Considerations for Starting a 3D Printing Lab in the Department of Radiology

Introduction

Medical three-dimensional (3D) printing provides essential services within the modern hospital system. Patient-specific 3D printed anatomic models are used clinically to provide improved understanding of anatomy, more exact pathology evaluation, and more precise surgical intervention. 3D printed anatomic models can be used for many applications including presurgical planning, intraoperative guidance, trainee education, and patient counseling. Furthermore, 3D printed anatomic guides created from the patient’s imaging data can positively impact patient care by improving procedural accuracy. ^, By allowing for improved understanding of anatomy, enabling precontouring of implants, and providing real-time guidance in the operating room (OR), 3D printed anatomic models and guides can reduce OR costs secondary to shortening procedure times.

Surgeons have a low margin of error when providing care. Their goal is to obtain the best outcome on the first attempt, which may be challenging for high-complexity cases, even for the most experienced surgeons. Although both radiologists and surgeons are skillful at interpreting two-dimensional (2D) images, errors in interpretation may occur. Furthermore, mentally reconstructing 2D images into 3D representations is challenging. In the operative setting, surgeons are frequently conflicted in both their assessment of 2D images as well as their 3D interpretation where having an incorrect understanding of the true anatomy may potentially lead to an inappropriate surgical approach, increased operative times, and complications which all may impair patient outcomes. ^, Putting a physical 3D printed model of the anatomy in the hands of the surgeons allows them to see the true anatomy first-hand, thereby minimizing the chance of interpretation error and allowing them to be better prepared for the procedure.

In order to create patient-specific 3D printed anatomic models and guides, Digital Imaging and Communication in Medicine (DICOM) images are first acquired from any volumetric imaging modality (see Chapter 2 ). Computed tomography (CT) is the most common modality due to the ease of post-processing CT data, but magnetic resonance imaging (MRI) and ultrasound imaging are also utilized. Using the DICOM images, the desired anatomical regions of interest (ROIs) are then segmented, and the segmented ROIs are converted to a computer-aided design (CAD) format in preparation for printing (see Chapter 3, Chapter 4 ). Since the source data for these 3D printed models are the volumetric radiological imaging data, it is logical for a point of care 3D printing lab to be housed in the radiology department. However, collaboration across specialties is necessary and perhaps even having a synergistic partnership with related departments will be key to a successful lab.

Many hospitals already have a 3D Imaging Lab within their Department of Radiology that produces 3D volume renderings of anatomy to assist with visualization and surgical planning. This type of 3D visualization on a 2D display is limited though and does not provide the same depth as actually seeing the anatomy in 3D. The human visual system is highly dependent on visualization spatial relationships in 3D. Compared to 2D displays, 3D displays have been demonstrated to improve depth perception, decrease surgical times, and decrease perceived workflow for laparoscopic surgeries.

The timing for expanding an existing 3D Imaging Lab to include 3D printing services or opening a dedicated radiology-centered 3D printing lab is ideal based on the academic landscape, technological advances in image processing and 3D printing, and the emerging regulatory and reimbursement environments surrounding 3D printing–based technologies. Including 3D printing in-house at the point of care will enhance best practices by making 3D printing more accessible for the end user, ultimately leading to improved patient care, increased patient satisfaction, and cost savings for the healthcare system. ^,

3D printing technology is impacting the practice of medicine in almost every way imaginable. High-profile cases like the separation of conjoined twins and face transplant have been in the media’s eye. More day-to-day applications such as planning for orthopedic tumor surgery, facial reconstruction, and total knee replacement are on the rise, impacting at least tens of thousands of patients across the United States. By providing 3D anatomical models, doctors gain an insight that cannot be replicated by any form of medical diagnostics. In addition, medical student and surgeon training programs across specialties can benefit from use of the latest digital and physical training tools, in many cases offsetting or augmenting cadaveric practice sessions.

In the past, as mentioned in Chapter 6 , 3D printing was available only through outside vendors, resulting in expensive costs ( Table 15.1 ). Recent innovations have made this technology accessible to many hospitals. In-sourcing this work in a centralized 3D printing and advanced visualization facility can enhance patient care by shortening the time between ordering a model and holding it. A dedicated 3D printing lab will also enhance trainee education, support cutting-edge research, and will bring increased revenue and grant funding to the institution. This chapter will discuss considerations for starting a 3D Printing Lab within a Department of Radiology.

Financial Plan

Establishing a dedicated hospital-based 3D printing lab requires administrative support, physician leadership and guidance, biomedical engineers and technologist champions, and printing technicians. The financial plan for the 3D printing lab must include personnel costs as well as expenses for the construction or renovation of physical space, hardware and software expenses including maintenance costs, and potential revenue for 3D printed models (including clinical revenue via Category III CPT codes and research revenue).

In regard to staffing, determining the number of full-time equivalent staff will be based on the anticipated case volume. It is likely that most labs will start out small and will grow over time. To start out with, a lab could have one physician who serves as the medical director, performing image segmentation of complex anatomy and overseeing the printing process, as well as one technologist who performs basic image segmentation, runs the printers, and post-processes the models. Over time, based on the needs of the lab, this would expand to include more physicians, more technologists, biomedical engineers to carry out the design work especially for anatomic guides, and potentially could include printer technicians to run and maintain the printing hardware.

For the first few years, it is expected that the 3D printing lab will not generate significant revenue. However, over time this is expected to change, especially once Category I CPT codes are established for 3D printed anatomic models and anatomic guides. Until there is full reimbursement for 3D printed models, in order to cover some of the costs of running the lab, it is possible that surgical departments requesting 3D printed models will contribute money for image postprocessing and material cost based on the number of models requested. Grant or donor funds may also be utilized to help sustain the costs of running the lab.

At the time of writing, the price of a 3D printed anatomic reference model could range from approximately $1,000 to $5,000 USD depending on the printing technologies and materials used, and 3D printed guides and templates could range from approximately $1,500 to $10,000 USD per case. Although there is an up-front cost to purchase software and hardware, over time the cost of 3D printing at the point of care is expected to be much more cost-effective than outsourcing to vendors. Indirect benefits of 3D printing for the hospital include decreased operation times, decreased complications, improved patient outcomes, and improved patient and physician satisfaction. Future work will be performed to evaluate which case types are created in-house, to quantitatively measure how these models can positively impact patient care, and to determine the actual cost savings.

Training

There are currently no formal training programs for physicians interested to take the lead on medical 3D printing, therefore physicians must self-train to become proficient in the 3D printing process. For technologists interested in medical 3D printing, there is now a certificate program through Clarkson College which prepares imaging professionals to be knowledgeable and proficient in the medical 3D printing process. However, most established technologists, especially those already doing 3D image post-processing in dedicated 3D imaging labs, would most likely not return to school at this time. Technologists such as those who are proficient in 3D image post-processing may learn to perform image segmentation and CAD work as an extension of their current 3D visualization methods. Biomedical engineers may also be interested to perform some of this work and specific training will depend on their knowledge of medical imaging modalities as well as anatomy.

Once a 3D printing technologist or biomedical engineer is hired, the learning curve for performing adequate image segmentation, CAD, and printing is generally steep. An organized training program with defined and measurable objectives for each step of the process will help to set clear expectations. In order to build the optimized workflow and to integrate in the hospital infrastructure, the employee will become proficient in the following: (1) Understanding of diagnostic medical image acquisition techniques required for advanced visualization techniques and 3D modeling, (2) Mastering 3D image post-processing methods including segmentation, registration, and creation of 3D surface meshes, (3) Proper documentation and storage of design files, (4) Printing processes and materials, (5) Working with clinical teams to deliver and view models, and (6) Quality Assurance (QA) of 3D models.

Software Considerations

Many hospitals and hospital systems may have institution-wide licenses for 3D image post-processing and these software packages might have the capability to export segmented ROIs into CAD file formats. Examples of commonly used 3D image post-processing platforms with these capabilities include the Intellispace Portal (Philips, Amsterdam, Netherlands), Aquarius (Terarecon, Durham, NC), Advanced Workstation (GE Healthcare, Waukesha, WI), and Vitrea (Vital Images, Minnetonka, MN). If possible, these platforms, which many technologists are already familiar with, could be utilized to create 3D printed anatomic models. Specific 3D printing software packages are also available and may provide added benefits in regard to ease of segmentation as well as CAD modeling following the segmentation process. It is also important to note that for 3D printed medical models used at the point of care, it is recommended that FDA-cleared software is utilized. Costs for this type of software are approximately $10,000–$20,000 USD per license and may be perpetual or yearly licenses. Floating licenses may also be available for an increased cost. Note that for certain software there are minimum hardware requirements.

The stereolithography (STL) file format and other file formats utilized to create 3D printed models do not contain DICOM information, which is important in order to properly store this information in a patient’s medical record. At this time, DICOM Working Group-17 has worked to establish the DICOM 2018b standard where 3D file information in an STL format is contained within a DICOM Information Object. To date, one vendor has incorporated the new DICOM-encapsulated STL standard into their image segmentation platform. Integration of the DICOM format will allow for STL files to be traceable and properly integrated into a patient’s medical record. It is expected that more image post-processing platforms and Picture Archiving and Communication System (PACS) vendors will support these file types in the future. Until then, in order to properly archive these files, the files should be stored on a secure, backed-up network, and limited representative images of the segmentation and model may be sent to PACS systems to be stored in the patient medical record.