Modern Brain Arteriovenous Malformation Models: A Review

BackgroundResearch of William Hunter's hypothesized (then discovered) arteriovenous varix (now arteriovenous malformation [AVM]) has developed exponentially over the previous quarter‐millennium. Virchow and Luschka's subsequent contributions (nearly 100 years later) by identifying an AVM of the brain and its congenital nature were 2 of the first significant developments made in the field. AVMs present as an erroneous connection (known as a fistula) between an artery and a vein that bypasses the capillary circulation. The arteries and arterioles contributing to the malformation are known as feeders which connect to the draining veins via a plexiform vascular network known as a nidus. Prior to the design of a synthetic anastomosis coupled with vessel ligation by Spetzler et al, animal models were largely based on embolization or study of the normal anatomy. The animal and early genetic models have been reported on at length and numerous times across the literature, but novel developments spanning the previous decade have ushered in a technological revolution of vascular modeling that warrants discussion and analysis.MethodsParameterization of a PubMed query to include all literature including the words “brain,” “arteriovenous malformation,” and “model” yielded 489 articles. After extraction of relevant literature and full‐text screening, 41 articles were chosen for detailed review.ResultsTechnological innovations outside of neurosurgery have greatly impacted the development of novel AVM models in the form of 3D flow models printed into silicon models and combined with advanced imaging technology such as 4D flow magnetic resonance imagin. Technological developments in preservation solutions, catheterization tools, and imaging technologies have also allowed for advent of the cerebrovascular placental model for testing of treatments such as radiosurgery, glue embolization, coiling, as well as histological assessment of tissue directly after intervention.ConclusionWe review the breadth of AVM models in the literature over the last almost 5 decades.