Applying 3D surface scanning technology to create photorealistic three-dimensional printed replicas of human anatomy

Abstract

Aim: To describe advances in 3D data capture and printing that allow photorealistic replicas of human anatomical specimens for education and research, and discuss advantages of current generation printing for replica design and manufacture. Materials & methods: We combine surface scanning and computerized tomography datasets that maximize precise color and geometric capture with ultra violet (UV) curable resin printing to replicate human anatomical specimens. Results: We describe the process for color control, print design and translation of photorealistic 3D meshes into 3D prints in durable resins. Conclusion: Current technologies allow previously unachievable ability to capture and reproduce anatomical specimens, and provide a platform for a new generation of 3D printed teaching materials to be designed and used in anatomy education environments.

Plain language summary

The teaching of human anatomy has undergone significant change in the last 30–40 years, especially in respect to the technologies available to augment or replace traditional teaching using dissection of human bodies. This has included plastic models, software teaching packages, digital visualization tables and virtual/augmented reality. Our group initially developed a range of 3D printed replicas (Series 1) of human anatomy dissections. Our method involved computed tomography scanning a dissected specimen to capture the geometry and then digitally coloring the model with a standardized color palette to ‘false color’ the resulting 3D prints (e.g., yellow for nerves and red for arteries). This present report details how advances in full-color, high-resolution surface scanning can create a true colored photorealistic model of preserved human anatomical specimens. When these surface scanned models are 3D printed with the current generation of UV curable resin-based printers, it is possible to achieve photographic quality replicas comparable to the original anatomy specimens. This new generation of 3D printed replicas resembling traditional anatomy specimens (Series 1.1), while simultaneously still allowing color augmentation to further enhance their educational value. These replicas have an advantage over plastinated cadaver specimens as they can be utilized in any teaching environment such as peripheral or rural medical school locations, teaching hospitals and clinical environments.

Plain language summary

Summary points.

• Advances in technology have changed the way that human anatomy is taught in today’s classrooms.

• Digitally colored computed tomography (CT) data formed the basis to our first anatomical replica series (series 1).

• Technological advances in 3D printer technology means that our new generation models have improved true-color reproduction and durability.

• Advances in full-color, high-fidelity surface scanning can help us reproduce a photorealistic replica.

• The current generation of UV curable resin-based printers make it possible to achieve photographic quality replicas comparable to the original prosections.

• In many jurisdictions, plastinated specimens are classified as human tissue and require special handling, housing and treatment.

• 3D printed replicas have an advantage over plastinated cadaver specimens as they can be utilized in teaching environments that do not hold an anatomy license such as peripheral or rural medical school locations, teaching hospitals and clinical environments.

• Series 1.1 consists of 59 3D printed anatomical replicas, spanning several body regions with a distinct emphasis on axial organs.

1. Introduction

The teaching of human anatomy in medical and allied health curricula has often been at the frontier of the adoption of new technologies compared with other preclinical and paraclinical sciences and basic biological sciences. We previously described in detail methods used to create three-dimensional (3D) printed replicas of anatomical dissections [1]. Since those initial studies, additive manufacturing, more commonly described as 3D printing [2], has increasingly entered the medical and healthcare arena, with particular impacts on surgery [3–7], and other disciplines such as dentistry [8] by allowing the production of bespoke prefabricated models for presurgical planning or the creation of patient-specific prostheses, or as patient educational tools [9–11]. There have also been reports exploring the usefulness of 3D replicas in medical and anatomical education [12–15] and a study showing that supporting topic explanations with a 3D replica is more motivating and attractive to students than two-dimensional material [16] and that they can facilitate a student’s understanding [14,17].

In our first study, we relied on computerized tomography (CT) imaging to develop 3D volumetric mesh files of anatomical specimens which were then digitally false-colored using a ‘traditional educational palette’ akin to that used in anatomy textbooks and conventional plastic models (e.g., red arteries, blue veins, yellow nerves, etc) [1]. We called this ‘Series 1’ and subsequently went on to demonstrate their value in teaching anatomy to medical students [18]. The decisions regarding coloration were based on several factors. For one, attempting to replicate the monotone and relatively uniform grey and brown tones of embalmed cadaver specimens was not visually distinct when printed using the then available powder-based full-color 3D printers. For another, CT scanners only capture topographic information, so resulting 3D volumetric meshes were monochromatic – eliminating any important visual cues for structure delineation. Finally, we thought that a color-coding scheme similar to textbook diagrams and atlases would aid early career medical and allied health students in their use of this new generation of 3D printed replicas. The ‘natural’ greys and browns of embalmed human tissues present a well-recognized challenge for unexperienced students in their learning. Since the initial descriptions of ‘Series 1’ we have gone onto use this technology to produce specialized 3D printed anatomy replicas for ophthalmology education [19], as well as replicas of a human fetal collection [20]. These experiences have equipped us with the technical skills and resources to overcome some of the challenges of accurately recreating and replicating color, fine topographic detail and 3D form; particularly when coupled with technical advances in commercial 3D data acquisition and additive manufacturing. With the advances in the quality and resolution of 3D scanners combined with the high quality of UV curable resin 3D printers, we decided to revisit the concept of producing more photographic quality 3D replicas. Our initial experiences were instrumental in the creation of replicas of human pathology specimens (which we termed ‘Series 2’), in which surface topography and accurate color rendition are an essential element of accurate pathological diagnosis [21]. This led us to then consider how we might produce replicas of further ‘normal’ (e.g., nonpathological) anatomy specimens with more photorealistic properties than ‘Series 1’. We term this new collection ‘Series 1.1’. Here we describe in detail the method used to create this new series and discuss why in certain situations these types of replicas are being chosen to augment our teaching of human anatomy.

2. Description of materials & methods

2.1. Selection of anatomical material for Series 1.1

Our primary mandate when selecting specimens for our first anatomy series (Series 1) was to collect a finite and balanced range of anatomical regions and structures in a variety of configurations to support comprehensive medical and allied health curricula [1]. As a result, we avoided or limited our sampling of some types of specimens (e.g., whole organs with minimal surface topography), regions (e.g., neuroanatomy) or large dissections (e.g., partial body dissections) which were constrained by the available data capture/printing technology at the time or the perceived utility of a 3D print relative to a dissection or traditionally manufactured plastic model. In selecting specimens for Series 1.1 (Table 1) we expanded our remit to include a greater range of whole organ and neuroanatomy-specific specimens, as well as revisiting larger and topographically complex dissections. Series 1.1 consists of 59 anatomy replicas generated from 32 specimens (e.g., 32 specimens presented with their original color, with a subset of 27 specimens with a secondary false color overlay version including four specimens with false-coloring of organ impressions). Series 1.1 thus spans several body regions with a distinct emphasis on axial organs and systems which complimented the strong emphasis of limbs in Series 1. Collectively, Series 1 and 1.1 consist of 119 anatomy replicas that address a diverse suite of anatomical structures and dissections/exposures that support both medical, allied health and general science education in the anatomical sciences.

Table 1.

List of systems represented and numbers of cases chosen for printing.

Head and neck	Sagittal section of the head with infratemporal fossa dissection (H11) Midsagittal section of the head with deep dissection (H12) Sinus pathways (H27) Parasagittal section of the head and neck (HW30) Superficial face (HW44) Superficial facial nerves + parotid gland (HW45) Transverse section of the head (HW49) Sagittal section of head and neck with infratemporal fossa and carotid sheath dissection (HW50) Parotid gland and facial nerve dissection (HW7)
Neuroanatomy	Brain stem (BRW7) Brain stem (BRW10) Brain (Hemi section) (BRW12) Brain (Cerebellum) (BRW18)
Thorax	Thoracic cross section at T6 (AW17A) Pericardial space (TW38) Thorax with heart and vessels (TW42) Heart (TW48) Hilum of the left lung (TW60) Lung slab, hilum removed (TW61) Right lung, hilum removed (TW62) Hilum of the right lung (TW63)
Abdomen	Abdomen with bilateral hernias (A8) Abdomen (AW17) Vasculature of the spleen (AW34) Abdomen with inguinal hernia (AW36) Stomach (AW42) Spleen and pancreas (AW43) Liver with vessels and gall bladder (AW45) Internal anterior abdominal wall (AW47)
Pelvis	Female hemipelvis and thigh (LW91) Female pelvis deep dissection (PW05) Male hemipelvis and thigh (PW10)

2.2. Image data acquisition & manipulation

High resolution digital capture of anatomy specimens was performed using an Artec Spider™ and/or Space Spider™ hand-held 3D scanner (Artec Group, Luxembourg) (Figure 1) with a manufacturer stated 3D point accuracy up to 0.05 mm and 3D mesh resolution up to 0.1 mm. This generation of handheld scanner captures geometry as well as texture (e.g., color) information from the specimen which is then applied to the mesh file in the associated software (Artec Studio 14 Professional, Artec Group, Luxembourg) [19,22]. Anatomy specimens that included internal geometry or deep spaces, which could not be captured by the Artec Spider, were also CT scanned to create a full internal model of the specimen. Specimens were CT scanned using a 32 slice Siemens GoUp™ CT scanner at Monash Biomedical Imaging (Monash University, Australia) with a reconstructed isometric voxel dimension of ∼0.4 mm. The reconstructed CT data were imported into the CT editing program Mimics 23 (Materialise, Belgium) where the CT images were segmented to form a 3D model and exported as a Stereolithography (STL) file. This file was imported into a CAD modelling program (Geomagic Wrap, 2015, 3D Systems, USA) to clean and export again as a STL. The geometrically-modified mesh was then imported into Artec Studio and aligned with a 3D mesh generated using the Artec Spider/Space Spider of the same specimen. This realigned CT-derived mesh was then progressed into texturing using the Artec Spider/Space Spider texture information using the default texture algorithm settings. The textured mesh was exported as a VRML with an associated JPEG texture map; this texture map file was then edited in Adobe Photoshop CS/2020 (Adobe, Inc., USA) to manually adjust the contrast, reduce saturation and color balance to match the original specimen. However, in cases where contrast in the original specimen was poor (likely due to the time it had been in the specimen bottle or ‘pot’), we were able to marginally increase some of these parameters to make the print appear slightly fresher.

Figure 1.

A demonstration of capturing anatomical specimen geometry and texture (color) using a hand-held Artec Space Spider scanner used in developing the Series 1.1 digital files and 3D prints.

An example of the direct comparison of the capture of 3D topography between a CT-based approach and use of the Artec Space Spider™ is provided in Figure 2. While both data capture methods can yield broadly similar 3D meshes, we generally note that the generation of a user-thresholded mesh from medical CT scans frequently requiring more aggressive mesh surface smoothing to remove step-artefacts introduced by voxel boundaries (Figure 2A). In contrast, the 3D meshes generated by the Artec scanner provide sharper relief (e.g., skin wrinkling, muscle fibres, very small diameter vascular and nervous branches) and topographic gradations that capture the natural surface contours and variation in the specimen (Figure 2B). When coupled with the capture of original texture data, 3D meshes generated by the Artec scanner that were then 3D printed imparted both a visual appearance and haptic feel that lacked the artificially smoothed qualities of those produced from a CT-derived mesh.

Figure 2.

Comparisons between a CT-derived 3D mesh (A) and Artec Space Spider™ 3D mesh (B) of the H11 right head and neck specimen. On the left, the overall capture of the topography by both methods is broadly equivalent (with the added capture of color information by the Artec Space Spider™). On the right, enlarged images (and with the color removed for better comparison) demonstrate the overall smoother appearance in the CT mesh and the fine detail/sharper topographic relief of features of the Artec scan mesh.

To aid in use in lab teaching settings where key structure highlights would be useful, a subset of the 29 selected specimens were subjected to partial false-coloring using 3D Coat (v4.0 Ukraine) (Figures 3, 4, 5 and 6A). This false-coloring step could be used to highlight or augment the color of vessels, nerves and unique anatomy present in the specimens to aid in their educational value. For four of our Series 1.1 specimens, we also generated a unique specimen version that used false-color to indicate adjoining organ impressions that are commonly taught in lab settings (e.g., the renal impression on the liver; see Figure 6A). The transparency of the colouring was set to 40% so that it did not overpower the ‘realistic’ elements of this newer approach and retained the original specimen feature texture gradations, and significant structures were manually ‘painted’ to add colour (Figure 3). The final files were exported from 3D Coat in VRML format ready for 3D printing.

Figure 3.

Screenshots of the A8 abdominal specimen demonstrating the (A) original color captured by the Artec Space Spider scanner after processing in Artec Studio and (B) after application of a selective, semi-transparent false-color overlay using 3D Coat.

Figure 4.

Comparisons between the new Series 1.1 3D prints and our original Series 1 3D prints for equivalent anatomical regions/specimens. (A) The Series 1.1 TW48 heart (left) in contrast to the Series 1 MAS156 heart (right) demonstrating the difference in color presentation in the two methods of 3D data acquisition, processing and printing. (B) An example of a 3D printed heart in Series 1.1 created using the technology described in this manuscript. Note the reproduction of the cadaver coloration. B, the Series 1.1 A8 abdominal specimen (left) in contrast with the Series 1 MAS102 posterior abdominal wall specimen.

Figure 5.

A selection of 3D prints from Series 1.1 as summarized in Table 1.

Figure 6.

Examples of specialized false-coloring overlay options to indicate organ impressions on our Series 1.1 liver (AW45) and spleen (AW34) replicas (A). The use of clear material for supporting structures (arteries ‘freed’ from supporting tissues via dissection) (B).

3. Results

3.1. Three-dimensional printing

Since our original work on producing anatomy replicas with additive manufacturing in 2014 there has been a significant expansion and diversification in the types of 3D printers commercially available which use a variety of media, substrates and printing techniques. In our previous studies [1,20] we utilized a 3D Systems Z650 (3D Systems, Inc., Rock Hill, SC) powder printer or 3D Systems Project 4500 plastic powder printer. However, we found these printers to be insufficient to capture the color detail and surface fidelity required for quality reproduction of specimens requiring photorealistic reproductions of textures (such as pathological specimens; [21]. Equally, the powder-based build materials lacked sufficient mechanical resistance to breakage with regular handling in the classroom which led to eventual damage to unsupported thin elements like peripheral neurovascular structures. To resolve this, we purchased a Mimaki 3DUJ-553 (Mimaki Engineering Co., Ltd., Nagano, Japan) full color, UV curable resin printer. This printer has a resolution and layer thickness of 42 μm and 600 dpi down to 22 μm and 1270 dpi depending on the printing mode chosen. Specimens were printed in Standard Mode which used a print layer height of 32 μm and color resolution of 800 dpi which sufficiently depicted the color accuracy and geometry of the original specimens (Figures 4 & 5).

Prior to printing all models were shelled to give a wall thickness of 4 mm using the mesh editing software Geomagic Wrap. This meant that all models were then filled internally with support ink. For specimens that were deemed too heavy due to the infill of the support material in the model, the backs or undersides not depicting anatomy were digitally “cut off” using Geomagic Wrap, to enable the support material to be dissolved away. The resultant model was thus just the 4 mm shell.

The Mimaki 3DUJ-553 also has the ability to print using clear resin. We took advantage of this feature to strengthen thin structures within our models which could easily be broken with handling; further reinforcing thin structures that already expressed greater robusticity with the use of UV curable resin over earlier prior powder materials (Figure 6B). Structures such as nerves and vessels were digitally selected in Geomagic and a small wedge was created to fill the space between the vessel and the geometry below. This wedge was exported as a .STL file and imported along with the model into the Mimaki 3D Link software where printing properties can be assigned. The wedge models were assigned clear ink, and the rest of the model assigned full color ink. The resulting 3D printed part is a full realistically colored replica, with hidden clear sections to support fragile structures.

4. Discussion

The modifications seen in this new series of 3D anatomy replicas (Series 1.1) have taken advantage of new technologies to improve and enhance our original work applying additive manufacturing within the realm of anatomy education. The primary improvement was the result of combining surface scanning technology with current UV curable resin printers to reproduce high-fidelity, photographic quality replicas of the original specimens chosen for this series. This obviates some of the perceived limitations of specimen fidelity with our Series 1 replicas, particularly where the false-coloring approach could be viewed as oversimplifying or artificially removing some challenges in learning environments (e.g., providing simple visual cues for structure identity that reduced utility in some teaching environments or assessments). We believe our new series (Series 1.1) described here, in using real surface color (albeit slightly augmented in some cases), provides a greater perception of fidelity and resemblance of these 3D printed replicas to the original donor tissues. That said, and as we noted in our original study, we all recognize that donor tissue color in an embalmed and preserved state is ultimately altered and artificial relative to the original tissue colors prior to chemical treatment [1]. Evaluation studies, similar to those that showed the effectiveness of 3D anatomy prints in the classroom [18], are planned as the next logical step to gather direct evidence of the educational value of these new 3D printed replicas. A comparison of false colored anatomy 3D prints (Series 1) versus realistic colored 3D prints (Series 1.1) should provide interesting insights into the role of photorealism in student perception and learning.

There are, of course, some continued limitations to any 3D printed replicas of human anatomy specimens using current technologies. First, the output of 3D imaging and printing is only as good as the data input, and can only reproduce structures defined through dissection and/or digital segmentation. This not only reinforces the underlying reliance on donor tissue quality, but also careful and strategic planning of dissections with the eventual goal of 3D scanning and printing (which influences the balance of feature and structure retention and dissection approaches). A further current limitation is the lack of pliability or haptic properties or qualities when developing 3D printed replicas (whether false colored as in our Series 1 or ‘true color’ as in Series 1.1). While the range of 3D printing materials has notably diversified in the last decade and is expanding into lower range shore hardness (that approach the pliability of biological tissues), there has yet to be a commercially available solution that circumvents the trade-offs between material properties and color in the resultant 3D print.

There are also financial issues to consider when pursuing the use of 3D printing for generating anatomical specimens. Broadly speaking, these fall into the major categories of equipment and facilities access (3D printer, surface scanner, CT facilities), materials costs (consumables, waste disposal), operations costs (service, utilities) and labor (data acquisition, post-processing, academic consultation, printing, equipment operation and maintenance, post-print processing). Because of the highly variable specific costs under each of these categories, particularly across international academic and research institutions, it is difficult to provide meaningful indicative or specific costs. From our observations and experience from our original, CT-based approach to Series 1 to our near-exclusive use of surface scanning in Series 1.1, we have found that data acquisition was both simpler and less expensive when using surface scanning. Use of CT facilities may require regulated transport of human tissues to offsite facilities, access to a (typically nonclinical) CT scanner amenable to imaging donor tissues, and reliance on the labor of trained CT operators. The overall data acquisition via CT is, however, generally faster allowing for the acquisition of multiple specimen datasets per hour relative to the need for multiple scan passes and angles with the surface scanner. Developing a completely false-colored model from a CT scan dataset also tends to be somewhat faster from a labor standpoint (given that it is essentially 3D paint-by-numbers) than the more refined process of adjusting and modifying false-color overlays on surface scanner-acquired color textures. Finally, the cost of a 3D printer itself is both the central and typically greatest expense to consider. At present, photorealistic color reproduction is only available in a limited range of 3D printers, so comparisons to desktop printers are not applicable for the method described here. We note that the current costs of UV curable resin inks used by our Mimaki 3DUJ-553 are nearly equivalent to the resin costs for ‘desktop’ SLA printers (e.g., Form printers, FormLabs, USA); however, while ‘desktop’ SLA printers generate low material volume structural supports, the full encapsulation of prints in a water-soluble support material (like in the Mimaki) does add additional costs to replica production.

Finally, we would emphasize that the ultimate financial considerations when pursing 3D printing for anatomical replica production may be very idiosyncratic – but the rise of local, national and even international 3D imaging and printing services can circumvent establishing in-house facilities and some of these costs. With the expansion of ‘outsourced’ 3D scanning and printing facilities from academic and research institutions, we anticipate reductions in the financial and logistical barriers to those considering 3D printing solutions to challenges in anatomy or biomedical education.

The new series of 3D anatomy replicas described herein, continues to reinforce the many advantages of 3D printed reproductions of normal prosected anatomical specimens over traditionally-produced anatomy models or reliance on plastinated cadaver specimens. First, with 3D printing it is possible to reproduce an indefinite number of 3D printed copies, which is ideal for the large class sizes found in many centers and for teaching with social distancing in the new COVID-19 era. Furthermore, although preparation of plastinated specimens from cadavers sourced from local bequest or post-mortem programs is a possible alternative to 3D printing, plastination involves considerable infrastructure costs as well as health and safety issues because of the large volumes of flammable solvents involved. In addition, plastinated specimens in many jurisdictions are still classified as human tissue and thus require approved licensed and regulated facilities. In contrast, 3D printed replicas can be deployed in any teaching environment, for example in peripheral or rural medical school locations and even teaching hospitals and clinical environments. Furthermore, it is also not widely appreciated that plastinated specimens are also false colored so are not entirely the colors of conventional embalmed cadavers [23–25].

Although there remains an ongoing discussion about the role of cadaver dissection in modern medical undergraduate training [26–28] all parties would agree there are now a plethora of other learning resources that were not available in the recent past. The continued advances in 3D imaging and additive manufacturing technology, as highlighted in the present description, further establishes 3D printing as a complement, supplement or alternative to dissection for anatomy educators and students. It thus joins a list of alternatives that includes plastination [23], two-dimensional and 3D imaging [29], body painting [30] and a diversity of digital radiographic visualization tables, software applications, virtual reality and augmented reality [31]. This ever-increasing portfolio of physical and digital anatomy resources is not only to the benefit of institutions maintaining cadaver dissection-focused curriculum, but also critical to those diversifying and improving the resources available to medical and allied health educators where human donor tissues are not an accessible option.

5. Conclusion

The resources available for anatomical education is constantly evolving, with a significant emphasis over the past decade on novel technological innovations that can enhance or supplant traditional methods. While representing a significant innovation, our original development of 3D printing methods and applications in anatomy education reflected the data acquisition and additive manufacturing capabilities available in the early 2010s. Here we have reported on the most recent generation of consumer-accessible surface scanners and 3D printers as the next major milestone in the direct application of these devices to develop practical content for use in learning environments. The combination of high-fidelity geometric capture, photorealistic reproduction and UV resin-based printing allows for significant innovation in replica design and use in modern anatomy teaching environments.

Author contributions

All authors contributed to the idea conception, writing and editing of this piece.

Financial disclosure

The 3D printed anatomy specimens described here as series 1.0, 1.1 and 2.0 are available commercially through agreements between Monash University and external partners. Royalty agreements conform with Monash University policy.

Ethical conduct of research

All cadaver tissues used in this described research are those catalogued and reposited for teaching purposes within the Centre for Human Anatomy Education in the Department of Anatomy and Developmental Biology, within the School of Biomedical Sciences at Monash University, Melbourne, Australia. These cadaver tissues include donations that have been professionally dissected across several decades of active collections development for teaching purposes. No additional human tissues or cadaver tissues were acquired from any other sources. Permissions to access and use these specimens were obtained according to the University's policy.