3D printed models are becoming increasingly useful tools in today’s practice of personalized, precision medicine. As cases become more complex and treatments truly unique, visual and tactile anatomical models can enhance understanding and communication in the process of reaching a patient-specific solution.

Healthcare professionals, institutions, and organizations across the globe are using 3D printed anatomical models as reference tools for preoperative planning, intraoperative visualization, and sizing or pre-fitting medical equipment for both routine and highly complex procedures that have been documented in hundreds of publications. 1 Physicians often save time preparing for and conducting surgeries, leading to drastically reduced operating room costs while lowering patient risk, anxiety, and recovery time.

This guide offers a practical, step-by-step walkthrough for physicians and technologists to go from patient scan to 3D printed model, reviewing best practices for setting up a CT/MRI scan, segmenting datasets, and converting files to a 3D-printable format.

Anatomical models printed on Formlabs machines are already being used across several sub-specialties of surgery, including but not limited to orthopedic, cardiothoracic, vascular, oral and maxillofacial (OMFS), oncology, plastics, reconstructive, urology, and pediatric. This tutorial provides an overview of steps that can be applied to any Digital Imaging and Communications in Medicine (DICOM) dataset, which is the standard format for image storage in any modern picture archiving and communication system (PACS).

CT or MRI scan and associated DICOM file

• Recommended slice thickness: 0.25 mm-1.25 mm2

Computer with medical imaging software for segmentation, or use of an outsourced
segmentation service

Form 2 3D printer and resin, for example:

• White Resin for orthopedics or OMFS
• Clear Resin for cardiovascular or urology
• Dental SG Resin for surgical guides (approved uses only)

Form Wash and Form Cure for post-processing 3D prints (recommended for medical models)

Depending on your institution, anatomical models may be used for patientor pathology-specific purposes:

• Preoperative planning and intraoperative reference models for surgeons
• Device sizing (e.g., mandibular plates) and surgical tool design*
• Molds for implantable material or prostheses*
• Trainee education and simulation labs
• Patient education and enhanced informed consent

*May be subject to regulation and/or require institutional approval

Numerous published case studies and ongoing efforts have demonstrated the potential of 3D printing in a clinical setting. Published use cases include preoperative planning³, intraoperative use⁴, patient-specific instrument guides⁵, custom-tailored implants⁶, molds for bone cement⁷ or polymethyl methacrylate (PMMA) implants⁸, prostheses⁹, and trays.

These use cases have seen success in a variety of fields and surgical specialties, including orthopedic, cardiac, oral and maxillofacial (OMFS), vascular, neurological, cardiothoracic, musculoskeletal, plastics and non-facial reconstructive surgery, oncology, pediatrics, interventional radiology, and others.

3D printed models can also have substantial benefit for educating medical residents, fellows, and students¹⁰. Patient-specific models are especially appealing when considered against the cost of cadaveric specimens and animal testing and their associated requirements for laboratory space, surgical instruments, and disposal.

Widespread use of advanced visualization in radiology has been key for diagnosis and communication among physicians. While these visualizations have traditionally lived in 2D planes from CT or MRI scans as DICOM files, software developers have recently created tools to reconstruct diagnostic images as 3D anatomic renderings.

3D printed models are a natural progression from these various 3D visualization options, and offer many additional benefits, such as tactile feedback and other tangible information, that visualizations cannot.

For example, examining a custom 3D printed model led an orthopedic surgeon in the UK to find a lower risk solution to a young boy’s abnormal forearm injury.

“Access to the model changed the standard treatment indicated by the CT scans from a four-hour complex osteotomy to a simple, much less invasive 30-minute soft tissue procedure,” said Dr. Michael Eames.

The successful surgery was completed in just under 30 minutes—a decrease of over three hours from the originally planned time in the operating room, saving the hospital an estimated $5,500. Subsequently, the patient spent less time in postoperative care and recovered faster¹¹.

Thresholds may need to be dynamic and involve more complex algorithms to account for factors
such as CT noise and beam hardening that may produce artifacts and other undesired results.
Another option is to use region growing to automate segmentation, where an algorithm assigns
voxels as belonging to one part or another based on similarity or deviation from its surrounding
voxels. Additional adjustments and refinements may be necessary.

After segmentation is finished, convert segmented objects into a file type that a 3D printer can
use. This file type is typically an STL or OBJ file format.

After the conversion, perform any necessary physical adjustments that are more common in 3D
modeling, such as smoothing out surfaces, filling holes, and fixing other minor features. These
adjustments can be performed by a technician on a variety of CAD or CAM software and should
always include physician review to ensure that the output is clinically useful.

Once the final file is ready, it can be exported and sent to a 3D printer for fabrication. There are
several key factors to consider when choosing the right 3D printing technology, including: the
cost of the printer, software, and materials; print speed, accuracy, and resolution; ease of use
and access to customer service; and the type of printing materials, including biocompatibility
and the ability to be sterilized for certain use cases.

After printing, parts must be washed in isopropyl alcohol (IPA) to remove excess resin and,
depending on the material and applications, post-cured in a cure chamber. All biocompatible
resins require post-curing before use, and Formlabs Standard Resins, such as White and Clear
Resin, increase in strength and stability after post-curing.

Note: only one software is required per section, which represent steps in the workflow. This list
is meant to facilitate reference and research only and is not an endorsement of any particular
software or vendor. For those who want anatomical models without segmentation or printing
on-site, specialized service providers offer conversion, segmentation, and/or printing services
for a fee (e.g., Anatomage, Armor Bionics, Axial3D, and Materialise).

Please consider your local regulatory guidelines and intended uses before selecting a
software solution.

* Formlabs and Materialise have partnered to offer an end-to-end bundle in the US, Europe, and Japan. Mimics inPrint is a dedicated software solution to create accurate virtual anatomical models based on medical images and to prepare files for 3D printing. Integrated into clinical environments (PACS), Mimics inPrint includes pre-defined workflows with a direct link to Formlabs 3D printers

Some conventional medical imaging scans cannot be converted into high-quality 3D models to produce clear, accurate anatomic structures of interest. Plain radiographs (X-rays) and ultrasounds are not commonly used for 3D printing, and these imaging modalities are not recommended.

The most commonly obtained imaging modalities for evaluation of internal structures are CT scans and MRI imaging. These scans generate a DICOM file. DICOM is the standard for storing and transmitting medical images and can be considered a series of slices.

DICOM images cannot be edited in 3D design software or sent directly to a printer. In order to convert a DICOM file into a format suitable for 3D printing, such as an STL or OBJ file, a separate software is required to calculate the surface area of the structure of interest. This surface will then become the 3D model.

Almost any DICOM file with sufficiently fine detail (e.g., thin slices) can be converted into a format that allows for 3D printing of the structure(s) of interest.

The first consideration when turning a CT or MRI scan into a 3D model is what needs to be shown; bones, vessels, and solid organs all need to be modeled in different ways. A model with obsolete structures not only distracts from the focus of the model, but will also be more difficult to manufacture. A scan with proper characteristics makes it easier to create a 3D printable model. Basic features to consider are intravenous contrast and slice thickness.

For 3D models of bone structures, noncontrast images will likely be sufficient for accurate, detailed printing. Contrast-enhanced scans are almost always required for models that include solid organs, tumors, or vascular structures.

In addition to contrast enhancement, slice thickness and resolution are equally important to consider when planning for 3D model creation. Most clinically useful scans are obtained with adequate resolution for 3D printing. However, if you try to 3D print an anatomical model from a scan with thick slices, your model will have a rough finish. According to numerous sources, it is very important to use scans with slices less than thinner than 1.25 mm when creating a model for 3D printing.

The thickness of slices captured by a CT or MRI scan translates directly into the detail generated in a 3D scan. Depending on the item of focus, image sections should be reconstructed with isotropic voxels of 1.25 mm or less¹⁵. According to a presentation by the Mayo Clinic in March 2016, slices as thick as 1.5–3 mm may be used for larger structures, whereas 0.75 mm may be used for fine bone¹⁶. Thicker sections may compromise model accuracy, while very thin sections (e.g., <0.25 mm) could require extensive segmentation and STL refinement, particularly in the presence of image artifact. Cardiac models demonstrate sufficient accuracy with 0.5 mm sections, but thin objects such as the orbital floor may require thinner sections¹⁷. Generally, thicker sections may produce unclear or less accurate printed models. Nevertheless, unnecessarily thin sections may create significantly more work in the post-processing phase.

Please consider local regulations, material data sheets, patient information, and institutional requirements before 3D printing or using anatomical models. For physicians in the US, refer to the latest FDA presentation summarizing its guidance document¹⁸. Please note that these documents and guidelines are subject and likely to change. Always confirm you comply with the latest recommendations.

Barriers to entry for personalized, precision medicine are rapidly lowering. The advent of affordable professional-grade desktop 3D printers has enabled healthcare providers to produce patient-specific anatomical models for a variety of specialities with inspiring results. In some cases, the initial cost of high-quality a printer was covered by the costs associated with time saved in the OR room after a surgeon used a 3D printed model to prepare for a single complex procedure¹⁹.

Understanding the workflows required to integrate 3D printing is key to a successful operation. This tutorial addressed what you need to get started, popular workflows and tools, and techniques to facilitate the progression from patient scan to 3D printable model.

Contact us to learn more about introducing 3D printing into your institution and join the community of innovators in precision medicine.