Fabrication of Anatomical Sonographic Phantoms Through Direct 3D-Printing
J Zejmo, L Gorman, C Myles, J F X JonesAbstract
Objectives
Sonographic anatomy classes can be curtailed by the availability of subjects for sensitive regional scanning. These limitations can be overcome by using anatomically realistic ultrasound phantoms (USPs). Simulation-based training with USPs also allows interventional radiology trainees to safely practice complex procedures. However, commercial phantoms are not patient-specific, and involve costly, lengthy production methods.
We describe a simple method for producing anatomically realistic USPs from 2D grey-scale ultrasound images using rapid direct 3D-printing technology.
Methods
Original sonographic images were obtained from an open-source database and by scanning student volunteers. Images were digitally modified using InkscapeTM, ImageJTM, BlenderTM, MeshmixerTM, and CHITUBOXTM software programs. Models were printed using resin photopolymer 3D printers. In total, four phantoms were made: a femoral trochlea, a paediatric appendix, an eye with retinal detachment and a radial nerve/brachioradialis model. The ImageJ Fiji plugin function, “Block Matching Correspondences” (BMCs) was used to assess similarity between source and phantom images.
Results
All phantom images displayed BMCs with the original images (14-25% similarity). The radial nerve model displayed the highest similarity. USPs produced with a hydrophilic swellable resin could be imaged throughout their full thickness and penetrated by a needle.
Conclusions
In summary, this approach can manufacture USPs from 2D ultrasound data capable of generating anatomically accurate, patient-specific sonographic images. The low cost of these models, alongside their reproducibility through rapid 3D-printing technology, makes them a highly accessible, effective tool for simulation-based interventional radiology training.
Advances in Knowledge
To our knowledge, this study presents the first anatomical USPs which have been directly 3D-printed from a single 2D ultrasound image source.