4D flow cardiovascular magnetic resonance consensus statement

Dyverfeldt, Petter; Bissell, Malenka; Barker, Alex J.; Bolger, Ann F.; Carlhäll, Carl-Johan; Ebbers, Tino; Francios, Christopher J.; Frydrychowicz, Alex; Geiger, Julia; Giese, Daniel; Hope, Michael D.; Kilner, Philip J.; Kozerke, Sebastian; Myerson, Saul; Neubauer, Stefan; Wieben, Oliver; Markl, Michael

doi:10.1186/s12968-015-0174-5

Journal of Cardiovascular Magnetic Resonance

Table 2 4D Flow CMR scan parameters

From: 4D flow cardiovascular magnetic resonance consensus statement

	Ideally	Reason	Limiting factor	Consensus value
Acquisition Parameters
Field of view	Max	SNR, coverage	Scan time, system imperfections	Cover region of interest
Spatial resolution^a	Maximum, at least 5–6 voxels across vessel diameter of interest^b, isotropic resolution.	Accuracy	Scan time, SNR	<2.5×2.5×2.5 mm³ for aorta or pulmonary artery
Spatial resolution^a		Accuracy	Scan time, SNR	<3.0×3.0×3.0 mm³ for whole heart and greater vessels
Velocity encoding timing (beat- vs. TR-interleaved)	TR-interleaved	Avoid inter-cycle variability	Temporal resolution	TR-interleaved
k-space segmentation factor	1	Accuracy (temporal resolution)	Scan time	2
Temporal resolution^c	Max	Accuracy	Scan time	<40 ms
ECG synchronization^d	Retrospective	Cover entire ECG cycle, avoid sequence interruption	Reconstruction complexity	If available: retrospective
ECG synchronization^d	Retrospective	Cover entire ECG cycle, avoid sequence interruption	Reconstruction complexity	Else: Prospective^e
Respiratory motion compensation^f	100 % acceptance, motion correction	Scan time, reduction of breathing artifacts	Reconstruction complexity, robustness, breathing artefacts (ghosting and blurring)	If available: Leading or trailing MR navigator on liver/diaphragm interface, 6 mm window size, typically resulting in 50 % acceptance rate.
Respiratory motion compensation^f	100 % acceptance, motion correction	Scan time, reduction of breathing artifacts		Otherwise: Bellows with 50 % acceptance rate.
Partial k-space coverage in phase- and slice-encoding directions	Full k-space coverage	SNR, resolution	Scan time	If available: Elliptical k-space
	Full k-space coverage	SNR, resolution	Scan time	Otherwise: Half scan 75 % × 75 % (y × z)
Flip Angle^g	Ernst angle: α = acos(e^-TR/T1)	SNR	Contrast vs. SNR	Ernst angle
Parallel Imaging	No parallel imaging	SNR	Scan time	R = 2-3 (depends on #channels in coil array)
k-t undersampling^h	No k-t under sampling	SNR	Scan time	If available: R = 4-5
Venc	Maximum expected velocity, multiple vencs	VNR, avoid aliasing	Scan time	Single venc, 10 % higher than maximum expected velocity
Postprocessing Parameters
Maxwell correction	Yes	Accuracy		Yes
Eddy current correction	Yes	Accuracy	Different methods and their validity and robustness	Yes
Phase unwrapping	Yes	Accuracy	Different methods and their validity and robustness	Yes
Gradient non-linearity correction	Yes	Accuracy	Availability	If available

^aAlways indicate the effectively acquired resolution in combination with the interpolated resolution
^bStudies have demonstrated that 5–6 voxels across the vessel diameter is sufficient for flow volume quantification [165]
^cAlways indicate the effectively acquired resolution. If a temporal interpolation is performed, also indicate the interpolated temporal resolution along with the interpolation method used
^dSo called self-gating techniques have been evaluated and may become an alternative to the ECG [32]
^eFor prospective gating, analyses that involve integration over the whole cardiac cycle needs to be accompanied with a description of how the incomplete temporal coverage was handled
^fDifferent types of respiratory navigators exist; variants include approaches that allow less motion in the central parts of k-space. Always describe the method that has been used and indicate the mean navigator efficiency in percent as well as the navigator acceptance window in mm. For fix window sizes and no k-space reordering, 6 mm navigator window is recommended, and this typically results in 50 % navigator efficiency
^gThe SNR is strongly dependent on the in-flow effect, therefore the flip angle can be and is often chosen higher than the Ernst angle. When using contrast agents, the Ernst angle further increases (due to lower T₁)
^h k-t undersampling factor 4–5 in combination with conventional parallel imaging factor 2–3 is not recommended

Back to article page

ISSN: 1532-429X

Contact us

Submission enquiries: Access here and click Contact Us
General enquiries: info@biomedcentral.com