945

Intracranial Cavernous : Neuroradiologic Review of 36 Operated Cases

Mario Savoiardo 1 Neuroradiologic studies in 36 cases of histologically verified intracranial cavernous Liliana Strada hemangiomas were reviewed. Radionuclide brain scans were positive in 17 of 19 Angelo Passerini examinations. , performed in 35 cases, usually showed an avascular area with absent or moderate mass effect. Capillary blush and / or early draining veins, often mentioned in single case reports, were observed in only seven cases. Computed tomography (18 cases) usually demonstrated a hyperdense, nodular, or irregular without significant mass effect and always enhancing after contrast injection. The combination of a long clinical history of focal epilepsy with computed tomographic and angiographic findings should suggest the diagnosis of cavernous . In all cases of so-called spontaneous hematoma with negative angiography, computed to­ mography should be repeated after a long interval to exclude the presence of a cavernous hemangioma or other cryptic .

Cavernous hemangiomas of the brain are well c irc umscribed vascul ar malfor­ mations composed of thin-wall ed sinusoidal spaces lined with endoth elium ; th ey have no elastic membrane, no muscular ti ssue, and no intervening nervous parenchyma. Their frequency seemed to be quite low simply because th ey often remain undiagnosed during life; however, th e path ologic materi al published by Mc­ Cormick et al. [1] in 1968 demonstrated th at they are the second most common cerebral vascular malformation after arteriovenous malform ati ons (AVMs). Fur­ thermore, the possibility of diagnosing them offered by computed tomography (CT) has increased their recognition during life . We revi ewed all the neuroradio­ logic studies performed in a series of 36 intracranial cavern ous hemangiomas operated on at the Istituto Neurologico of Milan from 19 54 to 1981. The clinical and surgical aspects of th e first part of this seri es (14 pati ents) were reported by Giombini and Morello [2], and preliminary neuroradiologic review has also been reported [3].

Materials and Methods

Thirty-six cases of hi stologicall y verifi ed intracranial cavern ous hemangiomas were reviewed. There were 20 males and 16 females. Th ey were 14 months to 6 1 years old. The cavern ous hemangiomas were less common under age 20 (seven cases) and over 50 (six cases), bein g alm ost equall y distributed in the th ird , fou rth , and fifth decades. The most common clinical presentati on was epil epsy (25 cases), usuall y focal (2 1 Received October 19 , 1982; accepted aft er revision Janu ary 28, 1983. cases). Focal neurologic signs were present in 12 pati ents. Th e c linical presentation was sometimes acute, caused by hemorrhage leading to immediate hospitali za ti on and diagnosis 1 All authors: Department of Neuroradiology , Is­ tituto Neurolog ico , Vi a Celori a 11 , 20133, Mil an, (six cases); however, in three of these cases, the had been preceded by a long Italy. Address reprint requ ests to M. Savoiard o. history of focal epil epsy. In one patient , the fi rst episode of bleeding led to the surgical removal of an intracerebral hematoma, but the cavern ous hemangioma was found 5 years AJNR 4:945-950, July/ August 1983 0 195- 6 108 / 83/ 0404-0945 $00.00 later during the second hemorrhage. At the ti me of operati on, the average duration of © Ameri can Roentgen Ray Soc iety symptoms was 4 .3 years, with a maximum of 27 years. 946 SAVOIARDO ET AL. AJNR:4, Jul. / Aug. 1983

The radiologic studies included skull films in all patients. Radio­ pathologic circulation was supplied by the temporal branch nuclide studies were avail abl e in 19 cases. Pn eumoencephalo­ of the middle meningeal artery; at surgery, a left temporal grams were obtain ed in 17 patients; in one case, air ventriculogra­ intracerebral cavernous hemangioma, also attached to the ph y had been performed. In all but one case, the appropriate dura, was found . angiographic study (carotid or vertebral or both) was perform ed; in Of the five cases with repeat angiography, two showed fi ve pati ents angiography was performed twice. CT scans were some differences. There was an early draining vein in one available in 18 pati ents; in 17, both pre- and postcontrast studies were performed; in one, only a precontrast scan was obtained. and pathologic circulation in the other (fig . 3). Of the supratentori al cavernous hemangiomas, 14 were located CT demonstrated the cavernous hemangioma in 17 of 18 in the left hemi sphere and 18 in th e ri ght hemisphere without cases in which it was performed. In one case, a very small preferred localization. In one case the locati on of the cavern ous cortical cavernous hemangioma not seen on CT was an hemangioma was extradural; th e lesion eroded the right lesser wing incidental operative finding in a patient with astrocytoma. of the sphenoid, bulging both into the middle fossa and the orbit. On precontrast study, the lesion appeared hyperdense in In three cases, the cavern ous hemangiomas were located in the 11 cases, usually as a nodular area without mass effect posteri or fossa. One was in the pons, one in the right cerebellar (figs. 4 and 5). In three cases, mi xed hyperdense and hemi sphere, and the third was attached to th e dura and represented hypodense areas were observed, with a ring pattern in two. an in cidental operative finding in a patient with chordoma of the In one of these two cases, the hypodensity was well defined cli vus. These occurred in associati on with other in and corresponded to a cyst. In three cases, the cavernous th ree oth er instances. One right frontal cavern ous hemangioma was hemangioma, isodense, was recognizable only after con­ a surgical incidental fi nding in a pati ent with an astrocytoma. trast injection. On postcontrast study, enhancement was An oth er pati ent, in whom a cavern ous hemangioma had been re­ observed in all cases, usually nodular, well delineated, and moved from the left hemisphere, was readmitted 9 years later with of various intensity. right hemi sphere symptoms. Three lesions were found ; one in the Calcifications were seen in five cases (fig. 6). In five roland ic area, which had bled, and two very small ones in the cases mass effect was present; it was marked in two (fig. temporooccipital region. They were all removed. Histologic exami­ 7). In seven cases, poorly defined hypodensity surrounding nati on demonstrated venous dyspl asia at the periphery of th e he­ the lesion was considered consistent with edema or gliosis. matoma and in the other two locati ons, without definite aspects of cavernous hemangioma. It is noteworthy that this same pati ent also had a cavernous fibrohemangioma removed from his leg. A third patient had a coexistent meningioma. Discussion

Results Cavernous hemangiomas are peculiar with respect to the other vascular malformations because they do not have Skull films showed fine granular or coarse calcifications intervening brain parenchyma among the vascular spaces in three cases. Sellar changes consistent with increased that form the malformation. They do not have a true capsule, intracranial pressure were found in three patients. In one of but appear, on gross pathology, as well delineated brownish these, however, they were attributable to a coexistent or reddish masses, often polylobulated. Cavernous heman­ parasagittal meningioma. In the only extradural lesion, bone giomas can be extremely small or a few centimeters in erosion of the right lesser wing of the sphenoid was ob­ diameter, but they do not reach the large size of some AVMs served. and they bleed less frequently than AVMs [1]. Radionuclide brain scans were positive, usually with a They are the second most common type of vascular marked uptake, in 17 of 19 examinations. The negative malformation, and, in the large pathologic series of Mc­ scans occurred in two small lesions with diameters less than Cormick et al. [1], they had a 1 :3.6 ratio with AVMs. By 2 cm. contrast, in the neuroradiologic literature, papers related to Pneumoencephalography was abnormal in 11 of 1 7 ex­ cavernous hemangiomas have been extremely rare until a aminations; th e location of the cavernous hemangioma was few years ago, thus seeming to be less common than usually indicated by mass effect upon the ventricular sys­ published pathologic series suggest. This was probably tem. In two cases, local cortical atrophy or focal absence of because this type of often remained undiagnosed fi lling of the sulci in th e area of the hemangioma was found. even after angiographic examination and was detected only The only air ventriculogram demonstrated a large right at surgery, after it had bled [4]. pari etal lesion. In the early 1970s, there were several reports in the Angiography was completely normal, even on retrospec­ neuroradiologic literature. Most comprised single cases or tive analysis, in seven of 35 cases in which it was performed. very small series. They indicated a capillary blush and/ or In 13 cases, displacement of vessels without pathologic an early draining vein as the diagnostic features of the circulati on was seen ; the displacement was usually mild or cavernous hemangiomas [5-10]. In the last few years, CT moderate. In eight other cases, the pathologic findings were has increased the possibility of detecting these often angio­ limited to an avascular area visible in the capillary phase graphically occult vascular malformations [3, 11-13]. (fig. 1). In seven cases, pathologic circulation was seen, Reviewing our series, previously common studies such generally consisting of a faint capillary blush and/ or early as pneumoencephalography and ventriculography ap­ draining veins (figs. 2 and 3). In one of these cases, marked peared to be of little value, and they are not discussed here. AJNR:4, Jul. / Aug. 1983 INTRACRANIAL CAVERNOUS HEMANGIOMAS 947

Fig. 1.-Primarily " avascular" le­ sion. Cavernous hemangioma in left frontal lobe. Carotid angiogram. Avas­ cular area with minimal displacements - more evident in late arterial or capillary phase (arrows). -

Fi g. 2.-0ccult hemangioma. Carotid angiogram, sequential subtraction. Early draining vein (arrows) is only pathologic finding. 1 2

Fig. 3.-Hemangioma with blush in epilepsy patient. A, Earl y draining vein. B, 22 months later. Early capillary blush also evident (arrows). A B

A B c

Fig. 4.-Hyperdense, yet "avascular," lesion. A and B, Precontrast CT scans. Hyperdense large lesion wi thou I mass eHect. C, Carotid angiogram. Avascular area without displacement in late arterial phase (arrows).

Skull films also were of little help, because calcifications are was remarkably small and not recognizable even on CT. We better detected by CT. beli eve that radionuclide scanning should now be consid­ The value of radionuclide brain studies has also greatly ered a complementary study [11], useful in cases of heavily diminished since the introduction of CT. Nevertheless, it calcified lesions. In fact, in one case appearing as a was negative in only two cases. In one of these, the lesion " brainstone," in which enhancement on postcontrast CT 948 SAVOIARDO ET AL. AJNR:4, Jul./ Aug . 1983

was doubtful because it was masked by the heavy calcifi­ inent than those seen in cavernous hemangiomas, and their cations, the radioisotope study demonstrated a readily de­ visualization may persist in late venous phase. tectable pathologic uptake. A peculiar finding described by Namaguchi et al. [1 8], Review of the angiographic studies in our series reveals and observed by us only in a 14-month-old girl, is sedimen­ th at the reports from the early 1970s are mi sleading. Cap­ tation of contrast medium in the large cavernous spaces of ill ary blush and earl y draining veins are an infrequent feature the angioma (fig . 6c). This is probably caused by slow in cavern ous hemangiomas (figs. 2 and 3); the most frequent circulation because of small feeding vessels and the large findings are mild mass effect and an avascular area in the vascular bed within the malformation. However, this feature parenchymatous phase of the angiographic series (fig. 1). was not detected in five cases, in which 15 ml of contrast An giography's efficacy is limited even considering the tech­ medium was injected slowly, as was suggested by these nical improvements that have occurred. Cavernous heman­ authors. Sedimentation of contrast medium is much more giomas usually appear as avascul ar areas even with sub­ common in cavernous hemangiomas of the orbit; we ob­ traction, 2: 1 magnification, and stereo views. In addition, served it in more than half of th e cases in this location after th e capillary blush and the early draining vein are very normal injection of the usual amount of contrast medium nonspecific findings, compatible with neoplastic, ischemic, [19]. and inflammatory lesions [1 3, 14]. Considering other types CT detected the lesion in all the cases (except the one of vascul ar malformati ons, an early draining vein may be found in cidentally at surgery), and in one case it also dem­ observed also in capillary telangectasias [15], in larg ely onstrated two additional asymptomatic vascular malforma­ thrombosed AVM s, and in venous malformations [1 6, 17]. tions. There are no pathognomonic features of cavernous In venous malformations, however, th e veins are more prom- hemangiomas, but the most common and characteristic findings are slightly hyperdense areas with enhancement after intravenous contrast injection, without significant mass effect [11]. CT was superior to angiography in defining the exact boundaries of the lesions. In a few cases, followed for years and studied by angiography before the advent of CT, the neurosurgeon felt safe in operating on high-risk areas only after CT demonstration of the limits of the lesion [3]. In two cases in our series, th e cavernous hemangioma caused a hematoma, but the hemangioma was not found at the time of surgery. Only after the second episode of bleed­ ing was the hemangioma removed. Therefore, in all cases of so-called " spontaneous hematoma" in which angiogra­ phy fails to show a vascular malformation , CT should be repeated after a long interval, after the disappearance of the hyperdensity of the hematoma and of its enhancing periph­ eral rim , because of the possibility of detecting the patho­ logic nodule. A B Even if CT findings of cavernous hemangiomas are not specific, the long clinical history of epilepsy and the absence Fig. 5. -A, Precontrast CT scan. Markedty hyperdense nodular lesion. B, Postcontrast scan at slightly lower level. Enhancement of leSion, which is of mass effect are quite helpful in differentiating them from surrounded by hypodensity. tumors. In th e 14-month-old girl mentioned above (fig . 6),

Fig. 6. -Frontal cavernous heman­ gioma in 14-month-old girl. A, Precon­ trast CT. Calcificati ons without mass ef­ feel. B, Postcontrast CT. Marked en­ hancement surrounding calcifications. C, Carotid angiogram. Pooling of con­ trast medium in late ve nous phase. A B c AJNR:4, Jul. / Aug . 1 983 INTRACRANIAL CAVERNOUS HEMANGIOMAS 949

A E

Fi g. 7.- Cavernous hemangioma simulaling a meningioma. Pre-(A and B) and posl- (C and 0) conlrast CT scans. Markedly enhanc ing lesion, adherenl to dura, mimics meningioma. Intratemporal cavernous hemangioma, sur­ rounded by marked edema, causes sig nificant mass effect. E, Extern al carotid angiog ram. Pathologic c irc ulation is patc hy ( arrows ), different from typical stain of meningiomas. c o

reactive or neoplastic astrocytic tissue was found, surround­ hemangiomas operated on at our institute; 18 were removed ing and intermixed with the cavernous hemangioma. A 4 in the 22 years before CT , and 18 were removed in th e only year follow-up, however, did not show recurrence of tumor. 6 years since CT became available. This case may represent a hamartoma or an angioglioma such as described by Fischer et al. [20]. In the group of vascular malformations, thrombosed REFERENCES AVMs may have similar CT features and may be indistin­ guishable from cavernous hemangiomas except on histo­ 1. McCormic k WF, Hardman JM , Boulter TR . Vascular malfor­ logic examination [12]. In the differential diagnosis from mati ons (" 'angiomas" ) of th e brain , with special reference to capillary and venous vascular malformations, which may those occurring in th e posterior fossa. J Neurosurg 1968;28:241-251 have similar features on angiography (i.e., a group of drain­ 2. Giombini S, Morell o G. Cavern ous angiomas of the brain . ing veins reaching a single larger collector), CT is very Account of fo urteen personal cases and review of the literature. helpful in showing the nodular component of the cavernous Acta Neurochir (Wien) 1978;40: 6 1-82 hemangioma. This has important clinical implications. The 3. Savoiard o M, Passerini A. CT, angiography, and RN scans in venous malformations, in which CT may only demonstrate intracranial cavern ous hemangiomas. Neuroradiology the venous channels [17, 21], should not be operated on 1978; 16: 256-260 because they rarely bleed and because the presence of 4 . Voigt K, Ya:?a rgil MG. Cerebral cavern ous hae mangiomas or normal intervening nervous tissue among the vascular chan­ cavern omas. Incid ence, pathology, localizati on, diagnosis, nels carries a higher risk of postoperative damage. The clinical features and treatment. Review of the literature and report of an unusual case. Neurochirurgia (Stutt g) capillary telangectasias are rare, and usually occur in the 1976; 19: 59- 68 brainstem , which is not accessible to preventive surgery; 5. Bogren H, Svalander C, Wickbom I. Angiograph y in intracran ial surgery, therefore, is justified only for the evacuation of a cavern ous hemangiomas. Ac ta Radiol [Oiagn] (Stockh) hematoma. 1970;10 :8 1-89 The best demonstration of the impact of CT in demon­ 6. Jonutis AJ , Sondheimer FK, Klein HZ, Wise BL. Intracerebral strating angiographically occult vascular malformations is cave rn ous hemangioma with angiographicall y demonstrated offered by the distribution over the years of the cavernous pathologic vasculature. Neuroradiology 1971 ;3: 57 - 63 950 SAVOIARDO ET AL. AJNR:4. Jul. / Aug. 1983

7. Namag uchi Y, Fukui M, Miyake E, et al. An giographic manifes­ skull and brain. SI. Louis; Mosby, 1974: 1878-1902 tations of intracerebral cavernous hemangioma. Neuroradiol­ 15. Roberson GH , Kase CS, Wolpow ER . Teleangiectases and ogy 1977;14:113-1 16 cavern ous angiomas of the brainstem: " cryptic" vascular mal­ 8. Segall HD, Segal HL. Teal JS. Rumbaugh CL. Bergeron RT. formations. Report of a case. Neuroradiology 1974;8: 83-89 Calcifying cerebral cavern ous hemangioma with brain scan 16. Michels LG , Bentson JR, Winter J. Computed tomography of and angiog raphic findings. Neuroradiology 1974; 7: 1 33-1 38 cerebral venous angiomas. J Comput Assist Tomogr 9. Di amond C. Torvik A. Amundsen P. Angiographic diagnosis of 1977;1 :149-154 teleangiectases with cavernous angioma of the posterior fossa. 17. Saito Y, Kobayashi N. Cerebral venous angiomas. Radiology Report of two cases. Acta Radiol [Oiagn] 1976; 17 : 281 -288 1981 ;139:87-94 10. Liliequist B. Angiography in intracerebral cavern ous heman­ 18. Namaguchi Y, Kishikawa T, Fukuy M, et al. Prolonged injection gioma. Neuroradiology 1975;9: 69-72 angiography for diagnosing intracranial cavernous heman­ 11. Bartlett JE. Ki shore PR S. Intracran ial cavernous angioma. AJR giomas. Radiology 1979;131 : 137-138 1977; 128 : 653-656 19. Savoiardo M, Strada L, Passerini A. Cavernous hemangiomas 12. Becker DH , Townsend JJ, Kramer RA . Newton TH . Occult of the orbit: value of CT, angiography and phlebography. AJNR cerebrovascular malform ations. A series of 18 histologically 1983;4: 741-744 verified cases with negative angiography. Brain 1979;102 : 20. Fischer EG , Sotrel A, Welch K. Cerebral hemangioma with glial 249-287 neoplasia (angioglioma?). Report of two cases. J Neurosurg 13. Leblanc R. Ethier R. Th e computeri zed tomographic appear­ 1982;56: 430-434 ance of angiographically occult arteriovenous malformations of 2 1. Partain CL, Guinto FC , Scatliff JH , Limbacher J, Janicki P, th e brain. Can J Neurol Sci 1981 ;8: 7 -1 3 Heindel CC . Cerebral venous angioma; correlation of radionu­ 14. Heinz ER . Pathology involving th e supratentorial veins and clide brain scan, transmission computed tomography, and dural sinu ses. In: Newton TH. Potts DG , eds. Radiology of the angiography. J Nucl Med 1979;20: 11 66-1169