SecretariatMedical Imaging & Technology Alliance (MITA)
Lisa Spellman, dicom@dicomstandard.org
ChairsAllan Noordvyk, Change Healthcare
Justin Ryan, Phoenix Children’s Hospital
wg17chairs@dicomstandard.org
Last strategy update2016-11-05
MinutesWG-17 minutes

Rationale

3D printing (also known as additive manufacturing or AM) is a technology that is seeing rapid adoption at medical institutions, with multiple applications directly related to patient care. There is a strong natural tie between 3D printing and medical imaging at these institutions, as the patient’s cross-sectional imaging is often used as the primary input into the design of the object or to validate a proposed object’s design relative to patient anatomy.

The Medical Imaging Technology Alliance (MITA) has been approached by members of the clinical community and software developers serving this community with concerns about gaps in interoperability. Of concern is that the DICOM standard currently does not well address a number of key specific needs of medical 3D printing’s practitioners. This issue will grow further as in-progress and novel applications for 3D printing make the transition from medical laboratory to clinical practice. No existing DICOM work group appears to have a specific mandate or member expertise to fill this gap. Nor are standards efforts outside of MITA able to address this gap due to the DICOM standard being integral to the data exchange and workflows involved.

There is a risk that the 3D printing medical imaging standards void will be filled with compensating behavior by the clinical community. Specifically, this includes storing imaging-derived data separate from the patient’s imaging record, without formal ties to the specific source imaging data. This is problematic for efficient patient care. Additionally, in some cases such “workarounds” could represent a patient safety risk due to increasing the likelihood of wrong patient errors involving custom guides and implants. This is a risk that can be greatly reduced simply by leveraging the mature and well designed approaches already well-established in imaging systems via the DICOM standard.

The quality and utility of the DICOM Standard will benefit from a Working Group of experts drawn from the medical 3D printing community (both end users and industry), working in tandem with medical imaging experts who are familiar with the nature and structure of DICOM.

Additional Context

For those who are unfamiliar it, 3D printing is an area of technology whereby a physical object is created by constructing it in an additive manner. That is, the object is created from 3D model data by depositing a miniscule amount of material in a specific location where it is required, leaving the other portions of the build volume empty of permanent material, repeating this thousands of times at other locations, all in a manner that allows the deposited material to fuse to any adjacent material. This additive manufacturing process is typically done layer by layer, with each layer resembling the output of an inkjet printer. Thus, “3D printing” has come to be the common description of the process.

The 3D printing process differs substantially from traditional manufacturing techniques such as milling (removing unwanted material from a larger block) and forming (filling a pre-formed mold with liquid material that hardens within it, prior to the mold being removed). The primary difference is that 3D printing economically addresses the need to create uniquely designed but precisely defined objects, with few limits on external or internal complexity.

3D printing may performed using a variety of materials. These include polymers (both biocompatible or not), ceramics, metals (including implantable titanium), and cells.

Current uses of 3D printing in medicine are myriad, with the most common being creation of:

  1. Patient-specific anatomical models for patient education, surgical simulation, or surgical planning
  2. Patient-specific implants or implant mount adaptors (including temporary scaffolds for tissue/bone growth)
  3. Patient-specific external prosthetics or mount adaptors (including casts)
  4. Patient-specific interventional instrument guides
  5. Improved test phantoms for medical device calibration/validation
  6. Patient-specific artificial organs and tissue
  7. Custom laboratory equipment
  8. Replacement components for instruments

Items 1 to 5 above are directly relevant to DICOM as medical images are typically the primary data source guiding the shape and dimensions of the printed objects. Item 6 may have relevance, related to sizing and validation of surgical plans. Items 7 and 8 are likely to remain out of scope (see Scope definition below).

Scope

The WG would operate on the following general mandate:

  • Extend and promote the use of DICOM for the creation, storage and management of 3D printing models in a healthcare setting, where the model is either (a) derived from medical images, or (b) expected to be compared / composited with medical images

Within this mandate the WG would have the following specific scope of activities:

  • Identify and maintain a roadmap of use cases and compatibility concerns that should be addressed (see Roadmap below)
  • Develop or consult on relevant change proposals (CPs) and Supplements
  • Serve as a liaison body between the multiple stakeholder groups involved
    • Clinical end-users of 3D printing
    • Vendors offering 3D modeling and printer control software for healthcare applications
    • Vendors of medical image acquisition, processing, and management (i.e. existing MITA members at this point)
    • Standards and professional organizations addressing 3D printing in general (e.g. SME, AMI)
  • Facilitate including data relevant to the 3D printing imaging community in DICOM objects
  • Provide best practice guidance and reference implementations to promote the use of DICOM in 3D printing applications
    • In particular, ensure that 3D printing vendors entering the medical domain find it easy to adopt the standard (and encourage such vendors to participate in MITA)

Roadmap

The WG-17 roadmap is based upon the analysis of the 3D printing data and workflow needs through a standard framework. The results of this analysis would then be used to specify information objects and services needed to best serve patients and the imaging-based 3D printing medical community.

As much as possible and compatible with DICOM guiding principles, the new specifications this would be proposed in a manner to (a) leverage the existing and growing ecosystem of DICOM-capable systems in use in healthcare institutions and (b) leverage standards already in use in the 3D printing industry.

The ultimate goal is a comprehensive, standards-based digital platform for 3D printing in the patient care setting, of which DICOM-based imaging would be a significant part.

Short-Term Objectives

  • Establish and organize the new workgroup
  • Establish a working relationship with:
    • SME industry association’s AM/3DP work group
    • RSNA’s 3D printing Special Interest Group
  • Establish a working relationship with other DICOM Work Groups currently responsible for related Information Object Definitions (IODs): Computed Tomography, Magnetic Resonance, Encapsulated PDF, and Secondary Capture
  • Propose a new IOD, with references to existing IODs, to address the two already identified highest priority needs of the medical 3D printing community
    • Store 3D printing models in the Picture Archiving and Communication Systems (PACS) and Vendor Neutral Archive (VNA) systems in a manner that allows direct association with the relevant patient (and any source images from which the model was derived)
    • Allow for clinical users to review 3D printing models in the context of relevant patient images, prior to printing
  • Perform further gap analysis of the existing DICOM standard with respect to potential 3D printing requirements (e.g. intermediate segmentation representation)
  • Establish priorities for filling gaps identified

Current Status

  • Proposal of formation.

Current Work Items

  • New IOD for 3D printable models (see Short-Term Objectives above)

Risks

  • 3D printing community is eager for fast progress.
    • Slow progress may result in DICOM proposals being ignored by non-standard, workarounds that become entrenched but hamper interoperability (e.g. private element embedding in secondary capture objects)
  • Might not be able to recruit enough 3D printing stakeholders.
    • However, current level of enthusiasm for a standards-based approach in both end-user and vendor community appears high.

Challenges and Opportunities

  • The current level of enthusiasm for a standards-based approach in both end-user and vendor community appears high. Moving quickly to put the WG in motion can leverage this to address both risks above
  • There are a number of competing native data format standards for high fidelity 3D models outside of medicine, and no “lingua franca” appears to be developing in the near term
    • Different vendors and end-user communities favor different data formats as their primary working one
    • DICOM would need to take fairly neutral approach with respect to working with these data formats (suggesting an encapsulation approach, similar to what was done with PDF)
    • The STL data format, although older and limited, is near universally supported and thus may be leveraged as a baseline, to maximize interoperability
  • A 3D shape visualization capability of PDF is used at many institutions as part of the preview workflow
    • This appears to be compatible with DICOM PDF encapsulation IOD as it exists today
  • Unlike the case with medical printers, there appears to be no good case for DICOM directly support communicating with 3D printers themselves
    • There exists a final stage of preparing a 3D model for a specific printer’s capabilities and build material options before initiating printing
    • This final stage is already well addressed by vendors in this space and has no dependency on whether the object to be printed is related to a patient or is for some other purpose
    • DICOM’s scope can thus stop after delivering a 3D model to this final prep software
  • Many institutions have extended their 3D visualization labs to encompass 3D printing requests from clinical staff
    • This presents a community with knowledge of other DICOM-enabled workflow that could be leveraged for WG activities.

Relationship to Other Standards

  • As explained above, a goal of the WG would be to allow for direct compatibility with existing file format standards of the larger (i.e. non-medical) 3D printing community and industry. The most commonly used non-proprietary file format standards for 3D printing are:
    • G-CODE: The G-CODE (RS-274) file format provides the typical means by which precise instructions are sent to 3D printer hardware to construct each layer of the model. Thus, is it mainly used as a final step representation intended for a specific printer, loaded with specific consumable materials. This is typically produced from STL (see below).
    • STL: The stereolithography (AKA standard tessellation language) file format is essentially the lingua franca of 3D printing. While lacking many of the more advanced concepts of other later formats and being very verbose, its simplicity and longevity has allowed it to become the most common format for interoperable exchange of designs between various pieces of 3D printing software, whether medically-oriented or not.
    • VRML: Virtual Reality Modeling Language (ISO/IEC 14772-1:1997) is favored by some software for its ability to represent structures at a more abstract object-oriented level than STL. Use of VRML has in many cases been superseded by X3D (see below). VRML is often used as an intermediate precursor to STL creation.
    • X3D: As defined by ISO/IEC 19775/19776/19777, X3D is an XML-based format built as an extension of VRML/WRL, to address more advanced concepts. It is favored by some software for its ability to represent structures at a more abstract object-oriented level than STL. However, many of its concepts (e.g. lighting control) are aimed more at creating 2D projection images than physical objects. X3D is often used as an intermediate precursor to STL creation. X3D is in current use at the NIH for medical 3D printing.
    • AMF: The Additive Manufacturing File Format (ISO/ASTM 52915:2013) is a relatively new XML file format gaining in use in the 3D printing community. Unlike STL, VRML and XML, AMF was designed with the intent to specifically to address 3D printing. AMF is favored by some software for its ability to represent structures at a more abstract object-oriented level than STL and also describe materials. While it is sometimes used as an intermediate precursor to STL and/or G-CODE creation, but some 3D printers are now able to use AMF directly, obviating the need for use of these other formats. There is currently active debate within the community as to whether AMF or 3MF (see below) should supplant STL as the new lingua franca of 3D printing.
    • 3MF: A competing file format to AMF with much the same capabilities, but much more compact in file size. Invented by Microsoft, it is now published and promoted by a multi-vendor consortium. Like AMF, some 3D printers are now able to use 3MF directly, obviating the need for use of these other formats. There is currently active debate within the community as to whether AMF (see above) or 3MF should supplant STL as the new lingua franca of 3D printing.
    • PDF (U3D): The Portable Document Format with 3D extension (supported in Adobe 8 and above) is used to allow medical staff to preview 3D models prior to printing, without recourse to platform-specific specialty software. It is not possible to use this format to actually produce a 3D printed object. In the 3D printing context its only purpose is preview. This format should be compatible with the existing DICOM Encapsulated PDF IOD, but the current lack of reference to source images in that IOD is a concern (that could be addressed in a manner similar to CP 1559 “Reuse reference mechanisms from General Image Module in other contexts”).
  • The potential future roadmap could involve representation of intermediate segmentation steps. This may touch on standards beyond those listed above and/or involve harmonization/extension of existing DICOM IODs related to 3D visualization