Investigating Associations Between Signal Intensity of Silent MR Angiography and Hemodynamic Information From 4D Flow MRI in Brain Arteriovenous Malformation.

Background

Whether signal intensity (SI) on Silent MRA reflects hemodynamics in brain arteriovenous malformations (BAVMs) is unclear.

Objectives

To evaluate the relationship between Silent MRA SI and 4D flow-derived hemodynamic parameters and examine whether Silent MRA-based metrics are associated with symptomatic hemorrhage in BAVM. STUDY TYPE Prospective and retrospective cohort study. SUBJECTS Patients with BAVM who underwent DSA and MRI, including a hemodynamic cohort with Silent MRA and 4D flow (40 MRI from 28 patients: 28 baseline and 12 post-embolization follow-up), an untreated hemorrhage analysis cohort with Silent MRA (69 patients, 26 with symptomatic hemorrhage), and 16 healthy participants. FIELD STRENGTH/SEQUENCE 3.0T MRI; Silent MRA (arterial spin labeling with zero echo time readout) and electrocardiogram-gated 4D flow. ASSESSMENT Silent MRA SI and 4D flow-derived average blood flow (flowavg), maximum through-plane velocity (Vmax), and average through-plane velocity (Vavg) were measured in intracranial arteries of healthy participants and in draining veins (DVs), feeding arteries (FAs), and non-feeding arteries (non-FAs) of patients. A venous burden index (VBI) was defined as normalized DV SI divided by the fourth power of DV radius. Associations between Silent MRA factors and hemorrhage were assessed before and after adjustment for age, sex, and angioarchitecture. Measurements were performed independently by two readers (15 and 5 years of experience). STATISTICAL TEST Pearson or Spearman correlation, ANOVA, logistic regression; significance level p < 0.05.

Results

Silent MRA SI correlated with flowavg in DVs (r = 0.53), non-FAs (r = 0.42), and healthy arteries (r = 0.61), but not in FAs (p = 0.38), which exhibited higher velocities than other vessels (Vavg: FAs 27.93 ± 10.32, non-FAs 18.59 ± 7.54, DVs 12.42 ± 5.59). In hemorrhage analysis, DV ectasia (OR 0.25; 95% CI 0.09-0.70) and VBI (OR 1.42; 95% CI 1.15-1.98) were significant on univariable analysis; after multivariable adjustment, only VBI remained significant (OR 1.38; 95% CI 1.08-2.05). DATA

Conclusion

Silent MRA SI reflected 4D flow hemodynamics in non-FAs and DVs, but not in higher-velocity FAs. Silent MRA-derived VBI showed potential for noninvasive hemorrhage risk stratification in BAVM. EVIDENCE LEVEL 3. TECHNICAL EFFICACY STAGE 2.