Dynamic quantitative nonenhanced magnetic resonance angiography of the abdominal aorta and lower extremities using cine fast interrupted steady-state in combination with arterial spin labeling: a feasibility study

Peripheral artery disease (PAD) is a major health issue worldwide, affecting approximately 202 million people globally [1]. Imaging evaluation for those under consideration for revascularization was traditionally performed using catheter-directed digital subtraction angiography (DSA) [2]. In efforts to reduce the invasiveness and risk, physicians have turned to contrast-enhanced magnetic resonance angiography (CEMRA) and computed tomography angiography (CTA) for non-invasive vascular evaluation [3‐5]. CTA is a static technique that provides little information about the hemodynamic status of the arteries, whereas time-resolved CEMRA can be used to provide a dynamic display of arterial flow patterns. However, time-resolved CEMRA has limited temporal resolution and primarily provides qualitative rather than quantitative evaluation of blood flow [6, 7]. Both CTA and time-resolved CEMRA require the administration of contrast agents, which may be contraindicated in patients with impaired renal function.

Cine fast interrupted steady-state in combination with arterial spin labeling (cFASL) is a recently-described nonenhanced MRA technique that provides a DSA-like display of in-plane flow patterns with high temporal resolution [8]. Rather than measure phase shifts that have been created by the flow of arterial spins through magnetic field gradients, which is the underlying principle for phase contrast methods, cFASL quantifies flow velocity based on measuring the temporally-resolved displacement of a radiofrequency (RF)-labeled bolus of arterial spins. Unlike 2-dimensional phase contrast (2DPC), cFASL is insensitive to the effects of partial volume averaging and enables in-plane flow patterns to be directly visualized. We hypothesized that cFASL could be used to efficiently display arterial hemodynamics over extensive lengths of the aorta and lower extremity peripheral arteries.

Scan time for the cFASL sequence was approximately 65 s per slice. cFASL of each vessel segment was typically acquired in 1–6 slices depending on the location and plane of imaging, which gives an overall scan time of 65–390 s. For unilateral evaluation, sagittal cFASL was more efficient than coronal imaging as fewer slices were needed to encompass the arteries, whereas imaging in the coronal plane permitted direct comparison of flow between limbs. In one subject, 2DPC through the upper thigh was not evaluable due to susceptibility artifact, so neither 2DPC nor cFASL data at this level was included in the analysis. Post-processing time was rapid (approximately 1 to 2 min per level) for both 2DPC and cFASL.

This preliminary study demonstrates that cFASL provides a physiologically accurate display of in-plane flow patterns over extensive lengths of the lower extremity peripheral arteries and abdominal aorta. The FISS readout which is key to this technique is a novel variant of the widely-used bSSFP pulse sequence. It combines a frequently-interrupted echo train with a radial k-space trajectory, while periodic application of gradient and radiofrequency spoiling suppresses flow artifacts from out-of-slice and off-resonant spins [11]. FISS builds upon prior work using interrupted steady-state sequences, such as the S⁵FP technique [12]. With this earlier technique, which utilized a Cartesian k-space trajectory, the frequent disruption of the steady-state signal could produce severe image artifacts. To avoid these artifacts, FISS uses a radial k-space trajectory with equidistant view angle increments. The benefit is decreased flow artifact and fat signal compared to traditional bSSFP, with greatly reduced flow saturation effects compared with fast low-angle shot (FLASH) [11]. The combination of FISS with ASL was recently described to enable the visualization of in-plane flow patterns within the major arteries of the chest and abdomen [8]. However, validation of cFASL for the measurement of in vivo flow velocities has not previously been described, nor has it been used to image vascular pathology involving the peripheral arteries or abdominal aorta.

In this study, both large and small vessels were evaluated with cFASL. In healthy subjects, the suprarenal abdominal aorta has a predominantly plug flow pattern whereas the smaller peripheral arteries have a predominantly laminar flow pattern [13]. In individuals without significant arterial pathology, we found that peak velocity measurements using cFASL correlated well with 2DPC for both plug and laminar flow patterns. This finding suggests that cFASL may be applicable for flow measurements in both large and small arteries.

For qualitative evaluation, cFASL reliably depicted in-plane flow patterns in both healthy and diseased vessels. For instance, visual analysis of time-resolved cFASL images in PAD patientsreadily demonstrated acceleration of the labeled bolus through hemodynamically-significant stenoses along with differential rates of bolus progression versus healthy vessels in the contralateral limb. Tortuous progression of the labeled bolus through collateral vessels, swirling flow in aneurysms, as well as differential progression of the labeled bolus through the true and false lumens of an aortic dissection were also readily apparent. However, for quantitative evaluation of flow velocities in the setting of significant arterial pathology, cFASL has limitations since the labeled bolus must remain substantially intact and trackable over the measurement period. For instance, due to dispersion of the labeled bolus from turbulence, we found that the technique was not well suited for flow measurements across hemodynamically-significant stenoses.

Unlike 2DPC, cFASL is largely insensitive to partial volume averaging due to the extreme level of background signal suppression. For instance, a single 27-mm thick slice allowed the semi-projective depiction of arterial flow patterns through extensive lengths of an aortic dissection despite the small caliber of the true lumen. Using such thick slices, one can image substantial lengths of a peripheral artery using a single cine image series, enabling a several-fold reduction in exam time compared with the predominantly multi-slice approach used in the current study. However, further study is needed to determine whether the accuracy of the flow velocity measurements is maintained with thick slices.

2DPC has been used since the 1980s for evaluation of pulsatile flow in the body [14, 15] and has been validated and applied to the lower extremities since the 1990s [16‐18]. While 2DPC allows accurate measurement of through-plane flow, considerations of scan time generally limit its use to a few slice locations. Moreover, it is suboptimal for imaging lower extremity arteries in a non-axial orientation due to partial volume averaging between flowing spins and stationary background tissue, which can cause erroneous velocity measurements.

We found that peak velocities measured by cFASL were consistently higher compared to 2DPC (at 28 of 38 stations). This discrepancy could be explained in part by the fact that, while 2DPC is a well-established technique for through-plane flow quantification [16‐18], it has been shown to consistently underestimate velocities compared to Doppler ultrasound [19]. Compared with 2DPC, 4D flow (i.e. 3D cine PC) would have provided a more accurate reference standard for flow quantitation. However, it was not used in the current study due to impractically long scan times (e.g. 10 to 15 min per station) and time-consuming image analysis [20]. Despite the limitations of 2DPC, we found very strong positive correlation and excellent agreement between cFASL and 2DPC for the quantification of peak flow velocities throughout the peripheral arterial territory in healthy vessels.

This feasibility study suggests that cFASL provides an efficient, high quality and physiologically accurate display of in-plane flow patterns over extensive lengths of the abdominal aorta and lower extremity peripheral arteries. Further studies will be required to determine the value of this technique in patients with peripheral arterial disease.