Hemodynamic Quantification in Brain Arteriovenous Malformations With Time-Resolved Spin-Labeled Magnetic Resonance Angiography
Background and Purpose—Unenhanced time-resolved spin-labeled magnetic resonance angiography enables hemodynamic quantification in arteriovenous malformations (AVMs). Our purpose was to identify quantitative parameters that discriminate among different AVM components and to relate hemodynamic patterns with rupture risk.
Methods—Sixteen patients presenting with AVMs (7 women, 9 men; mean age 37.1±15.9 years) were assigned to the high rupture risk or low rupture risk group according to anatomic AVM characteristics and rupture history. High temporal resolution (<70 ms) unenhanced time-resolved spin-labeled magnetic resonance angiography was performed on a 3-T MR system. After dedicated image processing, hemodynamic quantitative parameters were computed. T tests were used to compare quantitative parameters among AVM components, between the high rupture risk and low rupture risk groups, and between the hemorrhagic and nonhemorrhagic groups.
Results—Among the quantitative parameters, time-to-peak (P<0.001) and maximum outflow gradient (P=0.01) allowed discriminating various intranidal flow patterns with significantly different values between feeding arteries and draining veins. With 9 AVMs classified into the high rupture risk group (whose 6 were hemorrhagic) and 7 into the low rupture risk group, the observed venous-to-arterial time-to-peak ratio was significantly lower in the high rupture risk (P=0.003) and hemorrhagic (P=0.001) groups.
Conclusions—Unenhanced time-resolved spin-labeled magnetic resonance angiography allows AVM-specific combined anatomic and quantitative analysis of AVM hemodynamics.
To better understand the clinical presentation and the bleeding risk of arteriovenous malformations (AVMs), and thus optimize individual patient management, AVM characterization needs to be further refined beyond usual anatomic features and rupture history.1 Characterization of the AVM flow hemodynamics could be a promising tool for improved AVM classification.
Several noninvasive imaging methods for hemodynamic quantification exist, such as time-resolved contrast-enhanced magnetic resonance angiography (MRA), yet with temporal resolution limited to ≈1 seconds,2,3 or unenhanced time-resolved phase-contrast MRA, yet with a long acquisition time and no hemodynamic visual analysis.4,5 Recently, unenhanced time-resolved spin-labeled MRA (4D-SL-MRA) has been reported to allow anatomic AVM characterization reliably3 and to overcome previous limitations by achieving high temporal resolutions of 50 to 100 ms,6–8 thus reducing vessel superposition9 and increasing the accuracy of further quantitative analysis.
Our study aimed at using 4D-SL-MRA to identify quantitative parameters that allow discrimination among different AVM components and to relate hemodynamic patterns with rupture risk.
Sixteen patients presenting AVM were included into the study after approval from the institutional review board. Anatomic AVM characterization was performed using digital subtraction angiography as the reference standard (associated with T1-weighted postgadolinium chelate for nidus size and localization). According to the Columbian Database,1 patients were assigned to the high rupture risk group if ≥1 of the following anatomic factors were present: history of ruptures (including asymptomatic bleeding signs on MRI), or exclusively deep venous drainage on digital subtraction angiography, or deep location. Other patients were assigned to the low rupture risk group. Another grouping distinguished hemorrhagic and nonhemorrhagic patients.10
Imaging was performed on a 3-T system (MAGNETOM Verio, Siemens Healthcare) with a 32-channel head array coil. The 4D-SL-MRA sequence used an optimized acquisition window for 2 cardiac cycles.3 The acquisition time was 7 minute, with a temporal resolution of 68 ms and an 1.5-mm isotropic spatial resolution (repetition time/echo time: 68.2/2.13 ms; field of view: 220×193×66 mm3).
Hemodynamic parameters were derived from the signal intensity curves in each voxel to generate color-coded maps (Figure 1), with a dedicated image processing software developed using Matlab (Mathworks Inc).
The following regions of interest were defined: the main arterial feeder and draining vein as close to the nidus as possible, the nidus center, and the contralateral artery corresponding to the main arterial feeder. Ratios of values were computed for time-to-peak ratio (TTP) and mean transit time.2
A 2-tailed paired t test was performed to compare, for each quantitative parameter, the values measured in the arterial feeder, the nidus and the draining vein, and in the contralateral artery. A 2-sided t test was used to compare patient groups for each quantitative parameter. P<0.05 was considered significant.
Clinical presentation and anatomic AVM characterization are presented in the Table.
Hemodynamic Discrimination Among AVM Components
Quantitative parameter color maps revealed various quantitative values among AVM components (Table I in the online-only Data Supplement). The difference was significant between feeding arteries and draining veins for TTP (P<0.0001) and maximum outflow gradient (P=0.010), and between feeding arteries and nidus for TTP (P<0.0001), as well as maximum outflow gradient (P=0.004) and mean outflow gradient (P=0.021). In addition, the values significantly differed between feeding arteries and normal contralateral arteries for all parameters except TTP, with the lowest P values noted for maximum outflow gradient (P=0.003) and maximum signal intensity (P=0.001; Figure 2A). Various intranidal flow patterns were revealed in large nidus, particularly when using TTP and maximum inflow and outflow gradient parameters (Figure 2A and 2B).
Hemodynamic Patterns and Rupture Risk
The comparison of hemodynamic patterns according patients grouping (Table II in the online-only Data Supplement) demonstrated that the computed draining-vein-to-arterial-feeder TTP ratio displayed significantly lower values in the high rupture risk group than in the low rupture risk group (1.6±0.2 versus 2.2±0.4; P=0.003), as well as in the hemorrhagic compared with the nonhemorrhagic group (1.5±0.1 versus 2.1±0.4; P=0.001). TTP ratio ≤1.7 was recorded in all patients with previous hemorrhage and TTP ratio ≤2.0 in all high rupture risk patients (Table). No significant difference between groups was observed for the other quantitative parameters.
4D-SL-MRA enables the combination of AVM-specific anatomic and quantitative analysis of AVM hemodynamics. TTP and maximum outflow gradients computation permit discrimination among different AVM components. This is allowed by the high temporal resolution, <70 ms, of 4D-SL-MRA sequence. Moreover, as suggested in previous reports,2,11 hemodynamic quantification could be related to the clinical presentation because a low draining-vein-to-feeder-artery TTP ratio was associated with the high rupture risk and hemorrhagic groups. This lower TTP ratio, <2.0 in all cases, may reflect higher velocities and exposure to higher vascular pressure12 within the nidus in patients with previous hemorrhage or known angioarchitectural bleeding risk factor. It may be hypothesized that outliers like the 2 patients classified in the low rupture risk group but exhibiting a low TTP ratio may present higher rupture risk than presumed based on anatomic characteristics. This illustrates the added value of hemodynamic criteria for assessing rupture risk.
The limitations of the study are the small sample size and the fact that the observed patterns in ruptured AVM may be related to the bleeding event rather than being predictive of bleeding risk.2 Large-scale prospective studies are required to confirm the relevance of the hemodynamic classification.
In conclusion, hemodynamic analysis with 4D-SL-MRA may be instrumental in improving AVM classification and enable individualized management.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.006080/-/DC1.
- Received May 8, 2014.
- Revision received May 8, 2014.
- Accepted May 30, 2014.
- © 2014 American Heart Association, Inc.
- Illies T,
- Forkert ND,
- Saering D,
- Wenzel K,
- Ries T,
- Regelsberger J,
- et al
- Ansari SA,
- Schnell S,
- Carroll T,
- Vakil P,
- Hurley MC,
- Wu C,
- et al
- Chang W,
- Loecher MW,
- Wu Y,
- Niemann DB,
- Ciske B,
- Aagaard-Kienitz B,
- et al
- Todaka T,
- Hamada J,
- Kai Y,
- Morioka M,
- Ushio Y