Research Article: Creating patient-specific anatomical models for 3D printing and AR/VR: a supplement for the 2018 Radiological Society of North America (RSNA) hands-on course

Date Published: December 30, 2019

Publisher: Springer International Publishing

Author(s): Nicole Wake, Amy E. Alexander, Andy M. Christensen, Peter C. Liacouras, Maureen Schickel, Todd Pietila, Jane Matsumoto.

http://doi.org/10.1186/s41205-019-0054-y

Abstract

Advanced visualization of medical image data in the form of three-dimensional (3D) printing continues to expand in clinical settings and many hospitals have started to adapt 3D technologies to aid in patient care. It is imperative that radiologists and other medical professionals understand the multi-step process of converting medical imaging data to digital files. To educate health care professionals about the steps required to prepare DICOM data for 3D printing anatomical models, hands-on courses have been delivered at the Radiological Society of North America (RSNA) annual meeting since 2014. In this paper, a supplement to the RSNA 2018 hands-on 3D printing course, we review methods to create cranio-maxillofacial (CMF), orthopedic, and renal cancer models which can be 3D printed or visualized in augmented reality (AR) or virtual reality (VR).

Partial Text

Advanced medical image data visualization in the form of three-dimensional (3D) printing continues to expand in clinical settings. Many hospitals have started to adapt 3D technology to aid in patient care, for use in medical student education, and for research applications. 3D printing originated in the 1980s and encompasses various processes intended to generate a physical model from a digital file [1–3]. Virtual Reality (VR) uses a computer to simulate an alternate 3D environment and allows for user interaction within this space. Augmented Reality (AR), which overlays 3D content in the users real environment, is another method of advanced image visualization that has great potential to transform how physicians access medical imaging data. 3D printed models and AR/VR experiences are expected to provide improvements in the visualization of medical images as compared to viewing medical images on a two-dimensional screen [4].

In general, the steps required for 3D anatomical modeling from DICOM data include the steps shown in Table 1. If imaging is performed with the intent to create an anatomic 3D model, the image acquisition parameters should be optimized for quality [31]. However, this remains challenging considering that imaging studies are typically performed before a model is ordered. Factors to consider include spatial resolution (approximately 1 mm3), reconstruction kernel, multi-phase contrast, metal artifact reduction, and sequence parameters for magnetic resonance imaging (MRI). Repeat imaging solely for the purposes of producing a 3D model is often not advisable because it is not cost-efficient and will increase patient radiation dose if a computed tomography (CT) scan is performed.
Table 1Stages of the anatomical modeling processStep 1: Image Acquisition - Select imaging modality - Set appropriate protocols - Export data to independent image post-processing workstationStep 2: Image Post-Processing - Isolate tissues and organs of interest - Prepare files for data visualization method of choice - Save and transfer data in an appropriate formatStep 3: 3D Visualization or Physical Reproduction - Create 3D computer model - Prepare model for AR, VR, or 3D printing

Converting DICOM data to printable formats is a complex process requiring multiple steps. This paper describes key steps to create 3D printed CMF, orthopedic, and renal models. Techniques described here may also be applied to other organs and anatomical regions of interest. The number of 3D printed and AR/VR models generated from DICOM images is growing exponentially at the point of care. It is essential that radiologists and other health care professionals understand this complex process.

 

Source:

http://doi.org/10.1186/s41205-019-0054-y

 

Leave a Reply

Your email address will not be published.