J Neurosurg 73:174-180, 1990
Recurrent intraparenchymal hemorrhages from angiographically occult vascular malformations HOWARD TUNG, M.D., STEVEN L. GIANNOTFA, M.D., PARAKRAMA T. CHANDRASOMA, M.D., AND CHI-SHING ZEE, M.D.
Departments of Neurological Surgery, Surgical Pathology, and Division of Neuroradiology, University of Southern California School of Medicine, Los Angeles, California ~" Thirteen patients with recurrent hemorrhages from angiographically occult vascular malformations are presented. Recurrent hemorrhages were indicated by the exacerbation of neurological deficits, the appearance of a new neurological deficit, or the onset of acute severe headache confirmed by evidence of recent hemorrhage on either computerized tomography or magnetic resonance imaging. Persistent neurological deficits correlated with an increased number of recurrent hemorrhages and their location. The median time from initial hemorrhage to the first recurrent hemorrhage was 12 months and the second rebleed generally occurred much earlier, with a median time of 2 months after the first. Eight patients underwent surgery with total excision and favorable results. One patient with a large pontine lesion underwent partial excision and has had a progressing neurological deficit from recurrent hemorrhages. Histopathological review confirmed the excised lesions to be cavernous angiomas. The authors conclude that angiographically occult vascular malformations are not the benign entity they were previously thought to be, and that they are prone to cause recurrent hemorrhages and persistent neurological deficits. Surgery can be effective and relatively safe in removing these lesions even in eloquent areas of the brain, but the necessity of occasional incomplete removal must be recognized in order to avoid creating an unacceptable neurological outcome. KEY WORDS rebleeding
A
"
occult vascular malformations
NGIOGRAPHICALLY Occult vascular malformations of the brain are increasingly being identified due to the widespread availability of sophisticated imaging techniques such as computerized tomography (CT) and magnetic resonance (MR) imaging. Recent reports have highlighted the CT and M R characteristics of these lesions and their corresponding clinical and histopathological profiles. 6A7'23"28However, the natural history of these lesions remains obscure. The incidence and impact of recurrent hemorrhage have not been analyzed in any systematic fashion. The authors present 13 cases of documented clinically relevant recurrent hemorrhages from angiographically occult vascular malformations of the brain. These patients were evaluated between January 1, 1984, and January 1, 1989. All cases of recurrent hemorrhage were indicated by an increase in a preexisting neurological deficit, the appearance of a new neurological deficit, or the acute onset of severe headache, and confirmed by evidence of recent hemorrhage on either CT or MR imaging.
174
9 cavernous angioma
9 hemorrhage
9
S u m m a r y of Cases
Clinical and Histological Features The clinical and histological features of 13 patients with angiographically occult vascular malformations are presented in Table 1. There were four men and nine women, with an age range of 18 to 59 years. Eight lesions (61.5%) were located in the posterior fossa (five in the pons or pontomedullary junction) and three were in the cerebellum (one in the vermis and two i n the cerebellar peduncles). Four lesions were located in the white matter or subcortical region of the cerebral hemispheres and one was in the thalamus.
Neuroradiographic Studies The radiographic features of the lesions in our 13 cases were typical of those described in recent publicati0ns. 5'11'13-17'22-24 Cerebral angiography was negative in all patients, five of whom underwent at least two angiograms during their clinical course. No abnormal vasculature or early venous drainage was identified on
J. Neurosurg. / Volume 73~August, 1990
Recurrent hemorrhage from occult vascular anomalies TABLE 1
Clinical summary in 13 cases of angiographically occult vascular malformation* Age Case (yrs), No. Sex 1 18, F
Location of Lesion
Neurological Histology Deficit
Clinical Symptomsof Rebleeding
pons,floor of 4th ventricle vermisof cerebellum
3 episodes acute HA, numbness rt side, rt hemiparesis
transient
3 episodes acute HA, lethargy, nausea, vomiting,ataxia
transient
pons,floor of 4th ventricle pons,floor of 4th ventricle rt posterior, parietal It frontal It thalamus
4 episodes acute vertigo, diplopia, nausea, It arm numbness;recurrent symptoms several months postop indicating recurrent hemorrhages 2 episodes acute HA, diplopia, unsteadygait; rt INO with 2nd rebleed
cavernous angioma persistent cavernous angioma transient no surgery
3 episodes severe ictal HA, followedby grand real seizures 3 episodes of HA, lethargy;dysphasia with 2nd rebleed 4 episodes HA, rt hemiparesis, 1 seizure
none transient persistent
2 episodes HA, rt facial numbness, It UE paresthesias, numbness
transient
3 episodes HA, vomiting,vertigo;new ataxia with 2rid rebleed
transient
3 episodes rt facial numbness; radiosurgeryafter 2nd rebleed: rebleed 3 months postop with dysphagia, blurred vision, & rt-sided ataxia 2 episodes lower cranial nerve palsies, rt-sidedcold dysesthesias, rt hemiparesis 2 episodes severe HA & It homonymous hemianopsia(transient after 1st rehleed) 2 episodes severe HA, nausea, vomiting
2
34, F
3
59, F
4
30, M
5 6 7
26, M 36, F 35, M
8
39, F
9
27, F
pons,floor of 4th ventricle rt brachium pontis
10
34, M
rt brachium pontis
11
34, F
12
40, F
pons,floor of 4th ventricle rt occipital
13
22, F
It medial
no histology
no histology no surgery cavernous angioma no surgery
cavernous angioma persistent cavernous angioma transient no surgery persistent cavernous angioma none cavernous angioma
* HA = headache; INO = internuclear ophthalmoplegia;UE = upper extremity.
either prospective or retrospective analysis of the angiograms. The CT scans disclosed areas of mixed density with m i n i m a l contrast e n h a n c e m e n t , indicative of hemorrhage of differing ages. I n Case 3, the recurrent hemorrhage appeared as a low-density ring adjacent to a n acute hemorrhage. This low-density area was confirmed to be chronic hemorrhage o n subsequent M R imaging. In one case of a thalamic cavernous angioma, the lesion produced a large cyst with a hematocrit level. The m a j o r portion of the cyst was shown on M R imaging to have a low signal intensity on Tl-weighted images a n d a high signal intensity o n T2-weighted images, consistent with free m e t h e m o g l o b i n from hemorrhage. The hematocrit level was of intermediate signal intensity consistent with intracellular m e t h e m o g l o b i n of red blood cells (Fig. 1). All other lesions d e m o n s t r a t e d mixed areas of increased a n d decreased intensity consistent with various stages of hemorrhage (Figs. 2 a n d 3).
Clinical S y m p t o m s Eleven of the 13 patients experienced exacerbation of a preexisting neurological deficit as evidence for one or more recurrent hemorrhages. T e n patients had associated headache with their recurrent hemorrhage, a n d two suffered a seizure. N o patient in this series died of recurrent hemorrhage (Table 2).
J. Neurosurg. / Volume 73/August, 1990
TABLE 2
Symptoms of rebleeding Symptom
No. of Cases
exacerbation of deficit headache seizure death
11 10 2 0
TABLE 3
Temporal profile of rebleeding in 13 patients Interval Between Bleeds (mos)
Rebleed No.
Range
Median
1
1-60
12
2
1-8
2
Timing and Impact o f Repeat Hemorrhage The interval from the initial hemorrhage to the first recurrent hemorrhage in this series of 13 patients ranged from 1 to 60 m o n t h s ( m e d i a n time 12 months). Generally, the second rebleed occurred m u c h earlier, with a m e d i a n time of 2 m o n t h s after the first; one patient experienced recurrent hemorrhages separated by less t h a n 1 week (Table 3). 175
H. Tung, et al.
FIG. 1. T2-weighted magnetic resonance image showing a cystic lesion in the left thalamus with a hematocrit level from recurrent hemorrhage. Five patients had one clinically relevant recurrent hemorrhage. O f those five, one patient suffered a persistent new neurological deficit, three had transient exacerbations of previous neurological deficits, and one had no neurological deficit except for acute severe headache with nausea and vomiting. Six patients had two recurrent hemorrhages; of these, two experienced persistent neurological deficits, three had transient ex-
FIG. 2. Tt-weighted magnetic resonance image showing a multilaminated lesion of various signal intensities. The peripheral high signal intensity (methemoglobin) corresponds to recent hemorrhage. acerbations of neurological deficits, and one had no neurological deficit as a result of the second recurrent hemorrhage. Two patients suffered a third recurrent hemorrhage; both of these developed a persistent neurological deficit. Thus, there seemed to be a relationship between the number of hemorrhages and the produc-
FIG. 3. Left: Proton-density weighted magnetic resonance (MR) image showing a predominantly low signal intensity lesion in the floor of the fourth ventricle. The small locus of high signal intensity at the superior aspect corresponds to an area of recent hemorrhage. Right: Repeat MR image, after exacerbation of the patient's neurological symptoms 5 months later, showing a large high signal intensity area corresponding to recent recurrent hemorrhage.
176
J. Neurosurg. / Volume 73/August, 1990
Recurrent hemorrhage from occult vascular anomalies TABLE 5 Impact of lesion location on occurrence of deficits
TABLE 4
Impact o f rebleeds in 13 cases Rebleed No.
Persistent Deficits
Transient Deficits
No Deficit
Total Cases
Locationof Lesion brain stem & thalamus cerebellum lobar
1
1
3
1
5
2
2
3
1
6
3
2
0
0
2
tion of a lasting deficit (Table 4). There was also a general correlation between the location of the lesion and persistent neurological deficit (Table 5). Patients whose lesions were located in the brain stem, cerebellar peduncle, or thalamus tended to have a more debilitating and persistent neurological deficit subsequent to each rebleeding episode. Surgical Management Of the nine patients who underwent surgery, eight had total excision of their lesions. The remaining patient, with a large pontine lesion, underwent partial excision and has continued to deteriorate due to radiographically documented recurrent hemorrhages. Four patients refused surgery, two of whom have subsequently suffered a recurrent hemorrhage. Microsurgical technique was used in all instances due to the relatively small size of these lesions (the majority being < 1 cm in diameter). At surgery, all lesions showed evidence of old and recent hemorrhage. The majority of lesions appeared as small tangles of delicate blood vessels under relatively low pressure, some with thrombosed or partially thrombosed lumens. One lesion had formed a large cyst which was filled with blood clot of varying ages. One lesion had a mulberry shape typical of a cavernous angioma. Five patients improved following surgical excision, two remained unchanged, and two experienced an increased neurological deficit as the result of surgery. One patient with increased deficit had a lesion in the superior cerebellar peduncle which had hemorrhaged for a second time 9 months after stereotactic radiosurgery using the Bragg peak method. He has a mild hemisensory deficit with a slightly detectable cerebellar hemiataxia. He has returned to full activity as a chemical engineer. The second patient who experienced neurological deficit subsequent to surgery harbors a large pontine vascular malformation. Although she was ambulatory after surgery, progressive intermittent neurological deficits have persisted due to recurrent hemorrhages from the unresected part of the lesion. None of the nine patients undergoing surgery died. In the majority of deep-seated lesions, ultrasound was utilized to identify and locate the lesion. In most circumstances, these lesions showed a mixed echogenic signal of high and low intensity similar to the signal from MR imaging. No patient underwent CT-guided stereotaxis to locate the lesion. J. Neurosurg. / Volume 73/August, 1990
Persistent Transient No Total Deficits Deficits Deficit Cases 3
3
0
6
1 1
2 1
0 2
3 4
Histological Findings Detailed pathological review was available for seven of the nine surgical specimens. Originally, four of these specimens were considered arteriovenous malformations (AVM's) but, upon further study, all cases were reclassified as cavernous angiomas. In all seven cases with adequate tissue for histopathological examination, there was evidence of recent and old hemorrhage. Old hemorrhage was characterized by fibrosis, the presence of macrophages, and the deposition of hemosiderin pigment. In all cases, the hemorrhage and the reaction to it had obliterated normal tissues and part of the vascular malformation, and careful evaluation of the edge of the hematoma was necessary to identify the residual part of the lesion. Diagnosis of cavernous angioma was based on the identification of large endothelium-lined vascular channels separated by varying amounts of collagen (Fig. 4). Except at the edges, brain substance was not present between the vascular spaces. Fibrosis of the walls of the vascular spaces was present at least focally in all cases (Fig. 4 right) and, when extensive, resulted in a marked thickening of involved vessels. This caused a superficial resemblance to an AVM, and required elastic and trichrome stains for accurate diagnosis. The absence of elastic fibers and smooth muscle in the thickened wall (Fig. 5) rules out arterial and venous components in the vascular malformation. The demonstration that the thick walls are composed of collagen alone favors the diagnosis of cavernous angioma. The performance of special stains for elastin and smooth muscle contributed to the reclassification of four cases originally considered AVM's as cavernous angiomas. Discussion
Terminology and Literature Review The susceptibility of small vascular malformations to hemorrhage was first described in 1951 by Margolis, et al., ~9 four of whose patients suffered fatal hemorrhage. Crawford and Russell 8 subsequently introduced the term "cryptic" to describe these small (< 2 to 3 cm) vascular malformations which caused spontaneous cerebral hemorrhage. Initially, the term "cryptic" was applied to the group of vascular malformations based on their small size alone. However, more recently the terms "cryptic" and "occult" have been applied to those vascular malformations that cannot be identified by cere177
H. Tung, et al.
FIG. 4. Photomicrographs of cavernous angioma specimens. H & E, x 100. Left: Specimen showing endothelium-lined spaces separated by scant connective tissue. Right: Specimen with evidence of fibrosis of the vascular wall resulting in compression of the vascular spaces and thickening of the wall.
bral angiography. 1,6.21,23A more appropriate term would be "angiographically occult." Thus, the designation "angiographically occult vascular malformation" represents a myriad of vascular malformations with variable histology and includes cryptic AVM's, thrombosed AVM's, cavernous angiomas, and other histological vascular lesions escaping detection by cerebral angiography. Recent reports have thoroughly reviewed the histopathology, radiographic findings, and surgical management of patients with angiographically occult vascular malformations of the brain. 6'17'23'28 In these and other series, recurrent hemorrhage has been mentioned as the pathophysiological mechanism for neurological deftcir. 3,4,6A~ A review of the literature reveals that 80% to 90% of angiographically occult vascular malformations showed evidence of hemorrhage. 17'28 Even in asymptomatic cases or those presenting with seizures, MR imaging and pathological examination have documented the presence of hemosiderin, presumably the residue of previously experienced subclinical hemorrhage. 3"6,12AS'17"2L22,24'25'28However, the clinical and temporal profiles of clinically relevant hemorrhages have not been highlighted in the past. Location and Histopathology o f Lesions Our series of 13 patients differs in some ways from previously reported cases. In a review of the literature in 1988, Lobato, et al., ~7 found that over 75% of angiographically occult vascular malformations were located in the supratentorial region. In our series, eight of the 13 patients had lesions in either the brain stem or the cerebellum. Likewise, in Lobato's review, 31% of the lesions were cavernous angiomas. All of the lesions in this series examined pathologically were cavernous angiomas. The routine use of trichrome stain for smooth muscle and collagen and stains for elastic 178
fibers can help differentiate among the various pathological subtypes of vascular malformations for a more accurate diagnosis and clinical correlation. In addition, the diagnosis of angiographically occult vascular malformation is not specific despite modern imaging techniques and suggestive clinical history. The diagnosis can only be made definitively by surgical pathological confirmation. 2'4'6'12'17'21'23'25It is possible, although certainly not proven here, that clinically relevant recurrent hemorrhages may be associated with either one or both of these characteristics: namely, location or histopathology. Recurrent Hemorrhage In our series there was good correlation between the number of recurrent hemorrhages, the location of recurrent hemorrhages, and the occurrence of persistent neurological deficit. In most cases (80%) the initial rebleed caused only a transient deficit; however, with each successive recurrent hemorrhage, the likelihood of a resultant persistent neurological deficit increased, and all of the patients with more than two rebleeds suffered persistent neurological deficits. It is difficult to determine from our small series the relative influence on outcome of location versus number of hemorrhages. It is probable that hemorrhages, even small ones, are more likely to produce clinical symptoms or neurological deficits in those lesions involving hazardous areas of the brain, such as the brain stem, thalamus, and eloquent cortex, and to bring them to recognition earlier. Furthermore, angiographically occult vascular malformations located in eloquent areas of the brain are often initially managed conservatively, as surgery in these areas is most difficult and carries a higher risk of morbidity compared to that of angiographically occult vascular malformations located in other a r e a s . 12A7'21'23' 28-30 This may partially explain the disproportionate J. Neurosurg. / Volume 73/August, 1990
Recurrent hemorrhage from occult vascular anomalies number of lesions in our series located in the posterior fossa. Although the full spectrum of the natural history remains to be elucidated, angiographically occult vascular malformations are prone to cause recurrent hemorrhages. 1'2'6'1~176 Histologically, all of our surgical specimens revealed evidence of chronic and recent hemorrhage. In reviewing the literature, Lobato, et aL,'7 found that 31% of patients with documented histologically verified angiographically occult vascular malformations showed evidence of both chronic and recent hemorrhage. Similarly, Wakai, et al., 28 reported 44% of patients with chronic and recent hemorrhage. Obviously, it is difficult to determine the exact rate of rebleeding from these lesions, and long-term retrospective or prospective population studies will be necessary. However, the rate of rebleeding appears considerable and may be similar to that of high-flow angiographically apparent AVM's. All patients in this series presented clinical symptoms of hemorrhage or exacerbation of a previous neurological deficit. However, not all hemorrhages from angiographically occult vascular malformations may be clinically evident. Lobato, et al., w reported that 24% of the patients they reviewed had a clinical course suggestive of recurrent hemorrhage; seven of their own series of 21 patients with angiographically occult vascular malformations suffered recurrent symptoms of rebleeding. Although there were no deaths as a result of a recurrent hemorrhage in the present series, it is well known that patients may suffer fatal recurrent hemorrhages from these lesions. 1'8~t~176 It has, therefore, become our policy to recommend removal when they are discovered in surgically accessible noneloquent areas of the brain. Lesions found in more hazardous areas are removed only if they produce recurrent hemorrhage. Surgical T r e a t m e n t The surgical management of angiographically occult vascular malformations is not as straightforward as some have implied. They tend to be small and are not infrequently located deep in the cerebral or cerebellar hemisphere or brain stem. Localization of the lesions at the time of surgery may prove difficult. Therefore, we have relied on intraoperative ultrasonography to assist in this endeavor; CT-guided stereotaxis may be appropriate in certain circumstances. 9 Unlike their larger high-flow counterparts, the endpoint of surgery for angiographically occult vascular malformations may be difficult to determine. The lesions tend to be small with very fine vasculature. Frequently, the vessels are not tightly compacted, making it difficult to be certain of total removal. Even lesions with predominantly cavernous elements on histopathology may not demonstrate the classical gross appearance of the easily removed cavernous angioma. Vigorous exploration and surgical manipulation of the cavity may prove hazardous in brain-stem lesions and other eloquent areas. 21 24,27 30 Total removal in certain J. Neurosurg. / Volume 73/August, 1990
FIG. 5. Photomicrograph of an cavernous angioma specimen, stained for elastin, showing absence of elastic fibers in the thickened vessel wall. Elastin, • 200.
circumstances would be associated with unacceptable complications, especially in the case of pontine lesions. Radiation Therapy Recent reports have shown the efficacy of radiosurgery in the treatment of AVM's. 7'18'26 Although of theoretical value in the treatment of angiographically occult vascular malformations, the use of radiosurgery has met with limited success in treating cavernous angiomas as opposed to high-flow AVM's (L Steiner~ personal communication, 1989). In fact, one of our patients suffered a recurrent hemorrhage with a persistent neurological deficit 9 months following stereotactic radiosurgery with Bragg peak radiation, thus emphasizing the principle that recurrent hemorrhages may occur before complete obliteration of the lesion by radiosurgery or any other technique is achieved. In summary, this series of cases underscores the propensity of angiographically occult vascular malformations to cause recurrent hemorrhages. Each successive hemorrhage tended to occur at shorter time intervals, and each hemorrhage tended to be associated with an increased incidence of permanent neurological deficit. Surgery can be effective and relatively safe in removing these lesions even in hazardous areas of the brain. Acknowledgment The authors thank Pat Sullivan for manuscript preparation. References 1. Beeker DH, Townsend J J, Kramer RA, et al" Occult cerebrovascular malformations. A series of 18 histologically verified cases with negative angiography. Brain 102: 249-287, 1979 2. Bell BA, Kendell BE, Symon L: Angiographically occult arteriovenous malformations of the brain. J Neurol Neurosurg Psychiatry 41:1057-1064, 1978 179
H. Tung, et al. 3. Bitoh S, Hasegawa H, Fujiwara M, et al: Angiographically occult vascular malformations causing intracranial hemorrhage. Surg Neurol 17:35-42, 1982 4. Britt RH, Connor WS, Enzmann DR: Occult arteriovenous malformation of the brainstem simulating multiple sclerosis. Neurology 31:901-903, 1981 5. Chin D, Harper C: Angiographically occult cerebral vascular malformations with abnormal computed tomography. Surg Neurol 20:138-142, 1983 6. Cohen HCM, Tucker WS, Humphreys RP, et al: Angiographically cryptic histologically verified cerebrovascular malformations. Neurosurgery 10:704-714, 1982 7. Colombo F, Benedetti A, Pozza F, et al: Linear accelerator radiosurgery of cerebral arteriovenous malformations. Neurosurgery 24:833-840, 1989 8. Crawford JV, Russell DS: Cryptic arteriovenous and venous hamartomas of the brain. J Neuroi Neurosurg Psychiatry 19:1-11, 1956 9. Daumas-Duport C, Mann M, Munari C, et al: Cryptic vascular malformations diagnosed by stereotactic biopsies - - a preliminary study. Appl Neurophysiol 48:440-443, 1985 10. Giombini S, Morello G: Cavernous angiomas of the brain. Account of fourteen personal cases and review of the literature. Acta Neurochir 40:61-82, 1978 11. Gomori JM, Grossman RI, Goldberg HI, et al: Occult cerebral vascular malformations: high-field MR imaging. Radiology 158:707-713, 1986 12. Hashim ASM, Asakura T, Koichi U, ct al: Angiographically occult vascular malformations. Surg Neurol 23: 431-439, 1985 13. Kendall BE, Claveria LE: The use of computed axial tomography (CAT) for the diagnosis and management of intracranial angiomas. Neuroradioiogy 12:141-160, 1976 14. Kramer RA, Wing SD: Computed tomography of angiographically occult cerebral vascular malformations. Radiology 123:649-652, 1977 15. Kucharczyk W, Lemme-Pleghos L, Uske A, et al: Intracranial vascular malformation: MR and CT imaging. Radiology 156:383-389, 1985 16. Leblanc R, Ethier R: The computerized tomographic appearance of angiographically occult arteriovenous malformations of the brain. Can J Neurol Sei 8:7-13, 1981 17. Lobato RD, Perez C, Rivas J J, et al: Clinical, radiological, and pathological spectrum of angiographically occult intracranial vascular malformations. Analysis of 21 cases and review of the literature. J Neurosurg 68:518-531, 1988
180
18. Loeffier JS, Eben A, Siddon RL, et al: Stereotactic radiosurgery for intracranial arteriovenous malformations using a standard linear accelerator. Int J Radiat Oncol Biol Phys 17:673-677, 1989 19. Margolis G, Odom GL, Woodhall B, et al: The role of small angiomatous malformations in the production of intracerebral hematomas. J Neurosurg 8:564-575, 1951 20. McCormick WF, Nofzinger JD: "Cryptic" vascular malformations of the central nervous system. J Neurosurg 24:865-875, 1966 21. McFerran D J, Marks PV, Garvan NJ: Angiographically occult arteriovenous malformations of the brainstem. Surg Neurol 28:221-224, 1987 22. New PFJ, Ojemann RG, Davis KR, et al: MR and CT of occult vascular malformations of the brain. AJNR 7: 771-779, 1986 23. Ogilvy CS, Heros RC, Ojemann RG, et al: Angiographically occult arteriovenous malformations. J Neurosurg 69:350-355, 1988 24. Rigamonti D, Drayer BP, Johnson PC, et al: The MRI appearance of cavernous malformations (angiomas). J Neurosurg 67:518-524, 1987 25. Shuey HM Jr, Day AL, Quisling RG, et al: Angiographically cryptic cerebrovascular malformations. Neurosurgery 5:476-479, 1979 26. Steiner L, Lindquist C: Radiosurgery in cerebral arteriovenous malformation, in Tasker RR (ed): Neurosurgery: State of the Art Reviews. Philadelphia: Hanley & Belfus, 1987, pp 329-336 27. Veerapen RJ, Sbeih IA, O'Laoire SA: Surgical treatment of cryptic AVM's and associated hematoma in the brain stem and spinal cord. J Neurosurg 65:188-193, 1986 28. Wakai S, Ueda Y, Inoh S, et al: Angiographically occult angiomas: a report of thirteen cases with an analysis of the cases documented in the literature. Neurosurgery 17: 549-556, 1985 29. Wilson CB, U HS, Domingue J: Microsurgical treatment of intracranial vascular malformations. J Neurosurg 51: 446-454, 1979 30. Yoshimoto T, Suzuki J: Radical surgery on cavernous angioma of the brainstem. Surg Neurol 26:72-78, 1986 Manuscript received October 17, 1989. Accepted in final form February 5, 1990. Address reprint requests to: Steven L. Giannotta, M.D., Department of Neurosurgery, LAC-USC Medical Center, 1200 North State Street, Room 5046, Los Angeles, California 90033.
J. Neurosurg. / Volume 73/August, 1990
Recurrent intraparenchymal hemorrhages from angiographically occult vascular malformations HOWARD TUNG, M.D., STEVEN L. GIANNOTFA, M.D., PARAKRAMA T. CHANDRASOMA, M.D., AND CHI-SHING ZEE, M.D.
Departments of Neurological Surgery, Surgical Pathology, and Division of Neuroradiology, University of Southern California School of Medicine, Los Angeles, California ~" Thirteen patients with recurrent hemorrhages from angiographically occult vascular malformations are presented. Recurrent hemorrhages were indicated by the exacerbation of neurological deficits, the appearance of a new neurological deficit, or the onset of acute severe headache confirmed by evidence of recent hemorrhage on either computerized tomography or magnetic resonance imaging. Persistent neurological deficits correlated with an increased number of recurrent hemorrhages and their location. The median time from initial hemorrhage to the first recurrent hemorrhage was 12 months and the second rebleed generally occurred much earlier, with a median time of 2 months after the first. Eight patients underwent surgery with total excision and favorable results. One patient with a large pontine lesion underwent partial excision and has had a progressing neurological deficit from recurrent hemorrhages. Histopathological review confirmed the excised lesions to be cavernous angiomas. The authors conclude that angiographically occult vascular malformations are not the benign entity they were previously thought to be, and that they are prone to cause recurrent hemorrhages and persistent neurological deficits. Surgery can be effective and relatively safe in removing these lesions even in eloquent areas of the brain, but the necessity of occasional incomplete removal must be recognized in order to avoid creating an unacceptable neurological outcome. KEY WORDS rebleeding
A
"
occult vascular malformations
NGIOGRAPHICALLY Occult vascular malformations of the brain are increasingly being identified due to the widespread availability of sophisticated imaging techniques such as computerized tomography (CT) and magnetic resonance (MR) imaging. Recent reports have highlighted the CT and M R characteristics of these lesions and their corresponding clinical and histopathological profiles. 6A7'23"28However, the natural history of these lesions remains obscure. The incidence and impact of recurrent hemorrhage have not been analyzed in any systematic fashion. The authors present 13 cases of documented clinically relevant recurrent hemorrhages from angiographically occult vascular malformations of the brain. These patients were evaluated between January 1, 1984, and January 1, 1989. All cases of recurrent hemorrhage were indicated by an increase in a preexisting neurological deficit, the appearance of a new neurological deficit, or the acute onset of severe headache, and confirmed by evidence of recent hemorrhage on either CT or MR imaging.
174
9 cavernous angioma
9 hemorrhage
9
S u m m a r y of Cases
Clinical and Histological Features The clinical and histological features of 13 patients with angiographically occult vascular malformations are presented in Table 1. There were four men and nine women, with an age range of 18 to 59 years. Eight lesions (61.5%) were located in the posterior fossa (five in the pons or pontomedullary junction) and three were in the cerebellum (one in the vermis and two i n the cerebellar peduncles). Four lesions were located in the white matter or subcortical region of the cerebral hemispheres and one was in the thalamus.
Neuroradiographic Studies The radiographic features of the lesions in our 13 cases were typical of those described in recent publicati0ns. 5'11'13-17'22-24 Cerebral angiography was negative in all patients, five of whom underwent at least two angiograms during their clinical course. No abnormal vasculature or early venous drainage was identified on
J. Neurosurg. / Volume 73~August, 1990
Recurrent hemorrhage from occult vascular anomalies TABLE 1
Clinical summary in 13 cases of angiographically occult vascular malformation* Age Case (yrs), No. Sex 1 18, F
Location of Lesion
Neurological Histology Deficit
Clinical Symptomsof Rebleeding
pons,floor of 4th ventricle vermisof cerebellum
3 episodes acute HA, numbness rt side, rt hemiparesis
transient
3 episodes acute HA, lethargy, nausea, vomiting,ataxia
transient
pons,floor of 4th ventricle pons,floor of 4th ventricle rt posterior, parietal It frontal It thalamus
4 episodes acute vertigo, diplopia, nausea, It arm numbness;recurrent symptoms several months postop indicating recurrent hemorrhages 2 episodes acute HA, diplopia, unsteadygait; rt INO with 2nd rebleed
cavernous angioma persistent cavernous angioma transient no surgery
3 episodes severe ictal HA, followedby grand real seizures 3 episodes of HA, lethargy;dysphasia with 2nd rebleed 4 episodes HA, rt hemiparesis, 1 seizure
none transient persistent
2 episodes HA, rt facial numbness, It UE paresthesias, numbness
transient
3 episodes HA, vomiting,vertigo;new ataxia with 2rid rebleed
transient
3 episodes rt facial numbness; radiosurgeryafter 2nd rebleed: rebleed 3 months postop with dysphagia, blurred vision, & rt-sided ataxia 2 episodes lower cranial nerve palsies, rt-sidedcold dysesthesias, rt hemiparesis 2 episodes severe HA & It homonymous hemianopsia(transient after 1st rehleed) 2 episodes severe HA, nausea, vomiting
2
34, F
3
59, F
4
30, M
5 6 7
26, M 36, F 35, M
8
39, F
9
27, F
pons,floor of 4th ventricle rt brachium pontis
10
34, M
rt brachium pontis
11
34, F
12
40, F
pons,floor of 4th ventricle rt occipital
13
22, F
It medial
no histology
no histology no surgery cavernous angioma no surgery
cavernous angioma persistent cavernous angioma transient no surgery persistent cavernous angioma none cavernous angioma
* HA = headache; INO = internuclear ophthalmoplegia;UE = upper extremity.
either prospective or retrospective analysis of the angiograms. The CT scans disclosed areas of mixed density with m i n i m a l contrast e n h a n c e m e n t , indicative of hemorrhage of differing ages. I n Case 3, the recurrent hemorrhage appeared as a low-density ring adjacent to a n acute hemorrhage. This low-density area was confirmed to be chronic hemorrhage o n subsequent M R imaging. In one case of a thalamic cavernous angioma, the lesion produced a large cyst with a hematocrit level. The m a j o r portion of the cyst was shown on M R imaging to have a low signal intensity on Tl-weighted images a n d a high signal intensity o n T2-weighted images, consistent with free m e t h e m o g l o b i n from hemorrhage. The hematocrit level was of intermediate signal intensity consistent with intracellular m e t h e m o g l o b i n of red blood cells (Fig. 1). All other lesions d e m o n s t r a t e d mixed areas of increased a n d decreased intensity consistent with various stages of hemorrhage (Figs. 2 a n d 3).
Clinical S y m p t o m s Eleven of the 13 patients experienced exacerbation of a preexisting neurological deficit as evidence for one or more recurrent hemorrhages. T e n patients had associated headache with their recurrent hemorrhage, a n d two suffered a seizure. N o patient in this series died of recurrent hemorrhage (Table 2).
J. Neurosurg. / Volume 73/August, 1990
TABLE 2
Symptoms of rebleeding Symptom
No. of Cases
exacerbation of deficit headache seizure death
11 10 2 0
TABLE 3
Temporal profile of rebleeding in 13 patients Interval Between Bleeds (mos)
Rebleed No.
Range
Median
1
1-60
12
2
1-8
2
Timing and Impact o f Repeat Hemorrhage The interval from the initial hemorrhage to the first recurrent hemorrhage in this series of 13 patients ranged from 1 to 60 m o n t h s ( m e d i a n time 12 months). Generally, the second rebleed occurred m u c h earlier, with a m e d i a n time of 2 m o n t h s after the first; one patient experienced recurrent hemorrhages separated by less t h a n 1 week (Table 3). 175
H. Tung, et al.
FIG. 1. T2-weighted magnetic resonance image showing a cystic lesion in the left thalamus with a hematocrit level from recurrent hemorrhage. Five patients had one clinically relevant recurrent hemorrhage. O f those five, one patient suffered a persistent new neurological deficit, three had transient exacerbations of previous neurological deficits, and one had no neurological deficit except for acute severe headache with nausea and vomiting. Six patients had two recurrent hemorrhages; of these, two experienced persistent neurological deficits, three had transient ex-
FIG. 2. Tt-weighted magnetic resonance image showing a multilaminated lesion of various signal intensities. The peripheral high signal intensity (methemoglobin) corresponds to recent hemorrhage. acerbations of neurological deficits, and one had no neurological deficit as a result of the second recurrent hemorrhage. Two patients suffered a third recurrent hemorrhage; both of these developed a persistent neurological deficit. Thus, there seemed to be a relationship between the number of hemorrhages and the produc-
FIG. 3. Left: Proton-density weighted magnetic resonance (MR) image showing a predominantly low signal intensity lesion in the floor of the fourth ventricle. The small locus of high signal intensity at the superior aspect corresponds to an area of recent hemorrhage. Right: Repeat MR image, after exacerbation of the patient's neurological symptoms 5 months later, showing a large high signal intensity area corresponding to recent recurrent hemorrhage.
176
J. Neurosurg. / Volume 73/August, 1990
Recurrent hemorrhage from occult vascular anomalies TABLE 5 Impact of lesion location on occurrence of deficits
TABLE 4
Impact o f rebleeds in 13 cases Rebleed No.
Persistent Deficits
Transient Deficits
No Deficit
Total Cases
Locationof Lesion brain stem & thalamus cerebellum lobar
1
1
3
1
5
2
2
3
1
6
3
2
0
0
2
tion of a lasting deficit (Table 4). There was also a general correlation between the location of the lesion and persistent neurological deficit (Table 5). Patients whose lesions were located in the brain stem, cerebellar peduncle, or thalamus tended to have a more debilitating and persistent neurological deficit subsequent to each rebleeding episode. Surgical Management Of the nine patients who underwent surgery, eight had total excision of their lesions. The remaining patient, with a large pontine lesion, underwent partial excision and has continued to deteriorate due to radiographically documented recurrent hemorrhages. Four patients refused surgery, two of whom have subsequently suffered a recurrent hemorrhage. Microsurgical technique was used in all instances due to the relatively small size of these lesions (the majority being < 1 cm in diameter). At surgery, all lesions showed evidence of old and recent hemorrhage. The majority of lesions appeared as small tangles of delicate blood vessels under relatively low pressure, some with thrombosed or partially thrombosed lumens. One lesion had formed a large cyst which was filled with blood clot of varying ages. One lesion had a mulberry shape typical of a cavernous angioma. Five patients improved following surgical excision, two remained unchanged, and two experienced an increased neurological deficit as the result of surgery. One patient with increased deficit had a lesion in the superior cerebellar peduncle which had hemorrhaged for a second time 9 months after stereotactic radiosurgery using the Bragg peak method. He has a mild hemisensory deficit with a slightly detectable cerebellar hemiataxia. He has returned to full activity as a chemical engineer. The second patient who experienced neurological deficit subsequent to surgery harbors a large pontine vascular malformation. Although she was ambulatory after surgery, progressive intermittent neurological deficits have persisted due to recurrent hemorrhages from the unresected part of the lesion. None of the nine patients undergoing surgery died. In the majority of deep-seated lesions, ultrasound was utilized to identify and locate the lesion. In most circumstances, these lesions showed a mixed echogenic signal of high and low intensity similar to the signal from MR imaging. No patient underwent CT-guided stereotaxis to locate the lesion. J. Neurosurg. / Volume 73/August, 1990
Persistent Transient No Total Deficits Deficits Deficit Cases 3
3
0
6
1 1
2 1
0 2
3 4
Histological Findings Detailed pathological review was available for seven of the nine surgical specimens. Originally, four of these specimens were considered arteriovenous malformations (AVM's) but, upon further study, all cases were reclassified as cavernous angiomas. In all seven cases with adequate tissue for histopathological examination, there was evidence of recent and old hemorrhage. Old hemorrhage was characterized by fibrosis, the presence of macrophages, and the deposition of hemosiderin pigment. In all cases, the hemorrhage and the reaction to it had obliterated normal tissues and part of the vascular malformation, and careful evaluation of the edge of the hematoma was necessary to identify the residual part of the lesion. Diagnosis of cavernous angioma was based on the identification of large endothelium-lined vascular channels separated by varying amounts of collagen (Fig. 4). Except at the edges, brain substance was not present between the vascular spaces. Fibrosis of the walls of the vascular spaces was present at least focally in all cases (Fig. 4 right) and, when extensive, resulted in a marked thickening of involved vessels. This caused a superficial resemblance to an AVM, and required elastic and trichrome stains for accurate diagnosis. The absence of elastic fibers and smooth muscle in the thickened wall (Fig. 5) rules out arterial and venous components in the vascular malformation. The demonstration that the thick walls are composed of collagen alone favors the diagnosis of cavernous angioma. The performance of special stains for elastin and smooth muscle contributed to the reclassification of four cases originally considered AVM's as cavernous angiomas. Discussion
Terminology and Literature Review The susceptibility of small vascular malformations to hemorrhage was first described in 1951 by Margolis, et al., ~9 four of whose patients suffered fatal hemorrhage. Crawford and Russell 8 subsequently introduced the term "cryptic" to describe these small (< 2 to 3 cm) vascular malformations which caused spontaneous cerebral hemorrhage. Initially, the term "cryptic" was applied to the group of vascular malformations based on their small size alone. However, more recently the terms "cryptic" and "occult" have been applied to those vascular malformations that cannot be identified by cere177
H. Tung, et al.
FIG. 4. Photomicrographs of cavernous angioma specimens. H & E, x 100. Left: Specimen showing endothelium-lined spaces separated by scant connective tissue. Right: Specimen with evidence of fibrosis of the vascular wall resulting in compression of the vascular spaces and thickening of the wall.
bral angiography. 1,6.21,23A more appropriate term would be "angiographically occult." Thus, the designation "angiographically occult vascular malformation" represents a myriad of vascular malformations with variable histology and includes cryptic AVM's, thrombosed AVM's, cavernous angiomas, and other histological vascular lesions escaping detection by cerebral angiography. Recent reports have thoroughly reviewed the histopathology, radiographic findings, and surgical management of patients with angiographically occult vascular malformations of the brain. 6'17'23'28 In these and other series, recurrent hemorrhage has been mentioned as the pathophysiological mechanism for neurological deftcir. 3,4,6A~ A review of the literature reveals that 80% to 90% of angiographically occult vascular malformations showed evidence of hemorrhage. 17'28 Even in asymptomatic cases or those presenting with seizures, MR imaging and pathological examination have documented the presence of hemosiderin, presumably the residue of previously experienced subclinical hemorrhage. 3"6,12AS'17"2L22,24'25'28However, the clinical and temporal profiles of clinically relevant hemorrhages have not been highlighted in the past. Location and Histopathology o f Lesions Our series of 13 patients differs in some ways from previously reported cases. In a review of the literature in 1988, Lobato, et al., ~7 found that over 75% of angiographically occult vascular malformations were located in the supratentorial region. In our series, eight of the 13 patients had lesions in either the brain stem or the cerebellum. Likewise, in Lobato's review, 31% of the lesions were cavernous angiomas. All of the lesions in this series examined pathologically were cavernous angiomas. The routine use of trichrome stain for smooth muscle and collagen and stains for elastic 178
fibers can help differentiate among the various pathological subtypes of vascular malformations for a more accurate diagnosis and clinical correlation. In addition, the diagnosis of angiographically occult vascular malformation is not specific despite modern imaging techniques and suggestive clinical history. The diagnosis can only be made definitively by surgical pathological confirmation. 2'4'6'12'17'21'23'25It is possible, although certainly not proven here, that clinically relevant recurrent hemorrhages may be associated with either one or both of these characteristics: namely, location or histopathology. Recurrent Hemorrhage In our series there was good correlation between the number of recurrent hemorrhages, the location of recurrent hemorrhages, and the occurrence of persistent neurological deficit. In most cases (80%) the initial rebleed caused only a transient deficit; however, with each successive recurrent hemorrhage, the likelihood of a resultant persistent neurological deficit increased, and all of the patients with more than two rebleeds suffered persistent neurological deficits. It is difficult to determine from our small series the relative influence on outcome of location versus number of hemorrhages. It is probable that hemorrhages, even small ones, are more likely to produce clinical symptoms or neurological deficits in those lesions involving hazardous areas of the brain, such as the brain stem, thalamus, and eloquent cortex, and to bring them to recognition earlier. Furthermore, angiographically occult vascular malformations located in eloquent areas of the brain are often initially managed conservatively, as surgery in these areas is most difficult and carries a higher risk of morbidity compared to that of angiographically occult vascular malformations located in other a r e a s . 12A7'21'23' 28-30 This may partially explain the disproportionate J. Neurosurg. / Volume 73/August, 1990
Recurrent hemorrhage from occult vascular anomalies number of lesions in our series located in the posterior fossa. Although the full spectrum of the natural history remains to be elucidated, angiographically occult vascular malformations are prone to cause recurrent hemorrhages. 1'2'6'1~176 Histologically, all of our surgical specimens revealed evidence of chronic and recent hemorrhage. In reviewing the literature, Lobato, et aL,'7 found that 31% of patients with documented histologically verified angiographically occult vascular malformations showed evidence of both chronic and recent hemorrhage. Similarly, Wakai, et al., 28 reported 44% of patients with chronic and recent hemorrhage. Obviously, it is difficult to determine the exact rate of rebleeding from these lesions, and long-term retrospective or prospective population studies will be necessary. However, the rate of rebleeding appears considerable and may be similar to that of high-flow angiographically apparent AVM's. All patients in this series presented clinical symptoms of hemorrhage or exacerbation of a previous neurological deficit. However, not all hemorrhages from angiographically occult vascular malformations may be clinically evident. Lobato, et al., w reported that 24% of the patients they reviewed had a clinical course suggestive of recurrent hemorrhage; seven of their own series of 21 patients with angiographically occult vascular malformations suffered recurrent symptoms of rebleeding. Although there were no deaths as a result of a recurrent hemorrhage in the present series, it is well known that patients may suffer fatal recurrent hemorrhages from these lesions. 1'8~t~176 It has, therefore, become our policy to recommend removal when they are discovered in surgically accessible noneloquent areas of the brain. Lesions found in more hazardous areas are removed only if they produce recurrent hemorrhage. Surgical T r e a t m e n t The surgical management of angiographically occult vascular malformations is not as straightforward as some have implied. They tend to be small and are not infrequently located deep in the cerebral or cerebellar hemisphere or brain stem. Localization of the lesions at the time of surgery may prove difficult. Therefore, we have relied on intraoperative ultrasonography to assist in this endeavor; CT-guided stereotaxis may be appropriate in certain circumstances. 9 Unlike their larger high-flow counterparts, the endpoint of surgery for angiographically occult vascular malformations may be difficult to determine. The lesions tend to be small with very fine vasculature. Frequently, the vessels are not tightly compacted, making it difficult to be certain of total removal. Even lesions with predominantly cavernous elements on histopathology may not demonstrate the classical gross appearance of the easily removed cavernous angioma. Vigorous exploration and surgical manipulation of the cavity may prove hazardous in brain-stem lesions and other eloquent areas. 21 24,27 30 Total removal in certain J. Neurosurg. / Volume 73/August, 1990
FIG. 5. Photomicrograph of an cavernous angioma specimen, stained for elastin, showing absence of elastic fibers in the thickened vessel wall. Elastin, • 200.
circumstances would be associated with unacceptable complications, especially in the case of pontine lesions. Radiation Therapy Recent reports have shown the efficacy of radiosurgery in the treatment of AVM's. 7'18'26 Although of theoretical value in the treatment of angiographically occult vascular malformations, the use of radiosurgery has met with limited success in treating cavernous angiomas as opposed to high-flow AVM's (L Steiner~ personal communication, 1989). In fact, one of our patients suffered a recurrent hemorrhage with a persistent neurological deficit 9 months following stereotactic radiosurgery with Bragg peak radiation, thus emphasizing the principle that recurrent hemorrhages may occur before complete obliteration of the lesion by radiosurgery or any other technique is achieved. In summary, this series of cases underscores the propensity of angiographically occult vascular malformations to cause recurrent hemorrhages. Each successive hemorrhage tended to occur at shorter time intervals, and each hemorrhage tended to be associated with an increased incidence of permanent neurological deficit. Surgery can be effective and relatively safe in removing these lesions even in hazardous areas of the brain. Acknowledgment The authors thank Pat Sullivan for manuscript preparation. References 1. Beeker DH, Townsend J J, Kramer RA, et al" Occult cerebrovascular malformations. A series of 18 histologically verified cases with negative angiography. Brain 102: 249-287, 1979 2. Bell BA, Kendell BE, Symon L: Angiographically occult arteriovenous malformations of the brain. J Neurol Neurosurg Psychiatry 41:1057-1064, 1978 179
H. Tung, et al. 3. Bitoh S, Hasegawa H, Fujiwara M, et al: Angiographically occult vascular malformations causing intracranial hemorrhage. Surg Neurol 17:35-42, 1982 4. Britt RH, Connor WS, Enzmann DR: Occult arteriovenous malformation of the brainstem simulating multiple sclerosis. Neurology 31:901-903, 1981 5. Chin D, Harper C: Angiographically occult cerebral vascular malformations with abnormal computed tomography. Surg Neurol 20:138-142, 1983 6. Cohen HCM, Tucker WS, Humphreys RP, et al: Angiographically cryptic histologically verified cerebrovascular malformations. Neurosurgery 10:704-714, 1982 7. Colombo F, Benedetti A, Pozza F, et al: Linear accelerator radiosurgery of cerebral arteriovenous malformations. Neurosurgery 24:833-840, 1989 8. Crawford JV, Russell DS: Cryptic arteriovenous and venous hamartomas of the brain. J Neuroi Neurosurg Psychiatry 19:1-11, 1956 9. Daumas-Duport C, Mann M, Munari C, et al: Cryptic vascular malformations diagnosed by stereotactic biopsies - - a preliminary study. Appl Neurophysiol 48:440-443, 1985 10. Giombini S, Morello G: Cavernous angiomas of the brain. Account of fourteen personal cases and review of the literature. Acta Neurochir 40:61-82, 1978 11. Gomori JM, Grossman RI, Goldberg HI, et al: Occult cerebral vascular malformations: high-field MR imaging. Radiology 158:707-713, 1986 12. Hashim ASM, Asakura T, Koichi U, ct al: Angiographically occult vascular malformations. Surg Neurol 23: 431-439, 1985 13. Kendall BE, Claveria LE: The use of computed axial tomography (CAT) for the diagnosis and management of intracranial angiomas. Neuroradioiogy 12:141-160, 1976 14. Kramer RA, Wing SD: Computed tomography of angiographically occult cerebral vascular malformations. Radiology 123:649-652, 1977 15. Kucharczyk W, Lemme-Pleghos L, Uske A, et al: Intracranial vascular malformation: MR and CT imaging. Radiology 156:383-389, 1985 16. Leblanc R, Ethier R: The computerized tomographic appearance of angiographically occult arteriovenous malformations of the brain. Can J Neurol Sei 8:7-13, 1981 17. Lobato RD, Perez C, Rivas J J, et al: Clinical, radiological, and pathological spectrum of angiographically occult intracranial vascular malformations. Analysis of 21 cases and review of the literature. J Neurosurg 68:518-531, 1988
180
18. Loeffier JS, Eben A, Siddon RL, et al: Stereotactic radiosurgery for intracranial arteriovenous malformations using a standard linear accelerator. Int J Radiat Oncol Biol Phys 17:673-677, 1989 19. Margolis G, Odom GL, Woodhall B, et al: The role of small angiomatous malformations in the production of intracerebral hematomas. J Neurosurg 8:564-575, 1951 20. McCormick WF, Nofzinger JD: "Cryptic" vascular malformations of the central nervous system. J Neurosurg 24:865-875, 1966 21. McFerran D J, Marks PV, Garvan NJ: Angiographically occult arteriovenous malformations of the brainstem. Surg Neurol 28:221-224, 1987 22. New PFJ, Ojemann RG, Davis KR, et al: MR and CT of occult vascular malformations of the brain. AJNR 7: 771-779, 1986 23. Ogilvy CS, Heros RC, Ojemann RG, et al: Angiographically occult arteriovenous malformations. J Neurosurg 69:350-355, 1988 24. Rigamonti D, Drayer BP, Johnson PC, et al: The MRI appearance of cavernous malformations (angiomas). J Neurosurg 67:518-524, 1987 25. Shuey HM Jr, Day AL, Quisling RG, et al: Angiographically cryptic cerebrovascular malformations. Neurosurgery 5:476-479, 1979 26. Steiner L, Lindquist C: Radiosurgery in cerebral arteriovenous malformation, in Tasker RR (ed): Neurosurgery: State of the Art Reviews. Philadelphia: Hanley & Belfus, 1987, pp 329-336 27. Veerapen RJ, Sbeih IA, O'Laoire SA: Surgical treatment of cryptic AVM's and associated hematoma in the brain stem and spinal cord. J Neurosurg 65:188-193, 1986 28. Wakai S, Ueda Y, Inoh S, et al: Angiographically occult angiomas: a report of thirteen cases with an analysis of the cases documented in the literature. Neurosurgery 17: 549-556, 1985 29. Wilson CB, U HS, Domingue J: Microsurgical treatment of intracranial vascular malformations. J Neurosurg 51: 446-454, 1979 30. Yoshimoto T, Suzuki J: Radical surgery on cavernous angioma of the brainstem. Surg Neurol 26:72-78, 1986 Manuscript received October 17, 1989. Accepted in final form February 5, 1990. Address reprint requests to: Steven L. Giannotta, M.D., Department of Neurosurgery, LAC-USC Medical Center, 1200 North State Street, Room 5046, Los Angeles, California 90033.
J. Neurosurg. / Volume 73/August, 1990
