Pre- and postembolization evaluation of a left parietal cortical-ventricular AVM

This 31 year old patient presented with his first generalized epileptic seizure.
The initial CT examination revealed a left parietal lesion, presumably an AVM. Subsequent MRI-MRA confirmed the diagnosis. Conventional catheter based cerebral angiography was also performed before therapeutic decision making.
Endovascular treatment was carried out, the first session resulting in a 60% reduction of the nidus. Further sessions are scheduled.

Left parietal AVM
Examination 1a (0.5 T)
Fig.1 Sagittal T1-weighted spin-echo images, demonstrating the typical MR appearance of the AVM in the left parietal region.
Fig.2 Transverse proton density-weighted fast spin-echo images, delineating the nidus of the AVM.
Fig.3 Transverse T2-weighted fast spin-echo images. Abnormal high signal intensity areas are intermingled with the vascular structures, suggesting associated parenchymal lesions.
Fig.4 Transverse turbo FLAIR images. Most of the previously high signal intensity areas appear signal void on these images, corresponding therefore to CSF spaces. However, other, abnormal signal intensity areas are clearly identified in proximity to the nidus (arrows), which were less apparent on the T2-weighted images and presumably represent parenchymal lesions.
Fig.5 Sagittal survey single-slice 2D PC MR angiogram (Venc: 45 cm/s, Tac: 36 sec). This rapid acquisition provides information about the location, geometry and extension of the AVM which is helpful in designing the subsequent 3D PC acquisition.
Fig.6 Sagittal averaged modulus (left) and corresponding magnitude of complex differences (right) type source images from a non-enhanced 3D PC MRA acquisition. The triangular shaped nidus extending from the cortex to the lateral ventricle is well demonstrated.
Fig.7 Sagittal collapsed MIP reconstruction from the non-enhanced 3D PC MRA acquisition data set (matrix: 256, Venc: 45 cm/s). Despite the probably inappropriate Venc value selection, all the components of the lesion (arterial feeders from the middle cerebral artery and draining veins to the superior sagittal sinus and left transverse sinus) are well depicted.
Fig.8 Collapsed (above left) and targeted MIP reconstructions from the non-enhanced 3D PC MRA acquisition data set. Limiting the reconstruction subvolume to the midline structures, the absence of arterial supply from the anterior cerebral artery is well demonstrated (above right). Targeted MIP reconstructions from more lateral volumes of interest (below) show the venous drainage of the lesion (blue arrows) to better advantage.

Examination 1b (1.5 T)
Fig.9 Sagittal 2D PC MR angiogram (matrix: 512, Venc: 65 cm/s). The arterial feeders of the AVM from the middle cerebral artery are not well seen (inappropriate Venc value selection?, blurring due to motion?, signal loss due to turbulent flow?).
Fig.10 Sagittal Gadolinium-enhanced 2D PC MR angiogram (matrix: 512, Venc: 65 cm/s). Although a frank increase in intravascular signal intensity and vessel conspicuity is observed in most of the intracranial vessels, the branches of the left middle cerebral artery remain suboptimally visualized.
Fig.11 Sagittal collapsed MIP reconstruction from a non-enhanced 3D PC MRA acquisition data set (matrix: 256, Venc: 45 cm/s).
Fig.12 Sagittal collapsed MIP reconstructions from the non-enhanced 3D PC MRA acquisition data sets at 0.5 (left) and 1.5 T (right), for comparison. The intravascular signal intensity and consequently the conspicuity of small vessels is enhanced at higher field strength, however no significant difference in the overall diagnostic value of the two images is seen. This illustrates that, especially with the Phase Contrast technique, clinically valuable MR angiograms can be obtained on mid-field MR systems as well.

Examination 2 (post-embolization 1)
Fig.13 DSA images (lateral views) with selective injection of the right internal carotid artery after the first embolization session. Approximately 40% reduction of the nidus is achieved.

Examination 3 (post-embolization 1, 1.5 T)
Fig.14 Sagittal T1-weighted spin-echo images. High signal intensity components are seen within the nidus corresponding to the thrombosed vascular compartments of the AVM.
Fig.15 Transverse proton density weighted fast spin-echo images. Less signal void (nidus) and extensive high signal intensity areas (intravascular thrombosis) are seen now in the left parietal region (compare with Fig.2 of Exam 1).
Fig.16 Transverse T2-weighted fast spin-echo images. Here again, it is difficult to differentiate intravascular thrombosis, parenchymal lesions and CSF spaces, all presenting as high signal intensity areas.
Fig.17 Transverse turbo FLAIR images. The CSF spaces (signal void with this technique) are well differentiated from parenchymal lesions (and intravascular thrombosis?).
Fig.18 Transverse T2-weighted fast spin-echo (above) and turbo FLAIR (below) images, for comparison. Same observations as on Fig.16 and Fig.17.
Fig.19 Sagittal non-enhanced 2D PC MR angiogram (matrix: 512, Venc: 65 cm/s). Some reduction of the nidus size is observed, compared to the pre-embolization status (Fig.9).
Fig.20 Sagittal collapsed MIP reconstruction from a non-enhanced 3D PC MRA acquisition data set (matrix: 256, Venc: 45 cm/s). Same observations as on Fig.19.
Fig.21 Sagittal collapsed MIP reconstructions from the non-enhanced 3D PC MRA acquisition data sets (matrix: 256, Venc: 45 cm/s) before (above) and after (below) embolization, for comparison. The comparison of these images clearly shows that besides the reduction of the nidus size, the ascending cortical draining veins appears to be less dilated, suggesting decreased flow in the residual nidus.

Examination 4 (post-embolization 2)
Fig.22 DSA images (lateral views) with selective injection of the right internal carotid artery after the second embolization session. Further reduction of the nidus is achieved.

Examination 5 (post-embolization 2, 1.5 T)
Fig.23 Sagittal T1-weighted spin-echo images. A significant reduction in the size of the signal void residual nidus is observed. The high signal intensity components of the AVM most probably correspond to the thrombosed compartments.
Fig.24 Transverse proton density weighted fast spin-echo images. Same observations as on Fig.23.
Fig.25 Transverse T2-weighted fast spin-echo images. Same observations as on Fig.23. and Fig.24.
Fig.26 Transverse turbo FLAIR images. The high signal intensity area around the nidus has also decreased, confirming the beneficial effect of embolization on the adjacent parenchymal lesions, even if only partial.
Fig.27 Sagittal non-enhanced 2D PC MR angiogram (matrix: 512, Venc: 65 cm/s). Further reduction of the nidus size is observed, compared to the previous pre- and first postembolization studies (Fig.9).
Fig.28 Sagittal collapsed MIP reconstruction from the non-enhanced 3D PC MRA acquisition data set (matrix: 256, Venc: 45 cm/s). Same observations as on Fig.27.
Fig.29 Sagittal averaged modulus type source images from the pre- (left) and the second postembolization non-enhanced 3D PC MRA acquisitions, for comparison. The reduction in the size of the nidus and the diameter of the draining veins is well demonstrated.
Fig.30 Sagittal collapsed MIP reconstructions from the pre- (left) and the second postembolization non-enhanced 3D PC MRA acquisition data sets, for comparison. Same observations as on Fig.29.

The ESNR CD-Rom Series