshave been classified as proximal M1 bifurcation aneurysms oras lenticulostriate aneurysms in the past. Early frontalbranches, especially ones that originate from the proximal halfof the M1 segment, are often very large in caliber and resemblepostbifurcation M1 trunks (Fig. 3D). However, the true bifurca-tion lies lateral to the takeoff of these branches and is definedas the terminal division of the M1 segment into a superiortrunk supplying the frontal and superior parietal lobes and aninferior trunk supplying the temporal and inferior parietallobes. In other words, a true postbifurcation inferior trunk willnot give off a second, more distal frontal-lobe arterial branch.

This seemingly semantic point implies important surgical con-siderations, particularly with regard to the relationship of LSAsto the aneurysm neck, location of the aneurysm, and surgicalapproach. LSAs were found to arise from 81% of early frontalbranches in our previous anatomic study (22). These LSAs orig-inate from the proximal portions of these arteries, run in a recur-rent, inferior direction toward the anterior perforated substance,and are often hidden behind the aneurysm dome (Fig. 3A). They

FIGURE 6. A, bifurcation aneurysms: 20 of 36 aneurysms projectedtoward the temporal lobe, nine of 36 projected down the line of the sylvianfissure, and six of 36 projected toward the frontal lobe. B, trifurcationaneurysms: two of three aneurysms projected toward the frontal lobe andone of three projected toward the temporal lobe. C, early frontal branchaneurysms: 33 of 39 aneurysms projected toward the frontal lobe, six of 39down the line of the sylvian fissure, and no aneurysms projected towardthe temporal lobe. D, early temporal branch aneurysms: 16 of 18aneurysms projected toward the temporal lobe, and two of 18 aneurysmsprojected toward the frontal lobe. E, LSA aneurysms: three of four pro-jected toward the frontal lobe, and one of four projected toward the tempo-ral lobe. F, distal M2, M3, M4: zero of 100 saccular aneurysms were iden-tified. A1, A1 segment of anterior cerebral artery; M2, M2 segment ofMCA; EFB, early frontal branch; Car.A., ICA; M1, M1 segment of MCA;Genu, genu of MCA; ETB, early temporal branch.

A B

C D

E F




ULM ET AL.

arise from larger, more proximal early frontal branches, but alsofrom smaller, more distal early frontal branches. Failure to recog-nize these hidden early frontal branch LSAs can lead to unex-pected postoperative deficits resulting from LSA infarctions (Fig.3C). In their review of superior-wall M1 segment aneurysms,Iwama et al. (14) reported a postoperative stroke rate of 28%,despite visual inspection to ensure patency of lenticulostriatebranches. On average, these aneurysms are located more proxi-mal to the ICA bifurcation than typical bifurcation aneurysmsare: 13.1 versus 20.2 mm in our study. The overwhelming major-ity of early frontal branch aneurysms project superiorly towardthe frontal lobe (84.6%), whereas a minority of bifurcationaneurysm project toward the frontal lobe (16.7%).

Surgery for Early Frontal Branch AneurysmsThe subtle anatomic differences between early frontal branch

aneurysms and short M1 segment or typical bifurcationaneurysms can have a significant impact on surgical technique.The takeoff of the M1 segment frontal branches is often acute,approaching 90 degrees, combined with the fact that theaneurysm neck is often incorporated into the proximal portionof the early frontal branch; these aneurysms require a compli-cated clip strategy more frequently than typical bifurcationaneurysms. Early frontal branch aneurysms project superiorly,whereas bifurcation aneurysms tend to project down the line ofthe sylvian fissure. This has implications for approaching andexposing these aneurysms. For superiorly projecting earlyfrontal branch MCA aneurysms, we avoid frontal lobe retrac-tion that could result in premature rupture given the projectionof the dome. Splitting the sylvian fissure is different when com-paring typical bifurcation and early frontal branch aneurysms.When approaching early frontal branch aneurysms, we tend tobegin the fissure split laterally, again to avoid the frontal loberetraction that would be required to split from proximal to lat-eral. The lateral fissure is split down to the level of the limeninsula, where the bifurcation is identified. The dissection isthen carried out along the M1 segment avoiding the superior

surface of the M1 segment where the aneurysm neck arises.Enough M1 segment is exposed to allow temporary clippingbefore any frontal lobe retraction that might be required tocomplete the medial fissure. The dissection around theaneurysm neck is critical. The recurrent LSAs that arise fromthe proximal portion of early frontal branches course directlybehind the aneurysm dome and neck. These LSAs take a differ-ent course and direction compared with those associated witha short M1 segment typical bifurcation aneurysm. Early frontalbranch LSAs course inferiorly en route to the anterior perfo-rated substance and lie deep to the aneurysm neck and dome(Fig. 3A). Great care is required to ensure that these arteries arenot included in clip reconstruction. Many early frontal branchaneurysms lie partially hidden from view beneath the limeninsula, sometimes burrowing into the substance of the limen.Adequate head rotation facilitates the exposure of theseaneurysms that are buried beneath the limen.

Early Temporal Branch AneurysmsEarly temporal branch MCA aneurysms accounted for 18%

of identified aneurysms. These aneurysms occur in relation toone of what may be several temporally directed branches aris-ing off of the M1 segment (Fig. 4). Like early frontal branchaneurysms, these aneurysms occur most commonly within thesphenoidal compartment of the sylvian fissure, proximal to thetrue MCA bifurcation, and are associated with LSAs thatemanate from the temporal branch from which they arise (Fig.4A). However, LSAs were found to arise from only 48% of earlytemporal branches in our anatomic study compared with 81%for early frontal branches (22). Early temporal branchaneurysms occurred at an average distance of 12.7 mm fromthe ICA bifurcation, significantly closer than typical bifurca-tion aneurysms. In contrast to early frontal branch aneurysms,89% of these aneurysms were found to project toward the tem-poral lobe. In 12 of 18 early temporal branch aneurysms, thetemporal branch was equal to or greater than the caliber of theM1 segment from which it arose.

a MCA, middle cerebral artery; AP, anteroposterior; LSA, lenticulostriate artery; M1–M2, M1–M2 junction; M1, M1 segment of MCA. Five subtypes of saccular MCAaneurysm were identified in the study: bifurcation, trifurcation, early frontal branch, early temporal branch, and lenticulostriate. The subtypes are distinct with regard tolocation, distance from the internal carotid artery bifurcation, projection of aneurysm dome on AP cerebral angiography, association with lenticulostriate arteries, andsurgical head positioning.b For AP projection of dome in trifurcation aneurysms, there were not enough events to draw a conclusion.

TABLE 1. Summary of middle cerebral artery aneurysm characteristicsa

Location within MCA AP projec- Intimately associ- Surgical headAneurysm type

sylvian fissure segment tion of domeb ated with LSA rotation (degrees)

Bifurcation Junction of sphenoidal M1–M2 56% temporal No 45and insular compartments

Trifurcation Junction of sphenoidal M1–M2 — No 45and insular compartments

Early frontal branch Sphenoidal compartment M1 85% frontal Yes 30–40

Early temporal branch Sphenoidal compartment M1 89% temporal Yes 30–40

Lenticulostriate artery Sphenoidal compartment M1 75% frontal Yes 30–40

Typical Bifurcation and Trifurcation Aneurysms

Typical bifurcation aneurysms occur at the terminal divisionof the M1 segment of the MCA into an inferior and a superiortrunk (Figs. 1B and 2). According to previous anatomic studies,the MCA bifurcation occurs at or distal to the genu in 94% ofcases and occurs within 10 mm of the limen insula (22). In ourprevious examination of 50 hemispheres, all of the bifurcationswere located between 12.9 mm proximal and 20.4 mm distal tothe limen insula, with an average distance of 9.4 mm. The genuof the MCA was found to be lateral to the limen insula in allcases, with a mean distance of 9.7 mm. In this study, typicalbifurcation aneurysms were located on average 20.2 mm fromthe ICA bifurcation. In the overwhelming majority of cases,these aneurysms were found at the junction of the sphenoidaland insular compartments of the sylvian fissure, at the level ofthe genu. Rarely were LSAs found associated with the M1bifurcation. In the previous anatomic study, only one of 476LSAs was found arising directly from the bifurcation (22). Themajority of true bifurcation aneurysms lay lateral to the LSAs,which arose from the more proximal M1 segment (Fig. 5A).Most bifurcation aneurysms (20 of 35) were directed toward thetemporal lobe, but a significant number (nine of 35) weredirected down the line of the sylvian fissure, and a minority(six of 35) toward the frontal lobe.

A minority of bifurcation aneurysms were associated with ashort prebifurcation M1 segment. These short M1 segmentbifurcation aneurysms had a more proximal location withinthe sphenoidal compartment of the sylvian fissure and residedin close proximity to LSAs. The postbifurcation M1 segmentswere a frequent site of origin for LSAs; 15% of LSAs arose frompostbifurcation M1 trunks in our previous study (22).Anatomically, MCA bifurcation aneurysms will most likely beintimately associated with LSAs when there is a short M1 preb-ifurcation segment, when the aneurysm is large, or when thedome projects toward the frontal lobe.

CONCLUSIONS

In contrast to previous reports, M1 segment aneurysms madeup the majority of MCA aneurysms in our retrospective reviewof 100 consecutive cases. Early frontal branch aneurysms weremore common than typical bifurcation aneurysms. We believethat many early frontal branch aneurysms, particularly thoseassociated with a large-caliber arterial branches, have been clas-sified as short M1 segment bifurcation aneurysms in the past.Early branch aneurysms have distinct anatomic features thatimpact surgical management. The projection of the dome, loca-tion within the sylvian fissure, and relationship to LSAs aredifferent for early branch aneurysms, as compared with typicalor short prebifurcation MCA aneurysms. The location of theaneurysm along the M1 segment affects surgical planning,head rotation, direction of sylvian fissure splitting, sequence ofM1 exposure during surgical dissection, safety of frontal versustemporal lobe retraction, location of critical perforating arteries,and clipping strategy. It is hoped that an improved understand-

Surgery for Early Temporal Branch Aneurysms

Again, subtle anatomical differences between early temporalbranch aneurysms and typical or short M1 segment bifurcationaneurysms have a significant impact on the surgical manage-ment of these lesions. Early temporal branch aneurysms arisemore proximal and require head rotation that facilitates expo-sure of the M1 segment. These aneurysms, like the early frontalbranch aneurysms, are associated with LSAs that are posi-tioned and oriented differently from short M1 segment bifurca-tion aneurysms. The early temporal branches from which theseaneurysms arise are associated with recurrent, superiorly pro-jecting LSAs that are often hidden behind the aneurysm dome(Fig. 4A). Great care must be taken to avoid injury to these“misplaced” LSAs. Because the overwhelming majority ofthese aneurysms project toward the temporal lobe, we avoidtemporal lobe retraction when exposing these lesions beforeobtaining proximal control. When splitting the fissure for theseaneurysms, we often approach with a medial to lateral direc-tion, which allows early proximal control and avoids temporallobe retraction. The inferior surface of the M1 segment isavoided until after proximal control has been obtained, toavoid the origin of the aneurysm neck.

Surgery for M1 Segment AneurysmsEarly frontal branch, early temporal branch, lenticulostriate,

and short M1 segment bifurcation aneurysms all arise withinthe sphenoidal portion of the sylvian fissure (Fig. 1A). Weapproach M1 segment MCA aneurysms in much the samemanner as we do ICA bifurcation aneurysms. Slightly less headrotation provides improved access to the more proximal M1segment: 30 degrees, as compared with the 45 degrees for typ-ical bifurcation aneurysms.

Temporary clipping of the proximal M1 requires extra care toavoid injury to perforators arising from the M1 segment.Identifying a segment of the M1 segment that is circumferen-tially free of perforating vessels is critical to avoiding ischemiccomplications. Care must be taken to avoid not only the originsof these vessels arising from the M1 segment, but also therecurrent LSAs that arise from the proximal segments of theearly frontal and early temporal branches.

Intracerebral HematomasAneurysms associated with intracerebral hematomas are

often difficult to expose as a result of the associated mass effect.This is particularly true for M1 segment aneurysms that have adeeper location within the sphenoidal compartment of the fis-sure. We tend to assess each case of hematoma individually. Ifthe mass effect is minimal or can be overcome with cere-brospinal fluid diversion, then we secure the aneurysm withoutclot removal. If the mass effect requires excessive retraction togain exposure, then we decompress the hematoma transcorti-cally and then assess the sylvian fissure. If there is still signifi-cant mass effect after evacuation, which prevents sylvian fis-sure dissection, then subpial dissection through the hematomacavity is used to expose the aneurysm.




ULM ET AL.

ing of these differences will allow neurosurgeons to avoid com-plications when dealing with theses lesions. In future studies,we will investigate what importance these differences havewith regard to surgical versus endovascular therapy and clini-cal outcome.

REFERENCES

1. Aydin IH, Takçi E, Kadioglu HH, Kayaoglu CR, Tüzün Y: The variations oflenticulostriate arteries in the middle cerebral artery aneurysms. ActaNeurochir 138:555–559, 1996.

2. Chyatte D, Porterfield R: Nuances of middle cerebral artery aneurysm micro-surgery. Neurosurgery 48:339–346, 2001.

3. Ciszek B, Aleksandrowicz R, Zabek M, Mazurowski W: Classification, topog-raphy and morphometry of the early branches of the middle cerebral artery.Folia Morphol 55:229–230, 1996.

4. De Long WB: Anatomy of the middle cerebral artery: The temporal branches.Stroke 4:412–418, 1973.

5. De Sousa AA, Filho MA, Faglioni W Jr, Carvalho GT: Unilateral pterionalapproach to bilateral aneurysms of the middle cerebral artery. Surg Neurol63 [Suppl 1]:S1–S7, 2005.

6. Duvernoy HM, Delon S, Vannson JL: Cortical blood vessels of the humanbrain. Brain Res Bull 7:519–579, 1981.

7. Forget TR Jr, Benitez R, Veznedaroglu E, Sharan A, Mitchell W, Silva M,Rosenwasser RH: A review of size and location of ruptured intracranialaneurysms. Neurosurgery 49:1322–1326, 2001.

8. Gibo H, Carver CC, Rhoton AL Jr, Lenkey C, Mitchell RJ: Microsurgicalanatomy of the middle cerebral artery. J Neurosurg 54:151–169, 1981.

9. Grand W: Microsurgical anatomy of the proximal middle cerebral artery andthe internal carotid artery bifurcation. Neurosurgery 7:215–218, 1980.

10. Han DH, Gwak HS, Chung CK: Aneurysm at the origin of accessory middlecerebral artery associated with middle cerebral artery aplasia: Case report.Surg Neurol 42:388–391, 1994.

11. Heros RC, Fritsch MJ: Surgical management of middle cerebral arteryaneurysms. Neurosurgery 48:780–786, 2001.

12. Heros RC, Ojemann RG, Crowell RM: Superior temporal gyrus approach tomiddle cerebral artery aneurysms: Technique and results. Neurosurgery10:308–313, 1982.

13. Hosoda K, Fujita S, Kawaguchi T, Shose Y, Hamano S: Saccular aneurysms ofthe proximal (M1) segment of the middle cerebral artery. Neurosurgery36:441–446, 1995.

14. Iwama T, Yoshimura S, Kaku Y, Sakai N: Considerations in the surgical treat-ment of superior-wall type aneurysm at the proximal (M1) segment of themiddle cerebral artery. Acta Neurochir 146:967–972, 2004.

15. Komiyama M, Nakajima H, Nishikawa M, Yasui T: Middle cerebral artery vari-ations: Duplicated and accessory arteries. Am J Neuroradiol 19:45–49, 1998.

16. Krayenbul HA, Yasargil MG: Cerebral Angiography. Philadelphia, J.B.Lippincott Co., 1968, p 185.

17. Lasjaunias P, Berenstein A: Surgical Neuroangiography. New York, Springer-Verlag, 1992, vol 3, pp 142–148.

18. Marinkovic SV, Kovacevic MS, Marinkovic JM: Perforating branches of themiddle cerebral artery: Microsurgical anatomy of their extracerebral seg-ments. J Neurosurg 63:266–271, 1985.

19. Marinkovic SV, Milisavljevic MM, Kovacevic MS, Stevic ZD: Perforatingbranches of the middle cerebral artery: Microanatomy and clinical signifi-cance of their intracerebral segments. Stroke 16:1022–1029, 1985.

20. Martellotta N, Gigante N, Toscano S, Maddalena GF, Tripodi M, SettembriniG, Stroscio C, Distefano G, Citro E: Unilateral supraorbital keyhole approachin patients with middle cerebral artery (M1–M2 segment) symmetricalaneurysms. Minim Invasive Neurosurg 46:228–230, 2003.

21. Pásztor E, Vajda J, Juhász J, Tóth S, Orosz E, Horváth M: The surgery of mid-dle cerebral artery aneurysms. Acta Neurochir 82:92–101, 1986.

22. Tanriover N, Kawashima M, Rhoton AL Jr, Ulm AJ, Mericle RA: Microsurgicalanatomy of the early branches of the middle cerebral artery: Morphometricanalysis and classification with angiographic correlation. J Neurosurg98:1277–1290, 2003.

23. Umansky F, Juarez SM, Dujovny M, Ausman JI, Diaz F, Gomes F,Mirchandani HG, Ray WJ: Microsurgical anatomy of the proximal segmentsof the middle cerebral artery. J Neurosurg 61:458–467, 1984.

24. Yasargil MG: Microneurosurgery. Stuttgart, Georg Thieme Verlag, 1984, vol 1,pp 184–308.

COMMENTS

This thought-provoking article examines the prevalence and signifi-cance of early branch aneurysms of the middle cerebral artery

(MCA) in a series of 84 patients with 100 aneurysms. Aneurysms werecategorized into six types according to their anatomical origin: MCAbifurcation, MCA trifurcation, early frontal branch, early temporalbranch, lenticulostriate artery, or distal M2–M4 segments. The moststriking finding was that the early frontal branch aneurysm was themost common type (39 aneurysms), even more common than the MCAbifurcation aneurysm (36 aneurysms). Considering that early frontalarteries are observed anatomically in only one-third of cadaver speci-mens whereas MCA bi- or trifurcations are seen in all specimens, thisfinding suggests an unusual predisposition to aneurysm formation atthe early frontal branch, which is difficult to explain. The authors con-cluded that early frontal branch aneurysms are more common thanexpected and that perhaps aneurysms at the bifurcation of a short M1segment were not properly named in the past. Alternatively, one couldconclude that early frontal branch aneurysms might have been countedenthusiastically in this review. As an example, the aneurysm in Figure3D in the article might be categorized as a short M1 segment bifurca-tion aneurysm by a less discriminating eye. A decreased prevalence ofearly frontal branch aneurysms would be more consistent with the lowprevalence of early frontal branches generally. Nonetheless, theauthors’ effort here to elevate our level of discrimination is importantbecause early frontal branches often have critical lenticulostriate arter-ies arising from their proximal segment that can be missed. The supe-rior projection of these aneurysms obscures their visualization.Occluded lenticulostriate arteries can produce devastating motor andsensory deficits, so they must be meticulously dissected and preserved.The technical maneuvers recommended by the authors relating to headrotation, direction of Sylvian fissure dissection, frontal lobe retraction,sequence of M1 exposure, and clipping strategy are all valuable. Thiswork challenges us to be more precise with our understanding of MCAanatomy and more analytical with our description of MCA aneurysms.

Michael T. LawtonSan Francisco, California

This anatomical study nicely outlines the common locations for MCAaneurysms. The integration of a clinical series with cadaver dissec-

tions is very helpful in understanding the anatomy of this class ofaneurysms. In contrast with previous studies, the neck region of themajority of MCA aneurysms was proximal to the MCA genu. I havealso noticed this phenomenon in my practice, and I think this is a realfinding and probably not related to referral bias. It is nice to have thispoint documented in a reasonably sized series, but it would be helpfulto resolve this point in a larger nonselected group of MCA aneurysms.

The major relevance of aneurysms proximal to the genu relates to thepotential dangers of surgical intervention. Aneurysms proximal to thelimen insula are much more likely to be intimate with critical lenticu-lostriate arteries. Wide sylvian exposure and careful inspection for hid-den small branches is crucial to successful outcome with more proximalMCA aneurysms.

Robert A. SolomonNew York, New York

Ulm et al. present an interesting series of 100 aneurysms of the MCAthat they use to describe the varying anatomic correlates gleaned

from their previous anatomic dissections as applied to the treatment ofaneurysms of this region. In this evaluation, they describe the prevalenceof early frontal and temporal branches from the MCA and their potentialrelationship with the presence of aneurysms. They conclude that frontalbranch aneurysms of the MCA are more common than true MCA bifur-cation aneurysms and that these carry important clinical correlates.

The crux of this series lies in the fact that recognition of early frontalbranches or early temporal branches as opposed to a true “early bifur-cation” is important in determining various aspects of the surgicalapproach and potential clip choice. In other words, the primary focusto be gleaned from this article is the subtle surgical differences thatwarrant the need to recognize the aforementioned anatomic variants.Recognizing these differences may help avoid injury to perforatingbranches and thus, potentially improve surgical outcome. This is aninteresting anatomic analysis.

Charles J. PrestigiacomoNewark, New Jersey

Ulm et al. present an anatomic study of MCA aneurysms with partic-ular emphasis on those lesions arising from early frontal and early

temporal branches and in so doing continue a long tradition of anatomicstudies by the group from the University of Florida. This study, in con-junction with their previous study on the early branches of the MCA (1),makes a distinction between aneurysms at this location and true bifur-cation aneurysms. Although some may find such a distinction a seman-

tic one only, these early branch M1 segment lesions tend to have a moreintimate association with the lenticulostriate arteries than do their coun-terparts at the true bifurcation. Because of this difference, the neurosur-geon must recognize these early branch aneurysms for what they are andapproach them accordingly at the point of operation. The exclusion of ananeurysm from the circulation while preserving the surrounding vascu-lature remains the primary goal of the aneurysm surgeon, regardless ofthe location of the aneurysm. If an early branch aneurysm is mistaken fora true bifurcation aneurysm, the surgeon might be tempted to be less vig-ilant at the time of neck dissection by assuming that only the associatedbranch and the M1 trunk are at risk. That assumption, as the authorspoint out, could prove disastrous.

Whether (or not) early branch aneurysms make up the majority ofMCA aneurysms remains to be seen as the present study representsonly 100 aneurysms. Further studies would definitely help further sortout this issue. Regardless, the aneurysm surgeon should recognize thedifference between an early MCA branch and the true bifurcation whendealing with an MCA aneurysm.

Jay U. HowingtonSavannah, Georgia

L. Nelson HopkinsBuffalo, New York

1. Tanriover N, Kawashima M, Rhoton AL Jr, Ulm AJ, Mericle RA: Microsurgicalanatomy of the early branches of the middle cerebral artery: morphometricanalysis and classification with angiographic correlation. J Neurosurg98:1277–1290, 2003.



Pioneers of Cancer Treatment: Lymphatic System. Gaspare Aselli (1581–1625) of Milan (left) used dog dissections in defining thelymphatic vessels. Aselli’s work was further clarified in 1651 when Jean Pecquet (1622–1674) demonstrated the thoracic duct in dogs(right). From, Shimkin MB: Contrary to Nature: Being an Illustrated Commentary on Some Persons and Events of Historical Importance inthe Development of Knowledge Concerning Cancer. Washington, D.C., U.S. Department of Health, Education, and Welfare, 1977.

microsurgical and angiographic anatomy of middle cerebral artery aneurysms

Documents