Department of Neurosurgery and Services of Pathology and Neuroradiology, Bellaria Hospital, Bologna, Italy
OBJECTIVE: It has recently been found that patients receiving cerebral irradiation can develop hemorrhagic dysangiogeneses simulating occult vascular malformations. To analyze this connection, we report on five patients with occult cerebrovascular malformations occurring after "standard" or focused irradiation performed for brain tumors in four patients and for a deep-seated cavernous angioma in one patient.
METHODS: All lesions were within the radiation ports. The time interval between irradiation and the detection of the occult vascular malformations varied from 3 to 9 years; the ratio of female to male patients was 4:1. Four patients were <15 years old when first irradiated. Four patients presented with acute symptoms (headache, vomiting, focal signs) and one was asymptomatic when the lesions were first detected. Serial magnetic resonance imaging scans were available in four patients and a computed tomographic scan in the other patient.
RESULTS: The initial appearance was that of a hypointense T1T2 focus; magnetic resonance imaging then revealed focal or multifocal T1 hyperintensity and T2 mixed signal intensity followed by a late ring of decreased signal intensity. Four patients were operated on and one was under neuroradiological monitoring. Histological features of these lesions included clusters of closely packed vascular spaces resembling cavernous malformations sometimes associated with a thrombosed thick-walled vein with intense hemosiderin deposition and fibroblastic proliferation; telangiectasic changes were also seen in the adjacent brain.
CONCLUSION: Increased awareness of occult cerebrovascular malformations is necessary, because their occurrence is not infrequent and they have hemorrhagic potential. Children receiving cerebral irradiation are at greater risk of this complication.
(Neurosurgery 39:677684, 1996)
Key words: Cavernous angioma, Hemorrhage, Infancy, Magnetic resonance imaging, Occult vascular malformation, Radiation therapy
Occult cerebrovascular malformations (OVMs) include cavernous angiomas, telangiectasias, and thrombosed arteriovenous malformations. The existence of transitional lesions probably representing a spectrum within a single pathological entity has recently been emphasized (2, 22). The majority of OVMs requiring clinical attention and surgical intervention are cavernous malformations.
Recent reports have provided new insights into the biological behavior of cerebral cavernous angiomas (8, 23). Although many patients run an indolent course characterized by minor focal symptoms and/or seizures, some cases exist of offensive courses including recurrent bleeding and progressive growth (19, 20, 28). Although cavernous angiomas are likely to occur during embryogenesis, they may appear in areas of apparently normal brain suggesting an acquired origin and an unexpected proclivity to grow (16, 19, 30). New vascular malformations may represent the growth of a very small nidus caused by capillary proliferation or focal hemorrhage or some combination of these two factors (32). From this aspect, a close relationship with telangiectasia has been suggested (22).
In the multiple/familial form of the disease, de novo development has been demonstrated (32). Again, a small number of de novo OVMs, including telangiectasias and cavernomas, has been found in patients who had received cerebral irradiation (11, 19, 30). In this article, we present five such patients treated at our institution in the last 10 years.
Patient | Age at Irradiation (yr) | Sex | Latency (yr) | Dose (cGy) | Reason for Irradiation | Onset | MRI Findings | Treatment | Pathological Findings |
---|---|---|---|---|---|---|---|---|---|
1 | 10 | F | 6 | 3000 | Cerebellar astrocytoma | Headache, vomiting | Only CT scan (17) | Surgery | Cavernoma |
2 | 43 | F | 9 | 5050 | Pituitary adenoma | Asymptomatic | T1 hyperintensity, T2 mixed intensity; late dark ring | Surgery | Cavernoma/vein |
3 | 15 | F | 7 | 5000 | Dysgerminoma | Headache, vomiting | T1 hyperintensity surrounding T2 hypointensity | Surgery | Cavernoma/vein |
4 | 15 | F | 9 | 25005000 | Previous cavernoma | Seizure | T1T2 hyperintense fluid level | Surgery | Cavernoma/telangiectasia |
5 | 12 | M | 3 | 5400 | Cerebral astrocytoma | Headache, vomiting | T1T2 hypointensity, then hemorrhage | Conservative |
Routine MRI control scans at 3 years showed a small hypointense T1 and T2 vermian lesion, not demonstrated before radiotherapy and consistent with chronic hemorrhage (Fig. 5, A and B). In 1995, he complained of sudden headache and vomiting; a CT scan showed a hemorrhagic hyperdensity corresponding to the previously identified vermian lesion. MRI confirmed a vermian subacute hemorrhage (Fig. 5C). Because the patient's symptoms fully resolved in a week, a clinicoradiological program was delivered without surgical therapy. At follow-up examination 6 months later, the patient remained asymptomatic.
Vascular alterations taking the form of large irregular telangiectases (3, 18, 25) are more often adjacent to areas of radiation necrosis but they may also occur in otherwise normal irradiated brain tissue (11). Recent radiological literature shows that radiation to the brain can cause vascular lesions that are radiologically and pathologically identical to vascular malformations (4) and resemble closely telangiectasia and cavernous angioma (11, 19). Gaensler et al. (11) emphasized that small telangiectasias may be induced with radiation therapy and may slowly or acutely hemorrhage, resulting in an appearance on MRI scan similar to that of occult vascular malformations. However, the term telangiectasia is somewhat restrictive because radiation changes probably represent a more complex pathological phenomenon. Moreover, the clinical behavior is often similar to that of cavernous angiomas. The relationships between these acquired OVMs and irradiation again emphasize the strict link among cavernous angioma, telangiectasia, and venous anomalies. Recently, we have witnessed an evolution of the concept of cavernous angioma. This vascular malformation is now defined apart from its compactness; moreover, capillary and venous components may be present (24, 29). In patients who had received irradiation, Gaensler et al. (11) correlated punctate hyperintense areas on MRI scans representing subacute methemoglobin with autoptically demonstrated dilated venous varix; the gradual narrowing of the lumen as the vessel approached the surface of the brain suggested an original element of venous occlusion. Okeda and Shibata (17) suggested that radiation may preferentially affect the endothelium of veins, producing veno-occlusive disease with subsequent development of exudation, hemorrhage, and necrosis. They think that the formation of vascular telangiectasia may represent a physiological attempt to form collateral drainage away from the area of venous occlusion and congestion. Progressive obstruction of the venous outflow may also lead to venous hypertension at the distal radicles promoting ischemia, microhemorrhages, and angiogenic factor production leading to an OVM (7). The intense hemosiderin deposition and fibroblastic proliferation found in the venous walls in some our patients confirm this mechanism. Our surgical specimens stress this "cascade" linking a thrombosed vein with the development of endothelial vascular spaces, which may further and variably "compact" mimicking the aspect of telangiectasia (11) and/or cavernoma.
A high association of venous malformation accompanying a cavernous angioma has been emphasized by Robinson et al. (24); their coexistence in the same location may indicate a common pathogenesis (21, 26). Ciricillo et al. (7) suggested that although a venous angioma is a congenital malformation, any associated OVM is a dynamic acquired anomaly with a tendency for recurrent hemorrhages and the possibility of forming new occult malformations that may evolve into lesions visible on MRI scans. Cavernous angiomas probably release angiogenic factors with transformation of normal capillaries, true budding, and growth of new vascular channels (16). Both platelet-derived and vascular endothelial growth factors may play a role in the development of cavernous malformations (24). Other factors, such as the beta chain of fibroblast growth factor, are currently being investigated as possible mediators of vascular growth (5). It has also been reported that levels of several growth factors, including basic fibroblast growth factor, are elevated in irradiated endothelial cells (27); this finding may represent a link between irradiation and vascular malformation induction. A genetic predisposition to the formation of cavernous lesions is also possible (9, 12, 16).
It was found that OVMs induced by exposure to radiation were often superficial in location near the gray-white matter junction (11), peculiar in their histological conformation (19), and even multiple (4, 11). The structure was characterized by back-to-back endothelial vascular spaces with minimal collagen support and was devoid of the longstanding changes (calcifications and organized thrombi) that are typical of mature cavernous angiomas. This configuration has been described as "en dentelles" ("lace-like") (15) and correlates with an original or primitive aspect of this malformation.
The benefits of radiosurgery for cavernous angiomas have not been clearly demonstrated. Kondziolka et al. (13) recently reported a significant reduction in the hemorrhage rate after radiosurgery in patients who had deep hemorrhagic cavernous malformations. However, a decrease in the hemorrhagic potential may also occur spontaneously after an initial virulent clinical course (13, 19). The response of the cavernomatous vessels to radiosurgery might consist of endothelial cell proliferation, vessel wall hyalinization, and eventual luminal closure (13). Conversely, a similar process may promote the growth of a dormant OVM stimulating the dominant endothelial structure and the process of intraluminal thrombosis and subsequent recanalization which may favor the growth of a cavernous angioma (28, 29). In patients receiving radiosurgery for a cavernous angioma (Patient 4), the risk for radiation-induced growth of an OVM may be greater considering that an apparently single lesion may represent an occult multiple form of the disease (32) not evident at the time of the irradiation.
Although OVMs related to radiation treatment have only recently been reported, they may become a more prevalent phenomenon in patients subjected to previous and even remote radiotherapy. These de novo lesions occur more often in patients irradiated in infancy. In a series of OVMs in childhood reported by Ciricillo et al. (6), seven patients developed OVMs at 1.5 to 16 years after radiation therapy for a preexisting malignancy. Wilson reported three patients with radiation-induced OVMs occurring some decades after the initial irradiation (30). Findlay et al. (10) reported a case of associated induction of meningioma and cavernoma occurring 20 years after prophylactic cranial irradiation for infantile acute leukemia. The shortest interval was found by Gaensler et al. (11); in their patients undergoing routine follow-up MRI after excision of a brain tumor, the average latency between radiation therapy and the appearance of hemorrhage related to "radiation-induced telangiectasia" was 32.5 months. No convincing dose response relationship was observed by these authors. In our patients, the time interval varied from 9 to 3 years; all but one were female patients, as occurred also in other reports (1, 10, 30), but it is unclear if females are more susceptible than males. In asymptomatic patients, initial MRI findings included hypointense foci on T1- and T2-weighted images indicative of chronic resolved hemorrhage with hemosiderin deposition (11) as occurred initially in our Patient 5. In symptomatic patients, the most frequent finding was that of a focal or multifocal hyperintense T1 and mixed T2 lesions secondary to subacute hemorrhage or thrombosis. As generally found in cavernous angiomas (32), a T1T2 hypointense ring denoting hemosiderin staining occurred in two patients who underwent follow-up examinations before excision (Fig. 2). An important question also relates to the clinical behavior of these de novo OVMs. Massive cerebral hemorrhage caused by "an aggregation of abnormal microscopic blood vessels" and unrelated to the original neoplasia has been also reported in long-term survivors of irradiated pediatric brain tumors (1). Intracranial hemorrhage secondary to radiation-induced telangiectasia occurred in 5 of 20 patients (25%) reported by Gaensler et al. (11). These figures seem worse when compared to the natural history attributed to the better known "spontaneous" cavernous angiomas (8, 23).
Received, March 5, 1996.
Accepted, May 7, 1996.
Reprint requests: Eugenio Pozzati M.D., Department of Neurosurgery, Bellaria Hospital, Via Altura 3, 40139 Bologna, Italy.
Pozzati et al. make similar observations in five additional patients, with four histological verifications. They articulate the intriguing hypothesis that venous occlusion induced by radiation might be associated with this phenomenon. This would be consistent with activation of the angiogenesis cascade at the venous side of the capillary bed (3), and with recent observations of angiogenesis factor expression within the lesions (8). Interestingly, the mechanism of venous side hemorrhagic dysangiogenesis could also account for the frequent association of cavernous malformation with venous angioma (or venous developmental anomaly).
Our group and others have recently published preliminary mapping of the gene of familial cavernous malformations (2, 5). This gene seems to account for all known cases to date among Hispanic Americans of Mexican descent and is associated with a preserved haplotype (common chromosomal segment) in the region of the gene, suggesting a founder effect from a common ancestor in this population (6). The same preserved haplotype has also been found in sporadic cases of cavernous malformation in Hispanic Americans of Mexican descent, suggesting that they also carry the same genetic predisposition (6). The gene itself has not yet been identified, nor has the specific gene product or mechanism mediating the dysvasculogenesis of cavernous malformation. Although non-Hispanic families in the United States and abroad have also been linked to the same 7q gene locus, other non-
In the absence of specific identification of the gene or its product, it is not currently possible to perform a blood test to uncover genetic predisposition to this lesion. In Hispanic Americans, it is possible to look for the preserved haplotype (in DNA extracted from blood sample) as an indirect indicator of carrying the gene (6). In the absence of such marker or specific knowledge of the gene itself, it is not presently possible to know if a specific non-Hispanic case of cavernous malformation is associated with genetic predisposition. As soon as it is feasible to test for such predisposition in non-Hispanic patients, we would be able to determine whether cases of cavernous malformation
presenting after irradiation occur solely in patients with genetic predisposition. To date, we do not know of a Hispanic American patient with this complication in whom this hypothesis might be tested. We are surprised that none of the cases presented by Pozzati et al. or by Gaensler et al. (4) manifest any family history of the lesion. It remains an intriguing hypothesis that brain irradiation might precipitate the genesis of cavernous malformation in genetically predisposed patients. Alternatively, a different type of host
predisposition (from that causing familial cavernous malformation) might be operative in these patients. The predominance of pediatric cases is consistent with such host predisposition.
The management stance toward such lesions must be individualized, as in all patients with multiple occult vascular malformations. Two of our patients have been followed expectantly for over 2 years, with none of the lesions manifesting growth or overt hemorrhage necessitating surgical intervention. One must surely remain cognizant of the limitations of surgery in a disease where multiple lesions might arise in the future. Surgical excision must be considered only in the case of symptomatic progression or overt growth or hemorrhage from a particular lesion.
Finally, we must remain skeptical that all such reported lesions actually represent de novo genesis of cavernous malformations. In fact, most of the published cases lack baseline radiological studies of similar sensitivity to the late generation magnetic resonance scans when lesions were discovered. At least some of the lesions might represent incidental revelation of cavernous malformation, as occurs in 0.5% of the population, and it would not be surprising if they would occur in a similar proportion of irradiated patients. We also cannot be certain whether some patients harbored preexisting capillary malformations (not detected on baseline imaging studies) that progressed to overt hemorrhagic lesions.
We concur with the authors that the term "telangiectasia" is misleading. From the Greek, it translates as telos = end, angeion = vessel, and ektasis = a stretching out; i.e., "dilation of small or terminal vessels." In the brain, the term has been used to refer to capillary malformations. Elsewhere in the body, it has been used to refer to a wide variety of vascular malformations, including the typical true arteriovenous malformation of Osler-Weber-Rendu (hereditary hemorrhagic telangiectasia). This disease is in fact associated with multiple brain arteriovenous malformations, and not multiple telangiectasias as would be suggested by its nomenclature. The systemic lesions in this disease are not true telangiectasias but arteriovenous malformations (1). Because of this confusion in semiology, we favor abandoning the use of the term "telangiectasia" except in the historical context of nomenclature of disease. Otherwise, the term "capillary malformation" would best describe the abnormal dilation of capillaries in the brain and elsewhere. Another advantage to the term "capillary malformation" is its classification within a spectrum of vascular malformations, including cavernous malformations (which, in many instances, actually consist of two mixed capillary-cavernous malformations or a mixed venous-cavernous malformation). This would not only avoid confusion with the term telangiectasia, but also avoid the use of another imprecise term, "angioma," which has been used to refer to the venous developmental anomaly (which could simply be called "venous malformation"). It is hoped that this simplification will allow us to move the terminology toward pathophysiological significance consistent with modern biological understanding of the lesions.
Issam A. Awad
The authors provide several additional cases of de novo cavernomas appearing after radiation therapy and a good review of the knowledge as well as theories concerning their origins and development. Patient 1 is not altogether convincing, inasmuch as magnetic resonance examinations were lacking and it is difficult to be certain that the right parietal cavernoma was a de novo development secondary to the posterior fossa radiation. Likewise, Patient 4 may well have a genetic predisposition to multifocal cavernomas that developed outside the focus of radiosurgery.
The pathophysiology of the origins of these lesions is most intriguing and considering the extensive knowledge that has been gained in recent years through their demonstration by magnetic resonance imaging, and more recently factors involved in their growth, it is realistic to think that this enigmatic lesion may eventually be fully understood.
David G. Piepgras
This report will further stir the cauldron of controversy related to the natural history and management for occult vascular malformations. This article, as well as other similar reports linking radiation to the emergence of these lesions, makes the concept of using radiation as treatment paradoxical. The Joint Section on Cerebrovascular Surgery in conjunction with the Stereotactic Radiosurgical Team at the University of Pittsburgh and John Kestle from the University of British Columbia are currently developing a protocol to look at the impact of such techniques on the rate of subsequent bleeding of cavernous angiomas. It is only this kind of careful science and cooperative effort that is going to ultimately decide the role of radiation in the management of these lesions.
Steven L. Giannotta
This is an interesting report on a condition that is not widely recognized, and its possible causative factor(s) remain controversial. Radiotherapy has been identified as one possible causal factor in the formation of these cerebrovascular malformations. Although the authors presented five patients who have received brain irradiation, in two of these five patients radiotherapy seems to be a relatively weak causative factor. This article is, however, of importance because it may increase the clinician's awareness of existence of this interesting and intriguing entity and perhaps help to diagnose more patients with cryptic cerebrovascular malformations.
Zbigniew Petrovich
Return to the Neurosurgery Home Page
New Haven, Connecticut
Rochester, Minnesota
Los Angeles, California
Radiation Oncologist
Los Angeles, CaliforniaPlease submit your thoughts or comments to William Chandler, Internet Moderator at wchndlr@umich.edu.
© 1996 Williams & Wilkins