Purpose: To investigate motion artefacts in dynamic CT applications and the impact of different protocols using a rotating phantom.
Methodology: To study image quality in dynamic CT applications, a circular polymethylmethacrylate (PMMA) phantom was constructed. The PMMA disk (70 mm radius) contained 9 mm-deep wells of varying radii (1 mm, 2 mm, and 3 mm). The holes were drilled at 9 mm, 24 mm, 44 mm, and 64 mm from the disk center1. The circular disk was made to rotate using a Trinamic electric motor (TRINAMIC Motion Control Hamburg, Germany) that was controlled with an Arduino uno board and custom written c++ code. Dynamic scans were obtained while the phantom was rotating at 720/s in a wide beam CT scanner. We investigated the influence on motion artefacts of two dynamic acquisition protocols: "cine" and cardiac" .
Results: Visual inspection revealed less motion artefacts with the cardiac protocol compared to the cine protocol. In future studies, we will quantify the artefacts and evaluate temporal resolution in the entire FOV of the CT scanner.

Introduction and Aim: Dynamic Computed Tomography is a promising tool to investigate several affections, predominantly vascular and perivascular diseases. Advances in CT technology allow using dynamic CT to evaluate also musculoskeletal diseases2. Wide beam CT scanners, have the potential to acquire images from the same anatomic area over lengths up to 16cm. As such, detailed analysis of joint morphology and dynamic phenomena can be investigated. With the aid of image processing techniques, motion kinematics can be obtained from these dynamic images. However, the image quality is of great importance for the image processing steps as well as for the diagnosis of potential pathologies. Motion artefacts are of major concern in dynamic acquisitions for musculoskeletal applications and are influenced by the acquisition protocol and the speed of motion. In this work a phantom was designed to undergo rotatory motion in the gantry of the CT while different acquisition protocols were investigated. Visual inspection of the resulting dynamic images was performed.
Materials and Methods:
Phantom design:
The phantom consisted of a PMMA disk (70 mm radius) with 9 mm-deep holes of varying radii (1 mm, 2 mm, and 3 mm). The holes were drilled at 9 mm, 24 mm, 44 mm, and 64 mm from the disk center. The individual holes we spaced out at 2 mm for the 1-mm holes, 4 mm for the 2-mm holes, and 6 mm for the 3-mm holes. A Trinamic 24V electric motor (TRINAMIC Motion Control Hamburg, Germany) was used rotate the phantom and an Arduino uno board (Arduino AG) controled the motor speed.

Image acquisition:
Dynamic CT images were acquired on a 256-slice CT (GE Healthcare) while the phantom, positioned in the iso-center, rotated at 720/s parallel to the gantry. The scan parameters were: 120kV, 430mA and 1.25mm slice thickness. Two different acquisition protocols were investigated: "cine" and "cardiac" . Under the Cine protocol, two tube rotation times were investigated, 0.28 and 0.5 s. The cardiac protocol required an ECG trigger which we overcame by simulating a heart rate of 32bpm on the CT and scanned from 0 to 300% of the cardiac phase. This resulted in a total scan time of about 6 seconds and allowed enough time to capture a longer period of motion. Figure 1 shows the phantom and the setup in the CT scanner.

Conclusions: We designed a phantom to investigate motion artefacts in dynamic CT acquisitions using different scan protocols. A higher tube rotation speed and the cardiac protocol produced better images with less motion artefacts. In the future we will try to confirm these finds using a moving ankle joint model and use the model to investigate the temporal resolution in the entire FOV of the scanner
Original languageEnglish
Publication statusUnpublished - 2019
EventBHPA 2019 congress - Aalst, Belgium
Duration: 1 Feb 20192 Feb 2019


ConferenceBHPA 2019 congress
Abbreviated titleBHPA2019

Fingerprint Dive into the research topics of 'A rotating phantom to investigate motion artefacts in 4D-CT applications'. Together they form a unique fingerprint.

Cite this