Many classification systems have been proposed for BAVMs, almost all of which are designed for surgical resection planning [18, 19]. For instance, the famous BAVM SM grading system, which was introduced by Spetzler and Martin in 1986, has proven to be useful in guiding surgical resection [20, 21]. The Lawton-Young Grading System improves the outcome prediction accuracy and is a feasible alternative to the SM system [22]. Recently, several grading systems for procedural risk in the EVT of BAVMs have been proposed, including the Buffalo, Puerto Rico, and BAVM embocure scoring systems [23,24,25,26,27]. However, these scoring systems are simply based on the number of feeders, AVM size, eloquence, and venous drainage. Therefore, Lv et al. chose factors of clinical, anatomical and angioarchitecture of the BAVM nidus and combined these factors to form a new Tsinghua grading system [28].
However, these grading systems did not aim at BAVMs in a certain location. In our previous reports, the classifications of BAVMs involving the ACA and PCA regions were proposed to evaluate the risk and EVT outcomes [2, 3]. Therefore, in our study, a similar classification should be proposed specifically for MCA-BAVMs to benefit EVT. When performing EVT for MCA-BAVMs, transarterial EVT remains the mainstream technique [29]. Therefore, classification of MCA-BAVMs based on the anatomical segment of the MCA is feasible because MCA‑BAVMs have a wide distribution along the MCA course.
MCA-BAVMs can be divided into four types, including types I, II, IIIa and IIIb. Because the distribution of the supplying areas of M1-4 segments varied, the incidence rate of MCA-BAVMs in each segment was different. For instance, in our study, Type III BAVMs were the most common, accounting for 77% of MCA-BAVMs, due to the large blood supply region of the M4 segment [30]. Compared with other grading systems, our classification system considers the unique characteristics of MCA-BAVMs along the MCA and is very individualized.
Due to the different anatomical characteristics of M1-4 segments of the MCA, such as branching, neighboring venous structures, and space, the angioarchitecture in different types of MCA-BAVMs is unique. In our study, no difference in age or sex was found among the types, which indicated that they did not affect the classification of MCA-BAVMs. In addition, in our study, the average age was 33.8 years, which is in accordance with the peak age of 30–34 years in Petridis et al.’s report on 6527 patients with BAVMs [31].
Under continually high pressure, BAVMs are not static, and their angioarchitectural features can develop with time. In adults, feeding artery aneurysms and ectasia of the draining vein are more often observed, suggesting that these particular features take time to develop [32].
Due to hemodynamic stress, the feeding artery and nidus may become dilated and thinned, and associated aneurysms can develop on the feeding artery or in the nidus of the BAVM. The incidence of associated aneurysms in all BAVMs was 20.2%, and the rate in supratentorial BAVMs was lower than that in infratentorial BAVMs [33]. Therefore, the rate of associated aneurysm in MCA-BAVMs should be less than 20.2%. In Stein et al.’s report of 409 patients with supratentorial BAVMs, 14.4% of patients suffered associated aneurysms [34]. In this study, the incidence of associated aneurysm was 14.8%, similar to that of Stein et al.’s report. Due to hemodynamic stress, remodeling of the primary draining vein is common, presenting with dilated or varicose morphology. In Pan et al.’s report of supratentorial BAVMs, the percentage of varicosities of the draining vein was 54.6% [35]. In this study, the morphology of the primary draining vein was varicose in 45.9% of BAVMs, similar to the above report.
Different types of MCA-BAVMs have unique imaging characteristics. Due to the limited space, type I BAVMs are smaller than type II and type IIIb BAVMs; in contrast, due to sufficient space, type IIIb BAVMs were larger than type I and type IIIa BAVMs. Because type I BAVMs have a deep location and type IIIb BAVMs have a large size, they can be close to the ventricle, and the deep vein is more often involved as the draining vein in them than other types of BAVMs. After rupture, IVH involvement was common. In addition, the remodeling of the primary draining vein in various types was different; the draining vein of Type IIIa BAVMs tended to retain normal morphology (P value > 0.05) (Table 5) because their only feeder is the MCA, the blood flow is low, and they are located on the hemisphere surface. The superficial veins often act as drainers; given their short course to the dural sinus, the low resistance reduces the probability of cortical vein remodeling [36]. However, in type IIIb BAVMs, multiple feeding arteries and a large nidus induced high-flow blood to remodel the primary draining vein.
For ruptured BAVMs and unruptured BAVMs with weak structures, intervention can be considered [10, 37]. EVT can be performed as a curative or useful adjunct treatment for MCA-BAVMs. Different types of BAVMs have different levels of difficulty and degrees of embolization during EVT. For Type I BAVMs, EVT poses a unique challenge due to the difficulty of LSA catheterization and the small diameter of the LSA [38, 39]. EVT is also difficult for Type II Sylvian BAVMs because the feeding arteries are often slim and disordered [16]. Therefore, type I and II BAVMs had low embolization degrees (P value > 0.05) (Table 5). However, Type IIIa BAVMs are easily treated with EVT and obtain a high embolization degree because of the simple angioarchitecture and ease of catheterization.
When performing EVT for BAVMs, the goal should not be an embolization degree, as the role of curative embolization is uncertain [40, 41]. In addition, in large BAVMs, the embolization of too large a volume in one stage may result in normal perfusion pressure breakthrough or occlusion of the draining vein, resulting in hemorrhagic complications [16, 17]. Therefore, in our study, the percentage of MCA-BAVMs with an embolization degree of > 2/3 nidus was only 36.3%. In addition, twenty risk-associated aneurysms were embolized. Therefore, the EVT results in this study were acceptable. In this study, 70.4% of patients with MCA-BAVMs suffered intracranial hemorrhage. In general, supratentorial hematoma can be tolerated. Therefore, EVT can be performed first, and then the hematoma can be managed with several alternatives, including conversative treatment and hematoma evacuation together without/with BAVM removal. In this study, 21.5% of patients experienced hematoma evacuation, and a good outcome with GOS scores of 4 and 5 was achieved in 97% of patients.
