Challenges in the Management of Ruptured and Unruptured Brainstem Arteriovenous Malformations: Outcome After Conservative, Single-Modality, or Multimodality Treatments

Abstract

Abstract BACKGROUND Brainstem arteriovenous malformations are challenging lesions, and benefits of treatment are uncertain. OBJECTIVE To study the clinical course of Brainstem arteriovenous malformations and the influence of treatments on outcome. METHODS We reviewed a prospective series of 31 brainstem arteriovenous malformations. Demographic, morphological, and clinical characteristics were recorded. Factors determining initial and final outcomes (modified Rankin Scale), results of treatments (cure rates, complications), and disease course were analyzed. RESULTS Brainstem arteriovenous malformations were symptomatic and bled in 93% and 61% of cases, respectively. Examination was abnormal and initial modified Rankin Scale score was > 3 in 71% and 86% of patients, respectively. The average follow-up time was 6.2 years, and 26% of patients rebled (5.9 %/y). Treatment modalities included conservative, radiosurgical, endovascular, surgical, and multimodality treatment in 13%, 58%, 35%, 16%, and 26% of cases, respectively. The obliteration rate was 60% overall and 39% after radiosurgery, 40% after embolization, and 75% after microsurgery, with respective complication-free cure rates of 71%, 50%, and 0%. Overall procedural mortality and morbidity were 2.3% and 18.6%, respectively. Final modified Rankin Scale score was > 3 in 77% of cases. Neurological deterioration (35%) was related to treatment complications in 74% of cases with a negative impact of surgery (P = .04), palliative embolization (odds ratio = 16), and multimodality treatments (odds ratio = 24). Radiosurgery was inversely associated with worsening (odds ratio = 0.06). CONCLUSION Brainstem arteriovenous malformations require individualized treatment decisions. Single-modality treatments with a reasonable chance of complete cure and low complication rate (such as radiosurgery) should be favored.