Blood flow velocity measurement with phase contrast magnetic resonance imaging (PC-MRI) is widely applied in clinical routine imaging. Usually, velocity and volumetric flow measurements are performed using unidirectional encoding of the through-plane velocity with a 2-dimensional (2D) acquisition. Single-slice acquisitions and measurements with unidirectional encoding, however, may lead to significant errors, especially in tortuous vessels, but might benefit from higher signal-to-noise ratios (SNRs). To evaluate the impact of volumetric acquisition and multidirectional velocity encoding, blood velocity measurements were performed at 3 locations in the distal internal carotid artery with a 3-dimensional, 3-directional time-resolved phase contrast (PC) sequence (4-dimensional [4D]) and a 2D acquisition with 3-directional (2D-3dir) and through-plane velocity encoding (2D-tp) derived from the same sequence.
Twenty carotid arteries of 10 healthy volunteers (24–37 years) were evaluated. For each volunteer, 1 4D acquisition and 3 2D 3-directional PC measurements were placed according to a time-of-flight angiography. Unidirectionally encoded through-plane velocities were derived from the multidirectionally encoded 2D scan by discarding the in-plane components. Regions of interest were identified on the slab after postprocessing and visualization for the 4D data set as well as directly on the digital imaging and communications in medicine images for the 2D measurement. Blood flow velocity, volumetric flow, and SNRs were measured at carotid segments C4, C5, and C7 on both sides obtaining 20 values per vessel location. The quantities were tested for significant differences between each modality at all 3 locations with paired t tests.
At the segments C5 and C7, the highest peak velocities (PVs) were measured with the 4D sequence, followed by 2D-3dir and 2D tp. The PV differences between the sequences were significant (P < 0.01) at both locations. At the proximal segment of the carotid siphon (C4), the PV values of the 2D-3dir sequence were significantly higher than the ones measured with 2D-tp. The mean PV value of the 4D sequence was located in between 2D-3dir and 2D-tp without significant differences to either of the 2D sequences. Volumetric flow measurements were also significantly different between 2D and 4D acquisitions, but without a discernible trend. The SNR analysis clearly favored 2D over 4D acquisitions because of higher inflow enhancement.
The results of the current study show that velocity measurements with a unidirectional encoded through-plane PC sequence lead to a significant underestimation of velocity values in tortuous vessels. In all 3 evaluated segments of the distal internal carotid artery, multidirectional velocity encoding revealed significantly higher PV values than those of unidirectional velocity encoding. These results indicate that multidirectional encoding should be preferred to unidirectional encoding for velocity measurements in tortuous vessels. Furthermore, 4D PC-MRI is superior to 2D-3dir in 2 of 3 locations. However, single-slice measurements with multidirectional velocity encoding have higher SNR and may be an alternative to 4D PC-MRI with a scan time of only approximately 90 seconds per slice.
From the Divisions of *Neuroradiology, and †Radiological Physics, Clinic of Radiology and Nuclear Medicine, Basel University Hospital, Basel, Switzerland.
Received for publication April 10, 2013; and accepted for publication, after revision, October 1, 2013.
Conflicts of interest and sources of funding: none declared.
Reprints: Tilman Schubert, MD, Division of Neuroradiology, Clinic of Radiology and Nuclear Medicine, Basel University Hospital, Petersgraben 4, 4031 Basel, Switzerland. E-mail: Tilman.Schubert@usb.ch.