A 42-year-old woman presents with multiple episodes of generalized tonic-clonic seizures that started recently. Despite medication, she experiences up to two seizures per week. Given the adult onset of her seizures, enhanced CT is performed, followed by MRI and angiography.

Fig. 15.1 (A) Contrast-enhanced axial CT demonstrates a dilated vessel in the left parietal region. (B) Axial T1-weighted contrast-enhanced and (C) T2-weighted MR images confirm the presence of an abnormal vascular structure at the left parietal region, seen as flow voids without evidence of a classic caput medusae. Dynamic CTA in (D) early arterial, (E) late arterial, and (F) venous phases show early filling of a cortical vein, indicating an AV shunt.

CT AND MRI

Small pial AVM with venous ectasia and stenosis

• Standard 5F access (puncture needle, 5F vascular sheath)
  
• A standard 5F multipurpose catheter (Envoy; Cordis, Warren, NJ) with continuous flush and a 0.035-in hydrophilic guidewire (Terumo, Somerset, NJ)
  
• A 0.012-in flow-directed microcatheter (Magic; Balt International, Montmorency, France) with a 0.008-in guidewire (Mirage; ev3, Plymouth, MN)
  
• A 10% glucose solution
  
• Histoacryl/Lipiodol (1 mL/1 mL)
  
• Contrast material
  
• Steroids

Fig. 15.2 DSA. Left ICA angiogram in (A) AP and (B) lateral views in arterial phase reveals a small AVM in the left parietal region, supplied by the anterior parietal branch of the left MCA, draining into two superficial cortical veins. A long pial course of the draining veins with a focal stenosis before entry into the superior sagittal sinus is noted.

Although there is little discussion about the necessity of treating pial AVMs that have bled because of the significant risk for rebleeding, pial AVMs that have not bled must be further analyzed to select those patients in whom therapy is indicated (i.e., in whom the therapeutic risk is lower than the natural history risk). In our practice, we first classify unruptured AVMs according to their pathologic mechanism and angioarchitecture because hemorrhage is not the only way an AVM may become symptomatic. Second, we try to determine the natural history risk by a careful angioarchitectonic evaluation of the AVM in relation to risk factors for future hemorrhage. Epilepsy is one of the more common presenting symptoms of AVMs, and it may be related to (1) venous congestion (due to a long pial course of a superficial draining vein, as in this case), (2) mass effect of a venous pouch, or (3) perinidal gliosis. Venous outflow obstructions (especially if already associated with venous ectasia, as in this case) are regarded in our practice as a possible risk factor for future hemorrhage related to an increase in intranidal pressure.

• AVMs may be clinically silent; however, a thorough neurologic evaluation has to be performed to detect subtle neurologic deficits that may add to the understanding of the pathologic mechanism of the AVM.

Fig. 15.3 (A) Superselective microcatheter injection and (B) plain radiography in lateral view demonstrate deposition of the glue cast within the AVM nidus. (C) Left ICA angiogram in lateral view after embolization shows complete occlusion of the AVM.

• AVMs that do not present with hemorrhage may be detected incidentally during a workup for seizures, headaches, or neurologic deficits. CT and CTA can demonstrate an abnormal tangle of vessels in the parenchyma. Attention should be paid to calcifications, which may suggest longstanding venous congestion.

MRI/MRA

• On MRI, T2-weighted images will show intraaxial and subarachnoidal rounded flow voids within the brain parenchyma. On T1-weighted images, the signal within the AVM is unpredictable because of flow turbulence, blood degradation products, and the flow rate in the veins. Attention should be paid to perifocal edema, gliosis, or hemosiderin staining because these imaging findings may shed light on the individual pathologic mechanism. Although “static” MRA techniques may well detect the lesion, they do not demonstrate the angioarchitecture, and information about flow-related or intranidal aneurysms or the feeding type of the arteries is unlikely to be obtained by MRI because of restrictions in the spatial resolution. Dynamic MRA techniques will be able to demonstrate early venous opacification.
  
• In some instances, functional imaging techniques may be helpful to determine the eloquence of the perinidal brain tissue and to further evaluate potential pathologic mechanisms, such as arterial steal and venous congestion.

• In patients with suspected brain AVMs, we classically perform six-vessel angiography to determine all potential feeders to the brain AVM and to determine whether multiple brain AVMs are present.
  
• For incidentally discovered AVMs, the report should include details about the feeding arteries, such as the presence of flow-related aneurysms and the number and type of feeding arteries (en passage versus direct). Details to be reported about the AVM nidus include the number of compartments, the presence of intranidal aneurysms, and plexiform versus fistulous type of nidus. Finally, concerning the veins, the number of draining veins per compartment, any venous ectasia or pouches, and any venous stenoses need to be evaluated. In addition, the normal brain parenchyma must be analyzed for venous congestion or arterial steal.

• In small unruptured AVMs, the major entities in the differential diagnosis are developmental venous anomalies and capillary telangiectasia. However, these do not classically demonstrate early venous filling (i.e., a shunt).
  
• In rare cases, dural AV shunts with cortical venous reflux may mimic micro-AVMs.
  
• If multiple AVMs are present, hereditary hemorrhagic telangiectasia (HHT) should be strongly considered and appropriate investigations be performed. Metameric syndromes (cerebrofacial arteriovenous metameric syndromes, or CAMS) must also be considered.

• Conservative or medical management should be considered if the treatment risks are estimated to be higher than those of the natural history of the lesion, particularly in elderly patients with minimal symptoms.

• In young patients with asymptomatic lesions and a large lifelong risk for AVM-related symptoms, radiosurgery should be contemplated, especially if no angioarchitectonic risk factors can be identified. Radiosurgery can result in a high cure rate with a rather low complication rate in well-selected cases when it is performed by a treatment team familiar with this modality for the management of AVMs. However, the long time until a cure is achieved (up to 4 years) may make one decide against this treatment if the risk for hemorrhage or neurologic deficits is considered high.

• In patients who have superficial small lesions with multiple feeders or en passage supply, surgery may be more suitable than endovascular therapies.

• In our practice, endovascular therapy is the method of choice for small AVMs when complete endovascular obliteration seems feasible (i.e., no en passage vessels, small number of feeding arteries and AVM compartments) and when a defined pathologic mechanism that is related to the clinical symptoms can be identified.
  
• We employ only liquid embolic agents to occlude the most distal arterial segment, the nidus, and the most proximal venous segment of the shunt.
  
• Periprocedural heparin is recommended if the procedure is anticipated to take longer than 30 minutes.
  
• In our practice, steroids are given following Histoacryl embolization to compensate for the exothermic effect of the glue.

• Standard angiographic complications may occur (at the puncture site: bleeding, false aneurysms, fistulae; in catheterized vessels: emboli, dissections; systemically: contrast reaction, renal failure).
  
• Migration of the embolic agent into the draining vein with potential occlusion (especially in stenosed venous segments) of the draining vein may result in hemorrhage or venous ischemia.
  
• Proximal occlusion of the arterial feeders may result in delayed recanalization of the nidus.
  
• Arterial ischemia may develop as the consequence of reflux of the embolic agent into arteries supplying the brain.

The treatment of incidentally discovered pial AVMs of the brain is controversial. Little is yet known about their natural history, their pathologic mechanisms, and the efficacy and risks of some of the proposed treatments (e.g., Onyx). The annual hemorrhagic risk for unruptured brain AVMs varies between 1 and 5%. It is well-known that only complete exclusion of an AVM can eliminate the risk for hemorrhage, and that the rates of curative endovascular embolization of AVMs with an acceptable periprocedural risk are ~20 to 50%. However, these values do not take into account the individual patient for whom the treatment plan has to be tailored. In the present example, a defined pathologic mechanism could be linked to the patient’s symptoms and explained by the angioarchitecture, which could be easily treated with a high probability of complete cure, given the type of feeding artery (terminal feeder), the size of the nidus, and its accessibility. Therefore, treatment strategies, including conservative management as well as all three treatment options and combinations thereof, are discussed in our practice within a multidisciplinary team and tailored to the individual patient.

• As with dural AV shunts, a proximal vessel occlusion is not sufficient to achieve a permanent cure despite initial angiographic occlusion because the AVM may recruit leptomeningeal collaterals.
  
• A long pial course seems to be associated with a higher risk for and incidence of epilepsy, especially if associated venous stenoses are present and causing venous congestion along the pial surface.
  
• In AVMs with venous outflow stenoses, venous penetration must be avoided because it may lead to complete occlusion of the already stenosed venous segment; therefore, the authors use a more concentrated glue mixture under these circumstances.
  
• Surgical therapies are readily available for superficial sulcus AVMs and should be considered if en passage feeding vessels are present, distal catheterization is impossible, or the safety margin (i.e., the distance from the catheter tip to the closest, upstream brain-supplying artery) is too small.

Krings T, Geibprasert S, Terbrugge K. Interventional therapy of brain and spinal arteriovenous malformations. In: Mast et al, eds. Stroke. Philadelphia, PA: Lippincott Williams & Wilkins. In press.

Ledezma CJ, Hoh BL, Carter BS, Pryor JC, Putman CM, Ogilvy CS. Complications of cerebral arteriovenous malformation embolization: multivariate analysis of predictive factors. Neurosurgery 2006;58(4):602–611, discussion 602–611

Willinsky RA, Goyal M, terBrugge K, Montanera W, Wallace MC, Tymiansky M. Embolization of small (< 3cm) brain arteriovenous malformations. Correlation of angiographic results to a proposed angioarchitecture grading system. Interv Neuroradiol 2001;7:19–27

A 21-year-old student presents with increasing memory problems, a decreased attention span, and progressive left hemiparesis and clumsiness of his left hand. On physical examination, he has difficulty performing voluntary motor acts with his left arm and leg, and a hemiapraxia is observed, although the muscle tone is maintained. The function of his left side is not affected in serial automatic motor activities (dressing and walking). A supplementary motor syndrome is therefore diagnosed. In addition, his short-term memory is reduced. MRI and CT perfusion are performed.

Fig. 16.1 (A) T2-weighted MRI demonstrates an abnormal tangle of flow-void structures at the right parasagittal region, corresponding to an AVM. (B) Perfusion CTA in a parasagittal view shows a decreased MTT (blue) within the AVM caused by rapid shunting and a markedly increased MTT (yellow and orange) within the SMA and anterior cingulate gyrus region.

Large pial AVM with high-flow shunt and venous congestion associated with progressive neurologic deficits

• Standard 5F access (puncture needle, 5F vascular sheath)
  
• Standard 5F multipurpose catheter (Guider Soft Tip; Boston Scientific, Natick, MA) with continuous flush and a 0.035-in hydrophilic guidewire (Terumo, Somerset, NJ)
  
• A 0.012-in flow-directed microcatheter (Magic; Balt International, Montmorency, France)
  
• A 10% glucose solution
  
• Histoacryl/Lipiodol/tantalum powder (2 mL/0.2 mL/one-half vial)
  
• Contrast material
  
• Steroids

Fig. 16.2 DSA. Left ICA angiogram in (A) AP and (B) lateral views reveals an AVM supplied mainly by the pericallosal and callosomarginal arteries of the ACA, with significant indirect leptomeningeal collaterals from the MCA branches.

Fig. 16.3 (A) Superselective microcatheter injection reveals a fistula within the AVM nidus. (B) Plain radiography demonstrates the location of the glue cast, and left ICA angiogram in (C) AP and (D) lateral views after embolization shows significantly decreased flow through the AVM. There is better perfusion of the SMA and anterior cingulate gyrus region on the follow-up CT perfusion (E).

Venous congestion may be related to decreased outflow (in the case of venous stenoses) or increased inflow (in the case of fistulous AVMs) and can be associated with progressive neurologic symptoms. Even if treatment is not complete following the first embolization session, endovascular treatment may play a role in reducing the shunt to relieve clinical symptoms before definitive treatment of the AVM. The strategy of defining a target (i.e., a pathologic mechanism that is related to a clinical symptom as verified by imaging of the angioarchitecture) before treatment has been termed partial targeted embolization. The role of endovascular treatment is therefore to “secure” the AVM, stabilize or ameliorate the symptoms, or reduce the size of the AVM so that additional therapies that may take longer to be effective are made safer. In the present case, the perfusion data, the recruitment of leptomeningeal collaterals from a different vascular territory (MCA) to reconstitute the distal ACA territory, and the progressive neurologic symptoms strongly suggested a fistulous compartment within the AVM, which could then be regarded as the target of embolization.

• Perfusion imaging can be done with CT or MRI and may add to the understanding of the pathologic mechanisms of brain AVMs. Following the intravenous bolus application of a contrast agent, the time to peak (TTP), mean transit time (MTT), cerebral blood volume (CBV), and cerebral blood flow (CBF) can be determined. TTP indicates the arterial blood flow velocity (decreased in shunts, increased in slow arterial flow), and the MTT represents the AV transit time (decreased in shunts, increased in edema or venous congestion). The CBV is an indirect parameter of the venous outflow (increased CBV indicates venous outflow obstruction). A combination of these parameters may add to our understanding of the remote effects of a brain AVM. In patients with neurologic symptoms, arterial steal and venous congestion have been proposed as potential pathologic mechanisms to explain neurologic deficits. In venous congestion, an increased MTT and a normal or even decreased TTP are seen (as in the present case), whereas in arterial steal, a decreased MTT and TTP will be seen (and therefore, a decreased time during which oxygen can be extracted by tissue and subsequent chronic ischemia). Functional imaging can detect subtle changes before structural imaging, and the results can be correlated with the imaging findings during angiography.

• Although prospective randomized trials are lacking, it is our opinion that conservative management in young patients with neurologic symptoms that can be attributed to a high-flow AVM is not a good option, especially if the risks of treatment are low.

• Intranidal high-flow or fistulous compartments of AVMs are thought to be less likely to become obliterated following radiosurgery. In addition, because of the longer time needed for obliteration following radiosurgery, a treatment modality with more rapid effects may be necessary.

• In general, surgical resectability and the clinical outcome depend on the size of an AVM, the drainage pattern, and the eloquence of the brain region that harbors the AVM. Although the flow rate therefore does not seem to play a role in surgical resectability, in everyday practice, most neurosurgeons will appreciate a flow reduction achieved by endovascular means.

• In our practice, endovascular therapies are the method of choice for AVMs that harbor fistulous components, in particular when they are perceived to be responsible for the clinical symptoms.
  
• In most instances, a liquid embolic agent that polymerizes quickly is used (in the present case, 2 mL of N-butyl-cyanoacrylate mixed with 0.2 mL of Lipiodol). To increase the visibility of the embolic agent, tantalum powder is added.
  
• In our practice, steroids are given following Histoacryl embolization to compensate for the exothermic effect of the glue.

• Standard angiographic complications may develop at the puncture site (bleeding, false aneurysms, fistulae), in catheterized vessels (emboli, dissections), and systemically (contrast reaction, renal failure).
  
• Migration of the embolic agent into the draining vein with potential occlusion may result in hemorrhage or venous ischemia, proximal occlusion with delayed reopening, and arterial ischemia due to reflux of the embolic agent into arteries supplying the brain.
  
• Seizures following successful treatment may result from significant alterations in cerebral hemodynamics.
  
• Delayed venous thrombosis may develop as a result of significant flow reduction.

• The liquid embolic agent must penetrate from the arterial to the venous side to ensure stable obliteration of the fistulous compartment.
  
• Following an embolization procedure, we generally wait for 6 weeks to 3 months before performing a second embolization procedure to allow the brain vasculature time to remodel.
  
• If flow changes are significant following embolization, headaches and seizures may develop transiently. If new neurologic symptoms occur, follow-up imaging is necessary to rule out excessive venous thrombosis, which may require anticoagulation and steroid therapies.

Back AG, Vollmer D, Zeck O, Shkedy C, Shedden PM. Retrospective analysis of unstaged and staged gamma knife surgery with and without preceding embolization for the treatment of arteriovenous malformations. J Neurosurg 2008;109(Suppl):57–64

Hartmann A, Mast H, Mohr JP, et al. Determinants of staged endovascular and surgical treatment outcome of brain arteriovenous malformations. Stroke 2005;36(11):2431–2435

Krings T, Hans FJ, Geibprasert S, Terbrugge K. Partial targeted embolization of brain arteriovenous malformations. Eur Radiol June 11, 2010 [Epub ahead of print]

Ledezma CJ, Hoh BL, Carter BS, Pryor JC, Putman CM, Ogilvy CS. Complications of cerebral arteriovenous malformation embolization: multivariate analysis of predictive factors. Neurosurgery 2006; 58(4):602–611, discussion 602–611

Natarajan SK, Ghodke B, Britz GW, Born DE, Sekhar LN. Multimodality treatment of brain arteriovenous malformations with microsurgery after embolization with onyx: single-center experience and technical nuances. Neurosurgery 2008;62(6):1213–1225, discussion 1225–1226

Söderman M, Andersson T, Karlsson B, Wallace MC, Edner G. Management of patients with brain arteriovenous malformations. Eur J Radiol 2003;46(3):195–205

Yuki I, Kim RH, Duckwiler G, et al. Treatment of brain arteriovenous malformations with high-flow arteriovenous fistulas: risk and complications associated with endovascular embolization in multimodality treatment. J Neurosurg Oct 16, 2009 [Epub ahead of print]

A 28-year-old right-handed woman has had migraine headaches for 3 months, which have progressed in intensity over the last weeks. In addition, she experiences transient ischemic attack (TIA)-like symptoms (motor aphasia) that are not related to the headaches and are not associated with a loss of consciousness. Outside imaging reveals a right parietooccipital arteriovenous malformation (AVM). Functional MRI and CT perfusion are performed.

Fig. 17.1 (A,B) Functional MRI during speech production. (C,D) Perfusion MRI in axial cuts.

Large pial AVM with prenidal flow-related aneurysms and arterial steal

• Standard 5F access (puncture needle, 5F vascular sheath)
  
• Standard 5F multipurpose catheter (Guider Soft Tip; Boston Scientific, Natick, MA) with continuous flush and a 0.035-in hydrophilic guidewire (Terumo, Somerset, NJ)
  
• A 3 × 0.012-in flow-directed microcatheter (Magic; Balt International, Montmorency, France) with a 0.008-in guidewire (Mirage; ev3, Plymouth, MN)
  
• A 10% glucose solution
  
• Histoacryl/Lipiodol (1 mL/1.2 mL)
  
• Contrast material
  
• Steroids

Fig. 17.2 DSA. Right ICA injection in (A) AP and (B) lateral views, arterial phase. The arrows point toward the flow-related prenidal aneurysm.

Fig. 17.3 DSA. Right ICA injection in (A) AP and (B) lateral views, arterial phase, at follow-up after 3 months. The nidus size is markedly reduced, and there is complete regression of the flow-related aneurysm.

MRI AND CT

• The size of previously untreated AVMs is in our experience best estimated with volumetric contrast-enhanced CT or multiplanar T2-weighted MRI. Still, it may be difficult to differentiate the true nidus from the draining veins and feeding arteries.
  
• Following embolization, artifacts will be present on both MRI and CT that make it nearly impossible to estimate the residual target volume.
  
• In most instances, the spatial resolution of high-quality MRA or CTA will allow the detection of prenidal flow-related aneurysms.