3D printing: A value-added teaching tool for radiologists

3D printing technologies can provide radiologists with an opportunity to serve as leaders in medical education and clinical care, according to radiologists at George Washington University Medical Center in Washington, D.C.

Writing in the September issue of Academic Radiology, lead author Ramin Javan, MD, assistant professor of neuroradiology, and colleagues recommend that radiologists working at academic and/or teaching hospitals utilize 3D anatomic models to enhance the experience of radiology trainees and nonradiologist subspecialists.

“A firm understanding of complex anatomy not only serves as the fundamental basis of identifying pathologic states, but also allows for providing accurate reports tailored toward nonradiologist subspecialists and surgeons,” they write.

Low-cost 3D anatomic models are the product of innovations in advanced 3D printing and the capabilities of cross-sectional imaging like high-resolution CT, MRI, and steady-state free precession and fast spoiled gradient-echo. Virtual anatomic data are processed by 3D reconstruction software into a virtual 3D mesh.

The Academic Radiology article also explains how radiologists can create and use low-cost 3D anatomic models. The authors describe how they made 3D anatomic models of the liver, lung, prostate, coronary arteries, and the Circle of Willis. The models they created are being used for visualization, handheld spatial reasoning, and for testing in regard to complex segmental and branch anatomy relevant to radiology.

Figure 1. Circle of Willis and coronary arterial system 3D printed models, demonstrating complex branch anatomy in various views.

Circle of Willis model

The authors believe that this model is best used create when first starting to use 3D printing technology. They used DICOM data from a MR angiogram without contrast using free software (OsiriX Lite, OsiriX-Viewer.com), which allows for easy export of Standard Tesselation Language (STL) into commercial graphic design software (Autodesk 3D Studio Max) to remove noise and smooth the surface of the model to increase its visual appeal.

The Circle of Willis model and the coronary arteries model demonstrate angiographic appearance in various orientations. This is because they only contain the arterial system, exactly simulating an arterial injection.

Model of coronary arteries

The authors used a predesigned, commercial digital 3D mesh (TurboSquid.com) to obtain a generic coronary artery model that was then modified using graphic design software (Autodesk 3D Studio Max). Septal branches of the left anterior descending artery were added and the diameter of all the vessels was increased to allow for 3D printability in a reasonably sized model.

Figure 2. Lung, liver and prostate 3D printed models, demonstrating complex segmental anatomy in multiple views. The lung model was colored after 3D printing with wax pastel. Note that the prostate model has a modular design allowing for one piece to fit into another.

Lung model

The authors advise that designs of 3D models can be obtained from 3D reconstructed cross-sectional imaging data or created exclusively with graphic design software. The lung model they created was made using the graphic design software obtained from an online library (3DCADBrowser.com) and subsequently modified with the same commercial graphic design software used for the other models to add pulmonary arteries to provide the connection between each lung and to create a surface color map representing the lobes and segments of the lung.

When describing the accurate location of pulmonary nodules and segmental pulmonary emboli as well as interpreting nuclear medicine scans, knowledge of lung segmental anatomy is essential. The authors state that a 3D lung model provides orientation of the actual views, and that it can be used as a reference tool when interpreting a CT or VQ scan.

Prostate model

The initial prostate model was also obtained from a free online source. It was already 3D reconstructed from MRI using open-source software (Slicer.org). Anatomic subdividing was performed using the commercial graphic design software This was the most difficult model to create because all pieces had to fit together like pieces of a puzzle, the authors said.

The authors stated that the model’s modular design is especially useful when segmental anatomy entails parts of the anatomy that are embedded within one another requiring disassembly. Accurate localization of lesions is critical for pathologic correlation and staging of prostate carcinoma. The model makes this easier.

Liver model

The liver model was the most time consuming to create the initial 3D mesh, due largely to the intricate relationship between the vasculature as the source of the actual anatomic segmentation methodology of the liver. It was necessary to hand-draw 2D diagrams to communicate cut planes of the liver with the graphic designer.

As for radiologists who do not have advanced computer skills or who do not want to invest in 3D printers, the authors recommend commercial 3D printing services. Besides eliminating the investment in 3D printers, these offer the option of selecting from a variety of materials to make anatomic models and a selection of quality levels and costs. Most important, reputable companies have the technical expertise and quality control methods to create accurate model dimensions. The authors point out that the technical expertise needed to ensure an STL file meets 3D printability requirements can be extremely challenging.

Should 3D models replace 3D reconstructed images on workstations? Dr. Javan told Applied Radiology that “being able to hold a model just not only adds a new dimension of touch but also allows a much better understanding of a complex three-dimensional structure and its spatial relationship with its surroundings.”

REFERENCE

  1. Javan R, Herrin D, Tangestanipoor A. Understanding Spatially Complex Segmental and Branch Anatomy Using 3D Printing: Liver, Lung, Prostate, Coronary Arteries and Circle of Willis. Acad Radiol. 2016 23;9:1181-1189.
© Anderson Publishing, Ltd. 2024 All rights reserved. Reproduction in whole or part without express written permission Is strictly prohibited.