The role of MR angiography in the pretreatment assessment of intracranial aneurysms: a comparative study

Abstract

Background and purpose: With developments in coil technology, intracranial aneurysms are being treated increasingly by the endovascular route. Endovascular treatment of aneurysms requires an accurate depiction of the aneurysm neck and its relation to parent and branch vessels preoperatively. Our goal was to estimate the clinical efficacy of MR angiography (MRA) in the pretreatment assessment of ruptured and unruptured intracranial aneurysms. We compared MRA source data (axial acquired partitions), multiplanar reconstruction (MPR) of these data, as well as maximum intensity projection (MIP) and 3D-isosurface images with intraarterial digital subtraction angiography (IA-DSA).

Methods: The study was performed in 29 patients with 42 intracerebral aneurysms. The MRA data were examined in four different forms--as axial source data, MPR images of the source data, and MIP and 3D isosurface--rendered images. A composite standard of reference for each aneurysm was then constructed using this information together with the IA-DSA findings by looking at aneurysm detection rate, aneurysm morphology, neck interpretation, and branch vessel relationship to the aneurysm. All techniques, including conventional IA-DSA, were then scored independently on a five-point scale from 1 (non diagnostic) to 5 (excellent correlation with the standard of reference) for each of the aneurysm components as compared with the composite picture. An overall score for each technique was also obtained.

Results: Of the 42 aneurysms examined, 34 were small (<10 mm), six were large (10-25 mm), and two were giant (>25 mm). Three aneurysms were not detected with MRA. These were smaller than 3 mm and either in an anatomically difficult location (middle cerebral artery bifurcation) or obscured by adjacent hematoma. Two large aneurysms were depicted as undersized by IA-DSA owing to the presence of intramural thrombus shown by MRA axial source data. IA-DSA received the highest scores overall and in three of the four subgroups. Three-dimensional isosurface reconstructions scored higher than did IA-DSA for depiction of the aneurysm neck, although this difference was not significant. The MPR and 3D-isosurface images were comparable to those of IA-DSA in all categories. MPR images were particularly useful for defining branch vessels and the aneurysm neck. MIP images scored poorly in all subgroups (P < .005) compared with IA-DSA findings, except for in aneurysm detection. Source data images were significantly inferior to those of IA-DSA in all categories (P < .005).

Conclusion: MRA is currently inferior to IA-DSA in pretreatment assessment of intracranial aneurysms, and can miss small lesions (<3 mm). It can, however, provide complementary information to IA-DSA, particularly in anatomically complex areas or in the presence of intramural thrombus. If MRA is used in aneurysm assessment, a meticulous technique with reference to both axial source data and MPR is mandatory. The axial source data should not be interpreted in isolation. Three-dimensional isosurface images are comparable to those of IA-DSA and are more reliable than are MIP images, which should be interpreted with caution.

fig 1.

Overall average score for each imaging technique. Note.—SD = source data; MPR = multiplanar reformat; IA DSA = intraarterial digital subtraction angiography; MIP = maximum intensity projection; 3D = three-dimensional isosurface reconstruction.

fig 2.

A 55-year-old woman presented with intractable pain and a complete right CN III palsy caused by pressure effect from a giant (25 mm), fusiform, intracavernous internal carotid artery aneurysm. The circle of Willis should be scrutinized as part of the assessment prior to endovascular balloon occlusion of the parent vessel. A, TOF-MIP MRA (100/20/1) from above and slightly from the right. The right posterior communicating artery (arrow) cannot be differentiated from aneurysm by MIP-MRA findings. The ipsilateral anterior cerebral artery is also poorly demonstrated. B, TOF-3D-isosurface MRA (100/20/1), similar projection to A. The posterior communicating artery (arrow) can now be clearly separated from aneurysm. The ipsilateral anterior communicating artery is of increased caliber compared with the MIP-MRA projection. C, TOF-MPR MRA (100/20/1), parasagittal view. The relationship of the posterior communicating artery (arrow) to the aneurysm can be clearly seen along its whole length on this projection. The artery is in contact with aneurysm at its apex. D, TOF-MPR MRA (100/20/1), coronal view. By scrolling along the posterior communicating artery (arrow) in the coronal plane, its relationship to the aneurysm is clearly depicted.

fig 3.

A 48-year-old woman presented with an SAH secondary to an anterior communicating artery aneurysm rupture. Diagnostic IA-DSA showed multiple aneurysms. A, TOF-3D-isosurface MRA (100/20/1), right periorbital view. There is some minor artifact due to patient movement. The anterior communicating artery aneurysm (short broad arrow) appears to arise from the right A1/A2 junction. A terminal left internal carotid artery aneurysm is also visible (long thin arrow). B, TOF-MPR MRA (100/20/1), coronal slice through the anterior communicating artery complex. Scrolling through, no communication is seen between the left A1 segment (thin short arrow) and aneurysm (short broad arrow). The aneurysm is continuous with the right A2 segment (long arrow). C, IA-DSA, left periorbital projection. The anterior communicating and left terminal internal carotid artery aneurysms are clearly seen (arrows). There are also two discrete middle cerebral artery aneurysms. The anterior communicating artery aneurysm only filled from a left-sided injection. If one were to have planned endovascular treatment on the basis of the MRA findings, the approach would have been via the right internal carotid artery.

fig 4.

A 62-year-old woman presenting with a left CN III palsy due to posterior communicating artery aneurysm. A, IA-DSA of left common carotid artery, lateral projection. A large left, saccular, posterior communicating artery aneurysm with a discrete narrow neck is seen well. B, TOF-MIP MRA (100/20/1), similar projection to that of A. Streamlining artifact is seen within the internal carotid artery, and there is signal loss within the aneurysm itself. C, TOF-3D-isosurface MRA (100/20/1), similar projection to that of A and B. The posterior communicating artery aneurysm (arrow) appears identical to its depiction by IA-DSA. D, Axial-source TOF image (100/20/1). The true lumen of the posterior communicating artery aneurysm is depicted as high signal, but this is surrounded by concentric mixed signal rims representing intramural thrombus (arrow). The presence of thrombus is not appreciated on the IA-DSA image. The aneurysm is significantly larger than the IA-DSA and postprocessed MRA images suggest.

fig 5.

A 45-year-old woman presented with SAH due to rupture of an anterior communicating artery aneurysm. A, Diagnostic IA-DSA of left common cartotid artery, frontal projection. The anterior communicating artery aneurysm is seen, but its relationship with the left A2 segment is not clear. A submental vertical projection was not obtained. B, TOF-MIP MRA (100/20/1), submental vertical projection. It can now be appreciated that there is a bifid anterior communicating artery (long arrow). The aneurysm (short arrow) arises from the junction between the left A2 and the upper limb of the anterior communicating artery. On the MIP-MRA image, the aneurysm appears tubular and of narrower calibre than is suggested by the IA-DSA image. C, TOF-3D-isosurface MRA (100/20/1), similar projection to that of B. The lower limb of the anterior communicating artery is again clearly visible (long arrow). The morphologic characteristics of the aneurysm (short arrow) depicted by this technique are more similar to those shown by IA-DSA. D, IA-DSA of the anterior communicating artery aneurysm post GDC embolization. Prior manipulation of the MRA has allowed the appropriate projection to be chosen. Only the distal part of the aneurysm would retain a coil, leaving a significant remnant.