- Email: [email protected]
MR imaging of giant intracranial aneurysm
Journal of Clinical Neuroscience (2003) 10(4), 460–464 ª 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0967-5868(03)00092-4
Neuroradiology study
MR imaging of giant intracranial aneurysm Michael Mu-Huo Teng1 Cheng-Yen Chang1 MD
MD,
S.M. Nasir Qadri2
MD,
Chao-Bao Luo1
MD,
Jiing-Feng Lirng1
MD,
Shin-Su Chen1
MD,
1 Department of Radiology, Taipei Veterans General Hospital, 201, Sect 2, Shih-Pai Road, Peitou, Taipei 11217, Taiwan; School of Medicine, National Yang-Ming University, Taiwan, ROC, 2New X-ray Department, Services Hospital, Jail Road, Lahore, Pakistan
Summary We reviewed the clinical features and findings of magnetic resonance imaging (MRI) in 17 cases of giant intracranial aneurysm (GIA). All were confirmed by digital subtraction angiography (DSA). Seven were non-thrombosed and 10 were partially thrombosed. All thrombosed parts showed no enhancement. The majority of the lumen showed good enhancement in the post-contrast study (89%). About 80% of the partially thrombosed aneurysms had an onion skin appearance on non-contrast T1WI. On T2WI about 80% of partially thrombosed aneurysms had a low signal inside the thrombus or its wall from the susceptibility effect of blood elements such as haemosiderin. Flow void sign was noted in 88% on non-contrast T1WI and T2WI. Flow artifacts along the phase-encoding direction from the lumen of the aneurysm were noted in 41% of non-contrast T1WI, 56% of contrast-enhanced T1WI, and 29% of T2WI. Non-enhancement inside the thrombus and an onion skin appearance were important signs of thrombosis of the aneurysm. Flow artifacts along the phase-encoding direction from the lumen of the aneurysm provided an important clue of a patent lumen inside an aneurysm. The following signs provided clues to the diagnosis of patency lumen instead of thrombosis with high sensitivity and specificity: flow void sign on noncontrast T1WI (sensitivity 88%, specificity 100%), enhancement on contrast-enhanced T1WI (sensitivity 89%, specificity 100%), presence of either flow void sign or enhancement on contrastenhanced T1WI (sensitivity 100%, specificity 100%). ª 2003 Elsevier Science Ltd. All rights reserved. Keywords: digital subtraction angiography, aneurysm, giant aneurysm, partially thrombosed aneurysm, magnetic resonance imaging
INTRODUCTION Giant intracranial aneurysm (GIA) means an aneurysm greater than 2.5 cm at its greatest diameter.1–10 An aneurysm larger than 4.5 cm in diameter is called hypergiant aneurysm,1 a very uncommon entity. GIA constitute approximately 3–13% of all intracranial aneurysms.2 ,68 Most authors agree that GIAs usually present as slow-growing space occupying masses rather than with subarachnoid haemorrhage. GIA cause subarachnoid haemorrhage usually when rupture of the fundus occurs.4 We reviewed our cases of GIA, their clinical features and findings on MRI. METHODS Seventeen cases of partially thrombosed or non-thrombosed GIA were diagnosed by MRI examination and confirmed by digital subtraction angiography. Ten were female and 7 were male. We reviewed the clinical features and findings of MRI in these cases (Table 1). The size and location of these aneurysms, clinical manifestations, and complications of haemorrhage and infarct found in MRI are listed in Table 1. MRI examinations were performed using 1.5 T Sigma (GE Milwaukee, USA) or 1.5 T Magneton (Vision, Siemens, Germany). T2-weighted fast spin echo axial images, T1-weighted spin echo axial images and proton density weighted axial images were performed. Five cases had post-contrast T1-weighted images in coronal, axial and sagittal views after intravenous administration of gadolinium-DTPA. Lumen and thrombus portions of the an-
eurysm on MRI were identified in each case by correlation with the angiographic picture. We reviewed the signal changes and patterns of enhancement on MRI. Signal changes inside the aneurysm were compared with brain, and called 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. We also evaluated susceptibility artifact on T2WI, onion skin appearance on T1WI, flow void sign inside the lumen, and flow artifacts along the phase-encoding direction arising from the lumen of the aneurysm. Susceptibility artifact was considered when low signal changes inside the thrombus portion of the aneurysm was found on T2WI. This was usually from various blood degredation products such as haemosiderin. Onion skin appearance was used to describe the multilayered thrombus wall from a partial or completely thrombosed aneurysm.14 We considered flow void sign when low signal was found inside the lumen of the aneurysm. According to the presence or absence of flow void sign and high signal flow artifact, we formulated 6 classifications (Table 2). We called it type A if part of the lumen had flow void sign with no high signal flow artifact, type B if the whole lumen of the aneurysm had flow void sign with no high signal flow artifact, type C if there was presence of both regions of high signal flow artifact and low signal flow void sign inside the lumen, type D if part of the lumen had high signal flow artifact inside with no flow void sign, type E if the whole lumen was filled with high signal flow artifact completely with no flow void sign, and type F if neither low signal flow void sign nor high signal flow artifact was present. RESULTS
Received 14 August 2002 Accepted 19 September 2002 Correspondance to: Dr. Michael Mu Huo Teng, Department of Radiology, Veterans General Hospital-Taipei, 201, Sect 2, Shih-Pai Road, Peitou, Taipei 11217, Taiwan. Tel.: +886-2-28718251; Fax: +886-2-28757348; E-mail: [email protected]
460
Seven patients (41%) had new or old haemorrhage demonstrated by CT or MRI (Table 1). Intraparenchymal haemorrhage in adjacent brain parenchyma was found in 4 patients, intracerebral haemorrhage in remote brain parenchyma in 1 patient, subarachnoid haemorrhage in 2 patients, intraventricular haemorrhage in 2 patients, and haemorrhage along the wall of the aneurysm in 5
MR imaging of giant intracranial aneurysm
Table 1 #
Clinical data
Age
Sex
Size
Location
Clinical manifestations
Complications of haemorrhage and infarct No Old infarct in posterior cerebral territory. Recent haemorrhage in brain parenchyma (brainstem) and along the aneurysmal wall New infarct in internal capsule and lentiform nucleus Subarachnoid, intraventricular haemorrhage, and along the aneurysmal wall No Haemorrhage in brain parenchyma and along the aneurysmal wall Intraventricular haemorrhage No
1 2
49 68
M M
3.2 cm 3.3 cm
Cavernous ICA, right Basilar artery
Trigeminal neuralgia, right III nerve palsy Dizziness, vertigo, general weakness
3
78
M
2.8 cm
Supraclinoid ICA, left
4
59
F
4.8 cm
MCA, left M1
RT limb weakness; infarct in internal capsule (new) Dizziness
5 6
37 31
F F
3.5 cm 2.5 cm
Cavernous ICA, right M1 of left MCA
7 8
70 67
M F
3.2 cm 3.0 cm
Supraclinoid ICA, right Supraclinoid ICA, left
9 10 11 12
57 56 62 58
F F F M
3.0 3.0 3.0 4.0
cm cm cm cm
13 14 15 16 17
78 62 63 56 68
M F F F M
2.6 2.6 2.9 2.6 2.7
cm cm cm cm cm
Cavernous ICA, right Supraclinoid ICA, left Cavernous ICA, right Right anterior communicating artery Cavernous ICA, right Supraclinoid ICA, left Cavernous ICA, right Supraclinoid ICA, right Supraclinoid ICA, left
Diplopia. Right III nerve palsy Right facial weakness, headache, upper limb weakness General weakness, incoherent speech Chronic headache, nuchal rigidity, visual field defect Headache, blurred vision Headache, visual field defect Right VI nerve palsy Tiredness, haemiplegia, hydrocephalus
Recent infarct in middle cerebral artery territory No Old haemorrhage in lentiform nucleus Haemorrhage in brain parenchyma and along the aneurysmal wall No No Old brain contusion in bilateral low frontal No Intracerebral, subarachnoid haemorrhage and along the aneurysmal wall
Headache, blurred vision Headache, visual field defect Headache, blurred vision, facial numbness Blurred vision Headache
Table 2 Appearance of flow void and high signal flow artifacts inside lumen of the aneurysm Types
Flow void sign
High signal flow artifact
A B C D E F
In part of lumen In whole lumen In part of lumen Not present Not present Not present
Not present Not present In part of lumen In part of lumen In whole lumen Not present
Table 3
461
patients. Three cases (18%) had brain infarction in the distal branches of the parent artery of the aneurysm. Ten aneurysms were partially thrombosed. MRI findings of partially thrombosed aneurysms are listed in Table 3. Eight cases had an onion skin appearance (multilayered thrombus wall) indicating partial or completely thrombosed aneurysm. Low signal (signal intensity of 1 or 2) inside thrombus on T2WI as a result of susceptibility effect from various stages of blood elements was found in 8 patients. Five aneurysms with thrombus inside had
MRI findings of partially thrombosed aneurysms
Case number Signal change inside thrombus on non-contrast T1WI Onion skin appearance on T1WI Enhancement inside thrombus in the post contrast T1WI Rim enhancement on margin of thrombus on T1WI Signal change inside thrombus on T2WI Low signal change inside thrombus or in its wall on T2WI
1
2
6
7
8
9
10
11
12
15
4&6 ) ) + 1&2 +
4&6 + X X 1&2 +
4&6 ) X X 1&6 +
4&6 + ) Part 1&6 +
4 + X X 6 )
6 + X X 6 )
6 + X X 1&6 +
6 + ) + 1&6 +
3&6 + ) ) 2&6 +
6 + ) ) 1&4 +
Following cases had no thrombus inside: 3, 4, 5, 13, 14, 16, 17, thus are not included in this table. Signal change: 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. +, present; Part, partly present; ), not present; X, not available.
Table 4
MRI findings in the lumen of aneurysm and signs related to flow
Case number Signal change inside lumen on non-contrast T1WI Radiating artifact from lumen on non-contrast T1WI Well enhancement inside lumen in the post-contrast T1WI Radiating artifact outside the lumen on contrast-enhanced T1WI Signal change inside lumen on T2WI Radiating artifact outside the lumen on T2WI Flow-void sign&high signal flow artifact on noncontrast T1WI Flow-void sign and high signal flow artifact on T2WI
1 1&4 ) Part ) 1 ) A B
2
3
4
5
6
7
8
9
10
11
2&4 2&4 1&4 3&5 1&4 2&4 1&4 1 1&4 1 ) + + ) + + + ) ) ) X + X X X + X X X + X + X X X + X X X ) 2&4 1&6 2&4 3&6 1&6 1&6 1 1 1&6 1 ) + ) ) + + ) ) ) ) A A A F A A A B A B A C A D C C B B C B
12
13
14
2&4 3&4 1 + ) + + + ) + ) + 1&4 6 1 ) ) + A F B A E B
15
16
17
1&4 2&5 1&4 ) ) ) + + X ) + X 1&4 1&6 1&6 ) + ) A C C A C C
Signal change: 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. +, definitely present; ), not present; X, not available. Part: part of the lumen had flow void sign, and part of the lumen had enhancement. A–E, see Table 2.
ª 2003 Elsevier Science Ltd. All rights reserved.
Journal of Clinical Neuroscience (2003) 10(4), 460–464
462 Teng et al.
postcontrast studies: all showed no enhancement inside the thrombus (Table 3). Complete rim enhancement at margin of thrombus (usually in the wall) was present in 2, and a partial rim was found in 1. All 14 cases had patent lumens inside the aneurysm. Seven cases (cases 3–5, 13, 14) had no evidence of thrombosis within the lumen. MRI findings from the aneurysm lumen and signs related to flow are listed in Table 4. All lumens of these aneurysms on non-contrast T1WI had signal intensity of 1–3 inside. On T1WI 15 cases had signal of 1–2 inside indicating flow void sign: 3 type B, 10 type A and 3 type C. In additional to this, 12 had signal intensity of 4, and 2 had signal of 5–6 suggesting high signal flow artifact inside. Two patients had neither low signal flow void nor high signal flow artifact (type F) (see Figs. 1 and 2). The patent lumens of the aneurysms on T2WI were of mixed signal in 11. All except 2 (cases 5 and 13) had low signal of 1–2 inside indicating flow void, 8 of them mixed with signal of 6 (hyperintense) and 4 cases mixed with signal of 4 (isointense) inside. In 5 cases, the lumen was filled with signal 1, compatible with type B flow void. In 6 cases there is signal of 1 and 6 inside lumen, compatible to type C. Two patients had no flow void, 1 had a small area of high signal flow artifact inside (type D), and
the other the whole lumen was filled with high signal flow artifact (type E). Nine patients had post-contrast T1WI, the whole lumen of their aneurysms showed good enhancement in 7 patients, part of the lumen had good enhancement and part had flow void sign in 1 patient, and the whole lumen had flow-void sign and showed no enhancement in 1 patient (Table 4). Flow artifacts along the phase-encoding direction from the lumen of the aneurysm was noted in 41% (7/17) on non-contrast T1WI, 29% (5/17) on T2WI, and 56% (5/9) on postcontrast T1WI. Ghost images of aneurysm were present in 1 case on both noncontrast T1WI and T2WI. DISCUSSION GIA has been divided into many types: giant and hypergiant, saccular, fusiform and serpentine, fully thrombosed, partially thrombosed and non-thrombosed, nontraumatic and traumatic. Ten cases were partially thrombosed aneurysms while 7 cases were non-thrombosed. According to Fisher and co-workers,8 nearly half of the GIA arise from the ICA. Most of our GIAs (76%) were from the ICA.
Fig. 1 Case 12. Partially thrombosed aneurysm. (A) Non-contrast T1WI. There is intraparenchymal haemorrhage in left frontal region (arrowheads). Flow void sign is present inside part of lumen of the aneurysm (arrow), compatible with type A. (B) Post-contrast T1WI. The lumen of the aneurysm showed homogeneous wall enhancement (arrowheads). Flow void sign is not seen in this postcontrast study. (C) T2WI: flow void sign is present inside part of the lumen (arrow), compatible with type A. (D) AP view of right carotid angiogram shows the lumen of the aneurysm (arrowheads) arises from the junction of the right anterior cerebral artery and right anterior communicating artery.
Journal of Clinical Neuroscience (2003) 10(4), 460–464
ª 2003 Elsevier Science Ltd. All rights reserved.
MR imaging of giant intracranial aneurysm
463
Fig. 2 Case 14. Non-thrombosed aneurysm (A) Sagittal view postcontrast T1WI: There is flow-void sign inside the whole lumen of the aneurysm (arrowheads), compatible with type B. Enhancement is seen only in the wall of the aneurysm. Radiating flow artifacts (small arrows) originating from the aneurysm is present along the phase-encoding direction in the anterior and posterior aspect of the aneurysm. (B) T2WI: Flow-void sign is present in the whole lumen (arrow) of the aneurysm. Its area is the same as its lumen, compatible with type B. (C) Anteroposterior view and (D) lateral view of left carotid angiogram: an aneurysm (arrow) can be seen in the left internal carotid artery.
Many factors have been considered responsible for the formation and growth of an aneurysm to a GIA. One is neovascularisation and recurrent haemorrhage within the wall of the aneurysm.3;4;9;12 Because these recurrent haemorrhages are confined to the aneurysm wall and do not reach the subarachnoid space, the aneurysm is not detected until it becomes a GIA. Seven patients had new or old haemorrhage. A relative high percentage of them (71%) had haemorrhage along the aneurysmal wall and a low percentage (29%) had subarachnoid haemorrhage compatible with this theory. In addition to the congenital defect in the tunica media, other factors are also responsible, such as atherosclerosis. Some genotypic connective tissue diseases such as Marfans syndrome, Ehlers–Danlos syndrome, elastic pseudoxanthoma and fibromuscular dysplasia are also associated with GIA.1;13 Aneurysms may even develop after a cranial injury, infection, or a surgical operation. Four cases had a previous history of head injury. The clinical symptoms vary according to the location of the aneurysm.1 Most of our cases (88%) had symptoms caused by mass effect, such as cranial nerve palsy, visual field defect, and limb weakness. Eight cases (47%) had headache. Seven cases (41%) had haemorrhage demonstrated by CT or MRI (Table 1). Our findings are compatible with previous reports that GIAs usually present either as a slow growing space occupying intracranial masses or with intracranial haemorrhage.1–13 Distal cerebral infarction in the territory of the parent artery is a commonly ignored complication of GIAs. Three cases (18%) had brain infarction of different chronicity in the distal branches ª 2003 Elsevier Science Ltd. All rights reserved.
of the parent artery harbouring an aneurysm. The cause of distal cerebral infarction in the territory of the artery with GIA could be thrombus arising from stagnant blood flow inside the GIA or compression of the parent artery by the GIA. The two basic constituents of GIA are laminated blood in various stages of thrombosis, and flowing blood. Rapidly flowing blood in the aneurysmal lumen appears as an area of signal void on MRI.2 “The flow void sign” had a high sensitivity on non-contrast T1WI (88%) and T2WI (88%). On T1WI, only part of the lumen had flow void sign in most patients (12 cases, 71%), and the whole lumen had flow void sign in 3 cases (18%). On T2WI, part of the lumen had flow void sign in most patients 59% (10/17), and the whole lumen in 29% (5/17). The absence of flow void sign did not exclude the diagnosis of aneurysm with a patent lumen. It was not present in 2 cases on non-contrast T1WI and in 2 cases on T2WI. The persistence of flow void in the post-contrast MRI is a specific sign of aneurysm.8 From our study, this sign in the post-contrast T1WI had a low sensitivity (22%) and high specificity (100%). Flow void sign may also be confused with the low signal inside thrombus from blood elements on T2WI, thus having a relative low specificity (20%) on T2WI. On post-contrast T1WI, 8 out of 9 patients had good enhancement in the lumen of their aneurysms, 1 in the whole lumen, and 7 in a part of the lumen. Combining flow void sign and internal enhancement to indicate lumen patency instead of thrombosis had high sensitivity (100%), high specificity (100%), positive prediction sign (100%), and negative prediction (100%) (Table 5). Journal of Clinical Neuroscience (2003) 10(4), 460–464
464 Teng et al.
Table 5
Different signs of aneurysm
Signs
Sensitivity
Specificity
Positive predictive Negative predictive value value
Flow void sign on noncontrast T1WI without contrast as an indication of aneurysmal lumen (low signal inside thrombus considered false positive) Flow void sign as an indication of aneurysmal lumen on post-contrast T1WI (low signal inside thrombus considered false positive) Enhancement as an indication of aneurysmal lumen on post-contrast T1WI (enhancement inside thrombus considered false positive) Presence of either flow void sign or enhancement as a combined indication of aneurysmal lumen on post-contrast T1WI Flow void sign on T2WI without contrast as an indication of aneurysmal lumen (low signal inside thrombus considered false positive) High signal artifact on noncontrast T1WI without contrast as an indication of aneurysmal lumen (high signal inside thrombus considered false positive) High signal artifact on T2WI without contrast as an indication of aneurysmal lumen (high signal inside thrombus considered false positive) Low signal change as an indication of thrombed portion of an aneurysm on T2WI (low signal change inside the lumen on T2WI considered false positive)
15/17 ¼ 88%
10/10 ¼ 100%
15/15 ¼ 100%
10/12 ¼ 83%
2/9 ¼ 22%
5/5 ¼ 100%
2/2 ¼ 100%
5/12 ¼ 42%
8/9 ¼ 89%
5/5 ¼ 100%
8/8 ¼ 100%
5/6 ¼ 83%
9/9 ¼ 100%
5/5 ¼ 100%
9/9 ¼ 100%
5/5 ¼ 100%
15/17 ¼ 88%
2/10 ¼ 20%
15/23 ¼ 65%
2/4 ¼ 50%
2/17 ¼ 12%
1/10 ¼ 10%
2/11 ¼ 18%
1/16 ¼ 6%
8/17 ¼ 47%
3/10 ¼ 30%
8/15 ¼ 53%
3/12 ¼ 25%
8/10 ¼ 80%
2/15 ¼ 13%
8/23 ¼ 35%
2/4 ¼ 50%
Sensitivity of more signs: (A) Onion skin appearance on noncontrast T1WI: 8/10 ¼ 80%. (B) Nonenhancement inside thrombus of an aneurysm 5/5 ¼ 100%. (C) Rim enhancement on margin of thrombus in contrast-enhanced T1WI 3/5 ¼ 60%. (D) Radiating flow artifact from the aneurysmal lumen: noncontrast. T1WI, 41%; T2WI, 29%; contrast-enhanced T1WI 56%.
Flow artifacts and ghost images extending to either side of the lumen of the aneurysm along the phase-encoding direction are also motion artifacts. Presence of a wide band of flow artifact or ghost images indicates presence of an arterial lumen, thus the diagnosis of GIA. Although the sensitivity was low, they provided definite clues to the diagnosis of the patent lumen of an aneurysm. According to Sarwar et al.,14 partially thrombosed giant intracranial aneurysms display an onion skin appearance of the thrombosed portion. Each layer represents a clot of different chronicity. Onion skin appearance was seen in 80% of our cases with partial thrombosis on non-contrast T1WI. Therefore this appearance is very helpful in the diagnosis of GIA with partial thrombosis. In the post-contrast T1WI of all patients, no enhancement could be found inside the thrombosed portion of the aneurysm (sensitivity 100%). In 3 out of 5 cases with partial thrombosis, the aneurysmal wall enhanced on contrasted T1WI. Presence of low signal intensity on T2WI in the region of the thrombosed portion as a result of susceptibility effect from various stages of haemoglobin degredation was found in 80% of our cases. Because of the flow void sign inside the lumen of an aneurysm, signal intensity was also low on T2WI. Presence of low signal intensity on T2WI indicating the thrombosed portion of the aneurysm had a low specificity (13%) (Table 5). In conclusion, most of our patients (88%) had symptoms caused by mass effect. About 41% of our patients had haemorrhage. 18% of our patients with aneurysms had brain infarction of different chronicity in the distal branches of the parent artery. All thrombosed parts of an aneurysm showed no enhancement. The lumen of our aneurysms had good enhancement or the flow void sign in the postcontrast T1WI study. Onion skin appearance was noted in 80% of our partially thrombosed aneurysms on non-contrast T1WI. The following signs provided the clue to the diagnosis of lumen patency as opposed to thrombosis with high sensitivity and specificity: flow void sign on noncontrast T1WI (sensitivity 88%, specificity 100%), enhancement on post-contrast T1WI (sensitivity 89%, specificity 100%), and either flow void sign or enhancement inside on postcontrast T1WI (sensitivity 100%, specificity 100%). Radiating flow artifact from the lumen of the aneurysm provided a definite clue to the diagnosis of lumen patency (41% of non-contrast T1WI and 56% of contrast-enhanced T1WI). Journal of Clinical Neuroscience (2003) 10(4), 460–464
ACKNOWLEDGEMENTS The study was partially supported by National Science Council of Republic of China grant number NSC 89-2320-B-075-001-M08. Dr Qadri was neuroradiology fellow at Taipei Veterans General Hospital, 1999–2000. REFERENCES 1. Castel JC, Chaabane M, Guibert-Truiner F, Piton J, Caille JM. Intracranial supratentorial “Hypergiant” aneurysms – diagnostic problems. J Neuroradiol 1985; 12: 135–149. 2. Atlas SW, Grossman RI, Goldberg HI, Hackney DB, Bilaniuk LT, Zimmerman RA. Partially thrombosed giant intracranial aneurysm: correlation of MR and pathological findings. Radiology 1987; 162: 111–114. 3. Whittle IR, Dorsch NW, Besser M. Spontaneous thrombosis in giant intracanial aneurysm. J Neurol 1982; 45: 1040–1047. 4. Schubiger O, Valavanis A, Wichman W. Growth mechanism of giant intracranial aneurysms; demonstrated by CT and MR imaging. Neuroradiology 1987; 29: 266–271. 5. Fodstad H, Liliequist B, Staffan W, Nilsson PE, Boquist L, Abdul-Rahman A. Giant serpentine intracranial aneurysm after carotid ligation. J Neurosurg 1978; 49: 903–909. 6. Belec L, Cesaro P, Brugieres P, Gray F. Tumor simulating giant serpentine aneurysm of the posterior cerebral artery. Surg Neurol 1988; 29: 210–215. 7. Schubiger O, Valavnis A, Hayek J. Computed tomography in cerebral aneurysms with special emphasis on giant intracranial aneurysms. J Comput Assist Tomogr 1980; 4: 24–31. 8. Fisher A, Som PM, Mosesson RE, Lidov M, Liu TH. Giant intracranial aneurysm with skull base erosion and extracranial masses: CT and MR findings. J Comput Assist Tomogr 1994; 18: 939–942. 9. Steinmetz H, Heib E, Mironov A. Traumatic giant aneurysms of the intracranial corotid artery presenting long after head injury. Surg Neurol 1998; 30: 305–310. 10. Wellman BJ, Loftus CM, Noh D, Barnhart WH, Howard 3rd MA. A combined surgical-endovascular device concept for giant aneurysm neck occlusion. Neurosurgery 1998; 42: 1364–1367. 11. Gerber S, Dormont D, Sahel M, Grob R, Foncin JF, Marsault C. Complete spontaneous thrombosis of a giant intracranial aneurysm. Neuroradiology 1994; 36: 316–317. 12. Maruishi M, Shima K, Chigasaki H, Tsuchiya K. Giant intracranial aneurysm with rapid thrombosis formation and intramural hemorrhage – Case report. Neurol Med Chir (Tokyo) 1994; 34: 829–831. 13. Lan MY, Liu JS, Chang YY, Lin SH, Chen WH, Chen SS. Fibromuscular dysplasia with intracranial giant aneurysm. J Formosa Med Assoc 1995; 94: 692–694. 14. Sarwar M, Batnitzky S, Schechter MM. Tumorous aneurysms. Neuroradiology 1976; 12: 79–97.
ª 2003 Elsevier Science Ltd. All rights reserved.
Neuroradiology study
MR imaging of giant intracranial aneurysm Michael Mu-Huo Teng1 Cheng-Yen Chang1 MD
MD,
S.M. Nasir Qadri2
MD,
Chao-Bao Luo1
MD,
Jiing-Feng Lirng1
MD,
Shin-Su Chen1
MD,
1 Department of Radiology, Taipei Veterans General Hospital, 201, Sect 2, Shih-Pai Road, Peitou, Taipei 11217, Taiwan; School of Medicine, National Yang-Ming University, Taiwan, ROC, 2New X-ray Department, Services Hospital, Jail Road, Lahore, Pakistan
Summary We reviewed the clinical features and findings of magnetic resonance imaging (MRI) in 17 cases of giant intracranial aneurysm (GIA). All were confirmed by digital subtraction angiography (DSA). Seven were non-thrombosed and 10 were partially thrombosed. All thrombosed parts showed no enhancement. The majority of the lumen showed good enhancement in the post-contrast study (89%). About 80% of the partially thrombosed aneurysms had an onion skin appearance on non-contrast T1WI. On T2WI about 80% of partially thrombosed aneurysms had a low signal inside the thrombus or its wall from the susceptibility effect of blood elements such as haemosiderin. Flow void sign was noted in 88% on non-contrast T1WI and T2WI. Flow artifacts along the phase-encoding direction from the lumen of the aneurysm were noted in 41% of non-contrast T1WI, 56% of contrast-enhanced T1WI, and 29% of T2WI. Non-enhancement inside the thrombus and an onion skin appearance were important signs of thrombosis of the aneurysm. Flow artifacts along the phase-encoding direction from the lumen of the aneurysm provided an important clue of a patent lumen inside an aneurysm. The following signs provided clues to the diagnosis of patency lumen instead of thrombosis with high sensitivity and specificity: flow void sign on noncontrast T1WI (sensitivity 88%, specificity 100%), enhancement on contrast-enhanced T1WI (sensitivity 89%, specificity 100%), presence of either flow void sign or enhancement on contrastenhanced T1WI (sensitivity 100%, specificity 100%). ª 2003 Elsevier Science Ltd. All rights reserved. Keywords: digital subtraction angiography, aneurysm, giant aneurysm, partially thrombosed aneurysm, magnetic resonance imaging
INTRODUCTION Giant intracranial aneurysm (GIA) means an aneurysm greater than 2.5 cm at its greatest diameter.1–10 An aneurysm larger than 4.5 cm in diameter is called hypergiant aneurysm,1 a very uncommon entity. GIA constitute approximately 3–13% of all intracranial aneurysms.2 ,68 Most authors agree that GIAs usually present as slow-growing space occupying masses rather than with subarachnoid haemorrhage. GIA cause subarachnoid haemorrhage usually when rupture of the fundus occurs.4 We reviewed our cases of GIA, their clinical features and findings on MRI. METHODS Seventeen cases of partially thrombosed or non-thrombosed GIA were diagnosed by MRI examination and confirmed by digital subtraction angiography. Ten were female and 7 were male. We reviewed the clinical features and findings of MRI in these cases (Table 1). The size and location of these aneurysms, clinical manifestations, and complications of haemorrhage and infarct found in MRI are listed in Table 1. MRI examinations were performed using 1.5 T Sigma (GE Milwaukee, USA) or 1.5 T Magneton (Vision, Siemens, Germany). T2-weighted fast spin echo axial images, T1-weighted spin echo axial images and proton density weighted axial images were performed. Five cases had post-contrast T1-weighted images in coronal, axial and sagittal views after intravenous administration of gadolinium-DTPA. Lumen and thrombus portions of the an-
eurysm on MRI were identified in each case by correlation with the angiographic picture. We reviewed the signal changes and patterns of enhancement on MRI. Signal changes inside the aneurysm were compared with brain, and called 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. We also evaluated susceptibility artifact on T2WI, onion skin appearance on T1WI, flow void sign inside the lumen, and flow artifacts along the phase-encoding direction arising from the lumen of the aneurysm. Susceptibility artifact was considered when low signal changes inside the thrombus portion of the aneurysm was found on T2WI. This was usually from various blood degredation products such as haemosiderin. Onion skin appearance was used to describe the multilayered thrombus wall from a partial or completely thrombosed aneurysm.14 We considered flow void sign when low signal was found inside the lumen of the aneurysm. According to the presence or absence of flow void sign and high signal flow artifact, we formulated 6 classifications (Table 2). We called it type A if part of the lumen had flow void sign with no high signal flow artifact, type B if the whole lumen of the aneurysm had flow void sign with no high signal flow artifact, type C if there was presence of both regions of high signal flow artifact and low signal flow void sign inside the lumen, type D if part of the lumen had high signal flow artifact inside with no flow void sign, type E if the whole lumen was filled with high signal flow artifact completely with no flow void sign, and type F if neither low signal flow void sign nor high signal flow artifact was present. RESULTS
Received 14 August 2002 Accepted 19 September 2002 Correspondance to: Dr. Michael Mu Huo Teng, Department of Radiology, Veterans General Hospital-Taipei, 201, Sect 2, Shih-Pai Road, Peitou, Taipei 11217, Taiwan. Tel.: +886-2-28718251; Fax: +886-2-28757348; E-mail: [email protected]
460
Seven patients (41%) had new or old haemorrhage demonstrated by CT or MRI (Table 1). Intraparenchymal haemorrhage in adjacent brain parenchyma was found in 4 patients, intracerebral haemorrhage in remote brain parenchyma in 1 patient, subarachnoid haemorrhage in 2 patients, intraventricular haemorrhage in 2 patients, and haemorrhage along the wall of the aneurysm in 5
MR imaging of giant intracranial aneurysm
Table 1 #
Clinical data
Age
Sex
Size
Location
Clinical manifestations
Complications of haemorrhage and infarct No Old infarct in posterior cerebral territory. Recent haemorrhage in brain parenchyma (brainstem) and along the aneurysmal wall New infarct in internal capsule and lentiform nucleus Subarachnoid, intraventricular haemorrhage, and along the aneurysmal wall No Haemorrhage in brain parenchyma and along the aneurysmal wall Intraventricular haemorrhage No
1 2
49 68
M M
3.2 cm 3.3 cm
Cavernous ICA, right Basilar artery
Trigeminal neuralgia, right III nerve palsy Dizziness, vertigo, general weakness
3
78
M
2.8 cm
Supraclinoid ICA, left
4
59
F
4.8 cm
MCA, left M1
RT limb weakness; infarct in internal capsule (new) Dizziness
5 6
37 31
F F
3.5 cm 2.5 cm
Cavernous ICA, right M1 of left MCA
7 8
70 67
M F
3.2 cm 3.0 cm
Supraclinoid ICA, right Supraclinoid ICA, left
9 10 11 12
57 56 62 58
F F F M
3.0 3.0 3.0 4.0
cm cm cm cm
13 14 15 16 17
78 62 63 56 68
M F F F M
2.6 2.6 2.9 2.6 2.7
cm cm cm cm cm
Cavernous ICA, right Supraclinoid ICA, left Cavernous ICA, right Right anterior communicating artery Cavernous ICA, right Supraclinoid ICA, left Cavernous ICA, right Supraclinoid ICA, right Supraclinoid ICA, left
Diplopia. Right III nerve palsy Right facial weakness, headache, upper limb weakness General weakness, incoherent speech Chronic headache, nuchal rigidity, visual field defect Headache, blurred vision Headache, visual field defect Right VI nerve palsy Tiredness, haemiplegia, hydrocephalus
Recent infarct in middle cerebral artery territory No Old haemorrhage in lentiform nucleus Haemorrhage in brain parenchyma and along the aneurysmal wall No No Old brain contusion in bilateral low frontal No Intracerebral, subarachnoid haemorrhage and along the aneurysmal wall
Headache, blurred vision Headache, visual field defect Headache, blurred vision, facial numbness Blurred vision Headache
Table 2 Appearance of flow void and high signal flow artifacts inside lumen of the aneurysm Types
Flow void sign
High signal flow artifact
A B C D E F
In part of lumen In whole lumen In part of lumen Not present Not present Not present
Not present Not present In part of lumen In part of lumen In whole lumen Not present
Table 3
461
patients. Three cases (18%) had brain infarction in the distal branches of the parent artery of the aneurysm. Ten aneurysms were partially thrombosed. MRI findings of partially thrombosed aneurysms are listed in Table 3. Eight cases had an onion skin appearance (multilayered thrombus wall) indicating partial or completely thrombosed aneurysm. Low signal (signal intensity of 1 or 2) inside thrombus on T2WI as a result of susceptibility effect from various stages of blood elements was found in 8 patients. Five aneurysms with thrombus inside had
MRI findings of partially thrombosed aneurysms
Case number Signal change inside thrombus on non-contrast T1WI Onion skin appearance on T1WI Enhancement inside thrombus in the post contrast T1WI Rim enhancement on margin of thrombus on T1WI Signal change inside thrombus on T2WI Low signal change inside thrombus or in its wall on T2WI
1
2
6
7
8
9
10
11
12
15
4&6 ) ) + 1&2 +
4&6 + X X 1&2 +
4&6 ) X X 1&6 +
4&6 + ) Part 1&6 +
4 + X X 6 )
6 + X X 6 )
6 + X X 1&6 +
6 + ) + 1&6 +
3&6 + ) ) 2&6 +
6 + ) ) 1&4 +
Following cases had no thrombus inside: 3, 4, 5, 13, 14, 16, 17, thus are not included in this table. Signal change: 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. +, present; Part, partly present; ), not present; X, not available.
Table 4
MRI findings in the lumen of aneurysm and signs related to flow
Case number Signal change inside lumen on non-contrast T1WI Radiating artifact from lumen on non-contrast T1WI Well enhancement inside lumen in the post-contrast T1WI Radiating artifact outside the lumen on contrast-enhanced T1WI Signal change inside lumen on T2WI Radiating artifact outside the lumen on T2WI Flow-void sign&high signal flow artifact on noncontrast T1WI Flow-void sign and high signal flow artifact on T2WI
1 1&4 ) Part ) 1 ) A B
2
3
4
5
6
7
8
9
10
11
2&4 2&4 1&4 3&5 1&4 2&4 1&4 1 1&4 1 ) + + ) + + + ) ) ) X + X X X + X X X + X + X X X + X X X ) 2&4 1&6 2&4 3&6 1&6 1&6 1 1 1&6 1 ) + ) ) + + ) ) ) ) A A A F A A A B A B A C A D C C B B C B
12
13
14
2&4 3&4 1 + ) + + + ) + ) + 1&4 6 1 ) ) + A F B A E B
15
16
17
1&4 2&5 1&4 ) ) ) + + X ) + X 1&4 1&6 1&6 ) + ) A C C A C C
Signal change: 1, very hypointense; 2, hypointense; 3, slight hypointense; 4, isointense; 5, slight hyperintense; 6, hyperintense. +, definitely present; ), not present; X, not available. Part: part of the lumen had flow void sign, and part of the lumen had enhancement. A–E, see Table 2.
ª 2003 Elsevier Science Ltd. All rights reserved.
Journal of Clinical Neuroscience (2003) 10(4), 460–464
462 Teng et al.
postcontrast studies: all showed no enhancement inside the thrombus (Table 3). Complete rim enhancement at margin of thrombus (usually in the wall) was present in 2, and a partial rim was found in 1. All 14 cases had patent lumens inside the aneurysm. Seven cases (cases 3–5, 13, 14) had no evidence of thrombosis within the lumen. MRI findings from the aneurysm lumen and signs related to flow are listed in Table 4. All lumens of these aneurysms on non-contrast T1WI had signal intensity of 1–3 inside. On T1WI 15 cases had signal of 1–2 inside indicating flow void sign: 3 type B, 10 type A and 3 type C. In additional to this, 12 had signal intensity of 4, and 2 had signal of 5–6 suggesting high signal flow artifact inside. Two patients had neither low signal flow void nor high signal flow artifact (type F) (see Figs. 1 and 2). The patent lumens of the aneurysms on T2WI were of mixed signal in 11. All except 2 (cases 5 and 13) had low signal of 1–2 inside indicating flow void, 8 of them mixed with signal of 6 (hyperintense) and 4 cases mixed with signal of 4 (isointense) inside. In 5 cases, the lumen was filled with signal 1, compatible with type B flow void. In 6 cases there is signal of 1 and 6 inside lumen, compatible to type C. Two patients had no flow void, 1 had a small area of high signal flow artifact inside (type D), and
the other the whole lumen was filled with high signal flow artifact (type E). Nine patients had post-contrast T1WI, the whole lumen of their aneurysms showed good enhancement in 7 patients, part of the lumen had good enhancement and part had flow void sign in 1 patient, and the whole lumen had flow-void sign and showed no enhancement in 1 patient (Table 4). Flow artifacts along the phase-encoding direction from the lumen of the aneurysm was noted in 41% (7/17) on non-contrast T1WI, 29% (5/17) on T2WI, and 56% (5/9) on postcontrast T1WI. Ghost images of aneurysm were present in 1 case on both noncontrast T1WI and T2WI. DISCUSSION GIA has been divided into many types: giant and hypergiant, saccular, fusiform and serpentine, fully thrombosed, partially thrombosed and non-thrombosed, nontraumatic and traumatic. Ten cases were partially thrombosed aneurysms while 7 cases were non-thrombosed. According to Fisher and co-workers,8 nearly half of the GIA arise from the ICA. Most of our GIAs (76%) were from the ICA.
Fig. 1 Case 12. Partially thrombosed aneurysm. (A) Non-contrast T1WI. There is intraparenchymal haemorrhage in left frontal region (arrowheads). Flow void sign is present inside part of lumen of the aneurysm (arrow), compatible with type A. (B) Post-contrast T1WI. The lumen of the aneurysm showed homogeneous wall enhancement (arrowheads). Flow void sign is not seen in this postcontrast study. (C) T2WI: flow void sign is present inside part of the lumen (arrow), compatible with type A. (D) AP view of right carotid angiogram shows the lumen of the aneurysm (arrowheads) arises from the junction of the right anterior cerebral artery and right anterior communicating artery.
Journal of Clinical Neuroscience (2003) 10(4), 460–464
ª 2003 Elsevier Science Ltd. All rights reserved.
MR imaging of giant intracranial aneurysm
463
Fig. 2 Case 14. Non-thrombosed aneurysm (A) Sagittal view postcontrast T1WI: There is flow-void sign inside the whole lumen of the aneurysm (arrowheads), compatible with type B. Enhancement is seen only in the wall of the aneurysm. Radiating flow artifacts (small arrows) originating from the aneurysm is present along the phase-encoding direction in the anterior and posterior aspect of the aneurysm. (B) T2WI: Flow-void sign is present in the whole lumen (arrow) of the aneurysm. Its area is the same as its lumen, compatible with type B. (C) Anteroposterior view and (D) lateral view of left carotid angiogram: an aneurysm (arrow) can be seen in the left internal carotid artery.
Many factors have been considered responsible for the formation and growth of an aneurysm to a GIA. One is neovascularisation and recurrent haemorrhage within the wall of the aneurysm.3;4;9;12 Because these recurrent haemorrhages are confined to the aneurysm wall and do not reach the subarachnoid space, the aneurysm is not detected until it becomes a GIA. Seven patients had new or old haemorrhage. A relative high percentage of them (71%) had haemorrhage along the aneurysmal wall and a low percentage (29%) had subarachnoid haemorrhage compatible with this theory. In addition to the congenital defect in the tunica media, other factors are also responsible, such as atherosclerosis. Some genotypic connective tissue diseases such as Marfans syndrome, Ehlers–Danlos syndrome, elastic pseudoxanthoma and fibromuscular dysplasia are also associated with GIA.1;13 Aneurysms may even develop after a cranial injury, infection, or a surgical operation. Four cases had a previous history of head injury. The clinical symptoms vary according to the location of the aneurysm.1 Most of our cases (88%) had symptoms caused by mass effect, such as cranial nerve palsy, visual field defect, and limb weakness. Eight cases (47%) had headache. Seven cases (41%) had haemorrhage demonstrated by CT or MRI (Table 1). Our findings are compatible with previous reports that GIAs usually present either as a slow growing space occupying intracranial masses or with intracranial haemorrhage.1–13 Distal cerebral infarction in the territory of the parent artery is a commonly ignored complication of GIAs. Three cases (18%) had brain infarction of different chronicity in the distal branches ª 2003 Elsevier Science Ltd. All rights reserved.
of the parent artery harbouring an aneurysm. The cause of distal cerebral infarction in the territory of the artery with GIA could be thrombus arising from stagnant blood flow inside the GIA or compression of the parent artery by the GIA. The two basic constituents of GIA are laminated blood in various stages of thrombosis, and flowing blood. Rapidly flowing blood in the aneurysmal lumen appears as an area of signal void on MRI.2 “The flow void sign” had a high sensitivity on non-contrast T1WI (88%) and T2WI (88%). On T1WI, only part of the lumen had flow void sign in most patients (12 cases, 71%), and the whole lumen had flow void sign in 3 cases (18%). On T2WI, part of the lumen had flow void sign in most patients 59% (10/17), and the whole lumen in 29% (5/17). The absence of flow void sign did not exclude the diagnosis of aneurysm with a patent lumen. It was not present in 2 cases on non-contrast T1WI and in 2 cases on T2WI. The persistence of flow void in the post-contrast MRI is a specific sign of aneurysm.8 From our study, this sign in the post-contrast T1WI had a low sensitivity (22%) and high specificity (100%). Flow void sign may also be confused with the low signal inside thrombus from blood elements on T2WI, thus having a relative low specificity (20%) on T2WI. On post-contrast T1WI, 8 out of 9 patients had good enhancement in the lumen of their aneurysms, 1 in the whole lumen, and 7 in a part of the lumen. Combining flow void sign and internal enhancement to indicate lumen patency instead of thrombosis had high sensitivity (100%), high specificity (100%), positive prediction sign (100%), and negative prediction (100%) (Table 5). Journal of Clinical Neuroscience (2003) 10(4), 460–464
464 Teng et al.
Table 5
Different signs of aneurysm
Signs
Sensitivity
Specificity
Positive predictive Negative predictive value value
Flow void sign on noncontrast T1WI without contrast as an indication of aneurysmal lumen (low signal inside thrombus considered false positive) Flow void sign as an indication of aneurysmal lumen on post-contrast T1WI (low signal inside thrombus considered false positive) Enhancement as an indication of aneurysmal lumen on post-contrast T1WI (enhancement inside thrombus considered false positive) Presence of either flow void sign or enhancement as a combined indication of aneurysmal lumen on post-contrast T1WI Flow void sign on T2WI without contrast as an indication of aneurysmal lumen (low signal inside thrombus considered false positive) High signal artifact on noncontrast T1WI without contrast as an indication of aneurysmal lumen (high signal inside thrombus considered false positive) High signal artifact on T2WI without contrast as an indication of aneurysmal lumen (high signal inside thrombus considered false positive) Low signal change as an indication of thrombed portion of an aneurysm on T2WI (low signal change inside the lumen on T2WI considered false positive)
15/17 ¼ 88%
10/10 ¼ 100%
15/15 ¼ 100%
10/12 ¼ 83%
2/9 ¼ 22%
5/5 ¼ 100%
2/2 ¼ 100%
5/12 ¼ 42%
8/9 ¼ 89%
5/5 ¼ 100%
8/8 ¼ 100%
5/6 ¼ 83%
9/9 ¼ 100%
5/5 ¼ 100%
9/9 ¼ 100%
5/5 ¼ 100%
15/17 ¼ 88%
2/10 ¼ 20%
15/23 ¼ 65%
2/4 ¼ 50%
2/17 ¼ 12%
1/10 ¼ 10%
2/11 ¼ 18%
1/16 ¼ 6%
8/17 ¼ 47%
3/10 ¼ 30%
8/15 ¼ 53%
3/12 ¼ 25%
8/10 ¼ 80%
2/15 ¼ 13%
8/23 ¼ 35%
2/4 ¼ 50%
Sensitivity of more signs: (A) Onion skin appearance on noncontrast T1WI: 8/10 ¼ 80%. (B) Nonenhancement inside thrombus of an aneurysm 5/5 ¼ 100%. (C) Rim enhancement on margin of thrombus in contrast-enhanced T1WI 3/5 ¼ 60%. (D) Radiating flow artifact from the aneurysmal lumen: noncontrast. T1WI, 41%; T2WI, 29%; contrast-enhanced T1WI 56%.
Flow artifacts and ghost images extending to either side of the lumen of the aneurysm along the phase-encoding direction are also motion artifacts. Presence of a wide band of flow artifact or ghost images indicates presence of an arterial lumen, thus the diagnosis of GIA. Although the sensitivity was low, they provided definite clues to the diagnosis of the patent lumen of an aneurysm. According to Sarwar et al.,14 partially thrombosed giant intracranial aneurysms display an onion skin appearance of the thrombosed portion. Each layer represents a clot of different chronicity. Onion skin appearance was seen in 80% of our cases with partial thrombosis on non-contrast T1WI. Therefore this appearance is very helpful in the diagnosis of GIA with partial thrombosis. In the post-contrast T1WI of all patients, no enhancement could be found inside the thrombosed portion of the aneurysm (sensitivity 100%). In 3 out of 5 cases with partial thrombosis, the aneurysmal wall enhanced on contrasted T1WI. Presence of low signal intensity on T2WI in the region of the thrombosed portion as a result of susceptibility effect from various stages of haemoglobin degredation was found in 80% of our cases. Because of the flow void sign inside the lumen of an aneurysm, signal intensity was also low on T2WI. Presence of low signal intensity on T2WI indicating the thrombosed portion of the aneurysm had a low specificity (13%) (Table 5). In conclusion, most of our patients (88%) had symptoms caused by mass effect. About 41% of our patients had haemorrhage. 18% of our patients with aneurysms had brain infarction of different chronicity in the distal branches of the parent artery. All thrombosed parts of an aneurysm showed no enhancement. The lumen of our aneurysms had good enhancement or the flow void sign in the postcontrast T1WI study. Onion skin appearance was noted in 80% of our partially thrombosed aneurysms on non-contrast T1WI. The following signs provided the clue to the diagnosis of lumen patency as opposed to thrombosis with high sensitivity and specificity: flow void sign on noncontrast T1WI (sensitivity 88%, specificity 100%), enhancement on post-contrast T1WI (sensitivity 89%, specificity 100%), and either flow void sign or enhancement inside on postcontrast T1WI (sensitivity 100%, specificity 100%). Radiating flow artifact from the lumen of the aneurysm provided a definite clue to the diagnosis of lumen patency (41% of non-contrast T1WI and 56% of contrast-enhanced T1WI). Journal of Clinical Neuroscience (2003) 10(4), 460–464
ACKNOWLEDGEMENTS The study was partially supported by National Science Council of Republic of China grant number NSC 89-2320-B-075-001-M08. Dr Qadri was neuroradiology fellow at Taipei Veterans General Hospital, 1999–2000. REFERENCES 1. Castel JC, Chaabane M, Guibert-Truiner F, Piton J, Caille JM. Intracranial supratentorial “Hypergiant” aneurysms – diagnostic problems. J Neuroradiol 1985; 12: 135–149. 2. Atlas SW, Grossman RI, Goldberg HI, Hackney DB, Bilaniuk LT, Zimmerman RA. Partially thrombosed giant intracranial aneurysm: correlation of MR and pathological findings. Radiology 1987; 162: 111–114. 3. Whittle IR, Dorsch NW, Besser M. Spontaneous thrombosis in giant intracanial aneurysm. J Neurol 1982; 45: 1040–1047. 4. Schubiger O, Valavanis A, Wichman W. Growth mechanism of giant intracranial aneurysms; demonstrated by CT and MR imaging. Neuroradiology 1987; 29: 266–271. 5. Fodstad H, Liliequist B, Staffan W, Nilsson PE, Boquist L, Abdul-Rahman A. Giant serpentine intracranial aneurysm after carotid ligation. J Neurosurg 1978; 49: 903–909. 6. Belec L, Cesaro P, Brugieres P, Gray F. Tumor simulating giant serpentine aneurysm of the posterior cerebral artery. Surg Neurol 1988; 29: 210–215. 7. Schubiger O, Valavnis A, Hayek J. Computed tomography in cerebral aneurysms with special emphasis on giant intracranial aneurysms. J Comput Assist Tomogr 1980; 4: 24–31. 8. Fisher A, Som PM, Mosesson RE, Lidov M, Liu TH. Giant intracranial aneurysm with skull base erosion and extracranial masses: CT and MR findings. J Comput Assist Tomogr 1994; 18: 939–942. 9. Steinmetz H, Heib E, Mironov A. Traumatic giant aneurysms of the intracranial corotid artery presenting long after head injury. Surg Neurol 1998; 30: 305–310. 10. Wellman BJ, Loftus CM, Noh D, Barnhart WH, Howard 3rd MA. A combined surgical-endovascular device concept for giant aneurysm neck occlusion. Neurosurgery 1998; 42: 1364–1367. 11. Gerber S, Dormont D, Sahel M, Grob R, Foncin JF, Marsault C. Complete spontaneous thrombosis of a giant intracranial aneurysm. Neuroradiology 1994; 36: 316–317. 12. Maruishi M, Shima K, Chigasaki H, Tsuchiya K. Giant intracranial aneurysm with rapid thrombosis formation and intramural hemorrhage – Case report. Neurol Med Chir (Tokyo) 1994; 34: 829–831. 13. Lan MY, Liu JS, Chang YY, Lin SH, Chen WH, Chen SS. Fibromuscular dysplasia with intracranial giant aneurysm. J Formosa Med Assoc 1995; 94: 692–694. 14. Sarwar M, Batnitzky S, Schechter MM. Tumorous aneurysms. Neuroradiology 1976; 12: 79–97.
ª 2003 Elsevier Science Ltd. All rights reserved.