Pre- and postembolization MRI-MRA of a left temporo-occipital cortical-subcortical AVM

This 45 year old male presented with sudden onset of headache and vomiting. He also complained of neck stiffness.
However, the first CT examination was not performed until two weeks later and was suggestive of a left temporo-occipital AVM.
MRI and subsequent conventional, catheter based cerebral angiography confirmed the diagnosis. Five associated saccular aneurysms were also found, possible due to the principle of high flow angiopathy pathomechanism.
Endovascular treatment of both the AVM and the aneurysms was proposed. Aneurysms were treated by Guglielmi detachable coils (GDC) coils. Several sessions of embolization of the AVM by histoacryl resulted in progressive and significant diminution of the lesion. A small residual of the nidus was finally cured by stereotactic radiosurgery. MRA was used as a non-invasive follow up technique between successive treatments.

Left temporo-occipital AVM
Examination 1 (preembolization 1)
Fig.1 Sagittal T1-weighted spin-echo images demonstrating the left temporo-occipital AVM.
Fig.2 Transverse proton density weighted fast spin-echo images. The topography of the nidus is accurately evaluated. Note the metallic artifact at the tip of the basilar artery, generated by the GDC coil (used for the occlusion of a saccular aneurysm). Another saccular aneurysm is seen at the anterior communicating artery.
Fig.3 Transverse T2-weighted fast spin-echo images. Same observations as on Fig.2.
Fig.4 Sagittal averaged modulus (left) and corresponding magnitude of complex differences (right) type source images from a 3D PC MRA acquisition. The nidus of the AVM is accurately identified. Note that the high-flow components (arterial feeders) of the AVM remain signal void on the anatomical images (arrows).

Video 1.
3D ciné demonstration of the AVM (created from 180 MIP reconstruction with 2 degrees of angular differences) 3D PC MRA acquisition (Venc: 65 cm/s). The spatial disposition of the feeding arteries and the draining veins is well appreciated. This presentation might be useful for pre-embolization planning of the possible pathways for superselectively approaching the nidus.

Examination 2 (preembolization 2)
Fig.5 DSA images (lateral views) with selective injection of the left vertebral (above left) and the left internal carotid (above right and below) arteries. Note the complex venous drainage of the AVM.

Examination 3 (postembolization 1)
Fig.6 DSA images (lateral views) with selective injection of the left internal carotid artery. Significant reduction of the nidus of the AVM and reduction of the flow velocity through the lesion is demonstrated.

Examination 4 (postembolization 2, 0.5 T)
Fig.7 Sagittal T1-weighted spin-echo images. The reduction of the nidus size is well appreciated.
Fig.8 Transverse proton density weighted fast spin-echo images. The intravascular embolizing agent and the patent vessels are impossible to differentiate, both exhibiting low signal intensity on these images. Note again the metallic artifact of the GDC coil in the basilar tip aneurysm.
Fig.9 Transverse T2-weighted fast spin-echo images. The high signal intensity areas in proximity to the nidus appear to be more marked compared to the pre-embolization study, however the patient was neurologically normal. They were felt to correspond to enlarged CSF spaces and not true parenchymal lesions. Otherwise same observations as on Fig.8.
Fig.10 Transverse averaged modulus (parenchymal) and corresponding magnitude of complex differences (flow) type source images from Gadolinium-enhanced 3D PC MRA acquisitions before (above) and after (below) embolization, for comparison. The intravascular embolizing agent is clearly identified on the postembolization parenchymal images, exhibiting low signal intensity without corresponding signal on the flow images (arrows). The PC MRA technique is based on gradient-echo sequences, which increase the magnetic susceptibility artifacts, such as that generated by the Tantalum or Tungsten used in the embolizing mixture as a radioopacifying additive. Note also that the GDC coil in the basilar tip aneurysm has a much larger artifactual signal void halo, compared to the proton density and T2-weighted fast spin-echo images.
Fig.11 Sagittal averaged modulus (left) and corresponding magnitude of complex differences (right) type source images from a Gadolinium-enhanced 3D PC MRA acquisition. The intravascular embolic agent is again demonstrated.

Video 1.
3D ciné demonstration of the AVM (created from 180 MIP reconstruction with 2 degrees of angular differences) 3D PC MRA acquisition (Venc: 65 cm/s). The significant reduction of the nidus and changes affecting both the arterial feeders and the draining veins are well appreciated.

Examination 5 (1 year follow-up after embolization)
Fig.12 Sagittal T1-weighted spin-echo images. Moderate cortical-subcortical atrophy is seen in the left occipital region, a normal finding after embolization therapy of AVMs.
Fig.13 Transverse proton density weighted fast spin-echo images. Some compartments of the nidus are still patent.
Fig.14 Transverse T2-weighted fast spin-echo images. The cortical-subcortical atrophy and the concomitant enlarged pericerebral CSF spaces are well seen.
Fig.15 Sagittal survey single-slice 2D PC MR angiogram (matrix: 512, Venc: 65 cm/s). Demonstration of the residual nidus.
Fig.16 Sagittal averaged modulus (left) and corresponding magnitude of complex differences (right) type source images from a Gadolinium-enhanced 3D PC MRA acquisition. Unchanged appearance of the intravascular embolic agent and the residual nidus compared to the previous examination (Fig.11)
Fig.17 Sagittal collapsed MIP reconstruction from a Gadolinium-enhanced 3D PC MRA acquisition data set. The residual nidus is well delineated. Arterial feeders are detected arising from a branch of the middle cerebral artery (red arrows), the posterior cerebral artery (yellow arrows) and even the external occipital artery (white arrows). The residual drainage of the AVM leads towards the ipsilateral transverse sinus (blue arrows) and the middle segment of the superior sagittal sinus (green arrows). Two of the previously identified main draining veins, leading to the distal segment of the superior sagittal sinus are not visualized at this time (compare with Fig.5).
Fig.18 Transverse targeted MIP reconstruction from the sagittal Gadolinium-enhanced 3D PC MRA acquisition data set. The residual nidus is demonstrated, however, arterial feeders and draining veins are difficult to identify in this projection. Moreover, vessel wall delineation and vessel conspicuity are markedly inferior to the sagittal MIP reconstruction, due to the anisotropic voxels employed.

The ESNR CD-Rom Series