Correlation of Cerebral Arteriovenous Malformation Flow with Magnetic Resonance-detected Microhemorrhage

Background: A cerebral AVM (arteriovenous malformation) is a congenital or acquired vascular anomaly that shunts blood from dysplastic arteries to veins through a vascular nidus. Both genetic and environmental triggers have been proposed to instigate the development of brain Arteriovenous malformations (bAVMs). Endothelial dysfunction, primarily induced by high shear stress from increased nidal blood flow, possibly promoting a cycle of inflammation, leading to instability and bAVM rupture. Macrophages are identified as key inflammatory components in bAVM pathology and can be identified with the common marker CD68. Intracranial hemorrhage (ICH) significantly contributes to morbidity and mortality in cerebral arteriovenous malformations (AVM) patients, with a history of prior hemorrhage being the most significant predictor of a future hemorrhage. Silent microhemorrhage as detected on magnetic resonance imaging (MRI) has been proposed as a potential risk factor for future hemorrhage in AVM patients. Silent microhemorrhage is associated with higher nidal velocity measured through color-coded DSA, raising the question of correlation between flow and microhemorrhage. Objective Assess the relationship between AVM flow (measured by quantitative magnetic resonance angiography, QMRA) in non-ruptured AVMs with MR-detected microhemorrhage. The pilot project’s objective is to evaluate the relationship of inflammation with bAVM flow measured and hemosiderin. Methods: All unruptured AVM patients with a baseline quantitative magnetic resonance (QMRA) imaging and gradient echo or susceptibility-weighted MRI acquired prior to any embolization, radiosurgery or microsurgery were retrospectively reviewed (2004-2022) (n = 89). AVM flow was calculated from the aggregate flow within primary arterial feeders relative to their contralateral counterparts. A review of the MRI determined the presence of microhemorrhages. Descriptive statistics, the X2 test, and a binomial logistic regression were performed to test the association of demographics, clinical, and AVM features with microhemorrhage. For the pilot project, adult patients at the University of Illinois Hospital (2002-2022) with baseline quantitative magnetic resonance (QMRA) imaging, and microsurgical resection prior to embolization and radiosurgery were retrospectively reviewed(n=17). Brain AVM sections were stained with CD68 to quantify vessel wall macrophage infiltration and hematoxylin and eosin stain to as a control and to quantify hemosiderin. CD68 was quantified by averaging the positive stained cell count over vessel wall area for the 15 largest vessel walls in each section. Hemosiderin was graded on a previously reported scale of 0 to 4. QMRA with noninvasive optimal vessel analysis (NOVA) was retrospectively reviewed and AVM flow was calculated from the aggregate flow within primary arterial feeders relative to their contralateral counterparts. Descriptive statistics, independent t-tests, Pearson’s correlation, and a one-way ANOVA were performed to test the association of demographics, clinical and AVM features with CD68. Results: Of 634 patients with cerebral AVMs at a single center, 89 patients met the inclusion criteria (54 with micro- hemorrhage and 35 without microhemorrhage). The calculated AVM flow was significantly higher in the group with a microhemorrhage (447.9 ± 193.1 ml/min vs 287.6 ± 235.7 ml/min, p = 0.009), and remained significant when controlling for venous anomaly in binary logistic regression(OR 1.002, 95% CI 1.000–1.004; p = 0.031). Venous anomaly, arterial ectasia, and diffuse nidus are significantly associated with microhemorrhage (p = 0.017, p = 0.041, and p = 0.041, respectively), but not significant when included in logistic regression model. The small sample size and collinearity amongst our independent variables included in the model limit our ability to detect significant associations of the variables in the model with microhemorrhage. Our pilot study enrolled seventeen patients, (ruptured = 9, unruptured = 8). Vessel wall macrophage infiltration exhibited a positive association with patients who presented with confirmed AVM rupture (163.8 +/- 46.7 vs 101.3 +/- 49.4, p=0.017). Furthermore, graded increases in vessel wall macrophage infiltration were found to positively correlate with higher grades of hemosiderin (p=0.023), except for grade 4 hemosiderin. Conclusion: Patients with higher AVM flow have a higher likelihood of MR-detected microhemorrhage that remains significant when controlling for venous anomalies and nidus compactness. AVM-associated venous anomalies and diffuse nidus are associated with MR-detected microhemorrhage. Our pilot project suggests a relationship between AVM vessel wall inflammation, hemosiderin, and AVM rupture. Further investigations with larger sample sizes are warranted to fully understand the role of AVM hemodynamics, microhemorrhage and vessel wall inflammation.