Microsurgical resection for eloquent arteriovenous malformations (AVMs) remains challenging. Currently, there are only two grading systems concerning pretreatment assessment of brain AVMs: the Spetzler-Martin grading system proposed by Spetzler and Martin in 1986 and the supplementary grading system proposed by Lawton in 2010. Controversies exist regarding the treatment timing and treatment modalities for eloquent AVMs. Till now, there is no clinical trial concerning the efficacy of multimodal magnetic resonance imaging techniques in assessing the surgical risk for eloquent AVMs. The investigators assume that multimodal imaging-based grading system is superior to the classic Spetzler-Martin grading system and the supplementary grading system in predicting the surgical risk for eloquent AVMs.

Microsurgical resection for eloquent arteriovenous malformations (AVMs) remains challenging. Currently, there are only two grading systems concerning pretreatment assessment of brain AVMs: the Spetzler-Martin grading system proposed by Spetzler and Martin in 1986 and the supplementary grading system proposed by Lawton in 2010. Controversies exist regarding the treatment timing and treatment modalities for eloquent AVMs. Till now, there is no clinical trial concerning the efficacy of multimodal magnetic resonance imaging techniques in assessing the surgical risk for eloquent AVMs. The investigators assume that multimodal imaging-based grading system is superior to the classic Spetzler-Martin grading system and the supplementary grading system in predicting the surgical risk for eloquent AVMs.

This study consists of two parts: a retrospective analysis and a prospective study. Firstly, the investigators will retrospectively review the 250 AVM patients that were surgically treated at Beijing Tiantan Hospital between June 2012 and June 2015. All data of these patients were prospectively maintained in our AVM database. All these patients had preoperative functional magnetic resonance imaging (fMRI) studies. The investigators will collect the patient demographic data, AVM features, the least distances from the AVMs to the activated cortex as well as to the fiber tracts. The investigators will also collect the patient preoperative functional status (modified Rankin scale score, mRS) and the functional status (mRS) six months after surgery. The changes between the two functional status in mRS will be classified into two groups: those with worsened mRS (mRS score six months after surgery - presurgical mRS score >0) and those with unchanged or improved mRS (mRS score six months after surgery - presurgical mRS score ≤0). The new grading system will be proposed based on the multimodal magnetic resonance imaging (MRI). The investigators will compare the predictive accuracy of the three grading systems (the Spetzler-Martin grading system, the supplementary system and our new grading system) in assessing functional status (worsened mRS). Then in the prospective study, the investigators will enroll 400 surgically treated AVM patients from five neurosurgical centers. All patients will meet the inclusion and exclusion criteria. The preoperative preparation, the surgical procedures, the follow-up period, the primary outcomes and statistical analyses will be the same as those in the retrospective study part. Multimodal imaging-based system will be verified in this prospective phase. If the results demonstrate that the multimodal MRI-based grading system is superior to the Spetzler-Martin grading system and the supplementary system in predicting the functional status six months after AVM surgery, the investigators will propose the new AVM grading system in the public.

Inclusion Criteria:

1. Patients aged between 12 and 60 years
2. Data from preoperative structural MRI, blood oxygen level dependent (BOLD)-fMRI, time-of-flight magnetic resonance angiography (TOF-MRA), and DTI of the motor tract,language tract or optic radiation were available
3. Patients with a definite diagnosis of AVM confirmed by preoperative digital subtraction angiography
4. Patients opting for surgical management at the five centers in this study
5. Patients without any interventional therapy (microsurgery, radiosurgery, endovascular embolization, or multimodality treatment) before admission to the five centers

Exclusion Criteria:

1. Patients with acute intracranial hematoma and resultant brain hernia prompting emergency surgery
2. Patients with a hemorrhagic incident in the past month
3. Nonavailability of BOLD-fMRI and DTI data
4. Patients with other severe diseases that prevented them from having surgery
5. Patients unwilling to participate in the trial