Multimodality Treatment of Deep-seated Cerebral Arteriovenous Malformations: The Experiences of One Center

Objective: Treatment of deep-seated cerebral arteriovenous malformations (AVMs) remains challenging for neurosurgeons or neuroradiologists. This study aims to review the experiences of one center in using multimodality treatment for deep-seated AVMs. Methods: The AVM database of Xuanwu Hospital, Capital Medical University was searched, and 96 patients who were diagnosed with a deep-seated cerebral AVM between 2010 and 2020 were identified. The following information was collected and analyzed: patients’ clinical features, treatment modality used, posttreatment complications, AVM obliteration rate, rebleeding rate, and functional outcome during follow-up. The patients’ posttreatment modified Rankin scale (mRS) scores were split into two groups: good outcome (mRS score ≤ 2) and poor outcome (mRS score ≥ 3). Univariate and multivariate logistic regression analyses were applied to test the predictors of clinical outcomes and AVM obliteration. Results: Eighty-eight out of 96 patients (91.7%) presented with initial hemorrhaging. The pretreatment mRS score was ≤ 2 in 80 patients (88.3%) patients and ≥ 3 in 16 patients (16.7%). Limb weakness was present in 42 patients (43.8%). In this sample, 210 hemorrhages occurred during 2056 person-years before diagnosis, yielding an annual hemorrhage rate of 10.2% per person-year. Angiographic obliteration was achieved in 29 patients (30.2%). At the last follow-up, 80 patients (83.3%) had good clinical outcomes, whereas 16 (16.7%) had a deterioration in their clinical presentation following treatment. Multivariate analyses indicated that pretreatment limb weaknesses and a high Spetzler–Martin grade predicted poor clinical outcomes (P = 0.003 and 0.008, respectively). Fewer feeding arteries were a predictor for AVM obliteration (P = 0.034). Conclusion: Good outcomes can be achieved through multimodal treatment of deep-seated AVMs. A single supplying artery is a predictor of AVM obliteration. Pretreatment limb weaknesses and high Spetzler-Martin grades predict poor clinical outcomes.