10ournalof Neurology, Neurosurgery, and Psychiatry 1992;55:1040-1045

Cerebral cavernous angioma: a potentially benigncondition? Successful treatment in 16 cases

A Churchyard, M Khangure, K Grainger

AbstractCerebral cavernous angioma (cavern-oma) has previously been treated byresection for all presentations when surgi-cally resectable. In this retrospectiveseries of 16 cases, it is demonstrated that,for those patients presenting with epilepsyalone, surgery is often unnecessary. Diag-nosis can be made radiologically andexcellent seizure control can be obtainedwith medications. In the authors' experi-ence, those cavernomas symptomatic asepilepsy rarely cause major haemorrhageand the need for surgery as prophylaxisagainst bleeding in this group is unpro-ven.

(7 Neurol Neurosurg Psychiatry 1992;55:1040-1045)

Queen Elizabeth IIMedical Centre,Nedlands, WesternAustralia, AustraliaDepartment ofNeurologyA ChurchyardK GraingerDepartment ofMagnetic ResonanceImagingM KhangureCorrespondence to:Dr A Churchyard,Department of Neurology,Austin Hospital, BurgundySt, Heidelberg, Victoria,NSW 3084, AustraliaReceived 24 January 1991and final revision14 April 1992.Accepted 24 April 1992

Cavernomas of the central nervous systemtypically present as intracranial haemorrhage,epilepsy or as a mass lesion.`9 Macroscop-ically these lesions have been likened to a ripemulberry9 and microscopically are seen toconsist of blood-filled cavities lined by a singlelayer of endothelium and separated by neu-

roglia, but not by neural tissue.9 Old haemor-rhage and calcification are common.9 Thereare no other local vascular abnormalities.9

Surgical excision has commonly been repor-ted as appropriate treatment of cavern-oma' 2 4-8 even though its clinical course

remains poorly understood. Clinical series ofthose patients presenting with complicationsmay overestimate the pathogenicity of thecondition. The advent of MRI allows thereliable detection, follow up and diagnosis ofasymptomatic and symptomatic cavernomas,thereby permitting a better understanding ofthe condition's natural course and appropriatemanagement.

In this retrospective review of 16 patients inwhom cavernous angiomas were diagnosedradiologically, we found that this lesion is oftenbenign and produces little disability. In our

opinion, resection is not necessarily the besttreatment of cavernomas presenting as epi-lepsy.

MethodsSixteen consecutive cases of cavernous

angioma diagnosed by MRI at our hospitalwere reviewed retrospectively. The clinical fea-tures, investigations and treatment of eachpatient were gleaned from the hospital case

notes or the records of the consulting physi-cian. All patients with an MRI diagnosis ofcavernoma were reviewed. All MRI scans weretaken with a 1-5 Tesla Philips Gyroscan fromwhich T1 and dual T2 weighted axial andcoronal images were obtained. This particularmachine was the only MRI scanner in the stateand so this series represents all MRI-provencavernomas in Western Australia over thereview period.

Cavernous angiomata were diagnosed whenMRI demonstrated a discrete lesion or lesionsdisplaying central areas of differing signalintensity on Ti and T2 weighted images,indicative ofhaematomas of varying ages, focalfibrosis, focal calcification and a rim of lowintensity at the periphery of the lesion."''3Haemosiderin was detected as areas of lowsignal intensity inT2 weighted images, therebyallowing discrimination from a flow void.""'2The presence of a feeding artery or drainingvein or neural tissue within the malformationwas taken to exclude the diagnosis ofcavernoma.9

This series is based upon the use of an MRIscanner which services the entire state and isused by all neurologists and neurosurgeons,and consequently accurately represents all ormost cavernomas presenting as epilepsy duringthe time of review. The absence of supraten-torial cavernomas presenting with cerebralhaemorrhage in this series presumably reflectsour entry criterion requiring diagnosis by MRIand, as such, will underestimate their inci-dence. Such cavernomas may be lethal, orresected after an initial CT scan only. Theywere not included in this series because of lackof systematic access to necropsy and pathologydata and their incidence in our communityremains unknown. As all cases with an MRIdiagnosis of cavemoma were reviewed, we areconfident that no patient initially presentingwith epilepsy sustained a subsequent, symp-tomatic cerebral haemorrhage.

ResultsPresenting complaint and initial investigationsSixteen cases consisting of eight males andeight females, with an average age at onset of25-5 years and an average duration of seizuresbefore diagnosis of 5'3 years (in the 15symptomatic patients) were studied. Eightpatients had had seizures for three or moreyears. Average follow up after diagnosis wasthree years (more than five years in four).Diagnosis was made by MRI in all andconfirmed histologically in the two patients

1040 on N

ovember 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 N

ovember 1992. D

ownloaded from

Cerebral cavernous angioma: a potentially benign condition?

Figure 1 Cavernoma andadjacent cerebrum (patient5; haematoxylin and eosin;magmfication 4-2) showingfibrous capsule (smallarrow), calcifiedhaematoma (large arrow)and thin-walled vascularchannels withoutintervening neuronal tissue.

who underwent resection (figure 1). CT wasperformed in all cases, and in all but patients 8and 14 revealed the symptomatic lesion. CTsuggested a low grade neoplasm in patients1-4, 9 and 16. Patients 1 and 2 had multiplecavernomas on MRI, although initial CTscanning demonstrated only 1 lesion. Cavern-oma was not definitively diagnosed by CT inany case. Angiography was performed in 10patients and was normal in all but case 2,where a possible tumour circulation wasthought to be indicative of a meningioma.Focal electroencephalographic abnormalitieswere generally localised to the region of thesingle or largest cavernoma.

Thirteen patients presented with epilepsywhich was longstanding in 10. Successfulmanagement with anti-epileptic drugs wasachieved in all but patient 5. Seizures havebeen abolished in patients 3 and 7 and are rare(less than one or two a year) in cases 1, 2, 4, 1 1and 13. In Australia, an unprovoked seizure insuch patients results in the suspension of thedriver's licence for a period of three to sixmonths. All patients with occasional seizurestherefore faced infrequent periods when theycould not drive for up to six months, but thiswas not sufficient to prompt consideration ofexcision, as all patients were able to remain inemployment and maintain normal personaland family lives. Six patients were controlledon carbamazepine alone and one on carbama-zepine and phenytoin without significant

Table 1 Clinical details of each patient with epilepsy

Presenting Age at Time to Follow up SeizurePatient Sex complaint onset (years) diagnosis (years) (years) Treatment control

1 M CPS 14 3 11 PHY/CBZ Rare2 M CPS 44 10 4 CBZ Rare3 F GTC* 45 3 8 CBZ None4 M GTC 15 0 5 CBZ Rare5 M CPS* 11 19 0 33 Excision7 F CPS 14 3 1 CBZ None10 F CPS 18 5 1 PHY ?11 M CPS/GTC 17 1 8 VAIJPHE Rare12 F CPS/GTC 33 2 0 CBZ/excise ?13 F CPS/GTC 22 7 1 CBZ Rare14 M CPS 29 16 0 CBZ ?15 F Episodic dysphasia* 42 0 5 0 CBZ ?16 M CPS 20 2 days 0 5 CBZ None

*Additional migraine; PHY = phenytoin; VAL = valproate; CBZ = carbamazepine; PHE = phenobarbitone.

Table 2 Investigations for each patient with epilepsy

Patient CT MRI Angiogram Histology EEG

1 L temporal calcification Multiple cavernomas Normal ND L temporal slow waves2 L temporal calcification Multiple cavernomas L parietal vascularity ND L temporal slow waves3 L temporal calcification L frontal cavernoma ND ND L temporal epileptiform4 R parietal calcification R parietal cavernoma Normal ND Generalised slow waves5 L temporal calcification L temporal cavernoma Normal Cavernoma L temporal slow waves7 L temporal calcification L temporal cavernoma ND ND Bitemporal epileptiform10 R occipital calcification R occipital cavernoma Normal ND R slow waves11 Unavailable R temporal cavernoma ND ND R temporal epileptiform12 R temporal calcification R temporal cavernoma ND Cavernoma R temporal epileptiform13 R temporal calcification R temporal cavernoma Normal ND Normal14 L subarachnoid cyst R temporal cavemoma ND ND R temporal epileptiform15 L parietal calcification L parietal cavernoma Normal ND Normal16 L frontal calcification L frontal cavernoma ND ND L frontal slow waves

ND No data.

1041 on N

ovember 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 N

ovember 1992. D

ownloaded from

Churchyard, Khangure, Grainger

Table 3 Clinical details of each patient without epilepsy

Presenting Age Time at Follow upPatient Sex complaint (years) diagnosis (years) (years)

6 F Incidental ND 9 18 F Brainstem haemorrhage 33 0 0-259 M Brainstem haemorrhage 50 4 3

ND = No data.

change to their regime over the time sincediagnosis. Patient 10 responded poorly to bothsodium valproate and carbamazepine and hasrecently started phenytoin but its efficacy, as

yet, is unknown. Patient 11 was allergic to bothphenytoin and carbamazepine, but obtainedgood control with sodium valproate andphenobarbitone. Four patients (10, 14-16)have had too short a follow up to allow an

accurate estimation of seizure control. Patient5 required excision of the lesion because of

Figure 2 Axial Tl weighted (TR SE 2310; TE 25 rms) MRI scan from patient Ishowing multiple cavernomas with central area of increased signal intensity andperipheral low intensity.

refractory epilepsy and the presence of a

cavernoma was confirmed. Patient 12 pre-

sented with epilepsy of two years standing andthe cavernoma was excised as prophylaxisagainst haemorrhage although she had had no

history consistent with previous symptomatichaemorrhage. Complete seizure control was

obtained with carbamazepine before surgery.

She has yet to be followed up postoperatively.Patient 6 remains asymptomatic after 10

years follow up. The lesion was found unex-

pectedly after CT to investigate mild mentalretardation and a personality disorder. She hasnever received antiepileptic medications.Two patients presented with haemorrhage

secondary to a brainstem cavernoma and bothhave been managed conservatively because ofthe site of the lesion. Patient 8 survived theinitial haemorrhage with severe residual defi-cits and has had no further episodes. Patient 9suffered a transient neurological episode fouryears before presentation with three months ofprogressively worsening bulbar dysfunctionand a hemiparesis. Serial CT scans showed a

resolving brainstem haemorrhage and he hassurvived with progressively worsening deficitsfor a further three years.The clinical and other details are outlined in

tables 1-4.

Asymptomatic haemorrhage and epilepsyNone of the 13 patients with epilepsy had hadsymptomatic haemorrhages, even though mosthad been symptomatic for many years. In 11 ofthese 13 and in the sole patient with an

asymptomatic cavernoma, MRI demonstrateda past haemorrhage or haemorrhages that hadpresumably been asymptomatic or manifestonly as seizures. Two patients presented within72 hours of their first seizure, but MRIdemonstrated haemosiderin and calcificationindicative of definite asymptomatic haemor-rhage in the past in both.

Case historiesCASE 1

At 14 years of age, this otherwise well maledeveloped complex partial seizures (CPS) andpresented three years later with an average ofone seizure per month. A CT scan showed a

left temporal lesion thought to be an oligoden-droglioma. An EEG showed a left temporalslow wave focus and he was started on pheny-toin and carbamazepine with good controlbeing achieved. Eleven years later he suffered a

flurry of seizures and repeat CT and MRIshowed multiple cavernomas (figure 2). Neu-rological examination remained essentiallynormal. Continued carbamazepine and pheny-toin has maintained excellent control with only

Table 4 Investigations for the patients without epilepsy

Patient CT MRI Angiogram EEG

6 L frontal calcification L frontal cavernoma Normal Generalised slow waves8 Brainstem haemorrhage Cerebellar cavernoma Normal ND9 Brainstem haemorrhage/calcification Brainstem cavernoma Normal ND

ND = No data.

1042 on N

ovember 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 N

ovember 1992. D

ownloaded from

1043Cerebral cavernous angioma: a potentially benign condition?

4) Axial Tl weighted (TR SE 2560; TE 25 rms) and (B) axial T2 weighted (TR SE 2560; iT it)temporal cavernoma with variable, central signal intensity and peripheral low intensity and multiple

infrequent seizures.

CASE 2At the age of 44, this previously well maledeveloped CPS and presented 10 years later.He suffered occasional CPS whilst on car-

bamazepine. An EEG showed a left temporalslow wave focus and a CT scan was thought toshow a left temporal meningioma. MRI per-formed seven years later showed multiplecavernomas and multiple lacunes, but nomeningioma or neoplasm (figure 3).

Figure 4 (A) Axial, non-contrast CT scan showing calcified left frontal lesion and (B) axial T2 weighted (TR SE 2310; TE 25 rms) MRI scan frompatient 3 showing heterogeneous signal intensity centrally and a rim of low intensity at the periphery.

on Novem

ber 16, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 Novem

ber 1992. Dow

nloaded from

Churchyard, Khangure, Grainger

CASE 3This previously well 48 year old womansuffered three nocturnal generalised seizuresover three years. An EEG showed a lefttemporal epileptiform focus and serial CTscans showed an unchanging left frontal lesion.She was started on carbamazepine and hasremained free of seizures. MRI performed 11years after her first seizure showed a cavernousangioma (figure 4).

DiscussionIn this series of 16 patients, 81% presentedwith epilepsy, two (12-5%) presented withbrainstem haemorrhage and one was asymp-tomatic. The average duration of epilepsybefore diagnosis was 5-3 years; three patientshad had epilepsy for more than 10 years andeight for more than three years. No patientwith epilepsy suffered any of the other compli-cations of cavernomas and only one requiredresection for refractory seizures. As our MRIscanner services the entire state, and is used byall neurologists and neurosurgeons, we believethat this series accurately represents all or mostcavernomas presenting as epilepsy during thetime of review.

Previous series have found that cavernomaspresent as epilepsy in 34-70% ofcases' 5 6 8 13 14 with the remainder presentingas a mass lesion or with an intracranialhaemorrhage. In this series 81% presentedwith epilepsy and of these 13, all but onesuffered CPS with or without secondarygeneralisation. The frequency of those present-ing with epilepsy is somewhat higher than inother series, possibly reflecting our entry cri-terion of MRI diagnosis, whereas earlier seriesused only CT. Others have noted a predom-inance of temporal lobe lesions,48 but this hasnot been invariable.' 5 6We found cavernomasto occur in the temporal lobes in most of ourpatients. Two of our patients had multiplecavernomas and both suffered from epilepsythat was apparently caused by one angiomaonly. Multiple cavernous angiomata have beendocumented in a minority of cases in otherseries. l 7 8

Recurrent haemorrhage, as shown by thepresence of haemosiderin,'-5 has been pre-viously thought common. Recurrent sympto-matic haemorrhages have also beendocumented,38 but the incidence of sub-sequent symptomatic haemorrhage in thosepresenting with epilepsy, a mass lesion orcerebral haemorrhage remains unknown. Anecropsy study4 found that all 22 patients withsymptomatic cavernomas had epilepsy, butthat only three had had symptomatic haemor-rhages, one presented as a mass lesion and 12had stable neurological deficits. Sixteen ofthese 22 had evidence of old blood, suggestingthat asymptomatic haemorrhage is both com-mon and benign in those with epilepsy. Areview of the literature'4 noted that only a fewof those suffering a symptomatic cerebralhaemorrhage had a history consistent withepilepsy and noted that "the role of surgery inpreventing future haemorrhage in patients with

seizures can only be speculated". Asympto-matic cavernomas may also show evidence ofpast bleeding at necropsy.4

Long-standing epilepsy uncomplicated bysymptomatic haemorrhage has been notedbefore.58 These findings accord with ourexperience that symptomatic haemorrhage isuncommon in patients presenting with epi-lepsy only.Cavernomas had customarily been resected

without regard to the presenting complaint inmany series. 1246-8 Follow up periods rangefrom being unstated2 to years.' 46-8 The out-come has generally been stated as good if thefrequency of seizures has been reduced, but noseries seen by us adequately documented post-operative seizure control, continued use ofanti-epileptic drugs ranged from 15-71% andfollow up was often less than five years.' 467Evidence that excision of symptomatic cav-ernomas improves seizure control is lackingdespite claims that surgery was generallybeneficial.The absence of supratentorial cavernomas

presenting with cerebral haemorrhage in thisseries may reflect our entry criterion requiringdiagnosis by MRI. Cavernomas presentingwith haemorrhage may be lethal or resectedafter an initial CT scan only. Such cases are notincluded in this series because of lack ofsystematic access to necropsy and pathologydata and their incidence in our community isunknown.

ConclusionCavernomas may be symptomatic as epilepsy,intracranial haemorrhage or a mass lesion.Asymptomatic cases occur, but their incidenceremains unknown and will be underestimatedin clinical series of symptomatic patients.Excision has been widely practised regardlessof presentation although its efficacy isunknown. In our experience, most patientswith epilepsy attain excellent seizure controlwith anti-epileptic medications. No patientwith epilepsy suffered from symptomatichaemorrhage or progressive neurological defi-cits, which appear to be rare in those present-ing with seizures. Asymptomatic andapparently benign haemorrhage, as shown byMRI, is common. In this series only onepatient required excision as treatment forrefractory epilepsy.We found no evidence for excision of cav-

ernomas presenting as epilepsy on the groundsof either prophylaxis against haemorrhage, orfor improved seizure control in all but refrac-tory epilepsy. We recommend that resection ofthose cavernous angiomata manifesting as epi-lepsy alone be considered only if an adequatetrial of anti-epileptic drugs fails. Excision ofaccessible, supratentorial cavernomas presen-ting with a single or recurrent cerebral haemor-rhage or haemorrhages is a valid, althoughunproven, treatment where surgery is unlikelyto increase the patient's disability.

We thank the following clinicians who kindly allowed us accessto their case notes: Dr S Gubbay and Dr R Edis of the Royal

1 044 on N

ovember 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 N

ovember 1992. D

ownloaded from

Cerebral cavernous angioma: a potentially benign condition?

Perth Hospital, Western Australia; Mr B A R Stokes, Dr JScopa, Dr G Groom, Mr M Lee and DrW M Carroll, all of theQueen Elizabeth II Medical Centre, Nedlands, Western Aus-tralia.

1 Tagle P, Huerte I, Mendez J, Del Villar S. Intracranialcavernous angioma: presentation and management.Neurosurg 1986;64:720-3.

2 Wakai S, VedaY, Inoh S, Nagai M. Angiographically occultangiomas: a report of 13 cases with analysis of the casesdocumented in the literature. Neurosurgery 1985;17:549-56.

3 Abe M, Kjellberg R, Adams R. Clinical presentation ofvascular malformations of the brainstem: comparison ofangiographically positive and negative types. NeurolNeurosurg Psychiatry 1989;52:167-75.

4 Farmer J, Cosgrove G, Villemure J, Meagher-Villemure K,Tampieri D, Melanson D. Intracerebral cavernousangioma. Neurology 1988;38:1699-1704.

5 Steiger H, Markwalder T, Reulen H. Clinicopathologicalrelations of cerebral cavemous angiomas: observations in11 cases. Neurosurgery 1987;21:879-84.

6 Vaquero J, Salazar J, Martinez R, Martinez P, Bravo G.Cavernomas of the central nervous system: clinicalsyndromes, CT scan diagnosis and prognosis after surgi-cal treatment in 25 cases. Acta Neurochir [Wien]

1987;85:29-33.7 Vaquero J, Leunda G, Martinez R, Bravo G. Cavernomas of

the brain. Neurosurgery 1983;12:208-10.8 Giombini S, Morello G. Cavernous angiomas of the brain:

account of 14 personal cases and review of the literature.Acta Neurochir 1978;40:61-82.

9 Russell DS, Rubenstein U. Pathology of tumours of thenervous system. Ed. 4, Baltimore and Wilkins1977:129-34.

10 Rigamonti D, Drayer B, Johnson P, Hadley M, Zabramski J,Spetzler R. The MRI appearance of cavernous malforma-tions [angiomas]. Neurosurg 1987;67:518-24.

11 Gomori J, Grossman R, Goldberg M, Hackney D, Zimmer-man R, Bilanick L. Occult cerebral vascular malforma-tions: high-field MR imaging. Radiology 1986;158:707-13.

12 New P, Ojemann R, David K, et al. MR and CT of occultvascular malformations of the brain. Am Radiol1986;147:985-93.

13 Requena I, Arias M, Lopez-Ibor L et al. Cavemomas of thecentral nervous system: clinical and neuroimaging mani-festations in 47 patients. Neurol Neurosurg Psychiatry199 1;54:591-4.

14 Simard J, Garcia-Bengochea F, Ballinger W, Mickle J,Quisling R. Cavemous angioma: a review of 126 collectedand 12 new clinical cases. Neurosurg 1986;18:162-72.

1045 on N

ovember 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.55.11.1040 on 1 N

ovember 1992. D

ownloaded from