There are three causes of motion artifact in CCTA: (1) cardiac/coronary motion, (2) respiratory motion, and (3) patient complete body motion. Another computational cause of apparent “motion artifact” may occur during multisegment reconstruction when interbeat variation in the location of coronary artery segments leads to superimposition failure.

All images in CCTA demonstrate an element of coronary/cardiac motion artifact. Coronary/cardiac motion is complex, comprising multiple additive movements, which include squeezing [1]. The degree of apparent coronary motion on a CCTA image is a complex interplay between: (1) intrinsic patient-specific coronary movement, (2) the patient’s heart rate, (3) timing of scanning, (4) hardware related capabilities of the scanner, and (5) software/postprocessing capabilities of the image data. Since only the first variable is fixed, strategies to reduce cardiac motion focus on the latter four.

The patient’s heart rate is lowered to preferably around 60 beats per minutes (bpm) for most single source scanners. Dual source and newer single source scanners may tolerate higher heart rates. Typically this is performed via the negative chronotropic effects of metoprolol via PO or IV routes.

Imaging during diastole, hopefully including diastasis (the time of maximal left ventricular dilatation and minimal coronary movement), is the ideal imaging time for CCTA. This is achieved typically by choosing a delay corresponding to a %R-R of around 70–80 % of the cardiac cycle, in patients with slow heart rates (around 60 bpm). In patients with higher heart rates, the phase of minimal coronary artery motion moves toward systole.

Technologically, improving the effective temporal resolution intrinsic to the hardware of the scanner is the second strategy. The minimal amount of data needed to create an image, typically 180° plus fan angle (in single source scanners) and the gantry rotation speed, dictates the temporal resolution of a scanner. Dual source scanners utilize two photon source-detector row elements set roughly 90° from each other in the x–y plane (same z-axis alignment), to allow only about 90° plus fan angle of rotation to acquire the minimum amount of data needed for image reconstruction. Newer single source scanners can rotate at 200 msec/revolution.

Using software to improve the effect of temporal resolution such as multisegment reconstruction and other proprietary computational algorithms may be performed in certain scans. Multisegment reconstruction uses partial data from different heartbeats to reconstruct an image. This can be performed utilizing both helical and step and shoot techniques. Unfortunately, it relies on a low interbeat variability of coronary artery location for good image quality, which is not always the case.

Although cardiac CT examination occurs during breath holding, some patients are unable to comply, and typical respiratory motion artifact occurs. Clearly, longer examinations, and more specifically, longer acquisitions, are at increased risk for this artifact. Manifestations include step off at the chest wall/sternum or skin on sagittal images, and indistinct or mis-registered pulmonary vasculature/parenchyma. The latter is lower lobe predominant due to the direction of most CT image acquisitions being craniocaudal.

One of the most severe artifacts occurs when the patients move their entire body during the study. Typically this results in extremely poor image quality during the portion of the scan when the patient moved. Patient coaching, reducing total scan time and/or medication/anesthesia as needed, will reduce the probability of spontaneous patient motion.

Utilizing “step and shoot”/axial/sequential technique results in multiple 3.2–4 cm slabs in the z-axis, which then may be “stitched” together to create the typical 10–16 cm z-axis coverage needed for the heart. At the stitch points, there may be misregistration due to patient random motion, respiratory motion, or more commonly slight differences in vessel location/%R-R interval acquisition due to a fixed trigger delay. In most 64 detector row scanners there will be 2–3 stitches. Retrospective gating may also demonstrate misregistration; however, it occurs at any point along the z-axis. Misregistration may also occur during multisegment reconstruction where interbeat variation in coronary artery or cardiac structure location results in the two ghost-like, faint structures being superimposed. Even volume scanners, with 16 cm of z-axis coverage per rotation, may demonstrate misregistration artifact in the z-axis when larger anatomy is scanned (e.g., the entire chest in a triple-rule out exam.)

“Blooming” artifact describes the exaggeration of a calcified plaque’s size resulting from the combined effects of coronary motion, beam hardening, and spatial resolution limitations. Evidence suggests that beam hardening plays a minimal role in calcium blooming, which is mostly an interplay between motion effects and spatial resolution [2]. For a given innate scanner temporal resolution, motion effects predominate at high heart rates and spatial resolution effects predominate at lower heart rates.