Multimodality Management of Spetzler-Martin Grade 3 Brain Arteriovenous Malformations with Subgroup Analysis

Objective: Spetzler-Martin grade 3 (SM3) lesions entail 4 distinct subtypes described based on size, eloquence, and deep venous drainage (3A-3D). The ideal management of each is contentious, and the results of A Randomized Trial of Unruptured Brain AVMs (ARUBA) introduced additional controversy and attention toward management strategies of unruptured brain arteriovenous malformations (bAVMs).

Methods: We retrospectively reviewed 114 patients with treated SM3 bAVMs, including both ruptured and unruptured lesions. Primary outcomes included modified Rankin score at most recent follow-up, angiographic cure, and permanent treatment-related complications (morbidity). Other outcomes included mortality, bAVM recurrence or rebleed, and transient treatment-related complications. We used univariate and multivariate modeling to determine whether any specific features were predictive of outcomes. For unruptured bAVMs, an "ARUBA eligible" subgroup analysis was performed. We also reviewed the literature on management of ruptured and unruptured SM3 bAVMs.

Results: Of the 114 identified SM3 bAVMs, 40% were unruptured. Most (43.5%) lesions in the unruptured group were type 3C, whereas most ruptured bAVMs (66.2%) were type 3A. Unruptured lesions were mostly managed with radiosurgery (47.8%) and ruptured ones with preoperative embolization and surgery (36.7%). Surgical intervention was predictive of angiographic cure in multivariate modeling, even after controlling for ≥2 years of follow-up, although associated with a slightly higher rate of morbidity. Focal neurological deficit was the only predictor of a worse (modified Rankin score ≥2) functional outcome in follow-up for unruptured bAVMs. For ruptured bAVMs, superficial and cerebellar locations were predictive of better outcomes in multivariate models, in the absence of a focal neurological deficit at presentation and new after surgery deficit. ARUBA SM3 bAVMs specifically underwent more embolization as a monotherapy and less microsurgical resection than the present series.

Conclusions: In spite of a heterogeneous array of angioarchitectural and anatomic features, SM3 bAVMs can be treated safely and effectively with surgery and radiosurgery either without or with pretreatment embolization. Ruptured lesions are more often type 3A, with smaller nidus, deep brain location, and deep venous drainage. Focal neurological deficit predicts worse clinical outcomes. Contemporary multimodality management of SM3 bAVMs is not adequately represented in the results of ARUBA, likely due to differences in treatment strategies.