Technical And Clinical Considerations For The Development Of 3d

Transcript

1 Technical and Clinical Considerations for the Development of 3D Printed UpperLimb Prostheses for Pediatric Patients Jorge M. Zuniga1,2, Matthew J. Major3,4, Jean L. Peck5, Rakesh Srivastava6, James Pierce1, Nicholas Stergiou1,7 1. University of Nebraska at Omaha, Department of Biomechanics, USA 2. Facultad de Ciencias de la Salud, Universidad Autónoma de Chile, Chile 3. Northwestern University Feinberg School of Medicine, Department of Physical Medicine and Rehabilitation, USA 4. Department of Veterans Affairs, Jesse Brown VA Medical Center, USA 5. CHI Health Creighton University Medical Center and an adjunct faculty at the Department of Occupational Therapy at Creighton University, USA 6. Innovative Prosthetics & Orthotics, USA 7. University of Nebraska Medical Center, Department of Environmental Agricultural and Occupational Health, USA 2 Abstract The development of 3D printing for the manufacturing of prostheses and orthoses has resulted in cost reduction strategies, better accessibility and customization of prosthetic designs. The widespread use of 3D printing and the existence of myriad prosthetic designs available on the internet allows clinicians and researchers from different disciplines to manufacture their own devices. Given the dearth of studies discussing the practical application of 3D printed upper limb prostheses, the current paper describes the technical and clinical considerations for the implementation of these devices in rehabilitation and research settings. Specifically, considerations on fitting procedures, assembly, durability, regulatory implications, and patient functional outcomes are discussed. Keywords 3D printing, computer-aided design, low-cost prostheses, FDA regulations, transitional prosthesis, prosthesis for children, prosthetic function. 3 Background Children’s prosthetic needs are complex due to their small size, constant growth, and psychosocial development.1 Socio-economical background and financial resources play a crucial role in prescription of prostheses for children, especially when private insurance and public funding are insufficient.2 Electric-powered units (i.e., myoelectric) and mechanical devices (i.e., body-powered) have improved to accommodate children’s needs, but their maintenance and replacement costs make access difficult for many families.3-5 Voluntary-closing upper-limb prostheses are more suitable for children and could improve gross motor development.2,6 Currently, the best cost-effective option for pediatric populations is a passive prosthetic hook;7 although functional, these devices have a high rejection rate, in part due to weight, cost and low visual appeal.8 Most clinically-recommended prostheses do not adapt to the typical growth of children’s limbs and require regular visits to health care providers for adjustments or replacement, which may ultimately lead to device abandonment.2,5 It is estimated that in the United States, more than 541,000 individuals lived with upper limb deficiencies in 2005.9 Worldwide estimates for upper limb reductions range from 4-5/10,000 to 1/100 live births10, while in specific parts of the world such as Australia, Finland, and Canada, reports indicate that 3.4 to 5.3 of 10,000 live born children experience upper limb abnormalities. However, only 1 in 9,400 children is considered for a prosthesis due to the complexity of the reduction and fitting process, lack of interest from the pediatric patient, and the prohibitive cost of upper limb prostheses.6 Furthermore, many children with upper limb deficiencies from resourcelimited countries are not being fitted with a prosthesis due to a lack of trained technicians able to provide these services and a local shortage of the necessary componentry for the fabrication of proper devices.12 Consequently, there is a clear and critical need for practical, easily fitted and maintained, customized, aesthetically appealing, low-cost prosthetic devices for children.1,2 Advancements in computer-aided design software and additive manufacturing techniques (i.e., three dimensional or 3D printing), offer the possibility of designing, printing, and fitting prosthetic hands and other assistive devices at a very low cost. Previous studies1,13-16 have described low cost prosthetic hands, arms and shoulders with practical and easy fitting procedures that can be performed remotely. Importantly, in children the durability of the 3D printed prostheses is challenged continuously due to their activity levels and outgrowth of the prostheses.17 Therefore, the cost effectiveness of 3D printing makes repairs and upgrades of prostheses substantially more affordable.17 In general, previous publications1,13-15,17-19 have presented different aspects of the development of 3D printed prostheses for children, and the consensus is that 3D printing is a promising manufacturing method for the development of these devices. There is a lack of information, however, with respect to the fitting procedures, assembly, durability, and regulatory implications, and improvements of relevant clinical outcomes. The Food and Drug Administration (FDA) recently drafted guidelines for the industry outlining technical considerations associated with the 3D printing processes and recommendations for testing and characterizing 3D printed devices.20 However, the 4 literature still lacks a comprehensive review that can provide practical technical and clinical considerations for the implementation of 3D printed upper limb prostheses in a clinical or research setting. The current review aims to address this knowledge gap and to also: (i) provide preliminary findings for remote prosthetic fitting methodologies; (ii) describe printing specifications, post-processing and assembly of 3D printed prostheses for pediatric patients; and (ii) propose future steps for improving the usage of 3D printed prostheses in rehabilitation. 1. Technical considerations The introduction of 3D printing for the manufacturing of orthoses and prostheses has resulted in the development of new cost reduction strategies and better accessibility and customization of prosthetic designs.1,13-15,17-19 The existence of 3D printing prosthetic designs available on the internet, allows clinicians and researchers from different disciplines to manufacture their own devices.17-19 The Cyborg Beast 3D printed hand prosthesis was the first open-source design described in the scientific literature and will be used as a reference in the current review.1 This design is characterized by being both inexpensive and easy to manufacture. This prosthesis requires simple anthropometric measures of the upper limb to fit the device, scale dimensions and implementation adjustments.1 Currently there are 58 different 3D printed prostheses designs, 51 available online and seven reported in the scientific literature.19 The majority of the designs that are available on the internet, can be downloaded as a stereolithographic or source files. stereolithographic files are the standard file type used by most additive manufacturing systems. stereolithographic files contain less information than the source files originated in the computer-aided design software, thus modifications to the mesh or structure of a triangulated model is time consuming or not possible with most computer-aided design software. Table 1 includes a list of the main open-source stereolithographic editing software that facilitate modifications to the mesh of stereolithographic files. Standard for the Exchange of Product (STEP) model data files are the most common source files and seem to be the most appropriate format for transferring computer-aided design data. STEP files are able to contain all the parametric data required to read and modify the size and geometry of the overall prosthesis model contained in the digital file and simplify mesh editing procedures. This type of files is the preferred option, if extensive modifications to the mesh of the prosthesis are required. STEP files can be also converted to stereolithographic files providing great versatility when sharing files with collaborators. Having a digital file that can be modified or scaled to fit each patient facilitates a number of activities, such as collaborations among clinicians, transfer of prosthetic designs and necessary modifications, as well as the development of remote prosthetic fitting procedures. However, the procedures can be cumbersome and may require a multidisciplinary team with experience using this software to manipulate these type of files. As 3D modeling and printing procedures have become more mainstream, the industry and the open-source community has committed significant efforts to simplify the procedures for editing and modifying different file formats for prosthetic applications (Table 1). 5 1.1 Distance Prosthetic Fitting Procedures Once the files of the devices are downloaded they need to be properly scaled to the dimensions of: 1) the patient’s residual limb for generating an appropriately sized socket, and 2) the non-affected limb for facilitating bilateral length symmetry and overall function. Parametric digital files allow the standardization of certain parameters, Figure 1. A. Illustration of an image imported as a plane and overlaid while other parameters are under the Cyborg Beast palm model scaled at 140% for a 16 years old adjusted to fit the affected research participant. B. Overlaid of a 3D printed arm prostheses scaled at 99% for a 7-year old child with a trans-radial reduction upper limb. However, the model in the left arm. majority of digital files of prosthetic designs currently available on the internet, are not parametric and require scaling to fit each patient’s residual limb. This step requires the supervision of a board certified prosthetist, orthotist or a hand therapist familiar with fitting upper limb prostheses. Although, remote fitting procedures can be performed using 2D photographs (Figure 1), the most accurate method for performing a digital fitting is by superimposing the prosthetic design over a 3D image of the scanned affected limb (Figure 2). There are several commercially-available high-end 3D scanners for prosthetic application but they are often cost-prohibitive. There are several relatively low-cost hand-held scanners that can create a 3D model representation of the upper limbs, such as the Sense 3D scanner (Sense, 3D Systems Inc, Rock Hill, SC). These low-cost options often require multiple attempts to obtain accurate results. The main challenge in obtaining a reliable scan using low-cost scanners are related to lighting (indoor versus outdoor lighting), motion (involuntary movements of the patient), and positioning (scanning outside of the optimal distance range of 38.1 to 152.4 cm). To improve scanning reliability and capabilities, our laboratory has developed a 3D printable rotational scanning system compatible with low-cost scanners or smart phones (Figure 2).21 Our rotational scanning system contains an electronically driven epicyclical gearing system attached to a shaft that holds the third-party scanner or smartphone. The arm rest is located in the center of the epicyclical gearing system to stabilize the arm for scanning in a fixed limb position. This scanning system helps to partially address the issue of motion and position artifacts while scanning the upper-limbs of pediatric patients. The blueprints and digital files of this device are open-sourced and available upon request from the corresponding author of this review. 1.1.1. Anthropometric Measurements 6 If the scanning option is not possible, the majority of the current open-source prostheses including the Cyborg Beast can be fitted by acquiring a few simple anthropometric measurements, including hand length (tip of the middle finger to center of the wrist joint, Figure 3C1and C5), palm width (widest region of the palm above the base of the thumb, Figure 3C2), forearm length (center of the wrist joint to center of the elbow joint, Figure 3C3 and C6), forearm width at three-fourths (width of the forearm at proximal threefourths of the length of the forearm proximal to the wrist, Figure 3C4 and C7), and rangeof-motion of the wrists Figure 2. Upper-limb scanning system. A) Arm support component. (extension and flexion, Figure B) Epicyclical gearing system configuration. C) Demonstration of device prototype attached to a third party optical scanner 3A1 and A1). The remote fitting scanning D) Forearm scan of traumatic amputee patient. procedure involves the application of photogrammetric methods to extract all the required measurements from three photographs of the upper limbs using open-source software (Figure 3; Table 1). Validity and reliability of the anthropometric measurements are crucial to achieving proper fit of the prosthesis. A previous investigation compared the anthropometric measurements (Figure 3) obtained by a hand therapist directly from a subject’s upper limbs to those extracted from photographs using photogrammetric methods and found negligible mean differences between these measurements.1 Furthermore, the Bland and Altman plots showed no major bias and narrow limits of agreements.1 To demonstrate the validity and reliability of these procedures, three photographs of the upper limbs were taken as shown in Figure 3. All photographs included a ruler for scale (i.e., calibration) and were taken directly above the arms with the entire forearm up to the elbow in the frame. To measure wrist range-of-motion, participants were requested to actively and maximally extend (Figure 3A) and flex (Figure 3B) at the wrist joint. In addition, a longitudinal reference line was drawn over the participant’s non-affected and affected hands connecting the wrist joint center and tip of the 3rd digit of the non-affected hand (Figure 3C; C1) or the midline of the affected hand, respectively (Figure 3C; C5). The remote-fitting procedure can also be used for trans-radial amputees using the elbow joint to assess the required range-of-motion needed to operate elbow-driven arm prostheses (Figure 5). An open-source image 7 editing program (Tracker, Open Source Physics, Arlington VA) can be used to obtain all anthropometric measurements (Figure 3). Figure 3. Three photographs of a participant’s upper limbs. A: wrist extension (A1: non-affected, A2 affected), B: wrist flexion (B1: non-affected, B2: affected), and C: Top view (C1: Non-affected hand length, C2: Non-affected hand width, C3: Non-affected forearm length, C4: Non-affected forearm width, C5: Affected hand length, C6: Affected forearm length, and C7: Affected forearm width). After the application of photogrammetric methods to extract all the required measurements, this image can be imported to a computer-aided design software. Importing calibrated images into a modeling or computer-aided design software is required to properly scale the prosthetic 3D model. After importing the calibrated imaged into a modeling software, such as Blender (Blender 2.77, Blender Foundation, Amsterdam, Netherlands), calibration of the metric scale can be performed by changing the default unit (meters to centimeters) by adjusting the scale to 0.01. The image plane can be resized to match the size of the 1 cm background grid on Blender using the ruler on the imported image plane. The accuracy of the calibrations is confirmed using the interactive ruler tool on Blender and performing several measurements over the ruler included in the image plane. Importing of the calibrated images can also be performed in proprietary computer-aided design software such as Autodesk Fusion 360 (Fusion 360, Autodesk, Inc., San Rafael, CA, USA). To validate the use of a proprietary computer-aided design software our research team compared the different measurements of the computer-aided design model of the hand prosthesis for several research participants estimated from Blender and Autodesk Fusion 360. Anthropometric measurements of 10 children with congenital partial hand reductions (8 male, 2 female, range of 3 to 17 years of age) showed no significant differences between estimates using Blender and Autodesk Fusion 360 when statistically analyzed with dependent ttests (critical α=0.05). Specifically, there were no statistically significant differences in means for length of the palm (Blender=73.56 ± 12.60 mm; Fusion= 73.63 ± 12.66 mm), width of the palm (Blender=79.91 ± 13.77 mm; Fusion=80.65 ± 13.78 mm), height of the palm (Blender=45.45 ± 7.66 mm; Fusion=45.61 ± 7.71 mm), and diagonal length of the palm (Blender=106.1892 = ± =18.12 mm; Fusion=107.06 ± 18.65 mm). After the digital fitting has been performed by a technician under the supervision and approval of a clinical team, the design can be exported as an stereolithographic file 8 and 3D printed. All the relevant steps to manufacture a 3D printed prosthesis are summarized in Table 1. 1.2 3D Printing Specifications and Materials The most common 3D printing method for 3D printed prostheses is fused deposition modeling. Fused deposition modeling is a form of additive manufacturing that involves melting thin layers of plastic over each other to form a 3D structure. The two most common 3D printed filament materials used to manufacture upper limb prostheses is polylactide filament and acrylonitrile butadiene styrene filament. Polylactide filament has similar properties to a thermoplastic and permits minor modifications through targeted heating once the device has been 3D printed. Acrylonitrile butadiene styrene filament however, does not offer homogenous thermoplastic properties making postprocessing modifications difficult. Thus, the preferred material for 3D printed prostheses is polylactide filament given the ability to perform post-processing modifications that may be required during assembly or clinical fitting. After selecting the printing material, the files need to be sliced and nested before 3D printing. Slicing can be performed using open-source software (see Table 1). Slicing is the process that converts 3D models into a format understood by 3D printers. The model is "sliced" into 2D cross-sectional layers which can be placed by the extruder. The slicing software often develops a G-code which is a set of numerical values representing positions in the system-based reference frame in which the extruder needs to follow to trace and recreate the design. Furthermore, nesting is the process of organizing the printed parts in the building platform in an effective and efficient manner. Most consumer desktop 3D printers provide access to software that allows the user to set the preferred parameters. These parameters include percentage of infill and pattern (i.e., hexagon pattern), print speed (mm/s), temperature of the heated bed (°F/C), shell thickness (mm) and the option of using rafts and supports to ensure printing feasibility. Most 3D printed prostheses can be manufactured using low-cost desktop 3D printers (Ultimaker 2, Ultimaker B.V., Geldermalsen, The Netherlands) or industrial grade 3D printer (Uprint SE Plus, Stratasys, Minnesota, USA). In general, the digital files of any upper-limb prostheses for children available on the internet can be 3D printed with a minimum building platform of 28.5 cm L x 15.3 cm W x 15.5 cm H. When using desktop 3D printers, the recommended settings are 40% infill (hexagon pattern), 35 mm/s print speed, 150 mm/s travel speed, 65°C heated bed for acrylonitrile butadiene styrene filament (room temperature for polylactide filament), 0.15 mm layer height, and 1 mm shell thickness. Rafts and supports are always recommended to avoid misprints. Furthermore, the application of light adhesive (e.g., glue stick) or painter’s tape over the building platform of desktop 3D printers can aid in improving the adherence of the first plastic layer and facilitates removal of the printed structure. For industrial 3D printers, this may not be necessary since the building tray is often disposable and provides sufficient adherence. Due to the inherent anisotropic characteristics of fused deposition modeling, orientation of the different 3D printed prosthesis components over the building tray can play an important role in the durability of the prosthesis. As characteristics and mechanical properties of 3D printed parts are dependent on printing orientation, it is important to consider the direction of operational loads on each component. 3D printed 9 parts using fused deposition modeling are much more likely to delaminate and fracture when placed in tension in the direction of build height compared to the orthogonal axes. For this reason, it is recommended to print 3D printed prostheses using a horizontal axis. There are only few exceptions to this rule specific to the layer height and dimension of the printed part. Lower layer height (higher resolution) typically results in a part being printed with smoother surfaces and less likelihood to delaminate or fracture. In the case of the palm section of the 3D printed Prosthesis Cyborg Beast 2 using a 0.15 mm layer height, we have experienced no adverse effects in the durability of the palm component of the prosthesis when subjected to real-world use. The duration of the print depends on the size of the prostheses, infill, layer height (resolution) and orientation of the printed part. For example, the printing time of a partial hand prosthesis for a 12 year-old child can be between 6 and 8 hours when using the recommended settings.1,13 1.3 Description, Post-processing and Assembly of 3D Printed Prostheses After each component of the prosthesis is 3D printed, post-processing may be required depending on the implemented design and print settings. This post-processing stage includes the removal of the support and rafts using tools to file and smooth areas exposed to friction, such as the plastic pins, wrist joint, and metacarpophalangeal joints. To avoid extensive post-processing, it is recommended to manually add support to the 3D printed parts using specialized slicing software (Simplify3D LLC, Blue Ash, OH). The use of a specialized slicing software provides a practical method for the inspection of each 2D cross-sectional layers and the addition of support structures in critical areas of the prostheses, such as the palm structure. 1.3.1 Description of various 3D Printed Prostheses from our Laboratory 1.3.1a The 3D Printed Hand Prosthesis Cyborg Beast 2 A modified version of the 3Dprinted transitional Figure 4. The 3D printed partial hand prosthesis (Cyborg Beast 2). A. The hand prosthesis in the open position. Elastic cords placed inside the dorsal aspect of the fingers provide passive finger extension. B. Finger flexion was driven by non-elastic cords along the palmar surface of each finger and was activated through 20-30° wrist flexion of the residual functional joint. The red arrow shows the direction of wrist flexion to close the fingers and produce a functional grasp. 10 hand prosthesis named “Cyborg Beast”22 (Figure 4) will be used as a reference design for the current discussion. The new version, Cyborg Beast 2, was designed using the modeling software program Autodesk Fusion 360 and manufactured in the 3D Printed Prosthetic, Orthotic & Assistive Devices Laboratory located in the Biomechanics Research Building of the University of Nebraska at Omaha. This design is available upon request. For manufacturing we used a combination of a desktop (Ultimaker 2, Ultimaker B.V., Geldermalsen, The Netherlands) and an industrial (Uprint SE Plus, Stratasys, Minnesota, USA) 3D printer. Specifically, the plastic pins to secure all the different components of the prosthesis, as well as the fingers and thumb, were made of acrylonitrile butadiene styrene filament using the industrial 3D printer. The palm, socket, forearm brace, and leveraging structure were made of polylactide filament. The Elastic cords placed inside the dorsal aspect of the fingers provide passive finger extension. Finger flexion was driven by non-elastic cords along the palmar surface of each finger and is activated through 20 to 30° of wrist flexion. The result was a composite fist (flexing the fingers towards the palm) for gross grasp. The finger and thumb were oriented in opposition to facilitate cylindrical grasp and tip pinch. A BOA dial tensioner system (Mid power reel M3, BOA Technology Inc., Denver, Colorado) allowed tension regulation of the cables controlling finger flexion. A brace leverage structure was included in the proximal aspect of the forearm to increase torque development and stability. Furthermore, a thermoplastic socket embedded in the palmar aspect of the hand prosthesis was added to facilitate fitting of the device. The prosthesis was also customizable to each child’s limb size and aesthetic requirements, such as colors and theme (e.g., comic characters). 1.3.1b The 3D Printed Arm (trans-radial) Prosthesis The hand was designed to possess 5 fingers with 2 degrees of freedom (Figure 5). The finger and thumb were oriented in opposition to facilitate cylindrical grasp and tip pinch. Silicone finger pads were added to provide increase friction for grasping activities. A rotation mechanism placed on the wrist allowed full pronation and supination, and a pivot system with internal components allowed wrist rotation without twisting the line. The rotation mechanism of the wrist consisted of an inner circular disc/shaft with a center opening (Figure 5B). A circle of embedded magnets with matching polarity was placed around the disc. A bi-valve circular sleeve with embedded magnets aligned to match the disc magnets, was placed over the disc. The magnets were placed with opposing polarity to assure mutual attraction. The disc and sleeve rotate independently and were stabilized in various positions by the attraction of the magnets. The magnets were sealed in a protective sleeve for safety. Elbow flexion and extension can be performed using a simple hinge mechanism. A BOA dial tensioner system allowed the regulation of the tension of the cables controlling the finger flexion. A Velcro strap secured the prosthesis to the arm and harnessing was not needed for suspension. 1.3.2 Hand Post-processing and Assembly 11 Careful assembly of the hand prosthesis was required as all components must fit together well and glide smoothly. The fingers, thumb and their point of attachment on the palm may need to be sanded or filed to assure easy movement of the fingers (i.e., minimal friction and binding) prior to attaching the fingers to the palm and threading the polyester cord through the palmar side of the fingers for generating finger flexion. After the fingers and thumb were pinned in place, the elastic cords for finger extension were threaded through the tip of one finger, through the palmar loop and into the adjacent finger. Figure 5. The 3D Printed arm prostheses. A. The trans-radial arm With all the fingers and prosthesis in the open position. Elastic cords placed inside the dorsal aspect of the fingers provide passive finger extension. B. Finger flexion thumb secured, the wrist was activated through 10-20° of elbow flexion of the residual functional gauntlet was attached to joint. The red arrow shows the direction of elbow flexion to close the the palm using printed fingers and produce a functional grasp. The wrist can be manually snap pins and lock adjusted. washers. The BOA dial tensioner system was adhered to the back of the gauntlet using a strong adhesive (Sugru, FormFormForm Limited, London, UK.). The line of tension for each finger was attached to the BOA mechanism via a plastic line guide. The tension of each finger was adjusted so that the index finger and the thumb formed a functional pinch with the remaining fingers forming a functional grasp. 1.3.3. Arm Post-processing and Assembly Similar to the hand prosthesis, all components must fit together well and glide smoothly. Sanding and filing of rough edges and tight fitting parts may be necessary. With each of the fingers and the thumb, the polyester cord was threaded through the palm side for generating finger flexion. The finger and thumb cords were thread into the 12 palm and secured with a knot. Before pinning the fingers and thumb to the palm, the elastic cord was threaded through one finger, through the loop of the dorsum of the palm and then through the adjacent finger. When both cords were loosely in place, the finger and thumb could be pinned to the hand. The elastic cords for finger extension were then adjusted to achieve the desired tension. For the wrist rotational mechanism, magnets were secured in the threaded section and in both fitted sleeves. All magnets must be placed with polarization in the same direction. After threading the rotational mechanism into the matching forearm threads, the opposing side was secured to the palm. The upper arm and forearm sections were attached with locking pins. Threading of the control cable begun at one side of the BOA mechanism, passed through the upper arm and forearm channels, through the swivel, up through the contralateral side of the arm, and then returned into the BOA mechanism. The BOA mechanism could be secured to the upper arm with a strong adhesive. The finger and thumb cords were then attached to the opposite side of the swivel with careful adjustment of the tension to achieve functional pinch and grip. 1.4 Current FDA Recommendations 1.4.1 FDA Draft Guidance One of the most useful and groundbreaking features of 3D printing is its ability to produce complex geometries both quickly and at low cost for the manufacturer. As 3D printing technology matures and continues to permeate varied industries and products, it has become apparent that some level of regulation must be developed to ensure consistent quality and reproducibility across fabrication types, manufacturers and materials. A recent FDA Draft Guidance titled “Technical Considerations for Additive Manufactured Devices”20 serves as a first step toward defining government policy regarding the use of 3D printing for medical devices. In the following sections of this review (1.4.1 to 1.4.2c), a general description of the FDA recommendations is provided, followed by comments regarding their implications specific to 3D printed prostheses and the numerous production methodologies, such as selective laser sintering, fused filament fabrication or fused deposition modeling, stereolithography and powder bed and polyjet. In general, the FDA recommends20 that all medical devices fabricated using additive manufacturing processes (either entirely or in part) be validated and tested with the similar procedures as performed on traditionally manufactured devices with adaptation of protocols to ensure proper functionality. The overall requirements for this additional evaluation depends almost entirely on the intended end-use of the product; some common factors listed include whether it is an implant, load bearing, and/or available in pre-specified standard sizes or is patient-matched. Based on these parameters, a series of recommendations for design, production and testing are then provided. These recommendations stem from areas of importance to the additive manufacturing process, such as materials, design, printing, and post-printing validation; printing characteristics and parameters; physical and mechanical assessment of final devices; and biological considerations of final devices (including cleaning, sterility, and biocompatibility). For example, an external assistive device (i.e. prostheses) would require far less material regulation, biocompatibility testing and dimensional accuracy 13 than an additive manufacturing-produced vascular stent. Despite the lesser requirements placed on prostheses, there are still a number of intensive and important inspection tools and process tasks that are recommended for all design and manufacturing stages. 1.4.1a Design Starting with the design of a device, the FDA recommends20 that a full production flow diagram be created to ensure repeatability in the process of product development and engineering. This flow diagram would include all critical steps from device design to the post-processing of the final part identified and documented. During design, it is also highly recommended to develop feasible size ranges and feature sizes based on the end-user requirements as well as the additive manufacturing machine capabilities. Because the relative scale of additive manufacturing medical devices can be modified so easily as compared to traditionally manufactured parts, it is important that versioning history be recorded as well as specific device scales and other identifying information on the device itself. This policy prevents future third parties from improperly modifying or adapting the additive manufacturing device due to lack of information. 1.4.1b Materials Once a design has been completed, it is important to properly characterize and select suitable materials for fabricating the device. As with any type of manufacturing method, the FDA recommends20 to document safety procedures, and chemical, mechanical and physical properties of the materials used. For 3D printed prostheses, necessary material properties include chemical composition, molecular formula, purity, chemical structure, molecular weight, molecular weight distribution, glass transition temperature, and melting temperature. Because material recycling is not commonly used with fused deposition modeling/stereolithography additive manufacturing processes, no additional analysis into material properties should be required. 1.4.1c Printing Characteristics/Parameters. Equally as important as material properties, the printer and parameters used in the production of 3D printed prostheses also affect durability, surface finish, interlayer adhesion, strength-to-weight ratio, and many other part characteristics. Due to the anisotropic mechanical properties of nearly all additive manufacturing processes, it is important to optimize physical part orientation when printing. Additionally, the infill density (the density of the interior reinforcing lattice structure in fused deposition modeling) can significantly affect strength, flexibility and opacity of finished parts.20 Many additive manufacturing machines incorporate a supporting external lattice (or support structure) to aid in the creation of “floating” geometries which could not otherwise be produced. If not carefully placed, these supplementary supports can cause structural damage and issues with surface finish. Finally, layer thickness can significantly impact surface finish (causing pronounced interlayer striations), interlayer adhesion, and production speed. By justifying the print parameter choices for a balance between speed, strength, and quality, the opportunity exists to create optimized additive manufacturing devices and workflows.20 14 1.4.1d Physical/Mechanical Assessment The FDA suggests20 that a series of mechanically destructive and nondestructive tests be performed to determine ideal printing parameters. The results of such an analysis can be used to optimize parameters. To test additive manufacturing parts non-destructively, many potential tests are proposed,20 including ultrasound, computed tomography, X-ray (for simple geometry), confocal microscopy and hyperspectral imaging. Each of these methods can return valuable information such as part porosity, structural integrity, dimension accuracy, and density. In addition to this battery of assessments, the FDA recommends that tensile test coupons be destructively analyzed to document the impact of the additive manufacturing process on mechanical properties. Parts can also be manually inspected for conformance to the intended design constraints. 1.4.1e Biological Considerations Because the focus of the FDA Draft Guidance20 is to determine additive manufacturing medical device requirements, the interactions that parts have (mechanically, chemically, etc.) with the patients’ physiology must be documented and verified before use. Cleaning protocols must also be established, especially for devices with complex geometries, such as engineered porosity, honeycomb structures, channels, and internal voids or cavities that cannot be produced by traditional manufacturing methods. The FDA recommends20 that the cleaning process should account for the complex geometry of the device under worst-case conditions (e.g., greatest amount of residual manufacturing materials for cleaning validation, and a combination of largest surface area, greatest porosity, and most internal voids for sterilization validation). In general, the cleaning protocols should follow similar precautions to those already established by the FDA20 and used for standard prostheses and orthoses. 1.4.2 Observations about the FDA Draft Guidance 1.4.2a Need for Methodology-Specific Guidelines One issue that arises from a broad analysis of the entire field of additive manufacturing is the fact that each production methodology is significantly different in terms of part quality and mechanical properties, material usage and properties, safety concerns, and overall applications. This leads to a series of generalizations that have the potential of mischaracterize the different types of 3D printing in the market. For example, when describing the validation requirements for the production of parts, the FDA Draft Guidance20 states that all machines have specific build volume regions where they function optimally, and that this necessitates part quality testing after any repositioning of parts in the build volume irrespective of build orientation for quality control. This statement is true of some types of additive manufacturing namely selective laser sintering and delta-gantry fused deposition modeling, but has no bearing on the other 3D printing methods. 1.4.2b Documentation of 3D Printing Parameters Another area which would impact the implementation of 3D printed prostheses is the proposed requirement for print material and parameter validation. The FDA Draft 15 Guidance20 states that printing parameters should be captured and validated for each machine and material combination, and that all combinations be documented. With the extensive levels of parameter customization such as infill density, layer height, wall thickness, top/bottom solid layers, and many others parameter that are innate to 3D printing, it is challenging to fully validate and document every useful set of values which may hinder the production and implementation of low-cost 3D printed prostheses. Instead of quantifying every possible parameter, one alternative option could be to characterize structural properties as a function of normalized print parameters (especially infill density) using standardized bench tests. The mechanical properties of each material could then be applied to these normalized responses to obtain a reasonable approximation of final part strength from regression equations. Alternatively, the mechanical and physical properties of the various print parameters could be standardized through testing by the 3D printer manufacturers. This would eliminate variability associated with separate analyses performed by each medical device manufacturer and would encourage the use of suitable standardized equipment. 1.4.2c Part and Machine Validation The validation processes explained by the FDA20 seem to be derived from other manufacturing methods, likely due to a basis in traditional (subtractive) manufacturing processes. Ideas like using test builds (printed duplicates of a part to be nondestructively and destructively tested) and overall process validation are reasonable and would likely benefit final product quality when production frequency is sufficiently low. On the other hand, “worst-case builds,” or parts which have been intentionally produced with weak orientations/parameters, pose a significant challenge to the efficient design of devices. The FDA suggested that parts must be validated in the worst-case scenario to minimize potential of correctly printed part failure. This suggestion may lead to over-engineered (too complex and needlessly robust) prosthetic designs due to the inherent anisotropy of 3D printed parts. 1.5 Bench Testing of a 3D Printed Partial Hand Prostheses 16 There is a lack of quantitative parameters that address the durability of 3D printed prostheses. Cyclical and compression failure loading was performed in an early version of the Cyborg Beast 2 partial hand prosthesis. The 3D printed prostheses was manufactured for a 12-year old boy using polylactide filament at a 120% scale and at 40% infill. 1.5.1 Bench Testing Methods Cyclical testing of the prosthesis was conducted to determine acute changes in performance from repeated flexion and extension. An image of the testing setup is displayed in Figure 6A and B. The prosthesis was fit onto a limb surrogate consisting of a steel rod representing the forearm and a wooden model of the hand which were joined by a 3mm rubber pad. Both the forearm and hand surrogate were covered with soft foam. The Velcro strap of the prosthesis forearm cradle was fixed to the forearm surrogate Figure 6. Cycling testing setup. A. Side view B. Front view C. Wrist joint angle and the hand and distance between tips of the 1st and 2nd digits for cyclical testing. For joint surrogate filled the angles, 180 degrees represents neutral position with the longitudinal axis of prosthesis hand the hand segment aligned with the forearm segment axis. cradle. The forearm surrogate was secured to a fixture that did not permit longitudinalaxis rotation of the prosthesis. Hooks were threaded into the hand surrogate and 17 protruded through the prosthesis hand cradle. The hooks were attached to a steel cable that ran through a pulley located directly inferior to the wrist joint and was placed into tension by linear actuation of a hydraulic-driven materials test machine (Instron, Norwood, MA). Cable tension flexed the wrist joint and extension position was reset manually with the cable relaxed. The prosthesis was subjected to 50 cycles of flexion/extension from a starting position of approximately 15 degrees extension. The linear stroke displacement occurred over 10 seconds and was set such that the prosthesis 1st and 2nd digits initially mated when in wrist joint flexion. For each cycle, initial and final wrist joint angle were measured with a digital goniometer, and distance between tips of the 1st and 2nd digits were measured with digital calipers. The BOA cable system was tensioned once at the start of testing so that the digits were in full extension apart from the distal phalanges which were in approximately 45 degrees of flexion and provided slight resistance when extended. Using the materials test machine, the compression failure load and mode were determined by subjecting the prosthetic hand (i.e., components distal to the wrist joint) to a compressive force at a linear actuation displacement rate of 1 mm/s. Prior to testing, the hand without a surrogate model was secured to the loading surface with the hand in a pronated position in an orientation where the loading vector approximately passed through the palmer plane and was preloaded to 10 N. The prosthesis was compressed between two flat surfaces with a roller-bearing plate beneath one surface to minimize shear forces. The prosthesis was observed for signs of failure at which point the test was concluded. Instantaneous linear compression force and stroke displacement were recorded during testing. 1.5.2 Bench Testing Results Wrist joint angle at the start and end of each cycle, and the distance between the 1st and 2nd digit at the start of each cycle are displayed in Figure 6. Through the 50 cycles of flexion and extension, the prosthesis Figure 7. A. Force versus displacement profile of the compression failure test with instanced of 1st digit and cradle failure noted. B. Dorsal view of the prosthesis after failure. Red circle indicated the area of failure. C. Palmar view of the prosthesis after failure. Red circle indicated the area of failure 18 consistently rotated approximately 36 degrees at a rate of 3.6 degrees/second into 21 degrees of flexion and the 1st and 2nd digits always met (i.e., closed pinch) at the end of each flexion phase. Upon return into wrist extension, distance between the 1st and 2nd digits decreased sharply and then plateaued after 7 cycles, suggesting that the BOA cable should be re-tensioned regularly to ensure that the digits extend fully with wrist extension. Overall, the prosthesis demonstrated highly repeatable performance when tested over a small number of cycles. The prosthetic hand ultimately failed at 1,746 N, which was characterized by sequential catastrophic failure of the 1st digit and then hand cradle (Figure 7). The 1st digit failed at the joint housing allowing for separation of the digit from the hand (Figure 7). The hand cradle failed by shattering on the dorsal side with a diagonal line fracture and separation of the 1st digit section from the palmar mesh which pulled apart the radial side wrist joint pivot. 2. Clinical Considerations Previous investigations have reported a lack of evidence with respect to the user acceptance, functionality and improvement of relevant clinical outcomes after using upper-limb 3D printed prosthesis.19 Studies reporting improvements in function and other relevant clinical outcomes are increasing as 3D printing becomes a more established practice in clinical and research settings. Currently, the use of 3D printed prostheses has demonstrated promise for restoration of function and range-of-motion of patients with upper-limb reductions. It has been previously indicated 15 that the main goal of 3D printed prostheses is not to substitute a standard prosthesis, but to offer another tool to clinicians interested in recommending a low-cost transitional prosthesis for their patients. The term “transitional prosthesis” has been widely used in the field of prosthodontics, specifically with hemimaxillectomy patients.23 For upper and lower limb prostheses, however, these types of devices are referred as a “temporary prosthesis,” “initial prosthesis,” or “immediate postoperative prosthesis.”24 In this study “transitional prosthesis” will be defined as temporary or training prosthesis that is used to introduced or re-introduced children that have rejected or abandoned standard prostheses.15 2. 1. Improvements in Clinical Outcomes 2.1.1 Function Recent investigations and preliminary data have reported significant increases in function using 3D printed prosthesis. Lee et al25 demonstrated that a 67-year-old patient using a 3D printed prosthetic thumb displayed several functional improvements following a one month accommodation period. Functional improvements were identified in four of the six items assessed by the Jebsen–Taylor hand function test.26 The purpose of this test is to assess a broad range of unilateral hand functions required for activities of daily living. The items that showed improvements included elapsed time and accuracy of writing after the accommodation period (before=39.52s; after=34.48s), times for card turning (before=14.18s; after=11.73s), picking up small objects (before=61.30s; after=57.45s), and simulating feeding (before=20.81s; after=18.15s). The Box and Block test27 which is a measure of unilateral gross dexterity resulted in higher scores after using the 3D printed thumb prosthesis (before=17 blocks per minute; 19 after=21 blocks per minute). Furthermore, the patient also reported favorable patient satisfaction. The Quebec User Evaluation of Satisfaction with Assistive Technology28 score was 36 of a total 40 points with most statements scored as “very satisfied” or “quite satisfied”. Additionally, most activities on the Orthotics and Prosthetics Users’ Survey29 were scored as “easy” or “very easy”. These improvements in function as reported by Lee et al25 are in agreement with findings from our laboratory using a similar thumb prosthesis (Figure 8). A 72 year-old patient with a left thumb amputation of his dominant hand was fitted with a 3D printed thumb prosthesis. After 6 weeks of usage, the patient’s Box and Block test scores improved for the affected hand (before=0 blocks per minute; after=38 blocks per minute), while the scores for the non-dominant intact hand were unchanged (before=37 blocks per minute; after=37 blocks per minute). The patient reported using the thumb prosthesis 6 hours per day mostly for oil painting that utilized a tripod prehensile grip. For other types of devices manufactured in our laboratory, we observed Figure 8. A 72-year-old patient with a left thumb amputation of his dominant hand was fitted with a improvements in Box and Block test 3D printed thumb prosthesis. scores for a sample of eleven children (five female and six male, 3 to 15 years of age) fitted with a 3D printed transitional partial hand (n=9) or arm (n=2) prosthesis. The results of the repeated-measures twoway ANOVA suggested a significant hand × time interaction effect for function [F(1,10) = 6.42; p = 0.03, ηp2= 0.39]. The main finding of the present investigation was that usage of a low-cost 3D-printed transitional hand prosthesis significantly improved function after 1 to 6 months of usage (before=6.3±12.3 blocks per minute and after=13.0±12.7 blocks per minute; p = 0.03) in children with upper-limb deficiencies. 2.1.2 Range of Motion, Forearm Circumference, and Strength. Previous investigations have used upper-limb transitional prosthetic devices, such as opposition poles and dynamic orthoses, with the objective of restoring and preserving function, strength and range-of-motion in children and adults with upper-limb reduction.30,31 Our laboratory has published preliminary data indicating that 6 months of usage of a wrist-driven 3D printed hand prosthesis resulted in increased peak wrist 20 flexion (before=54±14° and after=68±14°, p=0.02), extension (before=40.38° and after=47±36° cm, p=0.04) and forearm circumference (before=16±2 cm and after=18±1 cm) in a sample of five children with upper-limb differences.15 range-of-motion was measured by a trained hand therapist using a standard measuring tape. The nonaffected side was assigned as the comparative limb to control for anatomical growth. Muscle strength of the wrist flexors that was measured with a hand dynamometer (microFET3, Hogan Health Industries, West Jordan, UT), was not statistically greater that the non-affected hand, but the lack of significance may be due to low statistical power resulting from the small sample size. The increase in circumference and wrist range-of-motion previously reported suggests that a low-cost 3D-printed prosthesis can be effectively used as a transitional or initial device in children with traumatic or congenital upper-limb differences. The increased circumference and active wrist rangeof-motion of the affected wrist after extended usage of a low-cost 3D-printed transitional prosthetic device suggest that transitional prostheses may play an important role in improving joint flexibility and patient rehabilitation.30,31 2.2 Applications of 3D Printed Prostheses in Rehabilitation 2.2.1 Need for Upper-limb Prosthesis Research Participants at the 2006 State-of-the-Science Meeting in Prosthetics and Orthotics sponsored by the National Institute of Disability and Rehabilitation Research identified the development and evaluation of new prosthetic devices as a high priority for the field.2 The design of low-cost functional upper-limb prosthetic devices that are acceptable to children has proven to be particularly challenging, as evidenced by the reported 10% to 58% rejection rate.6,32-34,2,35 Clinicians, parents and children have reported that the main factors related to dissatisfaction and abandonment of the prosthetic device include electronic failure, weight, replacement cost, discomfort, and control.11,13,14 Usability is a concept originated by the aerospace and automotive industries. The concept of device usability (i.e., usefulness) is a qualitative attribute that assesses the ease of the device-user interface and can be defined as the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specified context of use.2 Resnik2 indicated that the most important attributes of usability for users of prosthetic devices include functionality achieved while using the device, customized to each individual, ease of initial setup of the device, use of good posture and body mechanics while operating the device, easy daily maintenance, and repairs. A recent investigation36 that examined amount of variability of upper-body kinematics in adult trans-radial prosthesis users have reported that variability can also lead to frustration and eventual device abandonment. Specifically, the use of shoulder and trunk movements by prosthesis users as compensatory motions to execute goaloriented tasks demonstrates the flexibility and adaptability of the motor system. However, increased variability in movement suggests that prosthesis users do not converge on a defined motor strategy to the same degree as able-bodied individuals. As compensatory dynamics may be necessary to improve functionality of trans-radial prostheses, users may benefit from dedicated training that encourages optimization of these dynamics to facilitate execution of daily living activities, and fosters adaptable but 21 reliable motor strategies. It is conceivable that kinematic repeatability may increase with prosthesis experience, or encourage continued device use. Future work is warranted to explore these relationships and the interaction of variability measures with different types of prosthetic designs. Recent technological advances in computer-aided design programs and 3D printing1, offer the unique possibility of designing and manufacturing customized upperlimb prostheses that can aid the assessment of prosthetic rehabilitation. The ability to customize these devices at a very low cost allows clinical and research settings to develop several prototypes with varied visual appeal and colors, weight or density, size, function and configuration according to the patient’s needs. 2.2.2 Influence of Upper-limb Prostheses on Brain Activation Unfortunately, even with the great advances in prosthetic technology, a large number of children with upper limb reductions still express dissatisfaction with the available technology.2,6,37 Some of the factors that contribute to this discontent include a late age of fitting, insufficient prosthetic training, high complexity of the devices and fitting procedures, and high costs of devices and repairs.6,32,33 In addition, 10% to 58% of children with access to a prosthesis reject it due to the excessive weight, lack of functional use, discomfort, low visual appeal of the device, and late age of fitting.6,3234,2,35 Thus, it is conceivable that the lightweight and visual appeal of 3D printed prostheses in combination with a cost-effective method of customization may contribute to reductions in the high rate of device rejection and abandonment. It has been generally accepted by clinicians that prosthesis use and early age fitting may facilitate a child’s motor development33,37,38. Additionally, it has been suggested33,38 that if a child with upper-limb reductions learns to use a prosthesis early in life, the prosthesis might be better incorporated into the body schema and integrated into sensory feedback and motor control strategies than if used at a later stage of life. The fundamental rationale for promoting early fitting and prosthesis usage in children is in line with the Neuronal Group Selection Theory developed by Edelman in 1989.39 According to this theory, the brain is dynamically organized into neuronal networks or neuronal groups. The structure and the function of these networks are influenced by development and behavior of the child. The Neuronal Group Selection Theory proposes that motor development is characterized by two phases of variation: primary and secondary.33,38,39 During the primary variability phase that occurs prior to age one year, motor activity is variable and not based on environmental conditions. In the secondary variability phase present after one year of age, the child learns to select the most efficient motor strategy for that specific scenario based on active practice from the “variable movement repertoire”. The transition from primary to secondary variability in reaching movements takes place at one year of age but not until adolescence does secondary variability of all motor functions obtain its adult configuration.33,38,39 Thus, the use of a 3D printed prostheses and exposure to a diverse variety of motor programs will enhance the development of a greater “variable movement repertoire”. 22 The Neuronal Group Selection Theory holds that children with unilateral upper limb reductions may lack representation of the missing part of the limb in the cerebral cortex.33,38,39 As a consequence, the child may have a limited “motor repertoire” for the affected upper-limb.33 Under this framework, the Neuronal Group Selection Theory suggests that intervention in these children at an early age, such as prosthetic fitting and use, may lead to an enlargement of the primary neuronal networks located in the cortical area involved with motor control of the affected limb. Ultimately, this might lead to a larger repertoire of motor strategies reflected in improved cortex activity and function resulting in a greater functional use of the prosthesis and ultimately improved quality of life and.33,38 The use of a reliable neuroimaging technique in conjunction with highly customized and visually appealing 3D printed prostheses could provide the unique opportunity to quantitatively assess the influence of upper-limb prostheses on the brain activation responses of the primary motor cortex in children. 2.2.2a Results of Pilot Study A 7-year-old child with a congenital unilateral left trans-radial reduction and two aged-matched control children performed the Box and Block test (Figures 9A, B, C) while measuring cortical activation of the C3 (left hemisphere) and C4 (right hemisphere) brain landmarks associated with the movement of the hands and arms (Figure 9A). The mean values of oxygenated hemoglobin for left hemisphere (controlling the right arm) and the right hemisphere (controlling the affected left arm when using the Figure 9. A. A 7-year-old child with a congenital upper-limb trans-radial reduction during data collection. B. Illustration of the hemodynamic responses of oxygenated (HbO) and deoxygenated (HbR) hemoglobin during the Box & Block tests (highlighted is the 60s window for the analysis). C. Bar graphs comparing the mean values of HbO during the Box and Block with the right arm (left hemisphere) and affected left arm with the prosthesis (right hemisphere). 23 prosthesis) were measured and analyzed sing a functional near-infrared spectroscopy system (Hitachi ETG-4000, Hitachi Medical Corporation, Tokyo, Japan). The right hemisphere showed a disproportionally higher cortical activation ([oxygenated hemoglobin]=0.96 a.u. Figure 9C) when compared to the left hemisphere ([oxygenated hemoglobin]=0.58 a.u; Fig. 9C). The cortical activation of the right hemisphere of the child with trans-radial reduction was also higher than aged-matched control children ([oxygenated hemoglobin]=0. 21±0.03 a.u; Fig. 9C). These results suggested that there is an increased hemodynamic response of the area of the brain controlling the affected arm with the prosthesis. This increased activity may reflect an imprecise and exploratory cortical activity during a gross dexterity task. More research is necessary to elucidate the influence of upper-limb prostheses on brain activation. However, this is an area of great promise and can clearly provide with invaluable information regarding usage of upper limb prostheses in growing children. 2.2.3 Influence of upper-limb Prostheses on Motor Performance Up to 31% of children with upper-limb reductions present spinal deviations of functional origin due to asymmetries in arms and trunk.40 The use of a prosthesis has been shown to stimulate symmetrical movements and may help children with upper-limb reductions engage in activities of daily leaving that are fundamental to children’s normal growth and development of motor skills.33,38,40 Furthermore, children with congenital upper-limb reductions have significantly higher sensitivity in the affected upper-limb than acquired amputees. This hypersensitivity has been associated with high prosthetic rejection rate (10% to 58%)6,32-34,2,35 and has been negatively correlated with the ease of performance of functional tests.35 It has been reported that children with congenital upper-limb reductions have higher functionality scores than children with acquired reductions.41 These higher functionality scores have been attributed to successful rehabilitation programs that have encouraged patient compliance to daily prosthetic use and therapy.41 The increase of functional use of the prosthesis has been associated with a decrease in rejection rate.33,38 Thus, it is conceivable that prosthetic use and early rehabilitation may contribute to enlargement of the primary neuronal networks and increase the repertoire of motor strategies resulting in a better integration of the prosthesis into the child’s body schema. It is not clear if a change in neural parameters would result in a greater functional use of the prosthesis during different motor tasks, such as discrete and continuous activities. Motor tasks classified as discrete or continuous may be controlled by different underlying neural mechanisms.42 Discrete movements have a defined beginning and an end point. Common examples are turning on a light, pushing a button, raising your hand and goal-directed reaching. Continuous or rhythmic motions are those with no clear start or end, such as drawing, walking, playing the drums, swimming, and riding a bicycle. It has been well established that continuous or rhythmic tasks are better retained than discrete tasks.42,43 Since continuous motor tasks are generally easier to learn and better retained than discrete tasks42, they may be a more appropriate motor evaluation technique for children. However, there is a lack of information regarding the motor performance of children with upper-limb reductions during discrete or continuous motor tasks. Furthermore, the performance of continuous motor tasks, such as sports and other recreational activities would require activity-specific prostheses. However, as 24 the durability of 3D printed prostheses improves, the versatility of the production methodology for these devices is promising for the development of cost-effective activity-specific prostheses. 3. Conclusions The development, testing, and implementation of 3D printed prostheses in clinical and research settings is at a very early stage. The open-source nature of the computer-aided design and stereolithographic files of most 3D printed prosthetic designs provides the opportunity to access and manufacture devices for evaluation, testing and development of new prosthetic designs. Health care professionals and clinicians working in prosthetic rehabilitation are likely to benefit from adding 3D printed prostheses in their treatment procedures. Some important developments in this topic are already taking place. The U.S. Department of Veterans Affairs (VA) is working with the industry to launch a collaborative 3D printing hospital network.44 Five VA hospitals are in the process of been equipped with industrial-grade 3D printers and promoting training and education in this area to implement the development of custom prostheses, orthoses, and anatomical models for personalized healthcare. The collaboration among prosthetists, 3D printing industry, open source community, as well as any other clinicians (e.g., orthotists and physical therapists) and researchers interested in prosthetic rehabilitation is crucial for the proper implementation of 3D printed prostheses in clinical and research settings. 4. List of Abbreviations Three-dimensional (3D), Food and Drug Administration (FDA), Standard for the Exchange of Product (STEP), Department of Veterans Affairs (VA). 5. Declarations 5.1 Ethics approval and consent to participate All subjects completed a medical history questionnaire. All parents and children were informed about the study and parents signed a parental permission form. For children age 6 to 10, an assent was explained by the principal investigator and signed by the children and their parents. In addition, detailed safety guidelines were given to the parents regarding the use and care of the prosthesis. Participants were asked to visit the laboratory on three occasions. The studies presented in this article were approved by Creighton University Institutional Review Board (IRB# 13-16909) and the University of Nebraska Medical Center Institutional Review Board (IRB# 614-16-FB). All function and strength testing were performed by a trained occupational therapist, certified hand therapist. The prosthetic fitting of all our devices were performed by a certified prosthetic and orthotic professional. 5.2 Consent for publication Individual and parental consent and release waiver were obtained for all figures in this paper. 5.3 Availability of data and material 25 Data sets analyzed during the current study are available from the corresponding author on reasonable request. 5.4 Competing interests Drs. Jorge M. Zuniga, Jean Peck, and Rakesh Srivastava lead the research team that designed all the 3D printed prostheses and other devices discussed in this article. James Pierce designed and fabricated the prototypes for this study. The rest of the researchers declare no competing interests. 5.5 Funding Funding for this study was provided by NSF Nebraska EPSCoR, NASA Nebraska EPSCoR, the Center for Research in Human Movement Variability of the University of Nebraska Omaha and the NIH (P20GM109090 and R15HD08682). 5.6 Authors' contributions Conceptualization, drafting and editing of the manuscript was performed by JMZ. The bench testing section and editing of the manuscript was performed by MJM. All functional testing and assembling of the devices section was performed and written by JLP. All fitting procedures and overall editing was performed by RS. The design of porotypes, review of the FDA section, and overall editing was performed by JP. Part of the conceptualization and overall editing was performed by NS. 5.7 Acknowledgements We would like to thank the parents and their children for participating in our study. We would like to thank Jennifer Murphy from the Department of Physical Medicine and Rehabilitation, Northwestern University Feinberg School of Medicine for her help with designing and preparing the cyclical bench test setup. Thanks to the students working in the 3D Printed Prosthetic, Orthotic & Assistive Devices Laboratory at the Department of Biomechanics who helped with data collection. 5.8 Authors' information Jorge M. Zuniga1,2, Matthew J. Major3,4, Jean L. Peck5, Rakesh Srivastava6, James Pierce1, Nicholas Stergiou1,7. 1. University of Nebraska Omaha, Department of Biomechanics, USA 2. Facultad de Ciencias de la Salud, Universidad Autónoma de Chile, Chile 3. Northwestern University Feinberg School of Medicine, Department of Physical Medicine and Rehabilitation, USA 4. Department of Veterans Affairs, Jesse Brown VA Medical Center, USA 5. CHI Health Creighton University Medical Center and an adjunct faculty at the Department of Occupational Therapy at Creighton University, USA 6. Innovative Prosthetics & Orthotics, USA 7. University of Nebraska Medical Center, Department of Environmental Agricultural and Occupational Health, USA Table 1. Summary of Steps, Recommendations, and Resources for the Development of 3D Printed Prostheses Steps Recommendations Resources Distance-fitting procedures: This step includes selecting a prosthesis that is appropriate for the patient, obtaining a representation (3D scan or photograph) of the residual limb, and modifying or scaling the files of the 3D prostheses (STL, STEP or any other file). Technical: Open-source files can be found on the internet. STL files are commonly available and are easy to 3D print. If 3D scanners are not available, photographs can be used to fit the prostheses. Clinical: The majority of the files shared on the internet may or may not be appropriate and ready for patient use. The prostheses selection, fitting, remote fitting, and file modifications need to be performed in collaboration with the medical team responsible for the patient care. Open-sourced 3D printed prostheses and tutorials: NIH 3D Print Exchange (https://3dprint.nih.gov/collections/prosthetics), Thingiverse (http://www.thingiverse.com/) Cyborg Beast.org* (http://www.cyborgbeast.org/#/). 3D Printing Specifications and Materials: This step includes selecting the manufacturing method, materials, slicing procedures, calibration of the 3D printer, 3D printing settings, building platform preparation, and building time considerations. Post-Processing and Assembly of 3D Printed Prostheses: This step includes post-processing and assembling considerations. Technical: Fused Deposition Modeling is the preferred additive manufacturing method for 3D printed prostheses. The ideal materials for 3D upper-limb printed prostheses are PLA or ABS printed with a 40% to 60% infill density. “Purement” is an anti-microbial PLA filament that has promising potential in the development of 3D printed prostheses. Other FDA compliant filaments, such as Nylon 680, PET and XT-copolyester are examples of high strength filaments with promising applications in prosthetics. Simplify3D seems to be the preferred slicer software due to the degree of control over speed and quality of the prints. Clinical: Consideration of FDA recommendations is needed when selecting materials for 3D printed prostheses. The supervision of the clinical team overseeing the care of the patient is crucial for the safety of the patient and appropriate implementation of 3D printed prostheses. Technical: The use of a specialized slicer software has a great impact on the amount of residual support material present after printing. The fingers, thumb and point of attachment on the palm may need to be filed to ensure minimal friction and binding. Clinical: Post-processing and assembly are critical for the proper function of the 3D printed prosthesis. Clinicians with expertise in the development of prostheses and orthoses should oversee and approve the terminal device and the fitting of the device to the patient. Open-sourced software for mesh editing: Blender (https://www.blender.org/), FreeCAD (https://www.freecadweb.org/), SketchUp (https://www.sketchup.com/), MeshMixer (http://www.meshmixer.com/), MeshLab (http://www.meshlab.net/), 3D Slash (https://www.3dslash.net/index.php), SculptGL (https://stephaneginier.com/sculptgl/). Open-sourced slicing software: Cura (for Ultimakers 3D printer; https://ultimaker.com/en/products/curasoftware), Makerbot print (for Makerbot 3D printers; https://www.makerbot.com/print/). Proprietary CAD software: Autodesk Fusion 360, AutoCAD, Maya, Inventor (https://www.autodesk.com), Solid works (http://www.solidworks.com/), Rhino (https://www.rhino3d.com/), CATIA (https://www.3ds.com), 3ds MAX (https://www.autodesk.com). Proprietary slicing software: Simplify3D (https://www.simplify3d.com/) Open-sourced Image Editing Software: Tracker (http://physlets.org/tracker/) ImageJ (https://imagej.nih.gov/ij/) Abbreviations: 3D: Three-dimensional, STL: stereolithographic, STEP: Standard for the Exchange of Product, FDA: Food and Drug Administration, PLA: Polylactide filament, ABS: Acrylonitrile butadiene styrene filament, PET: Polyethylene terephthalate. *Please contact the corresponding author of this manuscript to access updated files and tutorials. 26 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Zuniga JM, Katsavelis D, Peck J, et al. Cyborg beast: a low-cost 3d-printed prosthetic hand for children with upper-limb differences. BMC research notes. Jan 20 2015;8(1):10. Resnik L. Development and testing of new upper-limb prosthetic devices: research designs for usability testing. Journal of rehabilitation research and development. 2011;48(6):697-706. Bray JJ, University of California Los Angeles. Prosthetics-Orthotics Education Program. Prosthetic principles, upper extremity amputations : fabrication and fitting principles. Los Angeles: Prosthetics-Orthotics Education Program, Division of Orthopedic Surgery, University of California, Los Angeles; 1970. Scotland TR, Galway HR. A long-term review of children with congenital and acquired upper limb deficiency. The Journal of bone and joint surgery. British volume. May 1983;65(3):346-349. Davids JR, Wagner LV, Meyer LC, Blackhurst DW. Prosthetic management of children with unilateral congenital below-elbow deficiency. J Bone Joint Surg Am. Jun 2006;88(6):1294-1300. Resnik L, Meucci MR, Lieberman-Klinger S, et al. Advanced upper limb prosthetic devices: implications for upper limb prosthetic rehabilitation. Arch Phys Med Rehabil. Apr 2012;93(4):710-717. Krebs DE, Edelstein JE, Thornby MA. Prosthetic management of children with limb deficiencies. Physical therapy. Dec 1991;71(12):920-934. Marshall JG. Partial hand and functional prosthesis. JACPOC. Dec 2000;6(1):1. Behrend C, Reizner W, Marchessault JA, Hammert WC. Update on advances in upper extremity prosthetics. The Journal of hand surgery. Oct 2011;36(10):17111717. Bethge M, von Groote P, Giustini A, Gutenbrunner C. The World Report on Disability: a challenge for rehabilitation medicine. American journal of physical medicine & rehabilitation / Association of Academic Physiatrists. Jan 2014;93(1 Suppl 1):S4-11. Canfield MA, Honein MA, Yuskiv N, et al. National estimates and race/ethnicspecific variation of selected birth defects in the United States, 1999-2001. Birth defects research. Part A, Clinical and molecular teratology. Nov 2006;76(11):747-756. ISPO consensus conference on appropriate orthopaedic technology for lowincome countries: conclusions and recommendations. International Society for Prosthetics and Orthotics. Prosthetics and orthotics international. Dec 2001;25(3):168-170. Zuniga JM, Carson additive manufacturing, Peck JM, Kalina T, Srivastava RM, Peck K. The development of a low-cost three-dimensional printed shoulder, arm, 27 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. and hand prostheses for children. Prosthetics and orthotics international. Apr 26 2016. Zuniga JM, Peck J, Srivastava R. Anthropometric, Range of Motion and Strength Changes After 6 Months of Using the Cyborg Beast, an Open Source Wrist Driven 3D-Printed Prosthetic Hand for Children. Midwest Chapter American Academy of Orthotists & Prosthetists 2015. Zuniga JM, Peck J, Srivastava R, Katsavelis D, Carson A. An Open Source 3DPrinted Transitional Hand Prosthesis for Children. JPO: Journal of Prosthetics and Orthotics. 2016;28(3):103-108. Gretsch KF, Lather HD, Peddada KV, Deeken CR, Wall LB, Goldfarb CA. Development of novel 3D-printed robotic prosthetic for transradial amputees. Prosthetics and orthotics international. Jun 2016;40(3):400-403. Tanaka KS, Lightdale-Miric N. Advances in 3D-Printed Pediatric Prostheses for Upper Extremity Differences. J Bone Joint Surg Am. Aug 03 2016;98(15):13201326. Burn MB, Ta A, Gogola GR. Three-Dimensional Printing of Prosthetic Hands for Children. The Journal of hand surgery. May 2016;41(5):e103-109. Ten Kate J, Smit G, Breedveld P. 3D-printed upper limb prostheses: a review. Disability and rehabilitation. Assistive technology. Feb 02 2017:1-15. FDA. Technical Considerations for Additive Manufactured Devices. Available from: http://www.fda.gov/downloads/medicaldevices/deviceregulationandguidance/guid ancedocuments/ucm499809.pdf. 2016. Systems D. Support: 3D scanners. Available from: https://www.3dsystems.com/shop/support/sense/faq. 2017. Zuniga JM. Cyborg Beast. 2014; http://www.thingiverse.com/thing:261462. Singh A, Kumar N, Singh V, Singh SK. Transitional prosthesis for a dentulous hemimaxillectomy patient. National journal of maxillofacial surgery. Jul 2010;1(2):173-175. Nallegowda M, Singh U, Khanna M, Babbar A. Temporary Upper Limb AIIMS Prostheses - A New Design. Indian Journal of Physical Medicine and Rehabilitation. 2010. Lee KH, Kim SJ, Cha YH, Kim JL, Kim DK, Kim SJ. Three-dimensional printed prosthesis demonstrates functional improvement in a patient with an amputated thumb: A technical note. Prosthetics and orthotics international. Dec 01 2016. Sears ED, Chung KC. Validity and responsiveness of the Jebsen-Taylor Hand Function Test. The Journal of hand surgery. Jan 2010;35(1):30-37. Mathiowetz V, Volland G, Kashman N, Weber K. Adult norms for the Box and Block Test of manual dexterity. The American journal of occupational therapy : official publication of the American Occupational Therapy Association. Jun 1985;39(6):386-391. Demers L, Weiss-Lambrou R, Ska B. Item analysis of the Quebec User Evaluation of Satisfaction with Assistive Technology (QUEST). Assistive technology : the official journal of RESNA. 2000;12(2):96-105. Heinemann AW, Bode RK, O'Reilly C. Development and measurement properties of the Orthotics and Prosthetics Users' Survey (OPUS): a 28 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. comprehensive set of clinical outcome instruments. Prosthetics and orthotics international. Dec 2003;27(3):191-206. Bryant MH, Donahue JR, Sweigart JF, Hunter JM. A prosthesis to restore opposition in children with congenital absence of the hand. The Association of Children's Prosthetic-Orthotic Clinics. 1979;17(6):5-10. Shim J-H, Lee Y-H, Lee J-M, Park JM, Moon J-H. Wrist-driven prehension prosthesis for amputee patients with disarticulation of the thumb and index finger. Archives of Physical Medicine and Rehabilitation. 7// 1998;79(7):877-878. Hadders-Algra M, Reinders-Messelink HA, Huizing K, van den Berg R, van der Sluis CK, Maathuis CG. Use and functioning of the affected limb in children with unilateral congenital below-elbow deficiency during infancy and preschool age: a longitudinal observational multiple case study. Early human development. Jan 2013;89(1):49-54. Huizing K, Reinders-Messelink H, Maathuis C, Hadders-Algra M, van der Sluis CK. Age at first prosthetic fitting and later functional outcome in children and young adults with unilateral congenital below-elbow deficiency: a cross-sectional study. Prosthetics and orthotics international. Jun 2010;34(2):166-174. Biddiss EA, Chau TT. Upper limb prosthesis use and abandonment: a survey of the last 25 years. Prosthetics and orthotics international. Sep 2007;31(3):236257. Reinkingh M, Reinders-Messelink HA, Dijkstra PU, Maathuis KG, van der Sluis CK. Stump sensibility in children with upper limb reduction deficiency. Journal of rehabilitation medicine. Jan 2014;46(1):51-58. Major MJ, Stine RL, Heckathorne CW, Fatone S, Gard SA. Comparison of rangeof-motion and variability in upper body movements between transradial prosthesis users and able-bodied controls when executing goal-oriented tasks. Journal of neuroengineering and rehabilitation. 2014;11:132. James MA, Bagley additive manufacturing, Brasington K, Lutz C, McConnell S, Molitor F. Impact of prostheses on function and quality of life for children with unilateral congenital below-the-elbow deficiency. J Bone Joint Surg Am. Nov 2006;88(11):2356-2365. Meurs M, Maathuis CG, Lucas C, Hadders-Algra M, van der Sluis CK. Prescription of the first prosthesis and later use in children with congenital unilateral upper limb deficiency: A systematic review. Prosthetics and orthotics international. Aug 2006;30(2):165-173. Sporns O, Edelman GM. Solving Bernstein's problem: a proposal for the development of coordinated movement by selection. Child development. Aug 1993;64(4):960-981. Postema SG, van der Sluis CK, Waldenlov K, Norling Hermansson LM. Body structures and physical complaints in upper limb reduction deficiency: a 24-year follow-up study. PloS one. 2012;7(11):e49727. Korkmaz M, Erbahceci F, Ulger O, Topuz S. Evaluation of functionality in acquired and congenital upper extremity child amputees. Acta orthopaedica et traumatologica turcica. 2012;46(4):262-268. Muratori LM, Lamberg EM, Quinn L, Duff SV. Applying principles of motor learning and control to upper extremity rehabilitation. Journal of hand therapy : 29 43. 44. official journal of the American Society of Hand Therapists. Apr-Jun 2013;26(2):94-102; quiz 103. Schmidt RL, T. . Motor Control and Learning A Behavioral Emphasis. 5th Edition ed. Champaign, IL: Human Kinetics; 2011. Innovation V. Hitting (3D) Print: Co-Creating Supportive Devices with & for Veterans. Available from: https://medium.com/vainnovation/hitting-3d-printcreating-devices-that-help-veterans-feel-whole-481657addf6b. 2017.