Dimensional accuracy and reproducibility of three different

methods of 3D printing using a known size object

Research Project

Submitted in partial fulfilment of the requirements for the degree of Master of Science

in Maxillofacial and Craniofacial Technology

Francisca Amanda Velasquez Cerda

Academic Centre of Reconstructive Science
Abstract

Objective: To assess the accuracy and reproducibility of FDM (Cubicon Single Plus,

Cubicon Inc., Korea) and Polyjet (Objet Eden 250, Stratasys Ltd., USA) 3D printers, by

comparing linear dimensions between 3D models, printed using photocurable resin

(RGD720 & SUP705, Stratasys Ltd., USA), ABS and PLA (Cubicon Inc., Korea), and

the original 3D image, to establish which one is more accurate and cost-effective.

Methodology: An irregular cube was designed using Autodesk Fusion 360 software and

exported as an STL file; 10 cubes of each material were printed and measured in 8

allocated dimensions and their 4 corners. These measurements were compared to the

dimensions of the original 3D image and accuracy was evaluated according to the

dimensional differences between the models and the 3D image.

Results: There was a highly significant difference between resin and ABS, and between

ABS and PLA; this difference was not clinically acceptable. On the other hand, there was

no significant difference between resin and PLA, and both materials are considered

clinically acceptable.

Conclusion: Both Polyjet’s resin and FDM’s PLA are equally accurate, therefore, the

selection of the manufacturing method would depend on the clinical use of the printed

object. Polyjet’s resin can be recommended for the fabrication of surgical guides and to

shape medical devices like fixation plates, as higher surface details and mechanical

properties are required. FDM’s PLA can be suggested for the fabrication of models for

surgical simulation, treatment planning and patient education, as it is equally accurate and

more cost-effective than resin.

Acknowledgments

I would like to express my deepest gratitude to my supervisor, Dr Trevor Coward, for his

unconditional support, valuable advice, encouraging words and incredible patience. Also,

I would like to thank him for his amazing role as a mentor for the past two years.

I would like to extend my gratitude to Dr Brendan Scott for his guidance and help with

the statistical analysis of this research and to Mr Alessandro Maffei for helping me with

the design of my experimental model and teaching me the intricacies of 3D printing.

I would like to thank Ms Naimesha Patel for her constant help during these years of

training, for her disposition, dedication and for being a great example of discipline and

effort. To Dr Lorenzo Veschini, Ms Caroline Reed, Ms Gabriela Bester and the entire

team at the prosthodontics lab for sharing their knowledge, their constant support and for

being an important part of my training.

I would like to thank Poppy Taylor and Chia-en Tsai for being the best classmates I could

ever ask for; thank you for the laughs, the emotional support, the camaraderie and for

making this stage of my life something memorable. To all my amazing friends (I cannot

list all the names here, but you know who you are) for always being present, especially

in the hardest moments, and for always making me feel morally supported, even in the

distance.

I want to thank my parents, for their fundamental role in my education and for never

doubting my abilities, making me feel like I was capable of achieving everything. Thank

you for your unconditional trust, for inspiring me to follow my dreams and pursue my

own destiny.

Last but not least, I would like to thank the National Agency for Research and

Development (ANID) for sponsoring my program (MAGISTER BECAS CHILE 2018 –

73190154).
Table of Contents

1. INTRODUCTION ....................................................................................................... 1

2. REVIEW OF THE LITERATURE........................................................................... 3

2.1 THREE-DIMENSIONAL PRINTING............................................................................... 4

2.1.1 Liquid-based systems: ...................................................................................... 6
2.1.2 Solid-based systems .......................................................................................... 8
2.1.3 Powder-based systems .................................................................................... 11
2.2 3D DATA ACQUISITION .......................................................................................... 14
2.2.1 Computed Tomography (CT) ......................................................................... 16
2.2.2 Magnetic Resonance Imaging (MRI) ........................................................... 17
2.2.3 3D Laser Scanning ........................................................................................ 18
2.2.4 Structured light scanning .............................................................................. 19
2.2.5 Digital stereophotogrammetry ....................................................................... 19
2.3 CLINICAL APPLICATIONS OF 3D PRINTING............................................................... 21
2.4 SELECTION OF THE APPROPRIATE 3D PRINTING SYSTEM ......................................... 27
2.5 CONCLUSIONS FROM THE LITERATURE ................................................................... 32

3. AIMS AND OBJECTIVES ...................................................................................... 33

4. NULL HYPOTHESIS ............................................................................................... 33

5. METHODOLOGY.................................................................................................... 34

5.1 DESIGN OF CUBE ..................................................................................................... 34

5.2 SELECTION OF 3D PRINTING TECHNIQUES ............................................................... 34
5.3 PILOT STUDY .......................................................................................................... 35
5.4 MAIN EXPERIMENT: 3D PRINTING OF CUBE ............................................................ 36
5.5 DIMENSIONAL MEASUREMENTS .............................................................................. 38
5.6 STATISTICAL ANALYSIS .......................................................................................... 41

6. RESULTS .................................................................................................................. 42

6.1 RESULTS OF PHASE 1 .............................................................................................. 42

6.2 RESULTS OF PHASE 2 .............................................................................................. 49

7. DISCUSSION ............................................................................................................ 54

7.1 REPRODUCIBILITY OF MEASUREMENTS ................................................................... 54

7.2 ACCURACY OF 3D PRINTING ................................................................................... 55
7.3 STATISTICAL ANALYSIS .......................................................................................... 55
 7.4 AGREEMENT OF FINDINGS WITH AVAILABLE LITERATURE ...................................... 56
7.4.1 Reasons for accuracy difference between techniques .................................. 56
7.4.2 Measuring limitations .................................................................................... 64
7.4.3 Dimensional accuracy ................................................................................... 64
7.5 CLINICAL APPLICATIONS......................................................................................... 66
7.5.1 Cost-effectiveness ........................................................................................... 67
7.6 NULL HYPOTHESIS .................................................................................................. 68
7.7 FUTURE RESEARCH ................................................................................................. 68

8. CONCLUSIONS ....................................................................................................... 69

9. GLOSSARY OF TERMS ......................................................................................... 70

10. REFERENCES ........................................................................................................ 73

11. IMAGE CREDITS .................................................................................................. 83

12. APPENDIX .............................................................................................................. 84

12.1 DIMENSIONS OF THE DESIGNED CUBE ................................................................... 84

12.2 IMAGES OF SAMPLES OF THE 3D PRINTED MODELS ............................................... 85
List of Figures

Figure 1. 3D printing workflow………………………………………………………….4

Figure 2. Stereolithography process……………………………………….…………….7

Figure 3. Polyjet Process………………………………………………………………...8

Figure 4. Fused Deposition Modelling…………………………………………………..9

Figure 5. Laminated Object Manufacturing……………………………………………10

Figure 6. 3D Systems Multi-jet Printing………………………………………………..12

Figure 7. Selective Laser Sintering……………………………………………………..13

Figure 8. Designed irregular cube……………………………………………………...34

Figure 9. Sample of 3D printed cubes in photopolymerizing resin (left), ABS (middle)

and PLA (right)…………………………………………………………………………37

Figure 10. Surface A showing D1 and surface B showing D2………………………….38

Figure 11. Surface C showing D3 and surface D showing D4 and D5………………….39

Figure 12. Top surface showing D6 and D7…………………………………………….39

Figure 13. Top and bottom surfaces showing the selected points for Central Cube…….40

Figure 14. Measured corners. Corner D is opposite to Corner B………………………..40

Figure 15. Mean difference values between the 8cm Dimension 1 image and the three

materials………………………………………………………………………………...43

Figure 16. Mean difference values between the 7cm Dimension 2 image and the three

materials………………………………………………………………………………...44

Figure 17. Mean difference values between the 5cm Dimension 3 image and the three

materials………………………………………………………………………………...44
Figure 18. Mean difference values between the 6cm Dimension 4 image and the three

materials………………………………………………………………………………...45

Figure 19. Mean difference values between the 3cm Dimension 5 image and the three

materials………………………………………………………………………………...46

Figure 20. Mean difference values between the 1cm Dimension 6 image and the three

materials………………………………………………………………………………...46

Figure 21. Mean difference values between the 2cm Dimension 7 image and the three

materials………………………………………………………………………………...47

Figure 22. Mean difference values between the 5cm Dimension Central Cube image and

the three materials………………………………………………………………………48

Figure 23. ABS cube showing a curvature in the bottom corners……………………….50

Figure 24. Mean difference values between the 8cm Corner images and the three

materials………………………………………………………………………………...51

Figure 25. Comparison of surface texture of the upper and lower faces of the cubes,

between Polyjet (top) and FDM (bottom)……………………………………………….62

Figure A1. Dimensions surface A (left) and B (right)…………………………………..84

Figure A2. Dimensions surface C (left) and D (right)…………………………………..84

Figure A3. Dimensions top (left) and bottom (right) surfaces………………………….85

Figure A4. Sample of 3D printed cube in photopolymerizing resin…………………….85

Figure A5. Sample of 3D printed cube in ABS…………………………………………86

Figure A6. Sample of 3D printed cube in PLA…………………………………………86

List of Tables

Table 1. Summary of 3D imaging systems……………………………………………..21

Table 2. Summary of 3D printing systems……………………………………………...28

Table 3. Details of Polyjet and FDM printers…………………………………………...35

Table 4. Main properties of RGD720, ABS and PLA materials………………………...37

Table 5. Coefficient of repeatability for the 8 allocated dimensions……………………42

Table 6. Two way ANOVA results for Phase 1…………………………………………48

Table 7. Bonferroni results for comparison between materials for Phase 1……………..49

Table 8. Coefficient of repeatability for corners A, B, C and D…………………………50

Table 9. Two way ANOVA results for Phase 2…………………………………………52

Table 10. Bonferroni results for comparison between materials for Phase 2……………53
1. Introduction

Surgical reconstruction of lost facial features due to trauma, cancer and congenital

anomalies is often restricted by insufficient remaining hard and soft tissue after surgical

treatment, and most of the time, there are vascular alterations subsequent to radiation

therapy, causing decay or failure of reparative mechanisms. Hence, in selected patients,

where aesthetic and functional necessities are bigger than the options of getting a surgical

reconstruction, a facial prosthesis should be considered as a viable alternative (Nuseir et

al., 2015).

Maxillofacial prostheses provide aesthetic and psychological improvements for these

patients, particularly considering that these alterations completely deteriorate their quality

of life, as these modifications in their facial anatomy directly affect their functional,

emotional and psychosocial well-being; especially because of the major impact on vital

functions such as breathing, eating and communication.

Historically, the fabrication of these prostheses has been a time-consuming process,

where the sculpting and colouring of the final prosthesis relies on the clinician’s artistic

abilities to precisely reproduce the shape and colour of the missing anatomy, as well as

the surrounding and opposite structures to conceal the prosthesis (Watson and Hatamleh,

2014). Nowadays, the advances in 3D Imaging, Virtual Treatment Planning and 3D

Printing have significant benefits compared to the conventional techniques and the

utilization of these new technologies results in a more time-effective fabrication process.

3D printing can have several uses, for example, improving treatment planning, surgical

simulation, patient education and fabrication of medical devices. Depending on the

clinical purpose of the element to be 3D printed, is necessary to select the most

appropriate technology according to cost, accuracy, reproducibility and available

1
materials. In order to do this, it is essential to compare the different products available in

the market and the characteristics of the printed element. Ideally, the selected printer

should be the one that is most cost-effective according to its clinical use.

The purpose of the present study was to evaluate the accuracy and reproducibility of three

different methods of 3D printing using a known size object, by comparing the dimensions

of the printed elements to the original 3D image, in order to elucidate which printing

system is more accurate and efficient in the fabrication of computer designed medical

items.

2
2. Review of the Literature

The use of three-dimensional (3D) printing technologies has increased considerably in

the last three decades, starting in 1986, when the first stereolithographic (SLA) systems

was introduced (Aimar et al., 2019). One of the areas that has seen the greatest growth of

these technologies is the field of medicine, where their applications include medical

research, education and training, development of personalized treatments and fabrication

of models for surgical planning or procedure rehearsal for both hard and soft tissues.

All these printing technologies are driven by computer data describing the 3D object to

be fabricated. For medical applications, the data may come from imaging systems such

as computed tomography (CT), magnetic resonance imaging (MRI), 3D surface scanning

or digital stereophotogrammetry, among others. The advances in medical imaging have

improved their potential as diagnostic tools, by reducing their invasiveness and upgrading

the quality of the image acquired. These imaging systems can also be used for research,

treatment planning, assessment of post-operative results and monitoring of morphological

changes due to growth and education.

These images can later be post-processed with a variety of technologies, such as

multiplanar reformation or 3D visualization, to be utilised as an important tool for

diagnosis and treatment planning. Additionally, with the integration of 3D printing

systems, these data can evolve from images on a flat screen into tangible 3D models that

can have a plethora of medical applications (Kim et al., 2016). The 3D printing workflow

can be observed in Figure 1.

3
 Figure 1. 3D printing workflow

2.1 Three-Dimensional Printing

To produce a physical model in medicine, two different methods have been utilized:

Subtractive and Additive Manufacturing.

Subtractive Manufacturing is a process by which three-dimensional (3D) objects are

fabricated by successively cutting material from an initial block to leave the desired

shaped part. This is usually accomplished with a Computer Numerical Control (CNC)

Machine. This machine utilizes multiple tools (like drills, lathes and mills), to cut around

at least three axes (X, Y, and Z), minimising the need for the designers to flip the material

block (Torabi et al., 2015).

Depending on its application, there are several materials that can be used in a CNC

machine. Common materials include metals such as aluminium, steel, and titanium, as

well as ceramics, fiberglass, and plastics such as polypropylene.

4
The subtractive process is used typically in small model-making machines, so their

clinical application is mainly in dentistry to fabricate metallic and/or ceramic crowns.

Subtractive manufacturing has some limitations when compared with additive techniques

(Bae et al., 2017; Torabi et al., 2015):

- The precision of the model depends of the size of the smallest usable tool for each

material and if the cutting tool has a bigger diameter than some parts of the

designed object, this could lead to a reduction in accuracy or inferior marginal

properties.

- The unused portions of the mono-blocks must be discarded; therefore, a

considerable amount of material is wasted.

- Milling tools can only be used for short periods of time, as they are exposed to

heavy abrasion and wear.

- Due to the fragility of ceramics, microscopic cracks can be generated in the

machining of this material.

- It is difficult to machine complex milling parts or undercuts that are inaccessible

to machining tools.

On the other hand, Three-dimensional (3D) printing, also known as Additive

Manufacturing, is described as the design of 3D computer models employed in the

reconstruction of a 3D object by the addition of thin layers of material, such as plastic,

metal, plaster, among others (Martelli et al., 2016). The layer thickness is determined by

the process that deposits the material, typically varying from 0.05 to 0.30mm.

The necessary time to build a model depends on the chosen 3D printing machine; it is

important to consider that with a thinner deposition layer, comes a higher resolution, but

a longer building time.

5
A growing diversity of 3D printing methods are available, varying in price, accuracy and

materials. A basic classification for this technology is based on the state of the raw

printing material: Liquid-based systems, Solid-based systems and Powder-based

systems.

2.1.1 Liquid-based systems:

[Link] Stereolithography Apparatus (SLA)

SLA is one of the most well-known 3D printing techniques. The system uses a vat filled

with liquid resin, which is polymerised layer by layer, and a build platform that is

immersed in the resin (Figure 2). A UV laser beam scans the section representing the

printing area and solidifies the first layer. A levelling blade is utilized to smooth the

surface, then the build platform is lowered by one-layer thickness (typically 0.05-0.07

mm) and the process is repeated. Overhanging areas are supported by scaffold structures

built at the same time as the model. When the whole model is built, the unused material

remains liquid, so it can easily be removed from internal spaces. Excess resin and the

supports are removed with manual finishing and solvents. Present medical applications

for this system include construction of anatomical models for surgical planning and

indirect fabrication of medical devices by using SLA printed moulds (Beumer III et al.,

2011; Nyberg et al., 2017; Robiony et al., 2008).

6
 Figure 2. Stereolithography process

[Link] Polyjet Process

Polyjet Process also uses ultraviolet cured resins, but these are deposited onto a build

platform by jetting heads, similar to those of inkjet printers (Figure 3). The system utilizes

two different types of liquid photo-polymerised resin, one for build material and one for

support material. The deposited material is solidified by a UV lamp, attached to the print

heads mechanism, and then a levelling blade generates a smooth surface, before the build

tray is lowered layer by layer. When the model is finished, the support material must be

washed off with water and no other finishing processes are required. Polyjet constructs

models in very thin layers (minimum layer thickness 0.015mm), providing high accuracy

and an even surface. Some printing materials include polymers such as polymethyl

methacrylate (PMMA), polypropylene, acrylonitrile butadiene styrene (ABS) and waxes.

7
Medical applications for this system include surgical planning through anatomical

prototypes and dental appliances (Beumer III et al., 2011; Nyberg et al., 2017).

Figure 3. Polyjet Process

2.1.2 Solid-based systems

[Link] Fused Deposition Modelling (FDM)

Plastic or wax materials in filament form are extruded through a heated nozzle in a semi-

liquid state to fabricate a model (Figure 4), the nozzle keeps the plastic at a temperature

just above its melting point, this way it flows easily to form a layer. The build table lowers

by one-layer thickness and the next layer is produced. The material solidifies and bonds

to the previous layer as a result of the lower temperature of the build chamber, in

comparison to the heated nozzle. Support areas are built with the model from a second

8
material, which can be removed by hand or immersion in detergent and water. Commonly

used materials are acrylonitrile butadiene styrene (ABS), polycarbonate, polylactic acid

(PLA) and polyphenylsulfone. The produced models are very strong and durable, which

makes them ideal for teaching; but on the other hand, accuracy and layer thickness are

determined by the nozzle´s diameter, which ranges from 0.33 to 0.13 mm. The relatively

simple process makes this technology widely available, as the machines and print

materials are more affordable compared to other systems. FDM has been reportedly used

to fabricate tissue engineering scaffolds and to print models for the rehabilitation of large

craniofacial defects (Beumer III et al., 2011).

Figure 4. Fused Deposition Modelling

9
[Link] Laminated Object Manufacturing (LOM)

In this system, parts are built up in sections from thin sheets of plastic, paper or polyester

composite. As the model is being constructed, each sheet is glued with adhesive to the

already built part, a laminating roller heats the adhesive to bond it and then the part is

trimmed with a laser, which only traces the layer outline and cross hatches any remaining

material (Figure 5). The building platform is lowered layer by layer (typically 0.05-0.5

mm) until the model is finished. LOM is fast compared with other 3D printing techniques

and uses simple and inexpensive technology. Nonetheless, the selection of materials is

limited and a strong bond between cross sections is necessary in order to form a solid

part, so the mechanical properties would depend on the strength of the adhesive (Chua

Chee, 1994).

Figure 5. Laminated Object Manufacturing

10
2.1.3 Powder-based systems

[Link] 3D Systems Multi-jet Printing (3D MJP)

The models are built using an inkjet that prints liquid binder solution onto a powder bed.

Two chambers are filled with powdered material; the build chamber and the feeder

chamber (Figure 6). The process begins with the building platform inside the build

chamber being raised by a piston while a levelling roller distributes a thin layer of powder

on the top of the build chamber, then an inkjet print head deposits liquid adhesive droplets

that binds the powder in selective areas. The loose powder works as a support to the part

that is being formed. The piston, powder bed and model are slightly lowered by one-layer

thickness (typically between 0.016-0.05mm), the levelling roller pushes a new layer of

powder into the build chamber and the surface is levelled; the process is repeated until

the entire model is completed. Commonly used materials in this technique are elastomers,

composites and ceramics. Some of the reported medical uses for this method include the

fabrication of porous scaffolds for tissue engineering, especially cartilage and bone

regeneration and anatomical models for surgical planning (Beumer III et al., 2011; Jardini

et al., 2014).

11
 Figure 6. 3D Systems Multi-jet Printing

[Link] Selective Laser Sintering (SLS)

In this technique, two chambers are filled with material (composites, elastomers,

thermoplastics such as nylon, polystyrene and polyamide) in powder form: the build and

the feeder chambers (Figure 7). In the build chamber, the material is maintained at high

temperature so that it fuses easily when exposed to the laser. The laser traces the first

layer and selectively fuses together particles of powder on the build platform. The powder

bed is then lowered by one-layer thickness and fresh powder from the feeding chamber

is laid over the previous layer. The second layer is traced by the laser beam, which

solidifies it and fuses it to the previous one, allowing the sequential building of the model

layer by layer (minimum layer thickness 0.1mm). The surrounding unused powder serves

as the support structure.

12
Direct Metal Laser Sintering (DMLS) is the same process as SLS, but using metals, such

as stainless steel, titanium and cobalt chromium. In plastic machines, the powder is heated

close to its melting point, but the particles don’t melt completely, leading to porous

models, which can make cleaning and sterilizing more difficult when the model is going

to be used for surgical applications, like the construction of anatomical models or porous

scaffolds for bone tissue engineering (Beumer III et al., 2011; Martelli et al., 2016;

Nyberg et al., 2017).

Figure 7. Selective Laser Sintering

13
2.2 3D Data Acquisition

All these printing systems are driven by computer data describing the 3D object to be

built. For medical applications, the data may come from sources such as computed

tomography (CT), magnetic resonance imaging (MRI), 3D surface scanning or digital

stereophotogrammetry, to name a few. These files are typically converted and saved in

Digital Imaging and Communications in Medicine (DICOM) or Wavefront object (OBJ)

format and later translated into a standard format that all 3D processes can utilize, the

Standard Tessellation Language (STL) file format (Beumer III et al., 2011; Flores et al.,

2017).

The utilization of 3D digital imaging has numerous assets, such as reduced storage

requirements, high acquisition speed, no degradation and fast transfer of data (De

Menezes et al., 2016). Additionally, in the case of 3D digital photography, one of the

major advantages is the possibility to evaluate the image’s settings, such as contrast,

brightness, subject position and so on; and if any of these characteristics are not

appropriate for the image’s purpose, they can be adjusted and the image can be repeated

instantly (Schaaf et al., 2009).

There are numerous clinical applications for 3D imaging techniques, such as: diagnosis,

anatomy measurements, prosthetic and surgical treatment planning, assessment of post-

operative results in maxillofacial surgery, fabrication of custom made surgical/cutting

guides, accurate implant position and patient education (Meulstee et al., 2017; Salazar-

Gamarra et al., 2016). Other important applications of 3D imaging involve medical

research, for example, studies about symmetry and evaluation of facial differences

between genders, racial groups, twins and craniofacial malformations through

superimposition techniques (Kau et al., 2019), and can also be used for the assessment of

14
growth or treatment-induced morphological changes, as 3D imaging systems can quantify

linear distances, surface areas, volumes and craniofacial angles; besides, they reduce the

magnification errors generated from geometric distortions that generally affect two-

dimensional imaging methods.

Patel et al., presented a 3D surface imaging method to automatically identify and quantify

soft-tissue facial asymmetry in adults by scanning the subject’s face and comparing the

surface difference between the original and mirrored 3D data using a colour map and

histogram. The proposed method was considered fast, simple and effective, and an

advantageous tool for diagnostic and treatment planning for maxillofacial surgery patients

with soft tissue asymmetry. It can also be a convenient way to improve clinician-patient

communication, as it can be used by the physician to explain the condition and treatment

plan to the patient in a more interactive way (Patel et al., 2015). Similarly, Coward et al.

developed a stereophotogrammetry technique to evaluate the effects of a palatal obturator

in restoring facial form in a patient with a maxillary defect on the right side after the

resection of a giant cell carcinoma; the results were acceptable, as the technique was able

to correctly identify changes in form and volume (Coward et al., 2011).

Papadopoulos et al. proposed a number of basic requirements for 3D imaging

(Papadopoulos et al., 2002):

1. Safety: limited or no ionizing radiation exposure.

2. High speed of data acquisition: lower risk of motion artefacts and patient

discomfort.

3. Low cost: higher clinical availability.

4. Simplicity: minimal or no training required to operate the equipment.

15
 5. Compatibility: exported data should be suitable for use with different computer

software packages and ideally they should use a standard file format.

6. Accuracy: indicates how exact is an image compared to the portrayed object. High

accuracy is necessary, so the difference of measurements between the image and

the object are minimal.

7. Precision: high precision is ideal, as it refers to how consistent repeated images of

an object are to each other.

Balancing these requirements provides us with the key to making the best selection of

imaging technique between some of the following:

2.2.1 Computed Tomography (CT)

Computed Tomography (CT) is an imaging acquisition technique that uses highly

collimated X-rays, to obtain information about a particular section of a subject recreating

tomographic images (slices). A detector is placed behind the object and opposite to the

radiation source, while the subject is in the centre of the rotating mechanism. A series of

cross-sectional slices can be acquired (with a resolution that can range from 0.5mm x

0.5mm to 2mm x 2mm), which can later be reconstructed in 3D images showing both

hard and soft tissues, represented by different levels of grey, as they possess different

radiodensity.

CT scan is an imaging technique that is performed routinely and presents a high

diagnostic value due to its high accuracy (around 0.2 mm) and good contrast between

soft and hard tissues, it can display internal and external anatomy and multiple views at

the same time (Coward et al., 2006).

16
Between its disadvantages, the patient is exposed to ionizing radiation, it has a long

acquisition time (10 to 30 minutes) which increases the risk of motion artefacts.

Moreover, the scan is taken in a supine position which can increase the risk of tissue

distortion. Metallic objects, like dental restorations, can cause artefacts in the image

which can be time consuming to remove or reduce the clarity of the image. It also requires

great clinical space, due to the size of the equipment, and qualified personnel, making it

a very expensive technique (Guttenberg, 2008; Plooij et al., 2011).

2.2.2 Magnetic Resonance Imaging (MRI)

MRI is a non-ionizing radiation imaging technique which uses a powerful rotating

magnetic field, disturbed by radiofrequency fields created by the hydrogen atoms in the

different fat and water content of the subject’s body. Depending on the strength of the

magnetic field, particular tissues of the body resonate, presenting outstanding contrast

between different kinds of soft tissues. It produces multiple axial images that can be

assembled into 3D models.

MRI is a highly accurate technique (around 0.2mm), with excellent diagnostic value of

soft tissues, presents multiple views and outstanding soft tissue contrast. The absence of

ionizing radiation makes it very safe for the patients (Coward et al., 2006).

On the other hand, it also requires a large clinical space, due to the size of the equipment,

and qualified personnel, making it a very expensive technique. It presents lower spatial

resolution when compared with CT scan, no textured surface and bone is not imaged.

This technique requires a long acquisition time (head MRI: 15-30 minutes) which

increases the risk of motion artefacts, and as the patient is scanned in a horizontal position,

17
presents a risk of tissue distortion. Metal objects like dental restorations or implants

generate artefacts in the image (Fullerton, 1982; Plooij, Maal et al., 2011).

2.2.3 3D Laser Scanning

Laser scanning allows the reconstruction of a 3D surface, capturing 2D profiles by

projecting laser lines (stripe) or individual laser dots (points) on the surface of an object.

The laser stripe is captured by a camera positioned at an established distance and angle

from the laser source (this principle is called triangulation), by rotating the object, or the

scanner around the subject, consecutive profiles of the surface can be registered and later

a 3D model can be reconstructed from this information.

In facial imaging, portable laser scanners can be used, which utilize a spectrum between

630nm and 780nm (red and far red), that is considered Class I eye-safe laser stripe

projection. Nevertheless, it is still suggested that the subject maintains his eyelids closed

during the laser scanning process to minimize the risk of eye damage.

Laser scanning presents a wide range of devices and availability of portable units, which

makes it a less expensive system when compared to CT or MRI. It is a very accurate and

precise technique (around 1mm accuracy). The absence of ionizing radiation makes it

very safe for the patient, and as the image is taken in a natural head position, the risk of

tissue distortion decreases.

However, it is considered a slow method (around 30sec for facial imaging), which

increases the risk of motion artefacts. It presents a limited capture of undercut areas and

some devices do not capture surface colour or texture (Bos et al., 2015; Coward et al.,

2000; Riphagen et al., 2008).

18
2.2.4 Structured light scanning

This technique consists of a variety of imaging systems that allow the reconstruction of

a 3D model by triangulation methods, through the projection of a light pattern onto the

subject’s surface.

Consecutive profiles of the desired surface can be registered by two techniques:

Sequential projections of different patterns or by frequently changing the pattern (phase-

shifting techniques).

This method has been applied since the late 1970’s, when the projected pattern was

generated using a physical grid and illumination, but lately the light pattern is computer

generated (Artopoulos et al., 2014).

Structured light scanning is an affordable imaging technique that presents a high

acquisition speed (around 10 seconds for facial imaging) and a high accuracy and

precision, that goes from 10µm up to 0.5mm. As this system does not use ionizing

radiation, is very safe for the patient. The image is taken in a natural head position,

therefore, the risk of tissue distortion is low.

Some disadvantages of this technique include the lack of portable units, the limited field

of view and the need for controlled lighting conditions in a dark environment (Riphagen

et al., 2008).

2.2.5 Digital stereophotogrammetry

Stereophotogrammetry is the science of measurement by using 2 or more images,

captured simultaneously, to calculate the position of a known point that has to be defined

in the recorded images. The photographs are taken under an established angle to obtain

19
topographic information and requires either the subject or the cameras to be stationary.

Digital stereophotogrammetry typically consists of 4-6 digital cameras and flash units on

a fixed frame and computer software that recognises common reference points on the

photographs and uses triangulation to identify the coordinates of a considerable number

of points, generating a 3D model of the subject, with reconstructed surface and texture

(Artopoulos et al., 2014; Davis, 2010; Riphagen et al., 2008; Salazar-Gamarra et al.,

2016).

Digital stereophotogrammetry is an affordable technique, with a wide range of devices

and portable units, it has an extremely high acquisition speed (< 2ms), high accuracy and

precision (0.25mm-0.6mm), captures surface’s texture and colour, and the absence of

ionizing radiation makes it very safe for the patient. The image is acquired in a natural

head position, therefore the risk of tissue distortion is minimal.

Some disadvantages of this technique include image distortion around the borders,

reflection of shiny surfaces (like earrings) and poor image of hair, highly curved small

areas can be difficult to reproduce, such as the eyes and nose. This method also requires

carefully controlled lighting conditions and daily calibration (Chen et al., 2015; de

Menezes et al., 2010; Dindaroğlu et al., 2016; van Loon et al., 2010).

As previously described, two of the main requirements for 3D imaging systems are

accuracy and rapid acquisition time. Acquisition time is relevant when it comes to record

images from living subjects, particularly children, since a slow capture speed will increase

the risk of motion artefacts. A shutter speed of 1/500 seconds or less has been suggested

as sufficient to eliminate inadequate data due to motion (Riphagen et al., 2008).

20
In addition, an inaccurate equipment can lead to dimensional errors and the obtained data

might not be suitable for clinical use, especially when it comes to evaluate landmark

positioning, symmetry and morphological changes. It has been reported in the literature

that a 1mm difference in accuracy is often clinically accepted (Nkenke et al., 2006),

however, Lübbers et al. proposed that for the evaluation of facial soft tissues a 1.5mm

discrepancy can be considered clinically acceptable, as differences of less than 1.5mm

are not detected by the naked eye, even for a skilled observer (Lübbers et al., 2012). Table

1 presents a summary of the accuracy and acquisition speed of different imaging methods.

Table 1. Summary of 3D imaging systems

Accuracy Acquisition speed
CT 0.2mm 10-30 minutes
MRI 0.2mm 15-30 minutes
3D Laser scanning 1mm 30 seconds
Structured light scanning 0.01mm - 0.5mm 10 seconds
Digital stereophotogrammetry 0.25mm-0.6mm < 2ms

2.3 Clinical applications of 3D printing

There is a plethora of applications for 3D technologies in the head and neck surgery and

rehabilitation areas (Day et al., 2018; Huang et al., 2016; Jardini et al., 2014; Msallem et

al., 2017; Pang et al., 2018; Schaaf et al., 2009; Sutradhar et al., 2016; Watson and

Hatamleh, 2014; Witjes et al., 2018). Some examples are:

1. Simulation of implant placement: through different software applications and

patient’s imaging (typically CT or CBCT), the practitioner can simulate the

implant’s position, allowing the 3D printing of models and surgical guides.

21
2. Fabrication of cranial implants: the clinician uploads CT images of the cranial

defect, virtually designs the reconstruction and gets a custom printed model

(usually in ABS or PLA), which is later used to shape the actual cranial implant.

3. Prostheses: a custom-made maxillofacial prosthesis can be fabricated by 3D

printing, either by printing a wax pattern or a copy of the defect’s mould. There is

currently research being carried out to directly 3D print a silicone prosthesis, in

the correct skin colour.

4. Burn masks: a resin burns mask can be directly 3D printed (it has been reported

the use of bio-compatible Polyjet photopolymer MED610), or an anatomical

model of the face can be fabricated by 3D printing in order to adapt a

thermoplastic burns mask.

5. Tumour margin assessment: visualization of a facial tumour in a 3D model has

been shown to be a precise method to distinguish the correct tumour margin,

reducing the risk of recurrence and the necessity of high postoperative radiation

therapy.

6. Radiation shields: studies have shown that intra-oral radiation shields can be 3D

printed directly through FDM using PLA.

7. Research: many research applications have been reported on the use of 3D

printing in clinical and laboratory studies. For example, polycaprolactone

scaffolds have been printed with SLS to be applied in bone tissue engineering.

8. Record keeping: a pre-operatively 3D printed model can be considered as a

permanent patient record, with high value for medico-legal reasons. It can also be

used for assessing post-operative outcomes and as a reference for future treatment.

9. Patient education: having an anatomical model for demonstration purposes can be

an important tool to gain informed consent.

22
 10. Training: 3D models of normal or pathological anatomy can be very useful for

educating healthcare trainees.

11. Treatment planning: anatomical models can be 3D printed in their actual size by

utilizing data from images acquired pre-operatively, allowing patient-specific

simulation and planning for surgical and prosthetic treatment.

The utilisation of 3D printing in surgical and prosthetic treatment planning has several

advantages (Dhima et al., 2013; Fernandes et al., 2016; Heller et al., 2016; Hoarau et al.,

2014; Huang et al., 2016; Mazzoni et al., 2013; Seok et al., 2017; Wang et al., 2018),

helping the clinicians to:

- Corroborate the feasibility of a treatment, facilitating the assessment of different

treatment options by studying a 3D model thoroughly, increasing the accuracy

and efficacy of the treatment.

- Evaluate post-operative outcomes.

- Practice the procedure, lowering the risk of complications by identifying possible

difficulties before the actual intervention, like proximity to vital structures.

- Reduce the surgical operating times, duration of anaesthesia, amount of blood loss

and rate of infection.

Depending on the complexity of the model and the selected printer and material, the

processing and fabrication of 3D models can take several hours; therefore, they would be

ineffective for emergency medical situations.

Nowadays, due to the availability of printing materials with different colours and

flexibilities, 3D printed objects can realistically imitate the physical properties of the soft

and hard tissues of the body. For example, in a study by Werz et al., upper and lower jaw

23
models were 3D printed in ABS and PLA filaments, coated with silicone rubber for soft

tissue simulation and used as training models for two basic oral and maxillofacial surgical

procedures in order to evaluate if these materials can genuinely simulate human bone. No

significant differences were found between ABS and PLA in relation to their performance

as surgery models, as both present a realistic bone simulation. The authors also concluded

that both materials can be used to create training models for the practice of surgical

procedures, however, ABS would be a better choice for procedures that require extensive

drilling with a non-cooled rotary instrument due to its higher resistance to melting (Werz

et al., 2018).

An important step in the integration of 3D printing into the medical field is the regulation

of the produced models. If the replicas are going to be used for diagnostic purposes, to

shape medical devices like catheters and fixation plates before a surgical intervention or

for the direct fabrication of a medical device like surgical guides, then the manufacturing

process, printer and materials must be reviewed by a regulatory entity (George et al.,

2017).

Although, the cost of 3D printing technology has been reduced over time, is essential to

always consider the cost of acquiring, employing and maintaining both the software and

hardware elements of this technology. Beumer (Beumer III et al., 2011) describes some

limitations to completely embrace this processes in the healthcare field:

- Many clinicians, especially those in the surgical disciplines, work either

individually or in small groups. Therefore, the cost to acquire this technology is

very high for these users.

24
 - In health care applications, usually only one or two anatomical structures will need

to be designed and printed for an individual patient, thus making the process more

expensive.

- Few of the 3D printing materials meet standards for medical applications.

- The use of technology takes time and is not funded as a medical procedure.

- Desktop 3D printing technologies are limited by their output size, the printers for

medical models are large and require considerable floor space.

The advances on medical applications for 3D printing have become relevant in the

maxillofacial prosthetics field; especially with the development of inexpensive printers,

multi-material printing and new modalities of 3D medical imaging.

Treatment planning in maxillofacial prosthetics is a mix of different digital virtual

processes, where the selection of the 3D imaging data set to be used varies if the treatment

involves hard tissues or soft tissues. For example, it is possible to get information of the

osseous anatomical structures and available bone volume for osseointegrated implants

through 3D images like CT scans, these can help to determine a suitable site for implant

placement and its orientation, which can be transferred to the patient via 3D printed

surgical guides. This is just one example of the many uses 3D printing technologies can

have in the maxillofacial prosthetics field.

In terms of soft tissue anatomical structures, it is possible to get information through 3D

images like 3D laser scanning or Digital stereophotogrammetry. In the case of ears and

orbits, for unilateral defects, the acquired imaging data of the unaffected side can be

mirrored and placed over the treatment site to corroborate the correct position of the

implants (Liacouras et al., 2011). In the case of a nasal prosthesis, if the nose was imaged

25
pre-operatively, the 3D data can be used as a prosthetic pattern (Ciocca et al., 2011), if

no records exist, a new nose can be designed digitally or a digital image of someone else’s

nose can be taken; the same principle applies for bilateral auricular prostheses. After these

images are obtained, it is possible to evaluate shape, size, angulation, aesthetics and

symmetry of the designed replica and this design can also be used to 3D print the wax

patter or a mould to be used in the fabrication of the future prosthesis (Davis, 2010).

In a study by Ciocca et al. (Ciocca et al., 2010), the authors designed bilateral prosthetic

ears for a patient affected by Treacher Collins syndrome, a craniofacial development

disorder, where the most common features include microtia, conductive hearing loss,

midface hypoplasia, micrognathia and cleft palate (Marszałek et al., 2002); in this case,

reconstructive surgery had failed and auricular prostheses were requested. The authors

used an Ear and Nose Digital Library, which contained digital models of different noses

and ears. The most suitable ear was selected according to its size and shape, in relation to

the patient’s anatomy, and later superimposed onto the corresponding defects and adapted

to the position of the patient’s implants. Finally, a hollow mould was 3D printed and

packed in the conventional silicone method.

The use of 3D technologies can considerably reduce the clinical time to fabricate a facial

prosthesis. In a study presented by Yadav et al.(Yadav et al., 2017), a process to

manufacture an auricular prosthesis in only two appointments using 3D technology was

described. The process starts with the acquisition of a CT scan of the patient’s head. Using

a 3D modelling software the authors extracted a 3D image of the normal ear, which was

later mirrored and located on the affected side. From that image, a hollow mould was

designed and 3D printed by selective laser sintering (SLS). Intrinsic colouring swatches

were prepared with two shades to colour the Room-temperature vulcanizing (RTV)

silicone that was used to pack the mould. After 48 hours, the prosthesis was trimmed and

26
finished. On the second appointment, the final replica was oriented on the patient and the

minor discrepancies in shade were corrected with extrinsic colouring; the patient was

satisfied with the result.

This imaging method eliminates the distortion due to compression of the soft tissues by

the impression material and also bypasses the time-consuming process of wax sculpting.

Moreover, it allows the refabrication of the prosthesis after deterioration or discoloration

of the silicone elastomer over time, as the printed mould can be used multiple times to

pack the silicone and the design can be digitally saved and reproduced. This is a

significant improvement, considering that is necessary to replace the silicone prosthesis

approximately every 18 months (Qiu et al., 2011).

Manufacturing facial prostheses with digital technologies can contribute to higher patient

comfort and production efficiency, but, it is important to consider the high initial

investment in equipment and the necessary training to learn how to work with particular

software (Palousek et al., 2014).

2.4 Selection of the appropriate 3D printing system

The precision and accuracy of the printed object would result from the combination of

the precision and accuracy of the acquired image, the suitable image processing for 3D

modelling and the accuracy of the 3D printing system (Kim et al., 2016).

Depending on the clinical purpose of the element to be 3D printed, is necessary to select

the most appropriate technology according to accuracy, reproducibility and available

materials. In order to do this, it is essential to compare the different products available in

the market and the characteristics of the printed element.

27
Another important factor to consider when selecting a 3D printer, is the cost of the

equipment. The price of the printers varies according to the manufacturing technology;

FDM has the lowest cost and it is possible to find professional performance FDM printers

starting from approximately £1,500. LOM, 3D MJP and SLS are in a higher price range

of over £5,000 and the most expensive technologies are SLA and Polyjet, with prices

higher than £10,000 (Imakr, 2016).

Table 2 shows a summary of the main 3D printing systems according to used materials,

main medical application, cost and layer thickness.

Table 2. Summary of 3D printing systems

Material Main Cost Layer
application thickness (mm)
SLA Photopolymer Surgical tools $$$ 0.05-0.07
Polyjet Photopolymer Models $$$ 0.015-0.030
FDM Thermoplastics Models $ 0.13-0.33
Elastomers
LOM Thermoplastics Models $$ 0.05-0.5
Paper
Composites
Metals
Ceramics
3D MJP Elastomers Surgical tools $$ 0.016-0.050
Composites
Ceramics
SLS Thermoplastics Medical $$ Minimum 0.1
Elastomers implants
Composites
$: refers to the price range of the 3D printer

Two of the most commonly used printing technologies are liquid-based Polyjet and solid-

based Fused deposition modelling (FDM); as previously described, Polyjet uses

successive layers of photocurable resin and support material to create a 3D model, each

28
layer is polymerised by a UV light immediately after being deposited. On the other hand

FDM utilizes a thermoplastic material that is extruded from a hot nozzle system and

deposited layer by layer to build a model (Kim et al., 2016).

So far, these methods have been compared only in a few studies, where their accuracy

has been evaluated through volumetric or linear changes in the fabrication of dental or

anatomical models; no study has evaluated their accuracy in relation to flat surfaces. Jin

et al. compared complete dental arch stone models with their replicas created with Polyjet

and FDM by evaluating their trueness and precision using 3D analysis software to identify

regional differences. They concluded that there were no significant differences in

accuracy between Polyjet and the stone models, FDM had a lower accuracy (probably

because of a higher layer thickness) but still acceptable clinically, and both Polyjet and

FDM had higher precision when compared to the stone models (Jin et al., 2018).

In a study by Rebong at al. the dimensional accuracy of the FDM, Polyjet and SLA

fabricated replicas was assessed by comparing them to traditional maxillary and

mandibular plaster models. All the models were measured with a digital calliper in several

parameters obtained in vertical, transverse, anteroposterior and mixed planes. The authors

concluded that there were no statistically significant differences between the 3D printed

models and the plaster casts in the majority of the parameters, however, FDM had the

least amount of variation when compared to Polyjet and SLA (Rebong et al., 2018).

Lee et al. scanned extracted molar teeth to obtain STL files and printed models using

FDM (material: ABS) and Polyjet (material: Vero White Plus). All the replicas were

scanned to obtain 3D images and computer software compared the replicas with the

original teeth, through linear and volumetric measurements to evaluate the accuracy of

the printing methods. The authors found that the differences between the original teeth

29
and the FDM and Polyjet replicas were statistically significant, however, they were

accepted clinically (Lee et al., 2015).

Murugesan et al. evaluated the dimensional accuracy and surface reproducibility of FDM

(material: ABS), 3D MJP (material: composite) and Polyjet (material: clear resin) by

comparing printed mandibular models with the original STL file. Measurements were

taken with a digital calliper in 5 anatomical points and the surfaces were visualised in a

scanning electron microscope. The results showed that Polyjet was more accurate,

followed by 3D MJP and FDM, and also presented a smoother surface with adequate

surface details (Murugesan et al., 2012).

Hong et al. designed a model based on a CT scan of a patient with thyroid cancer and

printed the structure using FDM (material: thermoplastic elastomer), colour-jet printing

(material: Visijet PXL) and Polyjet (material: Vero White Plus). The accuracy of these

methods was evaluated by measuring five landmarks with a digital calliper and comparing

these dimensions with the original STL file. The authors concluded that colour-jet

printing was the most accurate method, followed by Polyjet and FDM. (Hong et al., 2019).

It can be observed that the results of these investigations have some variations and they

do not reach the same conclusion. Furthermore, an important limitation to compare the

results of these studies, is that some of them excluded the type of material used in each

printing method, making it impossible to evaluate if the selection of the material has a

role in the accuracy of the 3D printing system.

FDM and Polyjet are two of the most advanced 3D printing technologies available, they

have the capacity to form complex shapes, in both positive (parts) and negative (moulds)

forms, and have a short and simple post-processing stage (Stanek et al., 2012).

30
Comparing these two types of 3D printing technologies is crucial to elucidate which

printing system is more efficient in the fabrication of computer designed medical items

according to their clinical function. The purpose of this study was to evaluate the

dimensional accuracy and reproducibility of three different methods of 3D printing using

a known size object, by comparing linear dimensions between the printed elements and

the original 3D image.

31
2.5 Conclusions from the literature

- The use of three-dimensional (3D) printing technologies has increased

considerably in the last decades.

- One of the areas that has seen the greatest growth of these technologies is the field

of medicine, where their main applications include research, education and

training, treatment planning and fabrication of custom made medical devices.

- A growing diversity of 3D printing methods are available, varying in price,

accuracy and materials.

- All these printing systems are driven by computer data describing the 3D object

to be built. The utilization of 3D digital imaging has numerous assets, such as

reduced storage, high acquisition speed, no degradation and fast transfer of data.

- The precision and accuracy of the printed object would result from the

combination of the precision of the imaging technique, the suitable image

processing and the accuracy of the 3D printing system.

- Depending on the clinical purpose of the element to be 3D printed, is necessary to

select the most appropriate technology according to cost, accuracy,

reproducibility and available materials.

- Two of the most commonly used printing technologies are liquid-based Polyjet

and solid-based Fused deposition modelling (FDM). These methods have been

compared only in a few studies, where their accuracy has been evaluated through

volumetric or linear changes in the fabrication of dental or anatomical models. No

study has evaluated their accuracy in relation to flat surfaces.

32
3. Aims and objectives

The aim of this study was to assess the accuracy and reproducibility of FDM (Cubicon

Single Plus, Cubicon Inc., Korea) and Polyjet (Objet Eden 250, Stratasys Ltd., USA) 3D

printers, by comparing linear dimensions between the 3D models, printed using

photocurable resin (RGD720 & SUP705, Stratasys Ltd., USA), ABS (Cubicon Inc.,

Korea) and PLA (Cubicon Inc., Korea), and the original 3D image, to establish which one

is more accurate and efficient to use clinically.

Objectives:

- To assess dimensional differences between 3D printed models fabricated with

Polyjet and FDM technologies and the original 3D image.

- To determine the accuracy of these printing technologies by comparing linear

measurements between the models and the original 3D image.

- To establish which of these printing methods is more cost-efficient to use

clinically according to the model’s purpose.

4. Null hypothesis

There are no differences in accuracy and reproducibility between the three different 3D

printing methods (Objet Eden 250-resin, Cubicon Single Plus-ABS, Cubicon Single Plus-

PLA) when comparing linear dimensions between the original 3D image and direct

measurements carried out using a digital calliper on the 3D printed models obtained with

each technique.

33
5. Methodology

The assessment of the 3D printing techniques was accomplished by comparing direct

measurements of allocated dimensions on the 3D printed models, obtained from three

manufacturing methods, to the correspondent linear measurements on the original 3D

image of a designed irregular cube.

5.1 Design of cube

An irregular cube was designed using Autodesk Fusion 360 (version 2.0.8749, Autodesk

Inc., USA) software (Figure 8) and exported as an STL file. The faces of the cube have

different dimensions, varying from 1 to 8 cm (See appendix 1).

Figure 8. Designed irregular cube

5.2 Selection of 3D printing techniques

The cube was 3D printed using three different materials and two additive manufacturing

systems. The group photocurable resin was printed using a Polyjet 3D printer (Objet Eden

34
250, Stratasys Ltd., USA) and compatible photopolymerizing resin (RGD720 & SUP705,

Stratasys Ltd., USA) in a layer thickness of 0.016mm. The group ABS included an FDM

3D printer (Cubicon Single Plus, Cubicon Inc., Korea) and compatible ABS filament

(Cubicon Inc., Korea). The group PLA involved the same FDM 3D printer and

compatible PLA filament (Cubicon Inc., Korea). Details of the Polyjet and FDM printers

used in this study are presented in Table 3.

Table 3. Details of Polyjet and FDM printers

Polyjet FDM
Machine Objet Eden 250 Cubicon Single Plus
Printer size 870 x 735 x 1200 554 x 579 x 524
(WxDxH)(mm)
Printer weight (kg) 280 24
Build volume 250 x 250 x 200 240 x 190 x 200
(WxDxH)(mm)
Resolution X-axis: 600 dpi XY-axis: 6.25 μm
Y-axis: 300 dpi Z-axis: 1.25 μm
Z-axis: 1600 dpi
Layer thickness (mm) High Quality: 0.016 0.1- 0.3
High Speed: 0.030
File format STL and SLC STL and OBJ
(Imakr, 2016; SMG 3D, 2011)

5.3 Pilot study

For the pilot study, 5 cubes of resin, 5 cubes of ABS and 5 cubes of PLA were printed

according to the additive manufacturer’s recommendations. Both ABS and PLA cubes

were printed in a layer thickness of 0.2mm with a 15% support interface and the resin

cubes were printed in a layer thickness of 0.016mm.

35
Each cube was measured in 8 allocated dimensions (refer to section 5.5) with a digital

calliper (Digitronic Caliper 110-DBL Series, Moore & Wright Europe, England). To

establish intrarater consistency, each model was measured 3 times on two separate

occasions, one week apart. These measurements were later compared with the original

3D image and their differences were recorded to be used in the necessary calculations.

The mean and standard deviation (SD) of these differences were calculated at every

recorded dimension for each material.

Consequently, sample size calculations were carried out using G*Power [Link] software

(Faul et al., 2007). For a power of 0.8 and pre-determined statistical significance of α =

0.05, a total of 10 specimens per group was calculated to be sufficient for the main

experiment.

5.4 Main experiment: 3D printing of Cube

10 cubes of resin, 10 cubes of ABS and 10 cubes of PLA were printed according to the

additive manufacturer’s recommendations (Figure 9). The resin cubes were printed in a

layer thickness of 0.016mm, and both ABS and PLA cubes were printed in a layer

thickness of 0.2mm with a 15% support interface.

36
Figure 9. Sample of 3D printed cubes in photopolymerizing resin (left), ABS (middle)
and PLA (right)

The main properties of the materials used in this study can be observed in Table 4.

Table 4. Main properties of RGD720, ABS and PLA materials

RGD720 ABS PLA

Tensile strength (MPa) 50-65 43 65

Elongation at break (%) 15-25% 22 8

Modulus of elasticity 2000-3000 MPa 1100-2900 3500

(MPa)

Flexural strength (MPa) 80-110 66 97

Flexural modulus 2700-3300 MPa 2348 3600

(MPa)

Heat deflection (°C) 45-50 78 56

Support material SUP705

(Imakr, 2018; Stratasys, 2015)

37
5.5 Dimensional measurements

In Phase 1 of the study, each cube was measured according to their length, width and

depth in 8 allocated dimensions. Dimension 1 (D1) of 8cm was recorded on surface A

(Figure 10A), dimension 2 (D2) of 7cm was recorded in surface B (Figure 10B),

dimension 3 (D3) of 5cm was recorded on surface C (Figure 11C), dimension 4 (D4) of

6cm and dimension 5 (D5) of 3cm were recorded on surface D (Figure 11D), dimension

6 (D6) of 1cm and dimension 7 (D7) of 2cm were recorded on the top surface (Figure

12). Dimension Central Cube of 5cm was recorded on the projection of the central points

of the top and bottom surfaces measured together (Figure 13).

Figure 10. Surface A showing D1 and surface B showing D2

38
Figure 11. Surface C showing D3 and surface D showing D4 and D5

D7 – 2cm
D6 – 1cm

Figure 12. Top surface showing D6 and D7

39
 Figure 13. Top and bottom surfaces showing the selected points for Central Cube

In Phase 2 of the study, the 4 corners of each cube (corners A,B,C,D of 8 cm) were

measured to observe any distortion in that area, as it was observed that the ABS cubes

had a small curvature in the bottom corners, while resin and PLA presented flat surfaces.

The direction and surface of each evaluated corner can be observed in Figure 14.

Figure 14. Measured corners. Corner D is opposite to Corner B

40
The measurements were recorded with a digital calliper (Digitronic Caliper 110-DBL

Series, Moore & Wright Europe, England). To establish intrarater reliability, each

measurement was performed 3 times on two separate occasions, one week apart. The

mean of these three readings was obtained for both week 1 and week 2, and later, the

average between them was calculated to establish a final mean for every measurement.

These final means were later compared with the 3D image (gold standard) and their

differences were recorded to be used in the subsequent calculations; a negative difference

(-) describes a dimension with increasing tendencies, that is, when the printed cube is

bigger than the original image.

5.6 Statistical Analysis

The mean and standard deviation (SD) of these differences for each material were

calculated at every recorded dimension. Repeatability of the measurements was assessed

using the coefficient of repeatability. The data was analysed using a two way ANOVA to

establish whether there were significant dimensional differences between the cube’s

digital image and the printed items, in the three different materials. Post-hoc multiple

comparison tests (Bonferroni) were executed to establish where the differences existed

between materials. A statistical software (IBM SPSS® Statistics version 25, IBM Corp.,

USA) was employed for statistical analyses.

41
6. Results

The results of this study are presented in two different phases, in Phase 1 each cube was

measured in 8 allocated dimensions and in Phase 2, the 4 corners of each cube were

measured to observe any distortion in that area. These measurements were later compared

with the original 3D image and their differences were recorded to be used in the

subsequent calculations. The results of both phases include the coefficient of

repeatability, the dimensional differences and the statistical analysis of the data.

6.1 Results of Phase 1

The coefficient of repeatability (Table 5) showed that the measuring technique was highly

reproducible. A value of zero in the coefficient of repeatability means that the difference

in measurements between the original image and the particular material for weeks 1 and

2 for all ten samples were identical. The coefficients are very small values in comparison

to the image size, suggesting the technique is highly reproducible.

Table 5. Coefficient of repeatability for the 8 allocated dimensions

Coefficient of Coefficient of Coefficient of

repeatability repeatability ABS repeatability PLA
Dimension Size (cm) Resin (rounded) (rounded) (rounded)
1 8 0.0063 0.0000 0.0000
2 7 0.0089 0.0063 0.0089
3 5 0.0000 0.0089 0.0000
4 6 0.0063 0.0063 0.0000
5 3 0.0000 0.0000 0.0000
6 1 0.0000 0.0000 0.0063
7 2 0.0089 0.0000 0.0000
Central Cube 5 0.0000 0.0063 0.0000

42
The differences between the linear measurements obtained on the 8 allocated dimensions

of the printed cubes in the three different materials and the original 3D digital image were

recorded. The mean and standard deviation (SD) of these differences were calculated for

each material at every recorded dimension and represented in bar charts.

For Dimension 1 (Figure 10A), the mean difference was 0.011 cm for resin (SD 0.003

cm), 0.061 cm for ABS (SD 0.009 cm) and 0.025 cm for PLA (SD 0.010 cm), as seen in

Figure 15.

Figure 15. Mean difference values between the 8cm Dimension 1 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image

For Dimension 2 (Figure 10B), the mean difference was 0.010 cm for resin (SD 0.002

cm), 0.042 cm for ABS (SD 0.011 cm) and 0.032 cm for PLA (SD 0.011 cm), this can be

observed in Figure 16.

43
 Figure 16. Mean difference values between the 7cm Dimension 2 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image

For Dimension 3 (Figure 11C), the mean difference was 0.010 cm for resin (SD 0.000

cm), 0.055 cm for ABS (SD 0.020 cm) and 0.022 cm for PLA (SD 0.009 cm), this can be

observed in Figure 17.

Figure 17. Mean difference values between the 5cm Dimension 3 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image

44
For Dimension 4 (Figure 11D), the mean difference was 0.009 cm for resin (SD 0.001

cm), 0.071 cm for ABS (SD 0.014 cm) and 0.031 cm for PLA (SD 0.009 cm), as seen in

Figure 18.

Figure 18. Mean difference values between the 6cm Dimension 4 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image

For Dimension 5 (Figure 11D), the mean difference was -0.006 cm for resin (SD 0.005

cm), 0.020 cm for ABS (SD 0.000 cm) and 0.030 cm for PLA (SD 0.000 cm), as seen in

Figure 19.

45
 Figure 19. Mean difference values between the 3cm Dimension 5 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image, conversely, a negative difference portrays a dimension with increasing
tendencies

For Dimension 6 (Figure 12), the mean difference was 0.011 cm for resin (SD 0.003 cm),

-0.002 cm for ABS (SD 0.004 cm) and -0.024 cm for PLA (SD 0.004 cm), as seen in

Figure 20.

Figure 20. Mean difference values between the 1cm Dimension 6 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image, conversely, a negative difference portrays a dimension with increasing
tendencies

46
For Dimension 7 (Figure 12), the mean difference was 0.016 cm for resin (SD 0.004 cm),

0.003 cm for ABS (SD 0.004 cm) and -0.020 cm for PLA (SD 0.000 cm), this can be

observed in Figure 21.

Figure 21. Mean difference values between the 2cm Dimension 7 image and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image, conversely, a negative difference portrays a dimension with increasing
tendencies

For Dimension Central Cube (Figure 13), the mean difference was 0.001 cm for resin (SD

0.005 cm), -0.040 cm for ABS (SD 0.007 cm) and -0.023 cm for PLA (SD 0.004 cm),

this can be observed in Figure 22.

47
 Figure 22. Mean difference values between the 5cm Dimension Central Cube image
and the three materials. A positive difference describes a cube’s dimension smaller than
the original image, conversely, a negative difference portrays a dimension with
increasing tendencies

The two way ANOVA showed that there was a statistically significant difference

(P<0.0005) between the dimensions of the cube’s digital image and the printed items, in

the three different materials (Table 6). Therefore, it was possible to proceed to the Post-

hoc multiple comparison tests (Bonferroni) to establish where the differences existed

between materials.

Table 6. Two way ANOVA results for Phase 1

Source F Sig.
Corrected Model 114.842 0.000
Intercept 770.980 0.000
DIMENSION 205.188 0.000
MATERIAL 131.669 0.000
DIMENSION * MATERIAL 67.264 0.000

48
 For the Bonferroni multiple comparisons test between materials, code 62 represents resin,

63 is ABS and 64 is PLA (Table 7). When comparing the differences between materials,

a highly significant difference (P< 0.0005) was observed between resin and ABS, and

between ABS and PLA. Conversely, there was no significant difference between resin

and PLA (P>0.05).

Table 7. Bonferroni results for comparison between materials for Phase 1

Multiple Comparisons
Dependent Variable: Measurement
Bonferroni
Mean 95% Confidence Interval
Difference Std. Lower Upper
(I) Material (J)Material (I-J) Error Sig. Bound Bound
*
62.000 63.000 -0.01844 0.001272 0.000 -0.02151 -0.01537
64.000 -0.00119 0.001272 1.000 -0.00426 0.00188
*
63.000 62.000 0.01844 0.001272 0.000 0.01537 0.02151
*
64.000 0.01725 0.001272 0.000 0.01418 0.02032
64.000 62.000 0.00119 0.001272 1.000 -0.00188 0.00426
*
63.000 -0.01725 0.001272 0.000 -0.02032 -0.01418
Based on observed means.
The error term is Mean Square(Error) = 6.47E-005.
*. The mean difference is significant at the .05 level.

6.2 Results of Phase 2

In Phase 2 of the study, corners A, B, C and D (Figure 14) of each cube were measured

to observe any deformation in that area as it was observed that the ABS cubes had a small

curvature in the bottom corners (Figure 23), while resin and PLA presented flat surfaces.

In the same way as in Phase 1, the differences between these linear measurements and the

3D digital image were recorded.

49
 Figure 23. ABS cube showing a curvature in the bottom corners

All coefficients of repeatability were zero for each reading between weeks 1 and 2. The

differences were absolutely identical for every sample and every material, therefore, the

technique is highly reproducible. The results can be observed in Table 8.

Table 8. Coefficient of repeatability for corners A, B, C and D

Coefficient of Coefficient of Coefficient of

repeatability Resin repeatability ABS repeatability PLA
Dimension Size (cm) (rounded) (rounded) (rounded)
Corner A 8 0.0000 0.0000 0.0000
Corner B 8 0.0000 0.0000 0.0000
Corner C 8 0.0000 0.0000 0.0000
Corner D 8 0.0000 0.0000 0.0000

The mean and standard deviation (SD) of the differences between the original image and

the printed models were calculated for each material at the four corners of the cube.

50
For Corner A, the mean difference was 0.013 cm for resin (SD 0.004 cm), 0.082 cm for

ABS (SD 0.018 cm) and 0.010 cm for PLA (SD 0.004 cm). For Corner B, the mean

difference was 0.010 cm for resin (SD 0.004 cm), 0.078 cm for ABS (SD 0.012 cm) and

0.007 cm for PLA (SD 0.004 cm). For Corner C, the mean difference was 0.010 cm for

resin (SD 0.000 cm), 0.094 cm for ABS (SD 0.016 cm) and 0.010 cm for PLA (SD 0.008

cm). For Corner D, the mean difference was 0.012 cm for resin (SD 0.004 cm), 0.072 cm

for ABS (SD 0.016 cm) and 0.007 cm for PLA (SD 0.006 cm). These results are

represented in Figure 24.

Figure 24. Mean difference values between the 8cm Corner images and the three
materials. A positive difference describes a cube’s dimension smaller than the original
image.

The two way ANOVA for this part of the study also showed a statistically significant

difference (P<0.0005) between the corners of the cube’s digital image and the printed

items, in the three different materials (Table 9). Consequently, Post-hoc multiple

51
comparison tests (Bonferroni) were performed to establish where the differences existed

between materials.

Table 9. Two way ANOVA results for Phase 2

Source F Sig.
Corrected Model 121.275 0.000
Intercept 1305.239 0.000
DIMENSION 3.403 0.020
MATERIAL 653.897 0.000
DIMENSION * MATERIAL 2.671 0.019

Similarly to the results in Phase 1, the Bonferroni multiple comparisons test between

materials showed a highly significant difference (P< 0.0005) between resin and ABS, and

between ABS and PLA. There was no significant difference between resin and PLA

(P>0.05). The results can be observed in Table 10, where once again code 62 represents

resin, 63 is ABS and 64 is PLA.

52
Table 10. Bonferroni results for comparison between materials for Phase 2

Multiple Comparisons
Dependent Variable: Measurement
Bonferroni
Mean 95% Confidence Interval
Difference Lower Upper
(I)Material (J)Material (I-J) Std. Error Sig. Bound Bound
62.000 63.000 -0.07025* 0.002288 0.000 -0.07581 -0.06469
64.000 0.00275 0.002288 0.696 -0.00281 0.00831
*
63.000 62.000 0.07025 0.002288 0.000 0.06469 0.07581
64.000 0.07300* 0.002288 0.000 0.06744 0.07856
64.000 62.000 -0.00275 0.002288 0.696 -0.00831 0.00281
*
63.000 -0.07300 0.002288 0.000 -0.07856 -0.06744
Based on observed means.
The error term is Mean Square(Error) = .000.
*. The mean difference is significant at the .05 level.

53
7. Discussion

The aim of this study was to assess the accuracy and reproducibility of three different

additive manufacturing techniques using FDM and Polyjet 3D printers, by comparing

linear dimensions between the 3D printed models and the original 3D image of a designed

irregular cube.

This was achieved by printing 10 cubes in photocurable resin, 10 cubes in ABS and 10

cubes in PLA. Subsequently, the models were measured in 8 allocated dimensions and

their 4 corners. These measurements were compared to the linear dimensions of the

original 3D image and accuracy was evaluated according to the dimensional differences

between the models and the original 3D image.

The cube was designed with this irregular shape to observe if considerable differences in

accuracy could be detected between different dimensions. Nevertheless, as it can be seen

in the results, all the dimensional differences were within micrometres.

7.1 Reproducibility of measurements

The measurements were taken with a digital calliper by one operator. Each model was

measured 3 times on two separate occasions, one week apart. The results from the

coefficient of repeatability showed that this measuring technique was highly

reproducible. All the values, for both the 8 allocated dimensions and the 4 corners were

close to zero, which means that the difference in measurements between the image and

the cubes in weeks 1 and 2 for all the samples were practically identical, suggesting the

technique is highly reproducible.

54
7.2 Accuracy of 3D printing

When comparing the accuracy of the different materials in the 8 allocated dimensions, it

was possible to observe that ABS presented the biggest dimensional differences in the

majority of the measurements. In dimensions 1, 2, 3, 4 and Central Cube, resin was the

most accurate material, followed by PLA and then ABS. In dimension 5, resin was still

the most accurate, but this time followed by ABS and then PLA. On the other hand, for

dimensions 6 and 7, ABS was the most accurate material, followed by resin and PLA. In

relation to the corners of the cube, PLA was the most accurate material, followed by resin

and ABS. The latter presented discrepancies higher than 0.007 cm on these dimensions.

7.3 Statistical analysis

In both phases of the study, the two way ANOVA analysis showed a statistically

significant difference (P<0.0005) between the cube’s digital image and the printed items,

in the three different materials. It was later observed in the Bonferroni multiple

comparisons test that a highly significant difference (P< 0.0005) existed between resin

and ABS, and between ABS and PLA, however, there was no significant difference

between resin and PLA (P>0.05).

Therefore, the results of this study showed that when compared with the original 3D

image, models fabricated with photopolymerizing resin in a Polyjet printer were the most

accurate replicas, followed by PLA, printed with FDM technology, and no statistically

significant difference was found between these two materials. Conversely, ABS was the

less accurate material and its difference with both Resin and PLA was statistically

significant.

55
7.4 Agreement of findings with available literature

These results are similar to the ones presented by Murugesan et al., where the models

fabricated with Polyjet were more dimensionally similar to the STL image when

compared to the ones printed in ABS using an FDM printer (Murugesan et al., 2012). The

same result was also obtained by Lee et al., where replicas of extracted teeth were printed

using FDM and Polyjet, the latter being the most accurate when compared to the original

teeth (Lee et al., 2015). Conversely, in a study by Rebong et al., the authors assessed the

accuracy of FDM, Polyjet and SLA produced models by comparing them to conventional

plaster casts, and they found that FDM models had the least amount of dimensional

differences and were the best option to replicate plaster models when compared to Polyjet

and SLA. The authors suggested that FDM models can be clinically useful and are not

inferior when it comes to dimensional accuracy in comparison to newer and more

expensive printing techniques. However, they failed to inform which material was

selected to print in each technique, therefore, it is not possible to make an thorough

comparison with the present study (Rebong et al., 2018).

7.4.1 Reasons for accuracy difference between techniques

[Link] Printing technology and materials

Some authors have explained that the most likely reasons for this difference between

printing technologies are the materials and manufacturing process involved in both

Polyjet and FDM. The polymers used in Polyjet have smaller molecules when compared

to the FDM filaments, being able to produce complex shapes and obtaining better

repeatability. On the other hand, FDM uses a semi-solid material that results in lower

accuracy, especially if ABS is used, since this material tends to warp and shrink when

56
cooling. In addition, during the hardening process of this material, is difficult to maintain

the necessary heat and humidity to avoid this contraction, which can result in inaccurate

models when compared with the reference image.

[Link].1 Polyjet

The higher accuracy of Polyjet models has been reported in previous studies. The

accuracy of surgical templates printed in Polyjet was evaluated in a study by Chen et al.,

here the authors concluded that the higher accuracy and reproducibility of the Polyjet

printed models can be explained by the higher printing resolution and smaller layer

thickness that this technology presents in comparison with simpler desktop printers (Chen

et al., 2019).

Due to its high dimensional accuracy, high resolution and the capacity to print in a variety

of modular strengths, Polyjet technology has been used widely in the medical field to

manufacture anatomical models for treatment planning and surgical simulation. One of

the main advantages of this technique is the lack of post-processing of the printed model,

this is due to the fact that the UV source is extremely close to the jetting nozzle,

instantaneously curing the resin after being deposited in the platform (Tappa and

Jammalamadaka, 2018). Additionally, Polyjet presents several advantages such as shorter

building time, higher precision and better surface quality, however, the cost of the

machine is higher when compared to FDM (Stanek et al., 2012).

In relation to the materials used in this study, RGD720 & SUP705 were utilised in the

Polyjet printer; RGD720 resin is a translucent photopolymer, with high dimensional

stability and surface smoothness, it also has the capacity to blend with other materials to

modify its opacity and hardness. SUP705 is used as a support material, is a gel-like

57
photopolymer that can be easily removed by hand, brushing or water pressure. Support

materials are designed to allow the fabrication of complex shapes, sustaining undercuts

and overhangs.

[Link].2 FDM

It has been reported in the literature that models fabricated through FDM technology are

prone to shrinkage during the thermoplastic cooling process, which can lead to

dimensional inaccuracies in the areas of the replica that are not correctly supported during

the building process (George et al., 2017).

FDM has been used in the medical field to fabricate custom-made medical devices, such

as anatomical models, prostheses, surgical guides and implants, due to the mechanical

properties of the materials and short fabrication time. In the case of ABS, as it is a non-

biocompatible material, it has mainly been used for the fabrication of medical models for

treatment planning and surgical simulation using FDM printers. These models have also

been proved a useful tool to explain surgical procedures to patients before the actual

intervention (Tappa and Jammalamadaka, 2018).

In relation to the materials used in this study, Acrylonitrile butadiene styrene (ABS)

alongside Polylactic acid (PLA) are the two most common materials for FDM printing;

both are thermoplastic polymers presented as a filament of 1.75mm diameter. ABS is a

very strong, flexible and lightweight material, safe for human use. ABS is not bio-

degradable, however it is recyclable. ABS is less brittle, more ductile, has better

mechanical properties and higher temperature resistance than PLA but is prone to warping

and to contract when cooling. ABS is sensitive to temperature changes and needs a heated

printing bed (80 - 115°C) for a successful outcome, is often printed at 210 - 260 °C.

58
On the other hand, PLA is bio-degradable to water and carbon dioxide, is produced from

corn starch or sugarcane, therefore is safe for human use. It is stiffer than ABS, however,

due to its lower printing temperature, it is less likely to shrink and warp. PLA is mainly

used when form is more important than function, this is particularly important when

printing sharp edges. PLA is not sensitive to temperature due to its low glass transition

and melting temperatures, therefore it does not require a heated bed (20 - 60 °C), and so

it is often printed at 180 - 220 °C.

Both ABS and PLA produce fumes during the printing process due to the heating of the

material. ABS fumes have a pungent smell and can be toxic if proper ventilation is not

used. Conversely, as PLA is made from sugar-based materials, it produces a semi-sweet

smell during printing, making it safer to use.

Due to its biocompatibility and biodegradable properties, PLA can be extensively used

for medical purposes. The pharmaceutical industry has taken advantage of the

characteristics of this material; PLA has the property to undergo hydrolysis in vivo, to

later be eliminated through the excretory system; in addition, it melts at 175 °C, therefore

it would not lose its bioactivity due to thermal degradation, making it ideal as a drug

delivery system for hormones, nanoparticles, chemotherapeutics, antibiotics and other

medications (Tappa and Jammalamadaka, 2018).

Other applications of PLA that have been studied in the medical field are the production

of surgical implants and porous scaffolds for tissue engineering; the main advantage is

that this biodegradable polymer would stay in the body temporarily and disappear upon

degradation, eliminating the necessity of a second surgery to remove the material after

the tissue involved has repaired (Ata et al., 2015).

59
In a study by Yang et al., the effect of utilizing a 3D printed model in the treatment

planning of complex elbow fractures and in physician–patient communication was

evaluated. They also determined which material was better for the fabrication of a surgical

model between ABS and PLA. The authors concluded that PLA presented several

advantages over ABS. PLA is a biodegradable and environmentally friendly material,

also it does not produce any fumes during the printing process and the final models

displayed greater glossiness and a more delicate surface when compared to ABS. In

addition, the models printed with PLA were sterilized using high-pressure steam and

showed no deformation after the process. On the other hand, ABS produced fumes during

the printing process and the fracture models that were fabricated with this material

presented curling edges, rougher surfaces, and after high-pressure steam sterilization the

models suffered distortion and were missing surface details. For all of these

characteristics, the authors suggested that models printed with PLA can be used on the

operating table as they can be sterilized and do not deform after the process, and also

presented a glossier and smoother surface, making them more suitable as surgical models

(Yang et al., 2017).

[Link] Layer thickness and surface quality

The properties of the utilised printing technology can cause differences in the dimensions

of the final model, for example, minimal thickness layer. In this study, the layer thickness

employed was 0.02 cm for the FDM printer and 0.0016 cm for the Polyjet printer, this

dissimilarity can also influence the variation in accuracy between models.

FDM models present thicker layers and rougher surfaces than Polyjet models, which can

result in less accurate replicas. The results of a study by Jin et al. corroborate this

statement, as the FDM models had significant thicker layers and were less accurate than

60
the Polyjet replicas; they used a layer thickness for FDM of 0.01cm and 0.0032cm for

Polyjet (Jin et al., 2018).

Layer thickness, printing velocity and dimensional accuracy are directly related, for

example, selecting a higher layer thickness, will allow a model to be printed in a shorter

time, however, small structures or details may not be appropriately reproduced, therefore,

it also affects the printing accuracy (George et al., 2017).

An important aspect that also needs to be taken into account is surface quality, which is

directly related with layer thickness (Stanek et al., 2012). In a study by Lee at al., Polyjet

models presented smooth surfaces and good surface details; on the other hand, the FDM

models were very rough, and the layer patterns were noticeable to the naked eye. From

this observation, the authors deduced that the differences of layer thickness between the

two technologies (in this case FDM layer: 0.330 mm, Polyjet layer: 0.016 mm) is one of

the main reasons for the different surface textures (Lee et al., 2015). Additionally,

Murugesan et al., concluded after observing FDM and Polyjet models under a scanning

electron microscope, that the Polyjet replicas also presented a uniform smooth surface,

which the authors attributed to the utilisation of a water jet to remove the supporting

material, generating a smooth surface and sufficient surface details (Murugesan et al.,

2012). This correlates with the surfaces observed in the cubes printed for this study, as

seen in Figure 25. When comparing two random samples it is possible to differentiate that

Polyjet presents a smoother surface in contrast with the visible layers of the FDM model.

61
 Figure 25. Comparison of surface texture of the upper and lower faces of the cubes,
between Polyjet (top) and FDM (bottom)

[Link] Material shrinkage and axis variations

Another important factor to consider is material shrinkage during both the building and

post-curing process, as this can also influence printing accuracy. Additive manufacturing

materials tend to suffer cure-related shrinkage and thermal contraction or expansion,

which causes dimensional distortions. The Polyjet system uses fast curing and drying,

and proper temperature and humidity inside the building chamber, reducing the risk of

distortion. Conversely, the models built through FDM need to be at room temperature for

curing and drying, causing bigger inaccuracies (Jin et al., 2018).

Another issue related to 3D printing techniques that build models layer by layer is the

stair-step effect, which results in irregular surfaces due to the accumulation or bonding of

multiple cross-sectional layers. As the layers pile up, the model becomes less accurate

and the printing method less effective. This effect can be reduced by building in thinner

layers, however, this would directly reduce the building speed, as there would be more

layers to deposit to manufacture the same structure (Jin et al., 2018). Consequently, this

62
layer by layer fabrication can produce inaccuracies in the vertical dimension (Z-axis)

(Hazeveld et al., 2014).

This corroborates the findings of other studies that suggested that rapid prototyping

technologies presented difficulties in the vertical dimension. As related by Rebong et at.,

the dimensions in the X and Y axis were prone to have increasing tendencies, conversely,

in the Z axis, they had decreasing tendencies (Rebong et al., 2018). This is directly

applicable to this study, where it can be observed that in dimensions 1, 2, 3 and 4, and

also in the four corners of the cube, the models had a decreasing tendency, being smaller

than the original 3D image. These dimensions were mainly fabricated in the Z axis and

were measured in the same direction, confirming a reduction in the vertical dimension.

These alterations were also reported in a study by Lee at al., where both Polyjet and FDM

replicas were smaller than the original group when comparing their height (Lee et al.,

2015).

In the same publication, the mean width of the replicas fabricated using a Polyjet printer

was significantly bigger when compared to the original group. This correlates with the

findings in this study, where dimension 5 of the resin cubes had an increasing tendency,

being on average 0.006 cm bigger than the original image. In the same way, the cubes

printed with ABS and PLA had an increasing tendency in dimension 6 and 7. These three

dimensions were mainly fabricated in the X and Y axis.

[Link] Model storage

Another factor that can influence the accuracy of 3D printed models and needs to be

considered is that when selecting acrylics and plastics as printing materials, due to their

characteristics, long-term storage of 3D printed models is not recommended. Most of

63
these materials tend to warp, deform or become brittle over time. In the same way, when

in contact with liquids for prolonged periods of time or stored in humid environments,

the majority if these models would swell and distort, making them inaccurate to be used

in any clinical setting (George et al., 2017).

7.4.2 Measuring limitations

When performing manual measurements with a calliper, the main sources of error are the

limitations related to the precision of the operator, such as difficulty in correctly

identifying and reproducing the measuring points. These operator errors can be

introduced, especially because the rough surfaces of the printed models make the

reference marks difficult to find (Hazeveld et al., 2014). Another source of error that has

been described in the literature, is that usually the amount of measurements is not enough

to perform a full quantification of the model. (George et al., 2017). This is the reason why

previous studies about dimensional accuracy have used 3D analysis software to identify

regional differences through multiple overlapping points. (Jin et al., 2018).

7.4.3 Dimensional accuracy

In this study, only one observer made all the measurements, therefore, the intrarater

reliability was extremely relevant. In this case, the coefficient of reproducibility was close

to zero, which indicates that all measurements were practically identical and there were

no errors that could have a significant influence in the results, making this study highly

reproducible.

Even though the results of this study establish a significant statistical difference between

ABS and the other two materials, resin and PLA, in order to integrate these results when

64
deciding on the technology to be used for 3D printing in the medical field, is paramount

to consider if this statistical difference translates into a clinical difference. An important

factor to consider is that no previous study has defined the clinically acceptable

dimensional difference between the 3D printed models and the original 3D image.

However, several studies comparing dental plaster models with replicas made by 3D

printing technologies have reported differences in the measurements of less than 0.025cm

as clinically acceptable. Jin et al., recommended that differences lower than 0.02 cm can

be clinically acceptable (Jin et al., 2018). In a study of orthodontic models, a difference

of 0.03cm was considered clinically accurate, as the tolerance for manual measurements

detected by the naked eye ranges from 0.01 to 0.03cm (Hazeveld et al., 2014).

Following this criteria, in a study by Lee et al., where the mean deviations of the replica

teeth fabricated with FDM and Polyjet were 0.004 cm and 0.003 cm respectively, the

authors concluded that both 3D printing technologies were considered clinically

acceptable, and even though the Polyjet replicas presented a statistically significant higher

accuracy, the difference with the other models was not clinically significant. (Lee et al.,

2015). However, the authors failed to report the materials utilized in each printing

method, an important factor to fairly compare their results with the ones in this research.

In the present study, on average, the mean differences were 0.01cm for resin, 0.05cm for

ABS and 0.02cm for PLA. If a difference of 0.03cm is considered clinically accurate,

then it can be concluded that the statistically significant difference of ABS is not clinically

acceptable. Conversely, both resin and PLA models dimensional difference can be

considered clinically acceptable.

65
7.5 Clinical applications

The main idea behind comparing the accuracy of these printing methods is to facilitate

the integration of these technologies into the clinical practice. The utilisation of additive

manufacturing has seen an important growth in the last few decades and the access to 3D

printing hardware and software has increased considerably due to a reduction in the cost

of equipment and materials.

Currently, this technology is mainly being used to manufacture anatomical models for

education, surgical simulation and intraoperative navigation, to build patient-specific

surgical instruments like surgical guides and to shape medical devices like fixation plates

before the actual intervention (George et al., 2017).

The utilization of 3D printed models for patient education and surgical simulation was

reported by Yang et al., in their investigation, printed models of patient-specific complex

elbow fractures were used by the surgeon to explain the medical condition and the

treatment plan to the patients and their families, receiving high satisfaction scores for the

quality of the patient-clinician communication. In addition, the 3D models were useful

for treatment planning and surgical simulation, but also for intra-operative navigation.

The utilization of a 1:1 size 3D printed model allowed a better evaluation of the fracture

type and involved structures, improving the precision of the fracture reduction. The group

that used the 3D model during surgery saw a reduction of the operative time, which may

additionally help reduce the risk of intra-operative bleeding and enhance functional

recovery (Yang et al., 2017). Therefore, an inaccurate 3D printed model for medical

purposes can result in inadequate treatment planning, which can lead to serious

consequences for both the patient and the clinician.

66
7.5.1 Cost-effectiveness

Cost-effectiveness in 3D printing is essential to determine which equipment provides the

highest “benefit for money” and is a relevant concept for the selection of the technology

that gets the most out of the available resources.

As it has been established that both resin and PLA models are clinically acceptable, in

order to determine which printing method is more cost-effective, is important to compare

the costs of both FDM and Polyjet systems, in relation to time and price. For the Polyjet

system, the printer Objet Eden 250 has a price of around £60,000, the materials RGD720

and SUP705 have a price of around £300 per cartridge, and it took 20 hours 21 minutes

to print 5 cubes. On the other hand, for the FDM system, the printer Cubicon Single Plus

has a price of £3,000 and each reel of filament has a price of £35; the printing time was

24 hours 25 minutes for 5 models.

Considering that the accuracy and printing time of both systems is similar, the main

difference between these methods comes from the major economic cost disparity, surface

details and mechanical properties. Although, FDM’s PLA is more cost-effective,

Polyjet’s resin presents better surface details and higher mechanical properties. Once

again, the selection of the manufacturing method would depend on the clinical use of the

3D printed object, but it can be concluded that both resin and PLA are equally accurate.

The utilisation of Polyjet’s resin can be recommended for the fabrication of moulds for

maxillofacial prosthesis, surgical guides and to shape medical devices like fixation plates

and cranial implants, as higher surface details and mechanical properties are required.

Due to its similar accuracy and lower cost, FDM’s PLA can be suggested for the

fabrication of models for surgical simulation, treatment planning, medical training and

patient education, and for the production of burn masks and radiation shields.

67
7.6 Null hypothesis

Based on the results, the null hypothesis was partially rejected, as there were statistically

significant differences in accuracy between ABS and Resin and between ABS and PLA,

and this difference was not clinically acceptable. However, no significant difference was

found between Resin and PLA.

7.7 Future research

Accuracy and reproducibility of 3D printed models for medical use is currently mentioned

in a small amount of publications. To the author’s knowledge, this is the first study that

compares the accuracy of FDM and Polyjet in relation to flat surfaces; all the printed

models in previous studies had irregular surfaces, like replicas of teeth, dental arches,

elbow fractures, among others. Future studies should investigate the volume differences

between these printing methods and ideally, more observers could be added to the

research, in order to evaluate both interobserver and intraobserver reproducibility. Also

further research is necessary on the impact of layer thickness in model accuracy.

68
8. Conclusions

The intrarater agreement was high for all measurements, the coefficient of repeatability

was close to zero, indicating that this measuring technique was highly reproducible.

There was a highly significant difference (P< 0.0005) between resin and ABS, and

between ABS and PLA; and for ABS, in some dimensions, the difference between the

original image compared with the cube measurements were judged to be not clinically

acceptable. On the other hand, there was no significant difference between Polyjet’s resin

and FDM’s PLA, and both printing methods are considered clinically acceptable.

Even though FDM’s PLA presents a lower economic cost and is considered equally

accurate as Polyjet’s resin, the latter presents better surface details and higher mechanical

properties, therefore, the selection of the printing method would depend on the clinical

purpose of the 3D printed model.

69
9. Glossary of terms

ABS: Acrylonitrile butadiene styrene, is a thermoplastic polymer generally used as the

build material in fused deposition modelling printers.

Additive Manufacturing: Also known as three-dimensional (3D) printing, is the process

of building a 3D model by the addition of thin layers of material.

Brittleness: A mechanical property of materials where, when subjected to stress, the

material breaks with little elastic deformation and without significant plastic deformation.

Extrude: The process of forcing out a layer of melted thermoplastic material onto a

building platform in in the fabrication of a 3D model.

Filament: General term given to the material used in fused deposition modelling printers,

it is commonly a thermoplastic polymer, like ABS or PLA, supplied in coils or reels. The

filament is fed to a print head as a solid, heated up and extruded through a small nozzle

into the building platform.

FDM: Fused deposition modelling, is a 3D printing technology where a heated

thermoplastic filament is extruded through a print head nozzle to build up layers of

material to fabricate a 3D model.

Glass transition temperature: The temperature range where a polymer changes from a

hard, rigid material to a soft, not melted material.

Nozzle : The part of a 3D printer’s head where the build material is extruded from.

PLA: Polylactic acid, is a biodegradable thermoplastic polymer commonly used as

building material in fused deposition modelling printers. PLA is easier to work with than

ABS, however, it is structurally more brittle.

70
Photopolymer: A polymer that changes its properties when exposed to certain types of

light. For 3D printing, this commonly refers to materials (polymers) that are in a

liquid/resin state and harden when exposed to UV light.

Polyjet: Is a 3D printing technology where ultraviolet cured resins are deposited onto a

build platform by jetting heads, layer by layer, similar to those of inkjet printers. The

system utilizes two different types of liquid photo-polymerised resin, one for build

material and one for support material.

Post processing: Any procedure to improve the appearance or properties of a 3D model

after it has been printed. It can involve processes like support removal, UV curing,

sanding, polishing and heat treating.

STL: Or “standard tessellation language”, is the most popular file format for 3D printing.

It represents a 3D object by only describing its surface geometry as a series of triangles,

without any details of colour or texture.

Subtractive Manufacturing: The opposite of additive manufacturing, is the process of

creating a 3d model by successively cutting material from an initial block, usually one

layer at a time.

Warping: Is a 3d printing defect that occurs due to material contraction when cooling

after printing. This contraction causes stress along the object’s lateral surfaces, especially

at corners where two sides meet; the stress on both sides causes the corner to deform

upwards and inwards.

71
X-Axis: Part of the Cartesian coordinate system used by 3d printers to move while

printing an object. The x-axis represents a horizontal, left to right movement.

Y-Axis: Part of the Cartesian coordinate system used by 3d printers to move while

printing an object. The y-axis represents a horizontal, front to back movement.

Z-Axis: Part of the Cartesian coordinate system used by 3d printers to move while

printing an object. The z-axis represents a vertical, top to bottom movement.

72
10. References

Aimar, A., Palermo, A. and Innocenti, B. (2019). The Role of 3D Printing in Medical

Applications: A State of the Art. J Healthc Eng 2019: 5340616-5340616.

Artopoulos, A., Buytaert, J. A. N., Dirckx, J. J. J. and Coward, T. J. (2014). Comparison

of the accuracy of digital stereophotogrammetry and projection moiré

profilometry for three-dimensional imaging of the face. International Journal of

Oral and Maxillofacial Surgery 43(5): 654-662.

Ata, R., Aladdin, A., Othman, N., Malek, R. A., Leng, O., Aziz, R. and El Enshasy, H.

(2015). Recent applications of polylactic acid in pharmaceutical and medical

industries. Journal of Chemical and Pharmaceutical Research 2015: 51-63.

Bae, E.-J., Jeong, I.-D., Kim, W.-C. and Kim, J.-H. (2017). A comparative study of

additive and subtractive manufacturing for dental restorations. The Journal of

Prosthetic Dentistry 118(2): 187-193.

Beumer III, J., Marunick, M. T. and Esposito, S. J. (2011). Maxillofacial Rehabilitation:

Prosthodontic and Surgical Management of Cancer-Related, Acquired, and

Congenital Defects of the Head and Neck. Third Edition Ed. Hanover Park IL,

Quintessence Publishing Co, Inc.

Bos, E. J., Scholten, T., Song, Y., Verlinden, J. C., Wolff, J., Forouzanfar, T., Helder,

M. N. and van Zuijlen, P. (2015). Developing a parametric ear model for

auricular reconstruction: A new step towards patient-specific implants. Journal

of Cranio-Maxillofacial Surgery 43(3): 390-395.

Chen, L., Lin, W. S., Polido, W. D., Eckert, G. J. and Morton, D. (2019). Accuracy,

reproducibility, and dimensional stability of additively manufactured surgical

templates. J Prosthet Dent 122(3): 309-314.

73
Chen, Z.-C., Albdour, M. N., Lizardo, J. A., Chen, Y.-A. and Chen, P. K.-T. (2015).

Precision of three-dimensional stereo-photogrammetry (3dMD™) in

anthropometry of the auricle and its application in microtia reconstruction.

Journal of Plastic, Reconstructive & Aesthetic Surgery 68(5): 622-631.

Chua Chee, K. (1994). Three-dimensional rapid prototyping technologies and key

development areas. Computing & Control Engineering Journal 5(4): 200-206.

Ciocca, L., De Crescenzio, F., Fantini, M. and Scotti, R. (2010). CAD/CAM bilateral

ear prostheses construction for Treacher Collins syndrome patients using laser

scanning and rapid prototyping. Computer Methods in Biomechanics and

Biomedical Engineering 13(3): 379-386.

Ciocca, L., Fantini, M., De Crescenzio, F., Persiani, F. and Scotti, R. (2011). Computer-

aided design and manufacturing construction of a surgical template for

craniofacial implant positioning to support a definitive nasal prosthesis. Clinical

Oral Implants Research 22(8): 850-856.

Coward, T., Richards, R., Fenlon, M. and J J Scott, B. (2011). Development of a

Stereophotogrammetry Technique to Assess Facial Change Following Surgery

for Head and Neck Cancer.

Coward, T., Scott, B., Watson, R. and Richards, R. (2006). A comparison between

computerized tomography, magnetic resonance imaging, and laser scanning for

capturing 3-dimensional data from a natural ear to aid rehabilitation. The

International journal of prosthodontics 19: 92-100.

Coward, T. J., Scott, B. J. J., Watson, R. M. and Richards, R. (2000). Laser scanning of

the ear identifying the shape and position in subjects with normal facial

symmetry. International Journal of Oral and Maxillofacial Surgery 29(1): 18-

23.

74
Davis, B. K. (2010). The role of technology in facial prosthetics. 18(4): 332-340.

Day, K. M., Gabrick, K. S. and Sargent, L. A. (2018). Applications of Computer

Technology in Complex Craniofacial Reconstruction. Plastic and Reconstructive

Surgery Global Open 6(3): e1655.

De Menezes, M., Cerón-Zapata, A. M., López-Palacio, A. M., Mapelli, A., Pisoni, L.

and Sforza, C. (2016). Evaluation of a Three-Dimensional

Stereophotogrammetric Method to Identify and Measure the Palatal Surface

Area in Children with Unilateral Cleft Lip and Palate. 53(1): 16-21.

de Menezes, M., Rosati, R., Ferrario, V. F. and Sforza, C. (2010). Accuracy and

Reproducibility of a 3-Dimensional Stereophotogrammetric Imaging System.

Journal of Oral and Maxillofacial Surgery 68(9): 2129-2135.

Dhima, M., Salinas, T. J. and Rieck, K. L. (2013). Virtual Surgical Planning for

Treatment of Severe Mandibular Retrognathia With Collapsed Occlusion Using

Contemporary Surgical and Prosthodontic Protocols. Journal of Oral and

Maxillofacial Surgery 71(11): 1923-1932.

Dindaroğlu, F., Kutlu, P., Duran, G. S., Görgülü, S. and Aslan, E. (2016). Accuracy and

reliability of 3D stereophotogrammetry: A comparison to direct anthropometry

and 2D photogrammetry. 86(3): 487-494.

Faul, F., Erdfelder, E., Lang, A. G. and Buchner, A. (2007). G*Power 3: a flexible

statistical power analysis program for the social, behavioral, and biomedical

sciences. Behavior research methods 39(2): 175-191.

Fernandes, N., van den Heever, J., Sykes, L. and Kluge, H. (2016). Nasal reconstruction

of a patient with complete congenital arhinia: A clinical report. The Journal of

Prosthetic Dentistry 116(6): 924-927.

Flores, R. L., Liss, H., Raffaelli, S., Humayun, A., Khouri, K. S., Coelho, P. G. and

Witek, L. (2017). The technique for 3D printing patient-specific models for

75
 auricular reconstruction. Journal of Cranio-Maxillofacial Surgery 45(6): 937-

943.

Fullerton, G. D. (1982). Basic concepts for nuclear magnetic resonance imaging.

Magnetic Resonance Imaging 1(1): 39-53.

George, E., Liacouras, P., Rybicki, F. J. and Mitsouras, D. (2017). Measuring and

Establishing the Accuracy and Reproducibility of 3D Printed Medical Models.

Radiographics 37(5): 1424-1450.

Guttenberg, S. A. (2008). Oral and Maxillofacial Pathology in Three Dimensions.

Dental Clinics of North America 52(4): 843-873.

Hazeveld, A., Huddleston Slater, J. J. and Ren, Y. (2014). Accuracy and reproducibility

of dental replica models reconstructed by different rapid prototyping techniques.

American journal of orthodontics and dentofacial orthopedics : official

publication of the American Association of Orthodontists, its constituent

societies, and the American Board of Orthodontics 145(1): 108-115.

Heller, M., Bauer, H.-K., Goetze, E., Gielisch, M., Roth, K. E., Drees, P., Maier, G. S.,

Dorweiler, B., Ghazy, A., Neufurth, M., Müller, W. E. G., Schröder, H. C.,

Wang, X., Vahl, C.-F. and Al-Nawas, B. (2016). Applications of patient-specific

3D printing in medicine. International journal of computerized dentistry 19(4):

323-339.

Hoarau, R., Zweifel, D., Simon, C. and Broome, M. (2014). The use of 3D planning in

facial surgery: Preliminary observations. Revue de Stomatologie, de Chirurgie

Maxillo-faciale et de Chirurgie Orale 115(6): 353-360.

Hong, D., Lee, S., Kim, T., Baek, J. H., Lee, Y. M., Chung, K. W., Sung, T. Y. and

Kim, N. (2019). Development of a personalized and realistic educational thyroid

cancer phantom based on CT images: An evaluation of accuracy between three

different 3D printers. Computers in biology and medicine 113: 103393.

76
Huang, Y.-H., Seelaus, R., Zhao, L., Patel, P. K. and Cohen, M. (2016). Virtual surgical

planning and 3D printing in prosthetic orbital reconstruction with percutaneous

implants: a technical case report. International Medical Case Reports Journal 9:

341-345.

Imakr. (2016). "Cubicon Single Plus." Available at:

[Link]

(Accessed: 12-08-2020).

Imakr. (2018). "Filament for FDM." Available at: [Link]

for-fdm (Accessed: 12-08-2020).

Jardini, A. L., Larosa, M. A., Filho, R. M., Zavaglia, C. A. d. C., Bernardes, L. F.,

Lambert, C. S., Calderoni, D. R. and Kharmandayan, P. (2014). Cranial

reconstruction: 3D biomodel and custom-built implant created using additive

manufacturing. Journal of Cranio-Maxillofacial Surgery 42(8): 1877-1884.

Jin, S. J., Jeong, I. D., Kim, J. H. and Kim, W. C. (2018). Accuracy (trueness and

precision) of dental models fabricated using additive manufacturing methods.

International journal of computerized dentistry 21(2): 107-113.

Kau, C. H., Wang, J. and Davis, M. (2019). A Cross-Sectional Study to Understand 3D

Facial Differences in a Population of African Americans and Caucasians. Eur J

Dent 13(4): 485-496.

Kim, G. B., Lee, S., Kim, H., Yang, D. H., Kim, Y.-H., Kyung, Y. S., Kim, C.-S., Choi,

S. H., Kim, B. J., Ha, H., Kwon, S. U. and Kim, N. (2016). Three-Dimensional

Printing: Basic Principles and Applications in Medicine and Radiology. Korean

J Radiol 17(2): 182-197.

Lee, K. Y., Cho, J. W., Chang, N. Y., Chae, J. M., Kang, K. H., Kim, S. C. and Cho, J.

H. (2015). Accuracy of three-dimensional printing for manufacturing replica

teeth. Korean journal of orthodontics 45(5): 217-225.

77
Liacouras, P., Garnes, J., Roman, N., Petrich, A. and Grant, G. T. (2011). Designing and

manufacturing an auricular prosthesis using computed tomography, 3-

dimensional photographic imaging, and additive manufacturing: A clinical

report. Journal of Prosthetic Dentistry 105(2): 78-82.

Lübbers, H.-T., Medinger, L., Kruse, A. L., Grätz, K. W., Obwegeser, J. A. and

Matthews, F. (2012). The influence of involuntary facial movements on

craniofacial anthropometry: a survey using a three-dimensional photographic

system. British Journal of Oral and Maxillofacial Surgery 50(2): 171-175.

Marszałek, B., Wójcicki, P., Kobus, K. and Trzeciak, W. H. (2002). Clinical features,

treatment and genetic background of Treacher Collins syndrome. Journal of

applied genetics 43(2): 223-233.

Martelli, N., Serrano, C., van den Brink, H., Pineau, J., Prognon, P., Borget, I. and El

Batti, S. (2016). Advantages and disadvantages of 3-dimensional printing in

surgery: A systematic review. Surgery 159(6): 1485-1500.

Mazzoni, S., Marchetti, C., Sgarzani, R., Cipriani, R., Scotti, R. and Ciocca, L. (2013).

Prosthetically Guided Maxillofacial Surgery: Evaluation of the Accuracy of a

Surgical Guide and Custom-Made Bone Plate in Oncology Patients after

Mandibular Reconstruction. 131(6): 1376-1385.

Meulstee, J. W., Verhamme, L. M., Borstlap, W. A., Van der Heijden, F., De Jong, G.

A., Xi, T., Bergé, S. J., Delye, H. and Maal, T. J. J. (2017). A new method for

three-dimensional evaluation of the cranial shape and the automatic

identification of craniosynostosis using 3D stereophotogrammetry. International

Journal of Oral and Maxillofacial Surgery 46(7): 819-826.

Msallem, B., Beiglboeck, F., Honigmann, P., Jaquiéry, C. and Thieringer, F. (2017).

Craniofacial Reconstruction by a Cost-Efficient Template-Based Process Using

3D Printing. Plastic and Reconstructive Surgery Global Open 5(11): e1582.

78
Murugesan, K., Anandapandian, P. A., Sharma, S. K. and Vasantha Kumar, M. (2012).

Comparative evaluation of dimension and surface detail accuracy of models

produced by three different rapid prototype techniques. J Indian Prosthodont

Soc 12(1): 16-20.

Nkenke, E., Lehner, B., Kramer, M., Haeusler, G., Benz, S., Schuster, M., Neukam, F.

W., Vairaktaris, E. G. and Wurm, J. (2006). Determination of Facial Symmetry

in Unilateral Cleft Lip and Palate Patients from Three-Dimensional Data:

Technical Report and Assessment of Measurement Errors. The Cleft Palate-

Craniofacial Journal 43(2): 129-137.

Nuseir, A., Hatamleh, M., Watson, J., Al-Wahadni, A. M., Alzoubi, F. and Murad, M.

(2015). Improved Construction of Auricular Prosthesis by Digital Technologies.

26(6): e502-e505.

Nyberg, E. L., Farris, A. L., Hung, B. P., Dias, M., Garcia, J. R., Dorafshar, A. H. and

Grayson, W. L. (2017). 3D-Printing Technologies for Craniofacial

Rehabilitation, Reconstruction, and Regeneration. Annals of biomedical

engineering 45(1): 45-57.

Palousek, D., Rosicky, J. and Koutny, D. (2014). Use of digital technologies for nasal

prosthesis manufacturing. 38(2): 171-175.

Pang, S. S. Y., Fang, C. and Chan, J. Y. W. (2018). Application of three-dimensional

printing technology in orbital floor fracture reconstruction. Trauma Case

Reports 17: 23-28.

Papadopoulos, M. A., Christou, P. K., Christou, P. K., Athanasiou, A. E., Boettcher, P.,

Zeilhofer, H. F., Sader, R. and Papadopulos, N. A. (2002). Three-dimensional

craniofacial reconstruction imaging. Oral Surgery, Oral Medicine, Oral

Pathology, Oral Radiology, and Endodontology 93(4): 382-393.

79
Patel, A., Islam, S. M. S., Murray, K. and Goonewardene, M. S. (2015). Facial

asymmetry assessment in adults using three-dimensional surface imaging. Prog

Orthod 16: 36-36.

Plooij, J. M., Maal, T. J. J., Haers, P., Borstlap, W. A., Kuijpers-Jagtman, A. M. and

Bergé, S. J. (2011). Digital three-dimensional image fusion processes for

planning and evaluating orthodontics and orthognathic surgery. A systematic

review. International Journal of Oral and Maxillofacial Surgery 40(4): 341-352.

Qiu, J., Gu, X.-y., Xiong, Y.-y. and Zhang, F.-q. (2011). Nasal prosthesis rehabilitation

using CAD-CAM technology after total rhinectomy: a pilot study. Supportive

Care in Cancer 19(7): 1055.

Rebong, R. E., Stewart, K. T., Utreja, A. and Ghoneima, A. A. (2018). Accuracy of

three-dimensional dental resin models created by fused deposition modeling,

stereolithography, and Polyjet prototype technologies: A comparative study. The

Angle orthodontist 88(3): 363-369.

Riphagen, J. M., van Neck, J. W. and van Adrichem, L. N. A. (2008). 3D Surface

Imaging in Medicine:: A Review of Working Principles and Implications for

Imaging the Unsedated Child. Journal of Craniofacial Surgery 19(2): 517-524.

Robiony, M., Salvo, I., Costa, F., Zerman, N., Bandera, C., Filippi, S., Felice, M. and

Politi, M. (2008). Accuracy of Virtual Reality and Stereolithographic Models in

Maxillo-Facial Surgical Planning. Journal of Craniofacial Surgery 19(2): 482-

489.

Salazar-Gamarra, R., Seelaus, R., da Silva, J. V. L., da Silva, A. M. and Dib, L. L.

(2016). Monoscopic photogrammetry to obtain 3D models by a mobile device: a

method for making facial prostheses. Journal of Otolaryngology - Head & Neck

Surgery 45: 33.

80
Schaaf, H., Malik, C. Y., Howaldt, H.-P. and Streckbein, P. (2009). Evolution of

photography in maxillofacial surgery: from analog to 3D photography – an

overview. Clinical, Cosmetic and Investigational Dentistry 1: 39-45.

Seok, H., Kim, S.-G., Park, Y.-W. and Lee, Y.-C. (2017). Postoperative Three-

Dimensional Evaluation of Mandibular Contouring Surgery Using Computer-

Assisted Simulation Planning and a Three-Dimensional-Printed Surgical Guide.

28(3): 768-770.

SMG 3D. (2011). "Stratasys Objet Eden 250." Available at:

[Link] (Accessed: 12-08-2020).

Stanek, M., Manas, D., Manas, M., Navratil, J., Kyas, K., Senkerik, V. and Skrobak, A.

(2012). Comparison of different rapid prototyping methods. International

Journal of Mathematics and Computers in Simulation 6: 550-557.

Stratasys. (2015). "RGD720." Available at:

[Link] (Accessed: 12-08-2020).

Sutradhar, A., Park, J., Carrau, D., Nguyen, T. H., Miller, M. J. and Paulino, G. H.

(2016). Designing patient-specific 3D printed craniofacial implants using a

novel topology optimization method. Medical & Biological Engineering &

Computing 54(7): 1123-1135.

Tappa, K. and Jammalamadaka, U. (2018). Novel Biomaterials Used in Medical 3D

Printing Techniques. Journal of functional biomaterials 9(1).

Torabi, K., Farjood, E. and Hamedani, S. (2015). Rapid Prototyping Technologies and

their Applications in Prosthodontics, a Review of Literature. Journal of dentistry

(Shiraz, Iran) 16(1): 1-9.

van Loon, B., Maal, T. J., Plooij, J. M., Ingels, K. J., Borstlap, W. A., Kuijpers-Jagtman,

A. M., Spauwen, P. H. and Bergé, S. J. (2010). 3D Stereophotogrammetric

assessment of pre- and postoperative volumetric changes in the cleft lip and

81
 palate nose. International Journal of Oral and Maxillofacial Surgery 39(6): 534-

540.

Wang, P., Wang, Y., Zhang, Z., Li, X., Ye, B. and Li, J. (2018). Comprehensive

consideration and design with the virtual surgical planning-assisted treatment for

hemifacial microsomia in adult patients. Journal of Cranio-Maxillofacial

Surgery 46(8): 1268-1274.

Watson, J. and Hatamleh, M. M. (2014). Complete integration of technology for

improved reproduction of auricular prostheses. The Journal of Prosthetic

Dentistry 111(5): 430-436.

Werz, S. M., Zeichner, S. J., Berg, B. I., Zeilhofer, H. F. and Thieringer, F. (2018). 3D

Printed Surgical Simulation Models as educational tool by maxillofacial

surgeons. European Journal of Dental Education 22(3): e500-e505.

Witjes, M. J. H., Schepers, R. H. and Kraeima, J. (2018). Impact of 3D virtual planning

on reconstruction of mandibular and maxillary surgical defects in head and neck

oncology. Current Opinion in Otolaryngology & Head and Neck Surgery 26(2):

108-114.

Yadav, S., Narayan, A. I., Choudhry, A. and Balakrishnan, D. (2017). CAD/CAM-

Assisted Auricular Prosthesis Fabrication for a Quick, Precise, and More

Retentive Outcome: A Clinical Report. 26(7): 616-621.

Yang, L., Grottkau, B., He, Z. and Ye, C. (2017). Three dimensional printing

technology and materials for treatment of elbow fractures. International

Orthopaedics 41(11): 2381-2387.

82
11. Image Credits

Figure 2. Adapted from [Link]

Figure 3. Adapted from [Link]

Figure 4. Adapted from [Link]

prototyping/[Link]

Figure 5. Adapted from [Link]

manufacturing

Figure 6. Adapted from [Link]

Figure 7. Adapted from [Link]

83
12. Appendix

12.1 Dimensions of the designed cube

Figure A1. Dimensions surface A (left) and B (right)

Figure A2. Dimensions surface C (left) and D (right)

84
 Figure A3. Dimensions top (left) and bottom (right) surfaces

12.2 Images of samples of the 3D printed models

Figure A4. Sample of 3D printed cube in photopolymerizing resin

85
Figure A5. Sample of 3D printed cube in ABS

Figure A6. Sample of 3D printed cube in PLA

86