Relationship of blood flow, angioarchitecture, and rupture in cerebral arteriovenous malformations

OBJECTIVE: Cerebral arteriovenous malformations (AVMs) have a 2%-4% annual risk of intracranial hemorrhage. Understanding hemorrhage risk is crucial for determining appropriate treatment. The degree to which cerebral blood flow within AVMs interplays with angioarchitecture and contributes to potential rupture represents an active area of research. Prior studies have shown conflicting results; those based on quantitative MR angiography (QMRA) have found lower flow associated with hemorrhage, while those using angiography-determined transit time associated higher flow with rupture. METHODS: A prospectively collected institutional database of 647 adult patients with cerebral AVMs (1995-2023) was queried. Patients with baseline QMRA studies were included in the analysis. Patient characteristics, angioarchitectural information, and hemodynamic data were collected for each patient and compared between those who presented with and those who presented without hemorrhage. RESULTS: A total of 171 patients met the inclusion criteria; 105 patients had unruptured AVMs and 66 had ruptured AVMs at presentation. The mean age did not differ between groups (39.4 years vs 39.9 years), but there was male predominance in the rupture group (67.7%, p = 0.023). A deep AVM location was more likely in those with rupture (p = 0.002). The mean calculated AVM flow was lower in the rupture group (284.3 mL/min vs 401.8 mL/min, p = 0.013). Additionally, the mean AVM volume was smaller in the rupture group (8.8 cm3 vs 14.5 cm3, p = 0.040). The mean number of arterial feeders did not differ between groups (2.4 vs 2.4, p = 0.986), but the mean size of the largest feeder was smaller in those presenting with hemorrhage (2.6 vs 3.0 mm, p = 0.038). Furthermore, ruptured AVMs were more likely to have deep venous drainage (p = 0.007) and fewer draining veins (mean 1.9 vs 2.4, p = 0.020) than unruptured AVMs. The presence of feeder aneurysms, intranidal aneurysms, and nidal compactness did not differ between groups. CONCLUSIONS: Patients with baseline QMRA studies who presented with ruptured AVMs were noted to have a deep location, deep venous drainage, lower flow, smaller nidal volume, smaller arterial feeders, and fewer draining veins compared with those with unruptured AVMs. These findings validate results from smaller prior studies based on QMRA.