Historical Perspective of Treatments of Cranial Arteriovenous Malformations and Dural Arteriovenous Fistulas

Cerebral arteriovenous malformations and intracranial dural arteriovenous fistulas represent two important classes of intracranial vascular lesions. This article recalls the history on which current technical advances, including diagnoses, characterization, and treatment, is based. It also describes modern therapeutic options, including microsurgical, endovascular, and radiosurgery techniques.

Cerebral arteriovenous malformations (AVMs) and intracranial dural arteriovenous fistulas (DAVFs) represent two important classes of intracranial vascular lesions. This article recalls the history on which current technical advances, including diagnoses, characterization, and treatment, is based. It also describes modern therapeutic options, including microsurgical, endovascular, and radiosurgery techniques.

Introduction to AVMs

Cerebral AVMs are complex, congenital arteriovenous shunts that consist of feeding arteries, an abnormal nidus of vessels, and draining veins. The true pathogenesis of AVMs is unknown and there is currently no animal model that accurately represents the histologic features of human cerebral AVMs. The two main hypotheses for AVM pathogenesis include (1) embryonic agenesis of the capillary system and (2) the retention of a primordial connection between arteries and veins. AVMs are thought to be dynamic and biologically active lesions, rather than static congenital vascular abnormalities. The high-flow state of the lesion also predisposes to arterialization of veins, vascular recruitment, and gliosis of brain tissue within and adjacent to the lesion.

The goal of definitive treatment of cerebral AVMs is complete obliteration of the nidus with preservation of normal neurologic function and elimination of the risk of future hemorrhage. AVM therapy has undergone many changes since the early descriptions of this pathologic entity in the 1700s and 1800s. Modern AVM management requires a multidisciplinary treatment approach. The main three treatment approaches are microsurgery, endovascular embolization, and stereotactic radiosurgery. These treatment options have been used as stand-alone therapies or in combination. The treatments available for AVMs today are based on the development and trialing of numerous techniques and technologies, the history of which is described in this article.

Historical perspective of treatment of AVMs (pre-1900)

John Hunter (1728–1793), the Scottish scientist and surgeon, described the clinical characteristics of extracranial AVMs in the mid-1700s, but it was not until Rudolf Virchow (1821–1902), the German pathologist, in 1863 published Die Krankhaften Geschwülste , his three-volume masterpiece on blood vessels, that many of the common intracranial vascular pathologic entities, including AVMs, were described and differentiated. Virchow’s book is considered the first landmark in the understanding of AVMs. The first clinicopathological correlate of AVMs likely came in 1888, 25 years after Virchow’s publication, when D’Arcy Power (1855–1941), the British surgeon, reported a 20-year-old patient with a right hemiplegia who subsequently died and was found on autopsy to have a left sylvian AVM with an associated massive hemorrhage. The first report of palliative treatment of a true cerebral AVM by ligation of a left parietal feeding artery was by Davide Giordano (1864–1954), the Italian surgeon, in 1889, 26 years after Virchow’s publication. Although cerebral AVMs were identified and occasionally treated before 1900, the tools for modern diagnosis, characterization, and treatment of these lesions came about almost exclusively during the 20th century, concurrent with the birth and formalized establishment of cerebrovascular neurosurgery.

Early Surgical AVM Treatment

In 1928, both Walter Dandy (1886–1946), the American pioneer neurosurgeon, and his mentor Harvey Cushing (1869–1939), the American surgeon and pioneer of neurosurgery, and Percival Bailey (1892–1973), the American neuropathologist and surgeon, independently reported series of AVMs treated before the introduction of angiography, with primarily catastrophic results in both series. In the 1928 article, Dandy described his treatment of eight cases of “arteriovenous aneurysms” at the Johns Hopkins Hospital during surgery for approximately 600 brain tumors during a 5-year span between 1922 and 1926. No successful radical AVM resection was reported out of the 16 cases of Cushing and Bailey or the eight cases of Dandy ( Fig. 1 ). At this time, Cushing was quoted as saying that surgical Cushing excision of an AVM was essentially unthinkable secondary to the significant risk of hemorrhage with even surgical exposure of the lesion.

Early surgical management of AVMs consisted of primarily decompressive procedures with goals of relieving elevated intracranial pressure. Ligation of feeding vessels was the next step in AVM surgery. Ligation of the internal carotid artery was reported and thought to result in some patient improvement, although this was only a palliative measure at the time. Extracranial vessel ligation was attempted first, and then abandoned for intracranial ligation. Intracranial vessel ligation had mixed results as a sole treatment modality, but it did give rise to the concept for embolization procedures.

Ultimately, Herbert Olivecrona (1891–1980), the surgeon regarded as the founder of Swedish neurosurgery, developed the surgical techniques that rendered AVM excision a reasonable treatment option for small to midsize AVMs that did not involve eloquent areas of brain. He was the first to successfully remove a cerebellar AVM on May 5, 1932, in a 37 year-old male. By 1954, Olivecrona had operated on 125 patients with intracranial AVMs. He achieved complete resection in 81 of these cases, with a remarkably low mortality rate of 9%. He introduced the technique of ligating superficial feeding vessels and then working in a circumferential fashion until the deep portion of the AVM was dissected and separated from the brain. The AVM draining veins were ligated as a final step. Gazi Yasargil (born 1925), the Turkish scientist and neurosurgeon, subsequently compiled information from the literature on 500 AVM patients who were treated between 1932 and 1957, and reported that operative mortality for “small” AVMs was 5% and for “moderate-size” AVMs was 10%. Surgeons such as Wilder Penfield (1891–1976) and Theodore Erickson (1906–1986), Lyle French (1915–2004) and Shelley Chou (1924–2001), and Charles Drake (1920–1998) all reported good early results with AVM surgery, each with some modifications of the techniques introduced by Olivecrona.

Development of neuroimaging

Diagnosis of AVMs before the introduction of cerebral angiography was essentially happenstance. There were no opportunities for the physician to clinically diagnose an AVM other than a jacksonian seizure and this finding was clearly not specific to AVMs. AVMs were discovered in some patients taken to the operating room for intracerebral hemorrhage or elevated intracranial pressure, but this finding was unexpected at the time of surgery. The subsequent advances in AVM treatment beyond surgery depended on the ability to visualize these lesions radiologically with angiography and tomography.

In the 1920s, Antonio Caetano de Abreu Freire Egas Moniz (1874–1955), the Portuguese neurologist, attempted visualization of intracranial vessels by injecting bromides into the carotid artery of patients. In his sixth patient, the intracranial vessels were barely visualized for the first time; however, the patient developed carotid thrombosis and subsequently died. Following this adverse outcome, Egas Moniz consulted with his staff and, after great debate, continued his investigations but used iodides rather than bromides. On the third case in June of 1927, Egas Moniz was able to visualize portions of the internal carotid artery in the sylvian fissure, thus performing the first successful cerebral angiogram. Once the technique was further refined with better tolerated contrast agents and better imaging systems, cerebral angiography was a major advance for the diagnosis and understanding of AVMs. Today this technique is crucial for the direct and accurate visualization of normal and abnormal cerebral blood vessels in patients with AVMs and other cerebral vascular lesions.

The era of modern neuroimaging was launched in the 1970s with the invention of the noninvasive modalities CT scanning and MRI. The first clinical CT scan on a patient was performed on October 1, 1971, at Atkinson Morley’s Hospital in England and the first clinical MRI on July 3, 1977, at Downstate Medical Center in New York. These imaging techniques are essential tools for determining the anatomic location of an AVM. In the 1990s, the development of functional MRI and diffusion tensor imaging further advanced cerebral AVM management by helping to map eloquent areas of brain and their relationship to the lesion. Given that sensorimotor and speech centers can be translocated from typical locations in patients with AVMs, these preoperative functional mapping studies are important. All of these imaging modalities can be of great clinical value in the pretreatment assessment of patients with AVMs and they should be used to predict and avoid postoperative deficits.

Intraoperative imaging is also of importance to the neurosurgeon during the resection of an AVM. Intraoperative digital subtraction angiography is an established adjunct tool during the microsurgical treatment of cerebral AVMs. Intraoperative angiography can evaluate for complete extirpation of the lesion, localize small AVMs, find missing or “hidden” feeding vessels, detect AVM in a patient undergoing evacuation of a hematoma, and evaluate the real-time hemodynamics of the lesion. Unexpected residual AVM is shown in 3.7% to 27.3% of intraoperative digital subtraction angiograms during AVM resection cases. Intraoperative fluorescence video angiography using indocyanine green (ICG) is another technique used in recent years to assist in AVM microsurgical resection. ICG is a safe, rapid, and noninvasive tool that is helpful for intraoperative evaluation of blood vessels visible in the surgical field. For AVMs, ICG is used to distinguish AVM vessels from normal vessels and arteries from veins.

The Birth of Microsurgery

Microsurgical extirpation of cerebral AVMs is an often lengthy and highly choreographed surgical process that requires meticulous and extensive dissection. The introduction of the operative microscope in 1960s and the resulting development of microsurgical instruments and microsurgical techniques significantly advanced the efficacy and safety of AVM surgery. In 1966, J. Lawrence Pool (1906–2004) the American neurosurgeon, and R.P. Colton, reported the first use of the operating microscope in an aneurysm patient. In 1969, Yasargil reported a series of AVM resections in 14 patients using microsurgical techniques. This was the first microsurgical AVM series, and his results were excellent. Surgeons, including Hans Pia (1921–1986), Charles Wilson (born 1929), Leonard Malis (1919–2005), Bennett Stein (born 1931), Dwight Parkinson (1916–2005), Thoralf Sundt (1930–1992), and Helge Nornes (born 1930), are other standout pioneers who further developed the microsurgical treatment of AVMs. During the ensuing decades, up to the modern time, dedicated cerebrovascular surgeons have advanced AVM microsurgery surgery so that small-to-medium size AVMs, particularly those in noneloquent cortex, can be excised with little to no morbidity and mortality. In addition to enhancing the surgeon’s operative ability to perform the AVM resection, the microscope also improved the learning curve for the neurosurgical trainees.

Another important development in surgical instrumentation that assisted the treatment of AVMs was the introduction of bipolar coagulation by Leonard Malis (1919–1995). The availability of improved aneurysm clips, which can be used for temporary clipping, feeding vessel ligation, and obliteration of feeding vessels aneurysms, also advanced AVM surgery.

The development of AVM classification schemes and the natural history of AVMs

The significant variability of their gross appearance and the complex structure of the vessels in AVMs made early attempts at classification and description of these lesions difficult and confusing. Cushing and Bailey described these “angiomatous malformations” as venous or arterial in nature, primarily based on whether the surface of the lesion appeared more like a tangle of veins or arteries. Dandy preferred the term “arteriovenous aneurysm.”

The development of useful classification schemes became important for the preoperative selection of patients with AVMs, particularly with regard to surgical risk. Alfred Luessenhop (1926–2009) and Thomas Gennarelli formulated an early anatomic grading scheme to correspond to the degree of surgical difficulty for total obliteration. This scheme was based on the number of named arterial feeders—those with “anatomic constancy and size to have acquired a generally accepted nomenclature.” The investigators mentioned the importance of the AVM location, but this was not considered a major surgical risk factor. Robert Spetzler and Neil Martin developed the next important AVM classification scheme based on lesion size, venous drainage, and eloquence of involved brain. It is currently the most widely used system for predicting the surgical risk of AVM treatment. Various modifications of the scheme by Spetzler and Martin have been reported. These modifications are described by Dr Lawton elsewhere in this issue.

Understanding the natural history of AVMs, particularly the risk of hemorrhage, has helped guide the indications for therapeutic management of AVMs over the years. In the first half of the 1900s, not much was known about the natural history of these lesions, and they were usually only discovered after a bleed or a seizure. In 1966, George Perret and Hiro Nishioka reported an analysis of 545 cases of cerebral AVMs and fistulae from the Cooperative Study of Intracranial Aneurysms and Subarachnoid Hemorrhage, and found a hemorrhage rate of 1.5% per year. Numerous later series, such as those by Crawford and colleagues, John Jane Sr, and colleagues. Robert Wilkins, and Stephen Ondra and colleagues suggested that the annualized risk of hemorrhage is between 2% to 4% per year. The study by Ondra and Troupp was particularly useful because it included a single population of patients with centralized medical care and the mean follow-up was 23.7 years. The studies by Ondra and Crawford also demonstrated that 6 months after a bleed, the risk of hemorrhage is the same for patients with prior hemorrhage and for patients who have never had a hemorrhage. This information, plus the data that a 10% to 15% mortality rate and an additional 20% to 30% risk of serious morbidity exists with each hemorrhage, provides an impetus to thoroughly evaluate and consider treatment of all patients with AVM, regardless of rupture status. Currently, the annual hemorrhage risk of all AVMs is estimated to be 2.4%, based on the work of Juha Hernesniemi and colleagues, from Helsinki, Finland.

Endovascular Techniques and AVM Treatment

Much of the groundbreaking early work on endovascular intracranial embolization was geared toward the treatment of aneurysms, although these techniques were subsequently adapted to the treatment of AVMs and other intracranial vascular pathologies. Endovascular techniques for treatment of AVMs largely developed in the early 1960s, with Luessenhop and William Spence (1908–1992) conceptualizing the blockage of hypertrophied, abnormal feeding vessels. This “artificial embolization,” as it was called, relied on the embolic material obstructing the enlarged feeding vessels but not penetrating the smaller vessels of the AVM nidus.

Early embolization procedures relied on open surgical approaches and direct puncture of target vessels to introduce catheters and embolic material. As such, they were fraught with the same morbidity and mortality of the open neurosurgical procedures of that time. The development of small, flexible catheters was a major technological advance because this allowed the introduction of these instruments and embolic agents from distal sites, obviating direct punctures. Catheters were initially navigated to their target location largely by flow, as is the case with balloon catheters. Balloon catheterization and occlusion of intracranial vessels was described by the Russian neurosurgeon Fedor Serbinenko (1928–2002) in 1974. In 1976, Charles Kerber described a calibrated leak balloon system for flow-directed catheter guidance and administration of contrast and cyanoacrylate embolic material. This idea of using the relatively high flow of abnormal AVM feeding vessels to direct catheters and embolic material was a driving force behind early endovascular AVM treatment.

Most modern catheter navigation relies on a torqueable guidewire system, in which a catheter is advanced over a wire to its target and the curve of the wire and catheter tip determine its steerability. Although guidewires were initially designed in the mid 1970s, the Tracker catheter system was the first guidewire and catheter system to achieve mass production in 1986. Advances in material science and engineering have subsequently led to an explosion in the development and widespread availability of various guidewires and catheters. Advances include radiopaque markers for better angiographic visualization, shaped tips for improved guidance, and construction with various materials for improved hydrophilic properties and tolerance to different embolic agents.

Embolic agents used for the endovascular management of AVMs, as with catheters and microwires, have undergone many advances over the years. Early attempts at AVM embolization were performed with numerous materials delivered by flow-directed catheters positioned proximal to the actual feeding vessels. Examples include methyl methacrylate spherules (2.5–4.2 mm), silk suture attached to Silastic emboli, gelatin particles, autologous blood clots, and Avitene (C.R. Bard, Inc, Murray Hill, NJ, USA). These early materials were limited by difficulty to control during embolization, high rates of vessel recanalization, and, in some cases, vasculitis. The adaption of acrylic tissue adhesives for embolization, as well as concurrent advances in catheter delivery systems, brought AVM embolization much closer to its current state.

Isobutyl 2-cyanoacrylate is an early generation tissue adhesive that was used for embolization because it polymerizes on contact with blood. However, of the rapid polymerization rate of isobutyl 2-cyanoacrylate led to complications including adhesion and incorporation of the catheter to the glue cast in the feeding artery, and this agent was replaced by N-butyl cyanoacrylate (NBCA). NBCA is currently a popular choice for AVM embolization. Onyx (ev3, Plymouth, MN), a combination of ethylene-vinyl alcohol copolymer and dimethyl sulfoxide, is another agent widely used for the treatment of AVMs. This agent is less adhesive and generally polymerizes slower than NBCA. Platinum coils are also used for AVM embolization to obstruct large feeding vessels.

The ultimate goal of endovascular AVM embolization is angiographic obliteration. However, this is often not achieved, despite attempted staged embolizations, and it may be contraindicated because of the high rate of complications. Many AVMs contain numerous small feeding vessels that cannot be safely catheterized for superselective embolization. In such cases, embolization is used as an adjunct tool before microsurgical resection or radiosurgery. Preoperative embolization can decrease the size of the AVM, reduce intraoperative blood loss, and target blood vessels that might not be surgically accessible. Before radiosurgery, the goals of embolization are to shrink the size of the AVM, particularly at the periphery; address feeding vessel aneurysms and other high-risk features; and, ultimately, to reduce the size of the radiation field required for treatment.

Radiotherapy for AVMs

The goal of radiotherapy for AVMs is complete obliteration of the AVM by delivering a high radiation dose to the AVM nidus while minimizing the radiation exposure to the surrounding normal brain. Various forms of radiation delivery have been developed over the years and used for AVM treatment. These include external beam x-rays from x-ray tubes or linear accelerators (LINAC), gamma rays from a cobalt source, and Bragg peak (proton beam or helium ion) therapy. Each of these radiation forms can be performed with stereotactic targeting (radiosurgery) and most can be done as single fraction or multifraction therapy. It is now known that the vascular changes seen after radiation are similar, regardless of whether x-rays, gamma rays, or proton beams are used. The limitations of these therapies are nonobliteration of lesions on angiogram, the risk of hemorrhage, and radiation-induced brain injury.

Vilhelm Magnus (1871–1929), Norway’s pioneering neurosurgeon, was likely the first to treat cerebral AVMs with radiation in 1914, using conventional fractionated radiation. Cushing and Bailey treated eight AVM patients with conventional radiation and there was an apparent benefit in three of these patients. One of the patients (Case 8) had complete resection of the AVM approximately 1 year after radiotherapy and, from the pathology specimen, it was first demonstrated that the radiation-induced occlusion of AVM vessels was secondary to intimal hyperplasia. This inflammatory change ultimately results in vessel thrombosis and obliteration. Although never truly popular, conventional fractionated radiotherapy was used for AVM treatment intermittently. R.T. Johnson had the largest series of AVMs treated with conventional radiation, with 100 patients treated over 20 years. Twenty of the patients had follow-up angiography. Complete obliteration was reported in nine patients and reduction in the size of the AVM was reported in five patients. Robert Laing and colleagues reported on the ineffectiveness of conventional fractionated radiotherapy in AVM patients. They demonstrated that the actuarial annual risk of bleeding in the treated patients was 2.3%, and this was similar to the risk in untreated AVMs. Ultimately, conventional radiation for AVMs was replaced by stereotactic radiosurgery.

Radiosurgery was first described in 1951 by Lars Leksell (1907–1986), the Swedish neurosurgeon, using a cyclotron. He subsequently introduced the first Gamma Knife using cobalt sources in 1968. Reports using Gamma Knife began appearing in the early 1970s. Ladislau Steiner, Christer Lindquist, and L. Dade Lunsford and Douglas Kondziolka were instrumental in improving the Gamma Knife technique and demonstrating its effectiveness. Radiosurgery for AVMs based on the Bragg peak phenomenon (proton beam and helium beam) was advanced by Raymond Kjellberg (1925–1993) and colleagues at Harvard and at Berkeley by Jacob Fabrikant (1928–1993) and colleagues and Richard Levy and colleagues. Radiosurgery using LINAC was introduced by Osvaldo Betti and colleagues, and then it was further improved on by William Friedman and F.J. Bova and other investigators. Radiosurgery using each of these 3 main techniques (Gamma Knife, Bragg peak phenomenon, and LINAC) have all been used successfully for treatment of cerebral AVMs. Current AVM management with radiosurgery results in obliteration in 1 to 3 years from the time of treatment, with obliteration rates of 54% to 92% and the risk of hemorrhage during this time period remaining unchanged from untreated patients. The details of radiotherapy for AVMs are described by Michael Lim elsewhere in this issue.

Only gold members can continue reading. Log In or Register to continue