Purpose or Case Report: Detecting foreign objects retained (RFO) during surgery presents unique challenges in imaging due to the growing complexity of materials and devices utilized in the operating room. This can include, but is not limited to, gauze, gloves, and suturing string. X-ray is the primary tool for detecting these objects, but the opportunity for radiologists to learn and practice can be limited. This project aimed to develop a cost-effective chest phantom and digital reconstructed radiograph (DRR) to allow for educational and practice.
Methods & Materials: The chest cavity was identified as the ideal preliminary anatomy to start with, as it is the most common area for RFOs. The anatomy includes the bones, airway, heart, and general soft tissue. Each is distinctly modeled from the reference anatomy of a 13-15 year old.
Programs such as Visage and 3D Slicer can digitally create artificial radiographs of existing anatomy imaging and place RFO. The programs are able to take objects modeled in the 3d space and project them in the 2D image and apply realistic radiopacity.
Physically, 3d printed material such as RadioMatrix and VeroWhite can be used to mimic the realistic radiopacity under X-ray. 2D slices of the anatomy were modeled and printed in the approximated matching radiopacity. They could be stacked along with the RFOs for imaging education.
Results: The DRR offered the potential for creating these artificial training images, but was limited to the understanding of how the RFO could look under true imaging. These could be improved over time with better references and design considerations, but hold early potential for training.
The physical trainer was able to create realistic images and offer flexibility in introducing many types of RFOs and in varied locations. This additional allowed for the education of the technologist in the process of scanning the RFOs. Further work can expand the detail of the anatomic slices and expand to other anatomic areas.
Conclusions: This preliminary work offers potential for creating both virtual and physical training images for the detection of RFOs. Further work can be done to refine the models and explore educational and clinical efficacy.