This course targets scientists and clinicians interested in MR Angiography (MRA). We will seek to offer a comprehensive overview of the limitations of MRA and the technical solutions implemented. We will go thought the identification of spurious artifacts in MRA so that the audience can understand the underlying sources of artifacts and the consequent proposals to mitigate their manifestations. There is no requirement for this course, but a minimum knowledge in MRA basic principles is preferable.

The course will cover both Contrast-enhanced MRA (CE-MRA) and Non-contrast enhanced MRA (NCE-MRA). In MR Angiography, it is essential to establish the reliability of the technique and pursue a robust assessment of the vascular tree. MRA is limited by the inherent acquisition capabilities of the technique, mainly: speed, resolution, noise tolerance and motion management. Additionally, the complexity of an MRA experiment prevents repetitive measurements so that clinicians set a zero tolerance for artifacts in MRA: in CE-MRA, a second injection is not acceptable that each bolus should succeed in the visualization of the vessels of interest; similarly in NCE-MRA, the duration of each acquisition renders each repetition often prohibitive within a complete MR examination.

Under these considerations, the occurrence of artifacts stands as an additional constraint that conditions the choice of the MRA technique and its parameters for the imaging of a target organ. The identification of artifacts in MRA can be split into 3 steps: 1/the understanding of physical interferences throughout the acquisition and/or reconstruction, 2/the identification of the artifact in the images and 3/ the proposal of a sustainable solution to circumvent the artifact in routine vascular evaluation.

The course will introduce artifacts through examples. The audience will be encouraged to participate through a little guessing game. Information regarding the acquisition technique will be provided for context. Each artifact will be thoroughly explained and basic physics principles will be covered whenever necessary. Technical proposed solutions will be explored and corresponding references will be provided. The non-exhaustive list of artifacts we will cover can be split into 4 categories: 1/ Set-up artifacts: timing, Gadolinium concentration, slice/slab positioning, B0 and B1 inhomogeneity, wrap around of outer-volume vessels, inaccurate trigger delays. 2/ Acquisition artifacts: susceptibility artifacts (metal, air interfaces), motion, flow and pulsatility effects, slices alignment, poor spoiling, missed triggers. 3/ Reconstruction artifacts: partial volume, parallel imaging and Compressed Sensing artifacts, view-sharing and mis-registration of multiple volumes. 4/Post-processing and interpretation: over aggressive Maximum intensity projection (MIP), poor orientation of visualization in 3D.

The set-up artifacts can be anticipated and avoided as long as careful attention is given to the preparation of the MRA examination: scout imaging is always recommended to confirm the positionning of the imaging volume. In CE-MRA, a tiny bolus injection can provide reliable timing delays for the high-resolution CE-MRA. In NCE-MRA, a complete set of anatomical references are required for the appropriate placement of each slice/slab.

Acquisition artifacts can be mitigated with routine B0 shimming, verification of susceptibility artifacts within scout images (see previously). It is also very important to consider and adjust acquisition parameters according to the organ of interest: flow velocity varies greatly across the entire vascular tree as well as between patients.

Reconstruction options should always be considered in pairs with their physical assumption: the resolution requested in regard to the lumen(s) diameter(s) of interest, parallel imaging acceleration is subject to the number of channels activated within range of the target area and Compressed Sensing imposes strong assumptions on the spatial and temporal behavior of signal intensity variations that might mask out precious information. View-sharing also triggers unequal artifacts throughtout time and vessels depending on their size. MRA reconstruction (especially NCE-MRA) typically involves the combination of multiple acquisitions that need precise superimposition, that is difficult to guaranty in the presence of motion.

Finally, clinicians always have the difficult task to navigate within one or multiple 3D datasets that can be relatively large in size. Attention to spatial orientation and scaling as well as the follow-up of a dynamic bolus travel is required and a double check is recommended for these multidimensional images.

At the end of this presentation, participants should be able to discuss the pros and cons of CE-MRA and NCE-MRA in terms of artifacts. MR-Angiography stands nowadays as a robust and precise mean for the assessment of cardiovascular pathologies.