CT angiography and its role inthe investigation of intracranial

haemorrhageRAD Magazine, 39, 458, 29-30

Dr M IgraRadiology SPR

Leeds General Infirmary

Dr I DjoukhadarResearch fellow

Wolfson Molecular Imaging CentreUniversity of Manchester

Dr T GoddardConsultant neuroradiologist

Leeds General Infirmary

IntroductionSpontaneous intracranial haemorrhage consti-tutes only 15%1 of acute stroke but remains themost devastating form, with a death and severedisability rate of more than 75%.2,3 Prompt iden-tification of a structural vascular abnormality, orlack of it, is a major factor in improving the clin-ical management of these patients and ultimatelytheir outcomes.

Digital subtraction angiography (DSA) hasbeen the gold standard imaging technique allow-ing for a detailed assessment of the circle ofWillis. However, it is an invasive examinationcarrying significant risks (1% stroke rate) to thepatient,4 is time consuming and requires a cer-tain level of experience limiting its availability.The wide availability of CT angiography, on theother hand, coupled with the introduction ofmulti-detector row CT angiography (MDCTA) andthe improved post processing techniques, hasresulted in an increased use of CTA as an imag-ing tool for the assessment of the intracranialvascular structure.CT angiography techniques and roleCT angiography is proving to be a suitable alternativemethod of assessing the cerebral vessels. This is due to itsnon-invasive nature, lower radiation dose,5 wide availabil-ity and the ability to perform the examination in conjunctionwith non-contrast CT. The introduction of multi-detector rowCT angiography has revolutionised the CTA technique withits quicker acquisition times and improved post-processingimage quality.

Following image acquisition, the volumetric data is trans-ferred to a working station for further processing. In ourinstitution this includes multi-planar reformation (MPR)with 1mm thickness in multiple planes, thin slab maximumintensity projections (MIP) and volume rendered technique(VRT) algorithm. A systematic and thorough approach isadopted in reviewing all the images.

The main role of CTA is to identify the culprit vascular

lesion and assess treatment options. With respect toaneurysms, this is achieved by examining the following fea-tures: size, shape, location in relation to the haemorrhage,calcification, number of vessels involved and the presenceof any other lesions or anatomical variants.

Subarachnoid haemorrhageIn our institution, once an unenhanced CT has demonstratedsubarachnoid haemorrhage (SAH), a CTA is usually acquiredin the same setting, saving the patient and the medical teaman extra visit to the radiology department. Intracranialaneurysms are responsible for most cases of SAH, approach-ing 90%.6 In addition to identifying the aneurysm, the radio-logist should look for certain features including location ofthe aneurysm in relation to the blood (figure 1), neck size,number of vessels incorporated within the aneurysm andthe presence of vasospasm. Occasionally, in the presence ofmultiple aneurysms, it might not be possible to identify theculprit lesion and the neuro-interventionist will need to treatmore than one lesion (figure 2).

Many authors have published data indicating the highsensitivity and specificity of CTA in detecting acutely rup-tured intracranial aneurysms. Using MDCTA, Agid et al7

quote a 98% sensitivity and 100% specificity for the detectionof aneurysm. Similar results are also published by Sideman,Lourenco and Byyny.8-10 A meta-analysis by Westerlaan etal11 showed a specificity of 99% for ruptured aneurysms anda sensitivity of 92%. Despite these impressive rates, CTAlags slightly behind conventional DSA and can occasionallyyield false negatives, especially in cases of tiny aneurysmsmeasuring less than 2mm12 and ones that are close to bonystructures.13

In late SAH presentation, the accuracy of unenhancedCT and lumbar puncture sensitivity and accuracy drop sig-nificantly. CTA is a very powerful tool in assessment of thissubgroup of patients. A negative unenhanced CT and CTAcan exclude aneurysmal SAH with a post-test probabilityof 99%.14

As interpretation is user dependent, technologicaladvances coupled with increased familiarity with the inves-tigation should yield greater detection rates. A systematicapproach to assessing the arterial tree, its branches and

Table 1CT angiography to aneurysm therapy chart.

MIPs in three orthogonal planes with different post-processing techniques should be implemented to ensureaneurysm detection.

Arteriovenous vascular malformationAVM are another well-recognised cause of SAH and intra-cerebral haemorrhage. The main modality for imaging theseabnormalities is DSA as it allows for dynamic assessment.However, when an AVM is encountered on CTA, certainradiological features should be actively assessed to try andstratify the risk of recurrent haemorrhage. Venous pouchesor intra-nidal aneurysms (figure 3) confer a higher risk forAVM rupture and can cause a diagnostic dilemma due totheir wide range and locations; hence identification of theselesions is essential.

Sanelli et al12,15 demonstrated that, while DSA remainsthe gold standard, CTA has an important role to play in theinitial diagnostic vascular assessment. CTA can assess botharterial supply of the nidus as well as its venous drainage,and is also useful in stereotactic localisation. A recent studyshowed that CTA was more sensitive (87%) than MRI andMRA (83% and 87% respectively) at identifying rupturedAVMs, and that sensitivity was 100% for AVMs larger than3cm.16 They also showed that associated aneurysms werebest detected with CTA, compared to these other imagingmodalities.

Although conventional cerebral angiography remains themodality of choice for the diagnosis of dural arteriovenousfistula, CTA can be useful with certain imaging featuressuggestive of the diagnosis. For example, tortuous feedingarteries with dilated cortical draining veins and dilatedexternal carotid artery branches are recognised imaging fea-tures (figure 4).

Lobar haemorrhageMultiple structural abnormalities can cause isolated intra-cerebral lobar haemorrhage. These include aneurysms, AVMas well as dural fistulae. The decision to investigate patientswith an intracerebral haematoma using CTA varies widelyfrom one institution to another. In our unit, a multi-disciplinary approach is followed, with each case assessedon an individual basis. Factors affecting the decision toinvestigate further include: patient age, haemorrhage loca-tion and the presence of risk factors such as hypertension.17,2

For example, in an over 60-year-old hypertensive patientwith a basal ganglionic haemorrhage, the yield of furtherCTA will be limited.

In 2009, Yoon et al18 compared the accuracy of MDCTAagainst conventional DSA in 78 patients. CTA successfullydetected the underlying vascular abnormalities in all butone patient (a 4mm AVM) with no false positive cases. Theauthors concluded that MDCTA is a promising investiga-tion in the assessment of lobar ICH but does not replaceDSA in all patients.

Primary intracranial haemorrhage refers to a spontaneoushaemorrhage where an underlying structural abnormalityis not identified. In addition to the obvious exclusionvalue of CTA, there is a potentially important prognosticvalue too.

Over the last two decades there have been an increasingnumber of publications relating to the spot sign in CTA. Theconcept has evolved but most authors refer to the spot signas tiny areas of enhancement in the haematoma on CTAsource data.19 Almandoz et al20 proposed a scoring systemreflecting certain imaging parameters including the num-ber, maximum axial dimension and Hounsfield unit attenu-ation values of the spot sign. It is important to rememberthat the spot sign is mainly used in primary intracranialhaemorrhage cases and its application in secondary haemorrhage is less clear.

Recently, Brouwers et al reviewed the available litera-ture highlighting the consistent ability of the spot sign inpredicting haematoma expansion, functional outcome andmortality.21

ConclusionThe increasing use of CTA can be attributed to its non-invasive nature, wide availability and lower dose whencompared with DSA. In addition, when reviewed by an expe-rienced operator (figure 5), CTA is a highly sensitive andspecific investigation and should be use as a gatekeeper lim-iting the need for DSA.

References1. Qureshi A I et al. Spontaneous intracerebral hemorrhage. N Engl J Med

2001;344(19):1450-60.2. Zhu X L, Chan M S, Poon W S. Spontaneous intracranial hemorrhage:

Which patients need diagnostic cerebral angiography? A prospective studyof 206 cases and review of the literature. Stroke 1997;28(7):1406-9.

3. Willinsky R A et al. Neurologic complications of cerebral angiography:Prospective analysis of 2,899 procedures and review of the literature.Radiology 2003;227(2):522-8.

4. Hankey G J, Warlow C P, Molyneux A J. Complications of cerebral angiog-raphy for patients with mild carotid territory ischaemia being consideredfor carotid endarterectomy. J Neurol Neurosurg Psychiatry 1990;(7)53:542-8.

5. Manninen A L et al. A comparison of radiation exposure between diag-nostic CTA and DSA examinations of cerebral and cervicocerebral vessels.Am J Neuroradiol 2012;33(11):2038-42.

6. Kirkpatrick P J. Subarachnoid haemorrhage and intracranial aneurysms:what neurologists need to know. J Neurol Neurosurg Psychiatry 2002;73(Suppl1):i28-33.

7. Agid R et al. Acute subarachnoid hemorrhage: Using 64-slice multidetec-tor CT angiography to “triage” patients’ treatment. Neuroradiology 2006;48(11):787-94.

8. Sidman R, Connolly E, Lemke T. Subarachnoid hemorrhage diagnosis:Lumbar puncture is still needed when the computed tomography scan isnormal. Acad Emerg Med 1996;3(9):827-31.

9. Lourenco A P et al. Does 16-detector computed tomography improve detec-tion of non-traumatic subarachnoid hemorrhage in the emergency depart-ment? J Emerg Med 2009;36(2):171-5.

10. Byyny R L et al. Sensitivity of noncontrast cranial computed tomographyfor the emergency department diagnosis of subarachnoid hemorrhage. AnnEmerg Med 2008;51(6):697-703.

11. Westerlaan H E et al. Intracranial aneurysms in patients with subarach-noid hemorrhage: CT angiography as a primary examination tool for diag-nosis – systematic review and meta-analysis. Radiology 2011;258(1):134-45.

12. Sanelli P C et al. CT angiography in the evaluation of cerebrovasculardiseases. Am J Roentgenol 2005;184(1):305-12.

13. Seruga T, Bunc G, Klein G E. Helical high-resolution volume-rendered 3-dimensional computer tomography angiography in the detection of intracra-nial aneurysms. J Neuroimaging 2001;11(3):280-6.

14. McCormack R F, Hutson A. Can computed tomography angiography ofthe brain replace lumbar puncture in the evaluation of acute-onsetheadache after a negative noncontrast cranial computed tomography scan?Acad Emerg Med 2010;17(4):444-51.

15. Sanelli P C, Mifsud M J, Stieg P E. Role of CT angiography in guidingmanagement decisions of newly diagnosed and residual arteriovenous mal-formations. Am J Roentgenol 2004;183(4):1123-6.

16. Gross B A, Frerichs K U, Du R. Sensitivity of CT angiography, T2-weighted MRI, and magnetic resonance angiography in detecting cerebralarteriovenous malformations and associated aneurysms. J Clin Neurosci2012; 19(8):1093-5.

17. Halpin S F et al. Prospective evaluation of cerebral angiography and com-puted tomography in cerebral haematoma. J Neurol Neurosurg Psychiatry1994;57(10):1180-6.

18. Yoon D Y et al. Multidetector row CT angiography in spontaneous lobarintracerebral hemorrhage; a prospective comparison with conventionalangiography. Am J Neuroradiol 2009;30(5):962-7.

19. Tanaka H et al. Initial experience with helical CT and 3D reconstructionin therapeutic planning of cerebral AVMs; comparison with 3D time-of-flight MRA and digital subtraction angiography. J Comput Assist Tomogr1997;21(5):811-7.

20. Romero J M et al. Accuracy of CT angiography for the diagnosis of vas-cular abnormalities causing intraparenchymal hemorrhage in youngpatients. Emerg Radiol 2009;16(3):195-201.

21. Brouwers H B et al. Clinical applications of the computed tomographyangiography spot sign in acute intracerebral hemorrhage: A review. Stroke2012;43(12):3427-32.

Figures 1A and 1B(A) Unenhanced CT brain showing predominantly posterior fossa haemorrhage suggesting a posterior cir-culation aneurysm. (B) CTA confirms a tiny left posterior inferior cerebellar artery aneurysm (2mm).

Figures 2A and 2BAxial (A) and coronal (B) CT angiography images demonstrating multiple vascular abnormalities.1. AVM in the left temporal region. 2. Left middle cerebral artery aneurysm, this is the culprit abnormalitygiven the surrounding hematoma (red arrow). 3. Small and incidental right middle cerebral artery aneurysm.

A B

A BHaematomaHaematoma

Figures 3A, 3B, 3C and 3D(A) Unenhanced CT brain demonstrating large intraventricular haemorrhage and a small adjacent leftcorona radiata haemorrhage. (B) CTA reveals a large left frontal AVM. (C) CTA also show a tiny aneurysmintimately related to the parenchymal haemorrhage. (D) 3D reconstructed images of a conventionalangiogram showing the aneurysmal lesion within the AVM.

A B C D

Figures 4A and 4B(A) CTA showing a right cerebellar tortuous vessels and, interestingly, the right temporal artery isenlarged raising concerns for a dural fistula. (B) Conventional cerebral angiogram confirming a dural arte-riovenous fistula.

A B

Figures 5A-DCT and initially CTA was interpreted as negative.

Figure 5EConventional cerebralangiogram showinganeurysm in the rightdistal internal carotidartery.

Figure 5FOn retrospective reviewof CTA, the lesion waspresent but difficult tosee.

A

B

C

E F

D