LAUSR

OBJECTIVE To examine a potential association between intranidal microhemodynamics and rupture using a phase-contrast magnetic resonance angiography (PCMRA)-based flow quantification technique in arteriovenous malformations (AVMs). METHODS We retrospectively collected data on 30 consecutive patients with AVMs (23 unruptured and 7 ruptured). Based on PCMRA data, maximal (Vmax) and mean (Vmean) intranidal velocities were calculated. Logistic regression analysis was performed to assess factors associated with previous AVM rupture. RESULTS All ruptures occurred within six months before PCMRA. The mean nidus volume 4.7 mL. Eleven patients (37 %) had deep draining vein(s), and six patients (20 %) had a deep-seated nidus. The mean ± standard deviation Vmean and Vmax were 9.6 ± 2.8 cm/s and 66.7 ± 26.2 cm/s, respectively. The logistic regression analyses revealed that higher Vmax (P = 0.075, unit odds ratio [OR] = 1.05, 95% confidence interval [95%CI] = 1.00-1.10) was significantly associated with prior hemorrhage. The receiver-operating curve analyses demonstrated that a Vmean of 10.8 cm/s (area under the curve = 0.671) and Vmax of 90.2 cm/s (area under the curve = 0.764) maximized the Youden Index. A Vmax > 90 cm/s was significantly associated with AVM rupture both in the univariate (P = 0.025, OR = 9.0, 95%CI = 1.3-61.1) and multivariate (P = 0.008, OR = 51.7, 95%CI = 2.8-968.3) analyses. CONCLUSIONS Presence of faster velocities in intranidal vessels may suggest aberrant microhemodynamics and thus be associated with AVM rupture. PCMRA-based velocimetry seems to be a promising tool to predict future AVM rupture.