3D printed phantoms mimicking cortical bone for the assessment of ultrashort echo time magnetic resonance imaging. Issue 2 (30th December 2017)

• Record Type: Journal Article
• Authors: Rai, Robba; Manton, David; Jameson, Michael G.; Josan, Sonal; Barton, Michael B.; Holloway, Lois C.; Liney, Gary P.
• Is Part Of: Medical physics. Volume 45:Issue 2(2018)
• Page Start: 758
• Page End: 766
• Publication Date: 2017-12-30
• Subjects: 3D printing -- anthropomorphic phantoms -- magnetic resonance imaging -- T2∗ decay -- ultrashort TE; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1002/mp.12727
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: Human cortical bone has a rapid T 2 ∗ decay, and it can be visualized using ultrashort echo time (UTE) techniques in magnetic resonance imaging (MRI). These sequences operate at the limits of gradient and transmit‐receive signal performance. Development of multicompartment anthropomorphic phantoms that can mimic human cortical bone can assist with quality assurance and optimization of UTE sequences. The aims of this study were to (a) characterize the MRI signal properties of a photopolymer resin that can be 3D printed, (b) develop multicompartment phantoms based on the resin, and (c) demonstrate the feasibility of using these phantoms to mimic human anatomy in the assessment of UTE sequences. Methods: A photopolymer resin (Prismlab China Ltd, Shanghai, China) was imaged on a 3 Tesla MRI system (Siemens Skyra) to characterize its MRI properties with emphasis on T 2 ∗ signal and longevity. Two anthropomorphic phantoms, using the 3D printed resin to simulate skeletal anatomy, were developed and imaged using UTE sequences. A skull phantom was developed and used to assess the feasibility of using the resin to develop a complex model with realistic morphological human characteristics. A tibia model was also developed to assess the suitability of the resin at mimicking a simple multicompartment anatomical model and imaged using a three‐dimensional UTE sequence (PETRA). Image quality measurements of signal‐to‐noise ratio (SNR) and contrast factor were calculated and these were compared to in vivo values. Results: The T 2 ∗ and T1 (mean ± standard deviation) of the photopolymer resin was found to be 411 ± 19 μ s and 74.39 ± 13.88 ms, respectively, and demonstrated no statistically significant change during 4 months of monitoring. The resin had a similar T 2 ∗ decay to human cortical bone; however, had lower T1 properties. The bone water concentration of the resin was 59% relative to an external water reference phantom, and this was higher than in vivo values reported for human cortical bone. The multicompartment anthropomorphic head phantom was successfully produced and able to simulate realistic air cavities, bony anatomy, and soft tissue. Image quality assessment in the tibia phantom using the PETRA sequence showed the suitability of the resin to mimic human anatomy with high SNR and contrast making it suitable for tissue segmentation. Conclusions: A solid resin material, which can be 3D printed, has been found to have similar magnetic resonance signal properties to human cortical bone. Phantoms replicating skeletal anatomy were successfully produced using this resin and demonstrated their use for image quality and segmentation assessment of ultrashort echo time sequences.