- Email: [email protected]
Spinal intramedullary cavernous angiomas: A literature metaanalysis
ELSEVIER
Spinal Intramedullary Cavernous Angiomas: A Literature Metaanalysis Sergio Canavero, M.D., Carlo Alberto Pagni, M.D., Sergio Duca, M.D., and Gianni Boris Bradac, M.D. Institute of Neurosurgery and Chair of Neuroradiology, University of Turin, Turin, Italy
Canavero S, Pagni CA, Duca S, Bradac GB. Spinal intramedullary cavernous angiomas: a literature meta-analysis. Surg Neurol 1994;41:381-8.
The present report reviews 57 (out of 65) cases of spinal intramedullary cavernomas collected from the literature, plus one personal patient. Almost 70% of all patients were women. Mean age at diagnosis for women was 36.4 years, with a peak in the third decade. More than threefourths of all women became symptomatic between the second and fourth decades, with a peak in the fourth decade. Unlike in men, cervical and thoracic lesions a r e almost equally represented, generally involving 1-2 vertebral levels. Mean size at diagnosis is 1.7 cm; no enlargement over time was seen. Symptoms are more frequently acute; pain and sensorimotor deficits are the usual complaints, but the clinical picture may simulate that of multiple sclerosis. The duration of history was less than 5 years in more than 80% of women. Bleeding was seen in 60% of women, with a risk of 1.6%/personyear of exposure globally. Cervical lesions have both a shorter course and increased frequency of bleeding. I f not immediately recognized, repeated cycles of bleeding are the norm, with a mean interval of 39.6 months between the first and second episodes. The preoperative status was the single most important factor bearing on outcome, whereas sex, age, size, location, duration of history and extent of removal were not. Magnetic resonance imaging was diagnostic in all cases, whereas angiography was 100% negative. Surgery should not be a necessary first option, as recovery from the first bleeding is apparently fairly frequent. K E Y W O R D S : Cavernoma; Cavernous angioma; Intramedullary; Natural history; Outcome; Spinal cord
Cavernous angiomas of the central nervous system (CNS) are berry-like vascular malformations in which
Address reprint requests to: Sergio Canavero, M.D., Via Morazzone 21, 10132 Torino, Italy. Received October 2, 1992; accepted March 3, 1993.
© 1994 by ElsevierScienceInc.
abnormally dilated blood spaces are closely clustered together, without intervening neural tissue. The surrounding parenchyma is often gliotic and hemosiderin-stained and may occasionally contain small lowflow feeding arteries and draining veins [44]. By far the great majority of CNS cavernomas is found in the cerebral hemispheres and brain stem. Many are entirely asymptomatic and up to 50°A may be multiple [41]. In 1990, we reviewed the literature on spinal subdural extramedullary cavernomas [35]; another case appeared in that year [38] and two possible cases were described by Bassett and colleagues [2]. At the time, we briefly reviewed intramedullary cases. The impression was that, in spite of many recent case reports and short series, spinal intramedullary cavernomas (SIC) remain a poorly known entity, probably overaggressively treated, without a proper definition of both their course and outcome. Highlights of our results have appeared in letter form [5a].
Case Report A 48-year-old woman, with a history of Graves' disease and ache rosacea, noticed at the time of her pregnancy (1977) a slow reduction of her right leg's sensitivity to temperature; this progressed over 3 years to involve the limb in toto and there were occasional episodes of urinary incontinence. In 1980, she gradually developed right leg weakness and gait disturbance. In January 1981, she complained of sudden, violent, boring pains in her lower abdomen, radiating to the right thigh, anteriorly. A lumbar puncture was followed by sudden spastic paraplegia, with gradual recovery 3 months later. A spinal computer tomography (CT) was normal. Discharge diagnosis was "demyelinating syndrome." Thereafter, she had recurrent, yearly (generally in winter) sensorimotor disturbances in her right leg; each time, steroids and rehabilitation brought about complete recovery. Since 1988, she suffered paresthesia only Ca feeling of a ball pressing on the sole") plus hypoesthesia. In March 1991, she noticed the onset of dysesthesia in [)090~ ~019/9~/$7 O0
382
Canavero et al
Surg Neurol 1994;41:381-8
Figure 1. MR images (T1Wls. left and center; T2WI. right) of the cavernousangioma (seecase report),
her left groin, with radiation to the sacrum. At this time, magnetic resonance imaging (MRI) disclosed a noncontrast-enhancing lesion with the characteristic neuroradiologic features of a cavernoma (Figure 1). On admission, a shuffling, spastic gait was apparent, with paraparesis and bilateral Babinski signs. The right leg was hypoesthetic to painful stimuli, and both legs lost vibratory sense. In November 1991, a D9 laminectomy was performed. Under the microscope, a softened, tawny cord appeared encased by a thickened arachnoid and many irregular vessels. A large draining venous channel was partially coagulated. The posterior columns were thin and yellow, under which dark-blue, solid lumps amid pitch-black blood and white-gray fibrotic nodules with many venules and arterioles were seen. Small transparent spherules ("crystal-like") were noted. The cleavage plane between this 1.5-cm mass and the surrounding white matter was found; yet, in several points, the two were tightly adherent, with fibrotic bridges. The lesion was removed, but several small portions were left, lest the cord could be damaged. Postoperatively, there was a global sensorimotor worsening with urinary retention. In February 1994, the result is unsatisfactory, as the patient cannot walk alone. The urinary retention resolved. The patient has developed central pain. Pathologically the lesion was a typical cavernoma.
gectatic [17] and another with an extramedullary component [14].
Incidence
W e have experienced three cases of SIC (of whom only the operated one is presently considered) since 1981; two from the Turin area. No case was seen at the Institute of Neurosurgery of the University of Turin in over 45 years. Krayenbuehl and Yasargil removed one single case over 25 years [22], Villani and colleagues three over 4 years [59], Stein six cases over 7.5 years [28], and Requena and co-workers one over 3 years [39]. However, no epidemiologic data were reported by these authors. Many case series of CNS cavernomas do not include any spinal cases [8,42]. W e do not know the prevalence of SIC in the general population.
Sex a n d Age
The female sex is affected in more than two-thirds of the cases (40 of 58). Mean age at diagnosis was 37.1 years (median: 35 years; range: 12-65), with women being diagnosed somewhat earlier than men (Figure 2). Both sexes developed their first symptom in the first decades of life, with a peak in the fourth (Figure 3). No correlation was found between level of the lesion and age of onset.
L i t e r a t u r e Meta-analysis From 1899 to 1991, 65 SICs have been reported in the literature [ 1-7,9-11,13-15,17-20,22-24,26-36,39, 4 3 - 4 8 , 5 0 , 5 2 , 5 4 - 6 0 , 6 2 - 6 5 ] , some with incomplete data as to age, sex, and site [7,13,15,20,22,29,39,54]. In this review, we included a few cases referred to as angiomas that on close scrutiny were most likely cavernomas [2,50,52,60], one case in which a few areas were telean-
Level a n d Extension
In female patients, cervical and thoracic lesions were equally represented (18 cervical, 19 thoracic, 1 cervicothoracic, 3 lumbar), whereas in male patients the proportion was 1:2 (5 cervical, 12 thoracic, 1 lumbar). Single-level lesions were the majority (n = 17, in the female sex), but two-level cases were also frequent (n =
Spinal Intramedullary Cavernous Angiomas
Surg Neurol 1994;41:381-8
5
(MALES) iH
I
21 7o.6~
I
I
I
i
10
10
J 9 (FEMALES)
1
<
383
then gradual (an acute episode which, after peaking, slowly worsened); (3) gradual (a history of slow onset and gradually worsening sensorimotor symptoms). The onset of symptoms was more frequently acute than gradual (56.5% vs 43.4%) in both sexes. Four cases gave no symptoms for a lifetime [14,24,63,64]. Thirty-three patients out of 54 (61%) reported pain sometime in their history, 26 (47%) as their initial symptom, alone or in association; this was more frequently acute than gradual (26 vs 17 cases). While in one case involving C1 it mimicked tic douloureux [46], according to the site of involvement; it affected the neck, the back, the groin, or the limbs; in fact, in simulated sciatica [49], a heart attack (upper thoracic or interscapular) [6,52] and acute appendicitis [28]. Pain may recede and disappear completely or be followed by other symptoms. Yet, opening complaints may be acute or gradual motor deficits (par-
@
I
Z
4
4
~,4%
I
1~19
3112
<
I
I 0-9
4
i 20-~g
I 30-3g
I 4~4g
(MALES) 2
I
I 50.69
Z i
AGE AT DIAGNOSIS BY DECADE
9F" 11
Figure 2. Relation between sex and age at diagnosis (n = 55). Mean age
at diagnosis was 36.4 years (median: 33 years; range: 12-65 years) for women and 39 years (median: 38 years; range: 15-58 years) for men. Note that when diagnosis was made in the fourth decade, history had been less than l year (women: 47%; men: 50%); when made in the third decade, of less than 5 years (women: 47%; men: 40%); and when made in the fifth (women: 50%) or sixth decade (men: 50%), this had been 5 years or more. A history of less than l year was seen in 30% (13) of women and 40% (6) of men; of less than 5 years in 44% (15) of women a n d 3 3 % (5) of men; of years or more in 18% (6) of women a n d 2 7 % (4) of men.
17, in the female sex). Three or more levels were involved exceptionally. In women, SIC cluster at C3-6, T2-3 and T7, although this does not seem to be the case in men.
i
10
(FEMALES)
~q
< L~
Z
2
2
Size Mean size was 1.7 cm (range: 7 m m - 3 cm). MRI did not detect any enlargement during follow-up of two cases (4 months to 3 years) [26,30], and the same applied to the two cases we are following (2 and 4 years). Myelography suggested possible enlargements, but this technique is not as specific.
Symptoms and Course We considered three patterns of presentation: (1) acute (single episode of less than 7-day duration); (2) acute,
0-9
10=19 80,-29 30--39 40-49
50-59
60-69
AGE AT ONSET OF SYMPTOMS BY DECADE F i g u r e 3. Relation between sex and age at onset of symptoms (n = 50). More than 74% of women and 73% of men (globally: 74%)first developed symptoms in the second to fourth decades, with a peak in the fourth in both sexes (31% of women, 33 % of men, 32% globally). No cases were found in the first decade or beyond the sixth (men) or seventh (women). In women, mean onset of cervical cases was at least 34 years, of thoracic lesions at 33.5 years.
384
Surg Neurol 1994;41:381-8
~ORACIC
Canavero et al
CBRVICAL
Females Malos
ioo~
o
~oo~
Figure 4. Gross bleeding according to location of cavernomas and gender, Cervical lesions bleed more in both sexes. Globally, 72. 7% of cervical lesions and 48.3% of thoracic lesions bled macroscopically. Women bled more often than men (60.5% vs 53%). Considering three lumbar cavernomas, the total rate of gross bleeding was 57. 1%. It should be noted that all these data are likely underestimated as several reports could not be defined.
aplegia, footdrop, others) or sensory disturbances other than pain (paresthesias). Although sphincteric problems are often seen, they never were the initial complaint. SIC may simulate multiple sclerosis (6 of 54:11%) [28,47, plus our personal case]. In a few cases, symptoms were precipitated by lumbar puncture (our case), minor trauma [27,28], and surgery [56]. Acute symptoms (both motor and sensory) completely, or, in case of paraplegia or severe paraparesis, almost completely receded. Symptoms presented during pregnancy in four cases (14% of fertile women; i.e. those from age 14 to 48) [26,55,59, plus our case]; in one patient, they occurred at two different pregnancies, thereafter completely recovering [55]. While all cases not immediately recognized and treated presented recurrences, in one patient submitted to biopsy, no relapse occurred over 15 years postoperatively [6]. The duration of history ranged from hours to decades in both sexes, but was generally less than 5 years (80% of all cases). Cervical lesions had a shorter history than thoracic cases in women.
Bleeding Gross bleeding was evaluated on the basis of M R I findings (acute or subacute blood outside the hemosiderin ring of the lesion) or surgical or pathologic evidence of a fresh clot outside the borders of the lesion. W o m e n bleed more frequently than men, and, in both sexes, cervical lesions are at greatest risk. These data are probably underrated, because several reports were equivocal (Figure 4). There was no correlation between size and risk of bleeding, as shown in cranial cases [42]. In this series, the risk of hemorrhage was 1.6%/person-year of exposure (at least 30 events/1867 years), assuming an equal risk throughout life and a congenital origin. However, (1) cavernomas may not be congenital in all cases, as new lesions may appear [49], (2) pregnancy increases the risk of hemorrhage, and
(3) cervical lesions bleed more frequently. In those cases in which the SIC was not immediately recognized, another bleeding episode was the rule, with a mean interval of 39.6 months (range: 2 - 1 6 8 months). One of the two cases with a 2-month interval was a pregnant women [26]. H e m o r r h a g e can occur with a gradual presentation. Subarachnoid hemorrhage, unlike subdural extramedullary lesions [35], is virtually unknown, although a xantochromic CSF may be encountered exceptionally.
Surgical Outcome W e assigned to each reported patient one of the following scores, both preoperatively and postoperatively: good (patient fully ambulatory, no sphincteric deficits, possible sensory disturbances), fair (patient with mild to advanced motor disturbance and/or gait restriction, sphincteric deficits, limited autonomy), and poor (patient bed-ridden or confined to a wheelchair; severe sphincteric disturbances). Intermediate scores (e.g., fairto-good) were used if authors' descriptions were equivocal. Total removal was possible in 24 women and eight men, partial in four women and three men, while a biopsy was done in three women and one man. Results were as follows: (1) women: deaths 7, poor 9, poor-tofair 1, fair 11, fair-to-good 2, good 4; (2) men: deaths 3, poor 2, fair 8, good 2. O f 11 cases where size was reported, a poor result was obtained in three lesions larger than 2 cm and in one of less than 2 cm, a fair result in three cases of lesions larger and smaller than 2 cm, whereas a good result in a case of less than 2 cm. Neither size nor extent of removal significantly influenced the outcome: the only appreciable difference between pre-1970 and post-1970 reports was that no deaths were seen after 1970. The most important single factor bearing on the final outcome was the preoperative status. W h e n this was poor, the patient made no or only minimal improvement, either with a total resection or a biopsy. Before 1970, "poor" patients often died, but two made a fair or good recovery, and three deemed fair later scored good or close to that. The only two patients we scored good both preoperatively and postoperatively were one man and one woman with a history of less than 1 year, who bled in their fourth and fifth decades; their average-sized lesions were either totally removed [46] or biopsied [6]. Both had no motor or spincteric deficits preoperatively. Good results did not increase after 1970 in this series. Site, age, duration of history, and pregnancy do not seem to influence the final outcome. In this series, mean follow-up has been 44 months (range: 20 days-15 years) for men and 27.9 months (range: 1 d a y - 1 6 years) for women.
Spinal Intramedullary Cavernous Angiomas
Figure 5. MR images of a spinal intramedullary cavernoma. This 48year-old man had his lesion discovered in 1990 after the onset of purely sensory deficits (left). This has not enlarged over 2 years (right).
M u l t i p l i c i t y , F a m i l i a l i t y , a n d Miscellaneous Associations
Multiple lesions were seen in two cases [10,26] and familiality in five [5,23,28,45,47]. Familial lesions more often affected women (80%). Craniospinal multiple cavernomas were seen in six cases [10,23,28,34,56, 63], mostly (80%) in men. SIC may be found in association with spinal cord tumors [24,32,34], angiomas in other body areas outside the CNS [52,63], cutaneous nevi [28], and acne rosacea [47, plus our reported case].
Neuroradiology
MRI, angiography, and myelography made a diagnosis in respectively 100%, 0%, and 60% of the cases. In cases followed up with MRI, the hemosiderin ring did not enlarge [26,30, see Figure 1]. Typical MRI of an SIC are also shown in Figure 5. Myelography cannot document an enlargement of the core, as suggested [5,11]. Discussion Cavernomas represent 3% [36], 12.8% [28], and 16% [4] of spinal vascular malformations. Although arteriovenous malformations (AVM) may be congenital, the pathogenesis of cavernomas is still obscure. Familial cavernomas may account for 50% of all cases of CNS cavernous angiomas [41]: their inheritance is autosomal dominant with variable expressivity [16,41]. The hypothesis that a mutant gene elaborates a factor (possibly an endothelial cell growth factor) locally or transported systematically that adversely affects the growth and function of CNS capillary vessels allowing for repeated hemorrhages and neoangiogenesis has been proposed [45]. Yet, in 1970, Gross induced cavernomas (and tumors) by inoculation of high doses of a PAPOVA virus in kittens [ 12]. There are many
Surg Neurol 1994;41:381-8
385
reports of brain cavernomas in association with tumors [reviewed in 59]; two cases from our collected series also associated with spinal tumors, simultaneously or at an interval [24,32]. Thus, an epigenetic mechanism may be at work. In addition, new lesions have appeared during follow-up of patients with cavernomas [49]. It can be speculated that two types of cavernomas exist--one congenital and another acquired. A few of these may increase in size over time, although no SIC apparently did so, through repeated bleedings, or, eventually, endothelial proliferation. The most important way SICs produce symptoms is through hemorrhage. Bleeding has been noted even with gradual patterns of clinical presentation. Although the mass effect is certainly important, the biochemical damage cannot be disregarded. The fact that acute deficits followed by worsening often completely clear off (e.g., our patient) can be explained with resorption and organization of the clot. As the clot disappears, ischemia in the surrounding cord gradually ebbs. The onset of seizures in cranial cases has been explained through the epileptogenic action of the Fe 3+ ion [59] of blood pigments. In spinal cases, repeated microhemorrhages or constant oozing might lead to a gradual iron-mediated generation of free radicals with attendant damage and progression of symptoms. The remarkable ability of steroids to markedly improve clinical conditions of spinal cases [28,30, plus our case] may stem from a reduction of free radical action. Intracranial cavernomas in neonates regressing following steroid treatment are on record [37]. It is possible that, in several cases in which the clinical history was mistaken for multiple sclerosis, one of the diagnostic criteria could have been the response to steroids. Direct flow measurements in cavernomas of the brain showed both their very low flow and a free communication with the venous circulation [25]. Those authors concluded that cavernomas are unlikely to produce symptoms of ischemia in the surrounding parenchyma and that a rapid increase in the central venous pressure possibly associated with a dependent head position could originate the bleeding [25]. Hypothetically, this very low flow might concentrate neurotoxic substances. Finally, hormones may play a role in the pathogenesis of cavernomas. Cavernomas increase in size during pregnancy [59]; fertile women in our series often (14%) presented symptoms during pregnancy, and so did cranial cases [42]. However, this could depend on pregnancy-related expansion of blood volume, and consequent increase of venous pressure inside the cavernoma. SICs are most frequent in females (69%). Estrogens may modulate some steps in the pathogenesis of cavernomas, either by altering coagulation or through a blood volume expansion.
386
Surg Neurol 1994;41:381-8
The Natural
HistoryofSIC
The typical patient is a woman in her thirties, with a cervical or thoracic lesion (mean size: 1.7 cm) spanning 1-2 vertebral levels, who develops a gross hemorrhage with acute pain and motor deficits. A prolonged history simulating multiple sclerosis (11% of the cases) can also be seen; symptoms may present during pregnancy. The estimated risk of hemorrhage is 1.6%/person-year of exposure, six times that of intracranial cases in a recent study [42], in which 86% of bleeding cases were women. When bleeding occurs, a recurrence is the norm after a mean interval of 39.6 months. Although it was pointed out that infratentorial cavernomas bleed more frequently than supratentorial cases, this has been denied [42]; perhaps, there is an increasing risk of bleeding along the craniocaudal axis of the CNS [see also 8]. W e have observed a yearly pattern of bleeding in several cases [30, plus this case report], shortening after several episodes. Thus, repeated cycles of hemorrhage mark the history of SICs in which an early diagnosis is not made [6,26,47,49,55]. A notable feature is the complete recovery witnessed by several patients after the first 2-3 bleeding episodes [5,6,28,30, plus our case], which also applies to intracranial cases [53]; certainly, repeated bleedings may lead to permanent deficits. Unlike spinal AVMs, in which the bleeding-induced deficit is maximal at onset, in SICs this worsens over time (explaining why we considered an episode of less than 7 days acute) [28].
Neuroradiology Today, high-field MRI (1.5 Tesla) has become the technique of choice in the evaluation of CNS cavernomas, with a "close-to-histologic" discriminative capability. MRI criteria for a cavernoma are an inhomogeneous hyperintense signal surrounded by a hypointense ring on T2WIs, with enhancement of the core with gadolinium-DTPA [40], this latter not seen by us and Mehdorn and Stolke [30]. Low-field machines (0.35 Tesla) occasionally missed a cavernoma [40]. The availability of MRI makes explorative surgery unjustifiable, and makes both myelography and spinal angiography obsolete. Cavernomas cannot be considered "cryptic" vascular malformations anymore [36].
Surgical Outcome and Treatment Strategies As detailed earlier, the most important prognostic factor bearing on the final outcome of SICs is the preoperative status (Karnofsky score). Although size might add to the surgical injury, comparable results have been obtained with both large (2 cm or more) and
Canavero et al
small less than 2 cm) lesions. Age, site, duration of history, and extent of removal had no significant impact. No clear-cut increase of good results occurred after 1970 in spite of the current use of sophisticated instrumentation (surgical microscope, electrophysiologic monitoring, Cavitron, Laser); instead, no deaths were recorded after that year, highlighting better anesthesiologic, physiotherapeutic, and pharmacologic management (control of infection, decubiti, others). The dead of pre-1970 may be some of the more recent "poor" cases. Some patients operated on over these last few years worsened after surgery and did not improve at 1 year follow-up [6,57]~ Although radiotherapy with LINAC or Gamma knife has been successfully used in the treatment of brain stem and hemispheric cavernomas [21], no data exist for SICs. External irradiation was of real benefit in some brain and brain stem cavernomas [61], but it worsened the clinical picture of SICs [2,6]. In the face of frequent complete recoveries from bleeding, the likely benefit of steroids, and a significant interval between hemorrhages in several cases, surgery, even if limited to clot evacuation, should, in authors' view, not be considered a first-line option, particularly when the patient complains only of sensory disturbances.
References 1. Barnwell SL, Dowd CF, Davis RL, Edwards MSB, Gutin PH, Wilson CB. Cryptic vascular malformations of the spinal cord: diagnosis by magnetic resonance and outcome of surgery. J Neurosurg 1990; 72:403-7. 2. Bassett RC, Peet MM, Holt JF. Pial Medullary angiomas. Clinicopathological features and treatment. Arch Neurol Psych 1949; 61:558-68. 3. Beraud R, Meloche BR. A propos de deux cas de malformations vasculaires. Un Med Canada 1965; 94:176-88. 4. Bergstrand A, H6ok O, Lidvall H. Vascular malformations of the spinal cord. 1964; 40:169-83. 5. Bicknell JM, Carlow TJ, Kornfeld M, Stovring J, Turner P. Familial cavernousangiomas. Arch Neurol 1978; 35:746-9. 5a. Canavero S. Intramedullary cavernous angiomas of the spinal cord: Cervical presentation, pathological features, and surgical management. Neurosurgery 1993;23:692-3. 6. Cosgrove GR, Bertrand G, Fontaine S, Robitaille Y, Melansos D. Cavernous angiomas of the spinal cord. J Neurosurg 1988; 68:31-6. 7. Cristofori L, Caudana R, Talacchi A, Vitale M, Bricolo A. La Risonanza Magnetica nei tumori intramidollari: correlazionifra immagini e reperti operatori. XXXIX Congresso della Societ~ Italiana di Neurochirurgia, Ed. Minerva Medica, 1990. 8. Curling DO, Kelly DL, Elster A, Craven TE. An analysis of the natural history of cavernous angiomas. J Neurosurg 1991; 75:702-8. 9. Fontaine S, Melanson D, Cosgrove R, Bertrand G. Cavernous hemangiomas of the spinal cord: MR imaging. Radiology 1988; 166:839-41.
Spinal Intramedullary Cavernous Angiomas
10. Gagel O, Meszaros A. Zur Frage der Myelopathia necroticans. Arch Psychiatry Neurol Sci 1947; 179:423-9. 11. Goran A, Carlson DJ, Fischer RG. Successful treatment of intramedullary angioma of the cord. J Neurosurg 1964; 21:311-14. 12. Gross L. Oncogenic viruses. 2nd ed. New York: Pergamon Press, 1970: p 991. 13. Gyula S, Matyas B. Hemangioma cavernosum intramedullary. Morphol Igazsagugyi Orv Sz 1973; 13:123-5. 14. Hadlich R. Ein Fall von Tumor Cavernosus des Rueckenmarks mit besonderer Beruecksichtigung der neueren Theorien ueber die Genesis des Kavernoms. Virchows Arch 1903; 172:4294-41. 15. Harbitz F. 1930: cited from: Hieke L, 1949. 16. Hayman LA, Evans RA, Ferrell RE. Familial cavernous angiomas. Natural history and genetic study over a 5-year period. A m J Med Genet 1982; 11:147-60. 17. Hieke L. Ueber Haematomyelie bei intramedullaeren Teleangiektasien. Beitr Pathol Anat 1949; 110:433-40. 18. Jellinger K. Pathology of spinal vascular malformations and vascular tumors. In: Pia HW, Djindjian R, eds. Spinal angiomas: advances in diagnosis and therapy. New York: SpringerVerlag, 1978. 19, Kitahara T, Miyasaki Y, Ohwada T, Yada K, Mera M. An operated case of spontaneous hematomyelia. Neurol Surg 1982; 10:675-9. 20. Kitamura K, Fukui M, Oka K, Matsushima T, Hasuo K, Fukushima T, Tomonaga M, Okundera T. Hemangiomas of the central nervous system in Japan: an epidemiological and clinicopathological study with special reference to venous and cavernous malformations. Neurosurg Rev 1986; 9:221-31. 21. Kondziolka D, Lunsford D, Coffey RJ, Bissonette DJ, Flickinger PA-C, Flickinger JC. Stereotactic radiosurgery of anglographically occult vascular malformations: indications and preliminary experience. Neurosurgery 1990; 27:892-990. 22. Krayenbuehl H, Yasargil MG. Die Varicosis spinalis und ihre Behandlung. Schweiz Arch Neurol Psychiatr 1963; 92:74-92. 23. Lee KS, Spetzler RF. Spinal cord cavernous malformation in a patient with familial intracranial cavernous malformations. Neurosurgery 1990; 26:877-80. 24. Lindboe CF, Nordal HJ. Multiple neurilemmomas of the cauda equina, cavernous hemangioma of the spinal cord and degeneration of the lateral corticospinal tracts in a man with the clinical diagnosis of multiple sclerosis. Clin Neuro Pathol 1985; 4:2604. 25. Little JR, Awad IA, Jones SC, Ebrahim ZY. Vascular pressures and cortical blood flow in cavernous angioma of the brain. J Neurosurg 1990; 73:555-9. 26. Lopate G, Black JT, Grubb RL. Cavernous hemangioma of the spinal cord: report of 2 unusual cases. Neurology 1990; 40:1791-3. 27. Lorenz O. Cavernoses Angiom des Rueckenmarks. Inaugural dissertation nr 375, Jena, 1901. 28. McCormick PC, Michelsen WJ, Post KD, Carmel PW, Stein BM. Cavernous malformations of the spinal cord. Neurosurgery 1988; 23:459-63. 29. McCormick WF, Nofzinger JD. "Cryptic" vascular malformations of the central nervous system. J Neurosurg 1986; 24:86575. 30. Mehdorn HM, Stolke D. Cervical intramedullary cavernous angioma with MRI-proven hemorrhages. J Neurol 1991; 238:420-6. 31. Meyer O, Kohler B. Ueber eine auf kongenitaler Basis entstandene kavernomaehnlische Bildung des Rueckenmarks. Frank Ztschr Pathol 1917; 20:37-57.
Surg Neurol 1994;41:381-8
387
32. Odom GL, Woodhall B, Margolis G. Spontaneous hematomyelia and angiomas of the spinal cord. J Neurosurg 1957; 14:192202. 33. Ogorodnikova LS. Hemangioma of the spinal cord. Zh Nevropat I Psikhiat 1958; 58:542~4. 34. Ohlmacher AP 1899: cited from: Van Reeth PC 1952. 35. Pagni CA, Canavero S, Forni M. Report of a cavernoma of the cauda equina and review of the literature. Surg Neurol 1990; 33:124-31. 36. Pia HW, Djindjian R (eds). Spinal angiomas: advances in diagnosis and therapy. New York: Springer-Verlag, 1978. 37. Prensky A, Gado M. Angiographic resolution of a neonatal intracranial cavernous hemangioma coincident with steroid therapy. Case report. J Neurosurg 1973; 39:99-103. 38. Ramos R Jr, De Toffol B, Aesch B, Jan M. Hydrocephalus and cavernoma of the cauda equina. Neurosurgery 1990; 27:13942. 39. Requena I, Arias M, Lopez-Ibor L, Pereiro 1, Barba A, Alonso A, Monton E. Cavernomas of the central nervous system: clinical and neuroimaging manifestations in {7 patients. J Neurol Neurosurg Psychiatry 1991; 54:590-4. 40. Rigamonti D, Drayer BP, Johnson PC, Hadley MN, Zabramski J, Spetzler RF. The MRI appearance of cavernous malformations (angiomas). J Neurosurg 1987; 67:518-24. 41. Rigamonti D, Hadley MN, Drayer BP, Johnson PC, HoenigRigamonti K, Knight JT, Spetzler RF. Cerebral cavernous malformations. Incidence and familial occurrence. N Engl J Med 1988; 319:343-7. 42. Robinson JR, Awad IA, Little JR. Natural history of the cavernous angioma. J Neurosurg 1991; 75:709-14. 43. Russell DS, Rubinstein LJ. Pathology of tumors of the nervous system. 4th ed. Baltimore: Williams & Wilkins, 1977. 44. Russell DS, Rubinstein LJ. Pathology of tumors of the nervous system. 5th ed. London: E Arnold, 1989. 45. Rutka JT, Brant-Zawadki M, Wilson CB, Rosenblum ML. Familial cavernous malformations. Diagnostic potential of magnetic resonance imaging. Surg Neurol 1988; 29:467-74. 46. Saito N, Yamakawa K, Sasaki T, Saito I, Takakura K. Intramedullary cavernous angioma with trigeminal neuralgia: a case report and review of the literature. Neurosurgery 1989; 25:97101. 47. Schroeder JM, Brunngraber CV. Ueber ein intramedullaeres cavernoeses angiom. Acta Neurochir (Wien) 1964; 12:632-41. 48. Schultze F. Weiterer Beitrag zur Diagnose und operativen Behandlung yon Geschwuelsten des Rueckenmarks. Deutsche Med Wochenschr 1912; 38:1676 9. 49. Scott RM, Barnes P, Kupsky W, Adelman LS. Cavernous angiomas of the central nervous system in children. J Neurosurg 1992; 76:3846. 50. Scott M. Lower extremity pain simulating sciatica. JAMA 1956; 160:528-34. 51. Simard JM, Garcia-Bengochea F, Ballinger WE Jr, Michele JP, Quisling RG. Cavernous angioma: a review of 126 collected and t 2 new clinical cases. Neurosurgery 1986; 18:162-72. 52. Szojchet A. Metameric spinal cord and skin hemangiomas. Case report J Neurosurg 1968; 29:199-201. 53. Tung H, Giannotta SL, Chandrasoma PT, Zee CS. Recurrent intraparenchymal hemorrhages from angiographically occult vascular malformations. J Neurosurg 1990; 73: 174-80. 54. Turner OA, Kernohan JW. Vascular malformations and vascular tumors involving the spinal cord. Arch Neurol 1941; 46:444-63.
388
Surg Neurol 1994;41:381-8
55. Tyndel FJ, Bilbao JM, Hudson AR, Colapinto EV. Hemangioma calcifications of the spinal cord. Can J Neurol Sci 1985; 12:321-2. 56. Van Reeth PC. Contribution a l'etude de l'angiomatose medullaire. Acta Neurol Psychiatr Belg 1952; 52:249-70. 57. Vaquero J, Martinez R, Martinez P. Cavernomas of the spinal cord: report of two cases. Neurosurgery 1988; 22:143-4. 58. Vaquero J, Salazar J, Martinez R, Martinez P, Bravo G. Cavernomas of the central nervous system: clinical syndromes, CT scan diagnosis and prognosis after surgical treatment in 25 cases. Acta Neurochir (Wien) 1987; 85:29-33. 59. Villani RM, Arienta C, Caroli M. Cavernous angioma of the central nervous system. J Neurosurg Sci 1989; 33:229-58. 60. Vraa-Jensen G. Angioma of the spinal cord, Acta Psychiatr Neurol Scand 1949; 24:709-21.
Canavero et al
61. Yamasaki T, Handa H, Yamashita J, Paine JT, Tashiro Y, Uno A, Ishikawa M, Asato R. Intracranial and orbital cavernous angiomas. A review of 30 cases. J Neurosurg 1986; 64:197208. 62. Wang AM, Morris JH, Fischer EG, Peterson R, Lin JCT. Cavernous hemangioma of the thoracic spinal cord. Neuroradiology 1988; 30:261-4. 63. Wood MW, White RJ, Kernohan JW. Cavernous hemangiomatosis involving the brain, spinal cord, heart, skin and kidney: report of case. Proc Staff Meet Mayo Clin 1957; 32:249-54. 64. Wyburn-Mason R. The vascular abnormalities and tumours of the spinal cord and its membranes. London: Kimpton, 1943. 65. Zentner J, Hassler W, Gawehn J, Schroth G. Intramedullary cavernous angiomas. Surg Neurol 1989; 31:64-8.
Spinal Intramedullary Cavernous Angiomas: A Literature Metaanalysis Sergio Canavero, M.D., Carlo Alberto Pagni, M.D., Sergio Duca, M.D., and Gianni Boris Bradac, M.D. Institute of Neurosurgery and Chair of Neuroradiology, University of Turin, Turin, Italy
Canavero S, Pagni CA, Duca S, Bradac GB. Spinal intramedullary cavernous angiomas: a literature meta-analysis. Surg Neurol 1994;41:381-8.
The present report reviews 57 (out of 65) cases of spinal intramedullary cavernomas collected from the literature, plus one personal patient. Almost 70% of all patients were women. Mean age at diagnosis for women was 36.4 years, with a peak in the third decade. More than threefourths of all women became symptomatic between the second and fourth decades, with a peak in the fourth decade. Unlike in men, cervical and thoracic lesions a r e almost equally represented, generally involving 1-2 vertebral levels. Mean size at diagnosis is 1.7 cm; no enlargement over time was seen. Symptoms are more frequently acute; pain and sensorimotor deficits are the usual complaints, but the clinical picture may simulate that of multiple sclerosis. The duration of history was less than 5 years in more than 80% of women. Bleeding was seen in 60% of women, with a risk of 1.6%/personyear of exposure globally. Cervical lesions have both a shorter course and increased frequency of bleeding. I f not immediately recognized, repeated cycles of bleeding are the norm, with a mean interval of 39.6 months between the first and second episodes. The preoperative status was the single most important factor bearing on outcome, whereas sex, age, size, location, duration of history and extent of removal were not. Magnetic resonance imaging was diagnostic in all cases, whereas angiography was 100% negative. Surgery should not be a necessary first option, as recovery from the first bleeding is apparently fairly frequent. K E Y W O R D S : Cavernoma; Cavernous angioma; Intramedullary; Natural history; Outcome; Spinal cord
Cavernous angiomas of the central nervous system (CNS) are berry-like vascular malformations in which
Address reprint requests to: Sergio Canavero, M.D., Via Morazzone 21, 10132 Torino, Italy. Received October 2, 1992; accepted March 3, 1993.
© 1994 by ElsevierScienceInc.
abnormally dilated blood spaces are closely clustered together, without intervening neural tissue. The surrounding parenchyma is often gliotic and hemosiderin-stained and may occasionally contain small lowflow feeding arteries and draining veins [44]. By far the great majority of CNS cavernomas is found in the cerebral hemispheres and brain stem. Many are entirely asymptomatic and up to 50°A may be multiple [41]. In 1990, we reviewed the literature on spinal subdural extramedullary cavernomas [35]; another case appeared in that year [38] and two possible cases were described by Bassett and colleagues [2]. At the time, we briefly reviewed intramedullary cases. The impression was that, in spite of many recent case reports and short series, spinal intramedullary cavernomas (SIC) remain a poorly known entity, probably overaggressively treated, without a proper definition of both their course and outcome. Highlights of our results have appeared in letter form [5a].
Case Report A 48-year-old woman, with a history of Graves' disease and ache rosacea, noticed at the time of her pregnancy (1977) a slow reduction of her right leg's sensitivity to temperature; this progressed over 3 years to involve the limb in toto and there were occasional episodes of urinary incontinence. In 1980, she gradually developed right leg weakness and gait disturbance. In January 1981, she complained of sudden, violent, boring pains in her lower abdomen, radiating to the right thigh, anteriorly. A lumbar puncture was followed by sudden spastic paraplegia, with gradual recovery 3 months later. A spinal computer tomography (CT) was normal. Discharge diagnosis was "demyelinating syndrome." Thereafter, she had recurrent, yearly (generally in winter) sensorimotor disturbances in her right leg; each time, steroids and rehabilitation brought about complete recovery. Since 1988, she suffered paresthesia only Ca feeling of a ball pressing on the sole") plus hypoesthesia. In March 1991, she noticed the onset of dysesthesia in [)090~ ~019/9~/$7 O0
382
Canavero et al
Surg Neurol 1994;41:381-8
Figure 1. MR images (T1Wls. left and center; T2WI. right) of the cavernousangioma (seecase report),
her left groin, with radiation to the sacrum. At this time, magnetic resonance imaging (MRI) disclosed a noncontrast-enhancing lesion with the characteristic neuroradiologic features of a cavernoma (Figure 1). On admission, a shuffling, spastic gait was apparent, with paraparesis and bilateral Babinski signs. The right leg was hypoesthetic to painful stimuli, and both legs lost vibratory sense. In November 1991, a D9 laminectomy was performed. Under the microscope, a softened, tawny cord appeared encased by a thickened arachnoid and many irregular vessels. A large draining venous channel was partially coagulated. The posterior columns were thin and yellow, under which dark-blue, solid lumps amid pitch-black blood and white-gray fibrotic nodules with many venules and arterioles were seen. Small transparent spherules ("crystal-like") were noted. The cleavage plane between this 1.5-cm mass and the surrounding white matter was found; yet, in several points, the two were tightly adherent, with fibrotic bridges. The lesion was removed, but several small portions were left, lest the cord could be damaged. Postoperatively, there was a global sensorimotor worsening with urinary retention. In February 1994, the result is unsatisfactory, as the patient cannot walk alone. The urinary retention resolved. The patient has developed central pain. Pathologically the lesion was a typical cavernoma.
gectatic [17] and another with an extramedullary component [14].
Incidence
W e have experienced three cases of SIC (of whom only the operated one is presently considered) since 1981; two from the Turin area. No case was seen at the Institute of Neurosurgery of the University of Turin in over 45 years. Krayenbuehl and Yasargil removed one single case over 25 years [22], Villani and colleagues three over 4 years [59], Stein six cases over 7.5 years [28], and Requena and co-workers one over 3 years [39]. However, no epidemiologic data were reported by these authors. Many case series of CNS cavernomas do not include any spinal cases [8,42]. W e do not know the prevalence of SIC in the general population.
Sex a n d Age
The female sex is affected in more than two-thirds of the cases (40 of 58). Mean age at diagnosis was 37.1 years (median: 35 years; range: 12-65), with women being diagnosed somewhat earlier than men (Figure 2). Both sexes developed their first symptom in the first decades of life, with a peak in the fourth (Figure 3). No correlation was found between level of the lesion and age of onset.
L i t e r a t u r e Meta-analysis From 1899 to 1991, 65 SICs have been reported in the literature [ 1-7,9-11,13-15,17-20,22-24,26-36,39, 4 3 - 4 8 , 5 0 , 5 2 , 5 4 - 6 0 , 6 2 - 6 5 ] , some with incomplete data as to age, sex, and site [7,13,15,20,22,29,39,54]. In this review, we included a few cases referred to as angiomas that on close scrutiny were most likely cavernomas [2,50,52,60], one case in which a few areas were telean-
Level a n d Extension
In female patients, cervical and thoracic lesions were equally represented (18 cervical, 19 thoracic, 1 cervicothoracic, 3 lumbar), whereas in male patients the proportion was 1:2 (5 cervical, 12 thoracic, 1 lumbar). Single-level lesions were the majority (n = 17, in the female sex), but two-level cases were also frequent (n =
Spinal Intramedullary Cavernous Angiomas
Surg Neurol 1994;41:381-8
5
(MALES) iH
I
21 7o.6~
I
I
I
i
10
10
J 9 (FEMALES)
1
<
383
then gradual (an acute episode which, after peaking, slowly worsened); (3) gradual (a history of slow onset and gradually worsening sensorimotor symptoms). The onset of symptoms was more frequently acute than gradual (56.5% vs 43.4%) in both sexes. Four cases gave no symptoms for a lifetime [14,24,63,64]. Thirty-three patients out of 54 (61%) reported pain sometime in their history, 26 (47%) as their initial symptom, alone or in association; this was more frequently acute than gradual (26 vs 17 cases). While in one case involving C1 it mimicked tic douloureux [46], according to the site of involvement; it affected the neck, the back, the groin, or the limbs; in fact, in simulated sciatica [49], a heart attack (upper thoracic or interscapular) [6,52] and acute appendicitis [28]. Pain may recede and disappear completely or be followed by other symptoms. Yet, opening complaints may be acute or gradual motor deficits (par-
@
I
Z
4
4
~,4%
I
1~19
3112
<
I
I 0-9
4
i 20-~g
I 30-3g
I 4~4g
(MALES) 2
I
I 50.69
Z i
AGE AT DIAGNOSIS BY DECADE
9F" 11
Figure 2. Relation between sex and age at diagnosis (n = 55). Mean age
at diagnosis was 36.4 years (median: 33 years; range: 12-65 years) for women and 39 years (median: 38 years; range: 15-58 years) for men. Note that when diagnosis was made in the fourth decade, history had been less than l year (women: 47%; men: 50%); when made in the third decade, of less than 5 years (women: 47%; men: 40%); and when made in the fifth (women: 50%) or sixth decade (men: 50%), this had been 5 years or more. A history of less than l year was seen in 30% (13) of women and 40% (6) of men; of less than 5 years in 44% (15) of women a n d 3 3 % (5) of men; of years or more in 18% (6) of women a n d 2 7 % (4) of men.
17, in the female sex). Three or more levels were involved exceptionally. In women, SIC cluster at C3-6, T2-3 and T7, although this does not seem to be the case in men.
i
10
(FEMALES)
~q
< L~
Z
2
2
Size Mean size was 1.7 cm (range: 7 m m - 3 cm). MRI did not detect any enlargement during follow-up of two cases (4 months to 3 years) [26,30], and the same applied to the two cases we are following (2 and 4 years). Myelography suggested possible enlargements, but this technique is not as specific.
Symptoms and Course We considered three patterns of presentation: (1) acute (single episode of less than 7-day duration); (2) acute,
0-9
10=19 80,-29 30--39 40-49
50-59
60-69
AGE AT ONSET OF SYMPTOMS BY DECADE F i g u r e 3. Relation between sex and age at onset of symptoms (n = 50). More than 74% of women and 73% of men (globally: 74%)first developed symptoms in the second to fourth decades, with a peak in the fourth in both sexes (31% of women, 33 % of men, 32% globally). No cases were found in the first decade or beyond the sixth (men) or seventh (women). In women, mean onset of cervical cases was at least 34 years, of thoracic lesions at 33.5 years.
384
Surg Neurol 1994;41:381-8
~ORACIC
Canavero et al
CBRVICAL
Females Malos
ioo~
o
~oo~
Figure 4. Gross bleeding according to location of cavernomas and gender, Cervical lesions bleed more in both sexes. Globally, 72. 7% of cervical lesions and 48.3% of thoracic lesions bled macroscopically. Women bled more often than men (60.5% vs 53%). Considering three lumbar cavernomas, the total rate of gross bleeding was 57. 1%. It should be noted that all these data are likely underestimated as several reports could not be defined.
aplegia, footdrop, others) or sensory disturbances other than pain (paresthesias). Although sphincteric problems are often seen, they never were the initial complaint. SIC may simulate multiple sclerosis (6 of 54:11%) [28,47, plus our personal case]. In a few cases, symptoms were precipitated by lumbar puncture (our case), minor trauma [27,28], and surgery [56]. Acute symptoms (both motor and sensory) completely, or, in case of paraplegia or severe paraparesis, almost completely receded. Symptoms presented during pregnancy in four cases (14% of fertile women; i.e. those from age 14 to 48) [26,55,59, plus our case]; in one patient, they occurred at two different pregnancies, thereafter completely recovering [55]. While all cases not immediately recognized and treated presented recurrences, in one patient submitted to biopsy, no relapse occurred over 15 years postoperatively [6]. The duration of history ranged from hours to decades in both sexes, but was generally less than 5 years (80% of all cases). Cervical lesions had a shorter history than thoracic cases in women.
Bleeding Gross bleeding was evaluated on the basis of M R I findings (acute or subacute blood outside the hemosiderin ring of the lesion) or surgical or pathologic evidence of a fresh clot outside the borders of the lesion. W o m e n bleed more frequently than men, and, in both sexes, cervical lesions are at greatest risk. These data are probably underrated, because several reports were equivocal (Figure 4). There was no correlation between size and risk of bleeding, as shown in cranial cases [42]. In this series, the risk of hemorrhage was 1.6%/person-year of exposure (at least 30 events/1867 years), assuming an equal risk throughout life and a congenital origin. However, (1) cavernomas may not be congenital in all cases, as new lesions may appear [49], (2) pregnancy increases the risk of hemorrhage, and
(3) cervical lesions bleed more frequently. In those cases in which the SIC was not immediately recognized, another bleeding episode was the rule, with a mean interval of 39.6 months (range: 2 - 1 6 8 months). One of the two cases with a 2-month interval was a pregnant women [26]. H e m o r r h a g e can occur with a gradual presentation. Subarachnoid hemorrhage, unlike subdural extramedullary lesions [35], is virtually unknown, although a xantochromic CSF may be encountered exceptionally.
Surgical Outcome W e assigned to each reported patient one of the following scores, both preoperatively and postoperatively: good (patient fully ambulatory, no sphincteric deficits, possible sensory disturbances), fair (patient with mild to advanced motor disturbance and/or gait restriction, sphincteric deficits, limited autonomy), and poor (patient bed-ridden or confined to a wheelchair; severe sphincteric disturbances). Intermediate scores (e.g., fairto-good) were used if authors' descriptions were equivocal. Total removal was possible in 24 women and eight men, partial in four women and three men, while a biopsy was done in three women and one man. Results were as follows: (1) women: deaths 7, poor 9, poor-tofair 1, fair 11, fair-to-good 2, good 4; (2) men: deaths 3, poor 2, fair 8, good 2. O f 11 cases where size was reported, a poor result was obtained in three lesions larger than 2 cm and in one of less than 2 cm, a fair result in three cases of lesions larger and smaller than 2 cm, whereas a good result in a case of less than 2 cm. Neither size nor extent of removal significantly influenced the outcome: the only appreciable difference between pre-1970 and post-1970 reports was that no deaths were seen after 1970. The most important single factor bearing on the final outcome was the preoperative status. W h e n this was poor, the patient made no or only minimal improvement, either with a total resection or a biopsy. Before 1970, "poor" patients often died, but two made a fair or good recovery, and three deemed fair later scored good or close to that. The only two patients we scored good both preoperatively and postoperatively were one man and one woman with a history of less than 1 year, who bled in their fourth and fifth decades; their average-sized lesions were either totally removed [46] or biopsied [6]. Both had no motor or spincteric deficits preoperatively. Good results did not increase after 1970 in this series. Site, age, duration of history, and pregnancy do not seem to influence the final outcome. In this series, mean follow-up has been 44 months (range: 20 days-15 years) for men and 27.9 months (range: 1 d a y - 1 6 years) for women.
Spinal Intramedullary Cavernous Angiomas
Figure 5. MR images of a spinal intramedullary cavernoma. This 48year-old man had his lesion discovered in 1990 after the onset of purely sensory deficits (left). This has not enlarged over 2 years (right).
M u l t i p l i c i t y , F a m i l i a l i t y , a n d Miscellaneous Associations
Multiple lesions were seen in two cases [10,26] and familiality in five [5,23,28,45,47]. Familial lesions more often affected women (80%). Craniospinal multiple cavernomas were seen in six cases [10,23,28,34,56, 63], mostly (80%) in men. SIC may be found in association with spinal cord tumors [24,32,34], angiomas in other body areas outside the CNS [52,63], cutaneous nevi [28], and acne rosacea [47, plus our reported case].
Neuroradiology
MRI, angiography, and myelography made a diagnosis in respectively 100%, 0%, and 60% of the cases. In cases followed up with MRI, the hemosiderin ring did not enlarge [26,30, see Figure 1]. Typical MRI of an SIC are also shown in Figure 5. Myelography cannot document an enlargement of the core, as suggested [5,11]. Discussion Cavernomas represent 3% [36], 12.8% [28], and 16% [4] of spinal vascular malformations. Although arteriovenous malformations (AVM) may be congenital, the pathogenesis of cavernomas is still obscure. Familial cavernomas may account for 50% of all cases of CNS cavernous angiomas [41]: their inheritance is autosomal dominant with variable expressivity [16,41]. The hypothesis that a mutant gene elaborates a factor (possibly an endothelial cell growth factor) locally or transported systematically that adversely affects the growth and function of CNS capillary vessels allowing for repeated hemorrhages and neoangiogenesis has been proposed [45]. Yet, in 1970, Gross induced cavernomas (and tumors) by inoculation of high doses of a PAPOVA virus in kittens [ 12]. There are many
Surg Neurol 1994;41:381-8
385
reports of brain cavernomas in association with tumors [reviewed in 59]; two cases from our collected series also associated with spinal tumors, simultaneously or at an interval [24,32]. Thus, an epigenetic mechanism may be at work. In addition, new lesions have appeared during follow-up of patients with cavernomas [49]. It can be speculated that two types of cavernomas exist--one congenital and another acquired. A few of these may increase in size over time, although no SIC apparently did so, through repeated bleedings, or, eventually, endothelial proliferation. The most important way SICs produce symptoms is through hemorrhage. Bleeding has been noted even with gradual patterns of clinical presentation. Although the mass effect is certainly important, the biochemical damage cannot be disregarded. The fact that acute deficits followed by worsening often completely clear off (e.g., our patient) can be explained with resorption and organization of the clot. As the clot disappears, ischemia in the surrounding cord gradually ebbs. The onset of seizures in cranial cases has been explained through the epileptogenic action of the Fe 3+ ion [59] of blood pigments. In spinal cases, repeated microhemorrhages or constant oozing might lead to a gradual iron-mediated generation of free radicals with attendant damage and progression of symptoms. The remarkable ability of steroids to markedly improve clinical conditions of spinal cases [28,30, plus our case] may stem from a reduction of free radical action. Intracranial cavernomas in neonates regressing following steroid treatment are on record [37]. It is possible that, in several cases in which the clinical history was mistaken for multiple sclerosis, one of the diagnostic criteria could have been the response to steroids. Direct flow measurements in cavernomas of the brain showed both their very low flow and a free communication with the venous circulation [25]. Those authors concluded that cavernomas are unlikely to produce symptoms of ischemia in the surrounding parenchyma and that a rapid increase in the central venous pressure possibly associated with a dependent head position could originate the bleeding [25]. Hypothetically, this very low flow might concentrate neurotoxic substances. Finally, hormones may play a role in the pathogenesis of cavernomas. Cavernomas increase in size during pregnancy [59]; fertile women in our series often (14%) presented symptoms during pregnancy, and so did cranial cases [42]. However, this could depend on pregnancy-related expansion of blood volume, and consequent increase of venous pressure inside the cavernoma. SICs are most frequent in females (69%). Estrogens may modulate some steps in the pathogenesis of cavernomas, either by altering coagulation or through a blood volume expansion.
386
Surg Neurol 1994;41:381-8
The Natural
HistoryofSIC
The typical patient is a woman in her thirties, with a cervical or thoracic lesion (mean size: 1.7 cm) spanning 1-2 vertebral levels, who develops a gross hemorrhage with acute pain and motor deficits. A prolonged history simulating multiple sclerosis (11% of the cases) can also be seen; symptoms may present during pregnancy. The estimated risk of hemorrhage is 1.6%/person-year of exposure, six times that of intracranial cases in a recent study [42], in which 86% of bleeding cases were women. When bleeding occurs, a recurrence is the norm after a mean interval of 39.6 months. Although it was pointed out that infratentorial cavernomas bleed more frequently than supratentorial cases, this has been denied [42]; perhaps, there is an increasing risk of bleeding along the craniocaudal axis of the CNS [see also 8]. W e have observed a yearly pattern of bleeding in several cases [30, plus this case report], shortening after several episodes. Thus, repeated cycles of hemorrhage mark the history of SICs in which an early diagnosis is not made [6,26,47,49,55]. A notable feature is the complete recovery witnessed by several patients after the first 2-3 bleeding episodes [5,6,28,30, plus our case], which also applies to intracranial cases [53]; certainly, repeated bleedings may lead to permanent deficits. Unlike spinal AVMs, in which the bleeding-induced deficit is maximal at onset, in SICs this worsens over time (explaining why we considered an episode of less than 7 days acute) [28].
Neuroradiology Today, high-field MRI (1.5 Tesla) has become the technique of choice in the evaluation of CNS cavernomas, with a "close-to-histologic" discriminative capability. MRI criteria for a cavernoma are an inhomogeneous hyperintense signal surrounded by a hypointense ring on T2WIs, with enhancement of the core with gadolinium-DTPA [40], this latter not seen by us and Mehdorn and Stolke [30]. Low-field machines (0.35 Tesla) occasionally missed a cavernoma [40]. The availability of MRI makes explorative surgery unjustifiable, and makes both myelography and spinal angiography obsolete. Cavernomas cannot be considered "cryptic" vascular malformations anymore [36].
Surgical Outcome and Treatment Strategies As detailed earlier, the most important prognostic factor bearing on the final outcome of SICs is the preoperative status (Karnofsky score). Although size might add to the surgical injury, comparable results have been obtained with both large (2 cm or more) and
Canavero et al
small less than 2 cm) lesions. Age, site, duration of history, and extent of removal had no significant impact. No clear-cut increase of good results occurred after 1970 in spite of the current use of sophisticated instrumentation (surgical microscope, electrophysiologic monitoring, Cavitron, Laser); instead, no deaths were recorded after that year, highlighting better anesthesiologic, physiotherapeutic, and pharmacologic management (control of infection, decubiti, others). The dead of pre-1970 may be some of the more recent "poor" cases. Some patients operated on over these last few years worsened after surgery and did not improve at 1 year follow-up [6,57]~ Although radiotherapy with LINAC or Gamma knife has been successfully used in the treatment of brain stem and hemispheric cavernomas [21], no data exist for SICs. External irradiation was of real benefit in some brain and brain stem cavernomas [61], but it worsened the clinical picture of SICs [2,6]. In the face of frequent complete recoveries from bleeding, the likely benefit of steroids, and a significant interval between hemorrhages in several cases, surgery, even if limited to clot evacuation, should, in authors' view, not be considered a first-line option, particularly when the patient complains only of sensory disturbances.
References 1. Barnwell SL, Dowd CF, Davis RL, Edwards MSB, Gutin PH, Wilson CB. Cryptic vascular malformations of the spinal cord: diagnosis by magnetic resonance and outcome of surgery. J Neurosurg 1990; 72:403-7. 2. Bassett RC, Peet MM, Holt JF. Pial Medullary angiomas. Clinicopathological features and treatment. Arch Neurol Psych 1949; 61:558-68. 3. Beraud R, Meloche BR. A propos de deux cas de malformations vasculaires. Un Med Canada 1965; 94:176-88. 4. Bergstrand A, H6ok O, Lidvall H. Vascular malformations of the spinal cord. 1964; 40:169-83. 5. Bicknell JM, Carlow TJ, Kornfeld M, Stovring J, Turner P. Familial cavernousangiomas. Arch Neurol 1978; 35:746-9. 5a. Canavero S. Intramedullary cavernous angiomas of the spinal cord: Cervical presentation, pathological features, and surgical management. Neurosurgery 1993;23:692-3. 6. Cosgrove GR, Bertrand G, Fontaine S, Robitaille Y, Melansos D. Cavernous angiomas of the spinal cord. J Neurosurg 1988; 68:31-6. 7. Cristofori L, Caudana R, Talacchi A, Vitale M, Bricolo A. La Risonanza Magnetica nei tumori intramidollari: correlazionifra immagini e reperti operatori. XXXIX Congresso della Societ~ Italiana di Neurochirurgia, Ed. Minerva Medica, 1990. 8. Curling DO, Kelly DL, Elster A, Craven TE. An analysis of the natural history of cavernous angiomas. J Neurosurg 1991; 75:702-8. 9. Fontaine S, Melanson D, Cosgrove R, Bertrand G. Cavernous hemangiomas of the spinal cord: MR imaging. Radiology 1988; 166:839-41.
Spinal Intramedullary Cavernous Angiomas
10. Gagel O, Meszaros A. Zur Frage der Myelopathia necroticans. Arch Psychiatry Neurol Sci 1947; 179:423-9. 11. Goran A, Carlson DJ, Fischer RG. Successful treatment of intramedullary angioma of the cord. J Neurosurg 1964; 21:311-14. 12. Gross L. Oncogenic viruses. 2nd ed. New York: Pergamon Press, 1970: p 991. 13. Gyula S, Matyas B. Hemangioma cavernosum intramedullary. Morphol Igazsagugyi Orv Sz 1973; 13:123-5. 14. Hadlich R. Ein Fall von Tumor Cavernosus des Rueckenmarks mit besonderer Beruecksichtigung der neueren Theorien ueber die Genesis des Kavernoms. Virchows Arch 1903; 172:4294-41. 15. Harbitz F. 1930: cited from: Hieke L, 1949. 16. Hayman LA, Evans RA, Ferrell RE. Familial cavernous angiomas. Natural history and genetic study over a 5-year period. A m J Med Genet 1982; 11:147-60. 17. Hieke L. Ueber Haematomyelie bei intramedullaeren Teleangiektasien. Beitr Pathol Anat 1949; 110:433-40. 18. Jellinger K. Pathology of spinal vascular malformations and vascular tumors. In: Pia HW, Djindjian R, eds. Spinal angiomas: advances in diagnosis and therapy. New York: SpringerVerlag, 1978. 19, Kitahara T, Miyasaki Y, Ohwada T, Yada K, Mera M. An operated case of spontaneous hematomyelia. Neurol Surg 1982; 10:675-9. 20. Kitamura K, Fukui M, Oka K, Matsushima T, Hasuo K, Fukushima T, Tomonaga M, Okundera T. Hemangiomas of the central nervous system in Japan: an epidemiological and clinicopathological study with special reference to venous and cavernous malformations. Neurosurg Rev 1986; 9:221-31. 21. Kondziolka D, Lunsford D, Coffey RJ, Bissonette DJ, Flickinger PA-C, Flickinger JC. Stereotactic radiosurgery of anglographically occult vascular malformations: indications and preliminary experience. Neurosurgery 1990; 27:892-990. 22. Krayenbuehl H, Yasargil MG. Die Varicosis spinalis und ihre Behandlung. Schweiz Arch Neurol Psychiatr 1963; 92:74-92. 23. Lee KS, Spetzler RF. Spinal cord cavernous malformation in a patient with familial intracranial cavernous malformations. Neurosurgery 1990; 26:877-80. 24. Lindboe CF, Nordal HJ. Multiple neurilemmomas of the cauda equina, cavernous hemangioma of the spinal cord and degeneration of the lateral corticospinal tracts in a man with the clinical diagnosis of multiple sclerosis. Clin Neuro Pathol 1985; 4:2604. 25. Little JR, Awad IA, Jones SC, Ebrahim ZY. Vascular pressures and cortical blood flow in cavernous angioma of the brain. J Neurosurg 1990; 73:555-9. 26. Lopate G, Black JT, Grubb RL. Cavernous hemangioma of the spinal cord: report of 2 unusual cases. Neurology 1990; 40:1791-3. 27. Lorenz O. Cavernoses Angiom des Rueckenmarks. Inaugural dissertation nr 375, Jena, 1901. 28. McCormick PC, Michelsen WJ, Post KD, Carmel PW, Stein BM. Cavernous malformations of the spinal cord. Neurosurgery 1988; 23:459-63. 29. McCormick WF, Nofzinger JD. "Cryptic" vascular malformations of the central nervous system. J Neurosurg 1986; 24:86575. 30. Mehdorn HM, Stolke D. Cervical intramedullary cavernous angioma with MRI-proven hemorrhages. J Neurol 1991; 238:420-6. 31. Meyer O, Kohler B. Ueber eine auf kongenitaler Basis entstandene kavernomaehnlische Bildung des Rueckenmarks. Frank Ztschr Pathol 1917; 20:37-57.
Surg Neurol 1994;41:381-8
387
32. Odom GL, Woodhall B, Margolis G. Spontaneous hematomyelia and angiomas of the spinal cord. J Neurosurg 1957; 14:192202. 33. Ogorodnikova LS. Hemangioma of the spinal cord. Zh Nevropat I Psikhiat 1958; 58:542~4. 34. Ohlmacher AP 1899: cited from: Van Reeth PC 1952. 35. Pagni CA, Canavero S, Forni M. Report of a cavernoma of the cauda equina and review of the literature. Surg Neurol 1990; 33:124-31. 36. Pia HW, Djindjian R (eds). Spinal angiomas: advances in diagnosis and therapy. New York: Springer-Verlag, 1978. 37. Prensky A, Gado M. Angiographic resolution of a neonatal intracranial cavernous hemangioma coincident with steroid therapy. Case report. J Neurosurg 1973; 39:99-103. 38. Ramos R Jr, De Toffol B, Aesch B, Jan M. Hydrocephalus and cavernoma of the cauda equina. Neurosurgery 1990; 27:13942. 39. Requena I, Arias M, Lopez-Ibor L, Pereiro 1, Barba A, Alonso A, Monton E. Cavernomas of the central nervous system: clinical and neuroimaging manifestations in {7 patients. J Neurol Neurosurg Psychiatry 1991; 54:590-4. 40. Rigamonti D, Drayer BP, Johnson PC, Hadley MN, Zabramski J, Spetzler RF. The MRI appearance of cavernous malformations (angiomas). J Neurosurg 1987; 67:518-24. 41. Rigamonti D, Hadley MN, Drayer BP, Johnson PC, HoenigRigamonti K, Knight JT, Spetzler RF. Cerebral cavernous malformations. Incidence and familial occurrence. N Engl J Med 1988; 319:343-7. 42. Robinson JR, Awad IA, Little JR. Natural history of the cavernous angioma. J Neurosurg 1991; 75:709-14. 43. Russell DS, Rubinstein LJ. Pathology of tumors of the nervous system. 4th ed. Baltimore: Williams & Wilkins, 1977. 44. Russell DS, Rubinstein LJ. Pathology of tumors of the nervous system. 5th ed. London: E Arnold, 1989. 45. Rutka JT, Brant-Zawadki M, Wilson CB, Rosenblum ML. Familial cavernous malformations. Diagnostic potential of magnetic resonance imaging. Surg Neurol 1988; 29:467-74. 46. Saito N, Yamakawa K, Sasaki T, Saito I, Takakura K. Intramedullary cavernous angioma with trigeminal neuralgia: a case report and review of the literature. Neurosurgery 1989; 25:97101. 47. Schroeder JM, Brunngraber CV. Ueber ein intramedullaeres cavernoeses angiom. Acta Neurochir (Wien) 1964; 12:632-41. 48. Schultze F. Weiterer Beitrag zur Diagnose und operativen Behandlung yon Geschwuelsten des Rueckenmarks. Deutsche Med Wochenschr 1912; 38:1676 9. 49. Scott RM, Barnes P, Kupsky W, Adelman LS. Cavernous angiomas of the central nervous system in children. J Neurosurg 1992; 76:3846. 50. Scott M. Lower extremity pain simulating sciatica. JAMA 1956; 160:528-34. 51. Simard JM, Garcia-Bengochea F, Ballinger WE Jr, Michele JP, Quisling RG. Cavernous angioma: a review of 126 collected and t 2 new clinical cases. Neurosurgery 1986; 18:162-72. 52. Szojchet A. Metameric spinal cord and skin hemangiomas. Case report J Neurosurg 1968; 29:199-201. 53. Tung H, Giannotta SL, Chandrasoma PT, Zee CS. Recurrent intraparenchymal hemorrhages from angiographically occult vascular malformations. J Neurosurg 1990; 73: 174-80. 54. Turner OA, Kernohan JW. Vascular malformations and vascular tumors involving the spinal cord. Arch Neurol 1941; 46:444-63.
388
Surg Neurol 1994;41:381-8
55. Tyndel FJ, Bilbao JM, Hudson AR, Colapinto EV. Hemangioma calcifications of the spinal cord. Can J Neurol Sci 1985; 12:321-2. 56. Van Reeth PC. Contribution a l'etude de l'angiomatose medullaire. Acta Neurol Psychiatr Belg 1952; 52:249-70. 57. Vaquero J, Martinez R, Martinez P. Cavernomas of the spinal cord: report of two cases. Neurosurgery 1988; 22:143-4. 58. Vaquero J, Salazar J, Martinez R, Martinez P, Bravo G. Cavernomas of the central nervous system: clinical syndromes, CT scan diagnosis and prognosis after surgical treatment in 25 cases. Acta Neurochir (Wien) 1987; 85:29-33. 59. Villani RM, Arienta C, Caroli M. Cavernous angioma of the central nervous system. J Neurosurg Sci 1989; 33:229-58. 60. Vraa-Jensen G. Angioma of the spinal cord, Acta Psychiatr Neurol Scand 1949; 24:709-21.
Canavero et al
61. Yamasaki T, Handa H, Yamashita J, Paine JT, Tashiro Y, Uno A, Ishikawa M, Asato R. Intracranial and orbital cavernous angiomas. A review of 30 cases. J Neurosurg 1986; 64:197208. 62. Wang AM, Morris JH, Fischer EG, Peterson R, Lin JCT. Cavernous hemangioma of the thoracic spinal cord. Neuroradiology 1988; 30:261-4. 63. Wood MW, White RJ, Kernohan JW. Cavernous hemangiomatosis involving the brain, spinal cord, heart, skin and kidney: report of case. Proc Staff Meet Mayo Clin 1957; 32:249-54. 64. Wyburn-Mason R. The vascular abnormalities and tumours of the spinal cord and its membranes. London: Kimpton, 1943. 65. Zentner J, Hassler W, Gawehn J, Schroth G. Intramedullary cavernous angiomas. Surg Neurol 1989; 31:64-8.