Clinical 3D Printing: Status, Challenges and Opportunities

KENNETH C. WANG, MD, PHD, is the MRI section chief at the Baltimore VA Medical Center and an adjunct assistant professor of diagnostic radiology at the University of Maryland School of Medicine (UM), both in Baltimore. He is a member of the RSNA Radiology Informatics Committee, where he serves as the chairman of the RadLex Steering Committee, as well as the liaison to the RSNA 3D Printing Special Interest Group. Dr. Wang earned his BS, MS and PhD degrees in electrical engineering from Stanford University, and his medical degree from the University of California, San Francisco. He completed his radiology residency and a musculoskeletal radiology fellowship at Johns Hopkins Hospital, and an informatics fellowship at UM. His research interests include radiologic applications of semantic computation, shoulder and ankle imaging, 3D printing and MR neurography.

During the past several years, 3D printing has become increasingly common in the clinical setting, with application to a wide range of clinical challenges. In many areas, radiologists, imaging scientists and other imaging professionals have been the driving force behind these clinical advancements. As this technology continues to transition into the clinical workflow, there will be many challenges and opportunities ahead. Through collaborations with other specialties on 3D printing applications, radiology departments and practices can deepen their engagement with other clinical services, raise the utility of imaging to referring providers and improve the care of patients.

History and Current Status

3D printing was invented in the 1980s, and found its first uses in engineering and manufacturing. Clinical applications were not far behind, and the first descriptions of 3D printed patient-specific models derived from radiologic images appeared in the 1990s. Since then, parallel advancements in imaging capabilities and 3D printing technologies have enabled a proliferation of clinical 3D printing applications.

Some of the first clinical applications of 3D printing were in craniomaxillofacial surgery, and this remains an important area for clinical 3D printing. Today, printed models are also used in many other domains such as orthopaedics, vascular surgery, cardiology and cardiothoracic surgery, transplant surgery, urology and neurosurgery. These models may be used for surgical planning (e.g., selection of a surgical approach or selection of surgical devices) and surgical simulation (i.e., the use of models to practice surgical techniques or test deployment of devices prior to performing an operation).

Several other perioperative and intraoperative applications have also been developed, including the use of printed models to enable preoperative bending of fixation plates to fit a specific patient’s bony anatomy (thereby reducing operative time), as well as the creation of patient-specific cutting guides and surgical tools (to increase the precision of surgical maneuvers). In addition, 3D printing has been used to create customized, patient-specific implants and prosthe-ses. 3D-printed models have also been used extensively for medical training and education. For example, trainees in spine surgery have reported benefits from learning pedicle screw placement with printed models. Printed models are also used to communicate with patients about diagnoses and procedures.

The patient-specific nature of these 3D printed models and devices derives from the use of cross-sectional imaging as the source of an individual patient’s anatomy. Furthermore, the construction of these models involves image interpretation and clinical judgment, such as with regard to lesion characteristics, the relevance of particular image findings and the assessment of image noise and artifacts. On the other hand, the specific clinical challenges which make printed models so useful are often most familiar to surgeons and other referring colleagues. Consequently, imaging departments and radiology practices are well-placed to serve a central but collaborative role in the development of 3D printing capabilities within a health care enterprise.

A typical workflow for creating image-based 3D printed models starts with isotropic or near-isotropic CT or MR images. Image filtering may be applied to reduce image noise. Target structures are identified in the images through the process of image segmentation. Segmented structures are then used to create surface models, where 3D objects are described by triangulated surfaces. Post-processing allows for operations such as surface smoothing, gap filling and error correction. Additional modeling at this stage enables the introduction of other features, such as a minimum wall thickness for hollow vascular models, cut planes to show internal structure, or engineered features such as drill holes through a surgical guide. Surface models are checked against the source images for fidelity, and these models may then be used as input to a 3D printer. Depending on model complexity and the type of printer used, various post-processing steps are typically required after printing is complete, such as the removal of sup¬port structures, cleaning of extraneous substrate material or the curing of printed models. Finally, printed models are checked for dimensional accuracy.