Patient specific ankle-foot orthoses using rapid prototyping
© Mavroidis et al; licensee BioMed Central Ltd. 2011
Received: 1 April 2010
Accepted: 12 January 2011
Published: 12 January 2011
Prefabricated orthotic devices are currently designed to fit a range of patients and therefore they do not provide individualized comfort and function. Custom-fit orthoses are superior to prefabricated orthotic devices from both of the above-mentioned standpoints. However, creating a custom-fit orthosis is a laborious and time-intensive manual process performed by skilled orthotists. Besides, adjustments made to both prefabricated and custom-fit orthoses are carried out in a qualitative manner. So both comfort and function can potentially suffer considerably. A computerized technique for fabricating patient-specific orthotic devices has the potential to provide excellent comfort and allow for changes in the standard design to meet the specific needs of each patient.
In this paper, 3D laser scanning is combined with rapid prototyping to create patient-specific orthoses. A novel process was engineered to utilize patient-specific surface data of the patient anatomy as a digital input, manipulate the surface data to an optimal form using Computer Aided Design (CAD) software, and then download the digital output from the CAD software to a rapid prototyping machine for fabrication.
Two AFOs were rapidly prototyped to demonstrate the proposed process. Gait analysis data of a subject wearing the AFOs indicated that the rapid prototyped AFOs performed comparably to the prefabricated polypropylene design.
The rapidly prototyped orthoses fabricated in this study provided good fit of the subject's anatomy compared to a prefabricated AFO while delivering comparable function (i.e. mechanical effect on the biomechanics of gait). The rapid fabrication capability is of interest because it has potential for decreasing fabrication time and cost especially when a replacement of the orthosis is required.
The unique advantages of rapid prototyping (RP) (also called layered manufacturing) for medical application are becoming increasingly apparent. Furthermore, developments in 3D scanning have made it possible to acquire digital models of freeform surfaces like the surface anatomy of the human body. The combination of these two technologies can provide patient-specific data input corresponding to anatomical features (via 3D scanning), as well as a means of producing a patient-specific form output (via RP). Both technologies appear to be ideally suited for the development of patient-specific medical appliances and devices such as orthoses.
This paper details a novel process that combines 3D laser scanning with RP to create patient-specific orthoses. The process was engineered to utilize surface data of the patient anatomy as a digital input, manipulate the surface data to an optimal form using Computer Aided Design (CAD) software, and then download the digital output from the CAD software to a RP machine for fabrication. The methods herein presented have the potential to ultimately provide increased freedom with geometric features, cost efficiencies and improved practice service capacity while maintaining high quality-of-service standards.
3D Scanning Technologies for Medical Modeling
Medical modeling is a process by which a particular part of the human body is re-created in the form of an anatomically correct digital model first and then as a physical prototype/model. Such models have had successful implementation in preoperative planning, implant design/fabrication, facial prosthetics post-surgery and teaching/concept communication to patients or medical students [1–3].
There are several 3D scanning technologies used to input the data necessary for medical modeling. Laser scanning is one method of capturing the anatomical data needed to create these models as exact replicas of the human body. 3D laser scanners use a laser beam normal to the surface to be scanned. The light reflected back from the surface is captured as a 2D projection by a CCD (charged-couple device) camera and a point cloud is created using a triangulation technique.
A second type of 3D scanner is based upon stereoscopic photogrammetry. 3D photogrammetric scanners use images captured from different points of view. Given the camera locations and orientations, lines are mathematically triangulated to produce 3D coordinates of each unobscured point in both pictures necessary to reproduce an adequate point cloud for shape and size reproduction.
Software packages that are used to create medical models for RP are unique in that they must take information from a 2D scan of the body and use that information to create a 3D model. They also have CAD functionalities to provide the possibility of optimizing the design of the model based on the application needs. The output file from the data analysis and design software is written in the standard tessellation language (STL) format, which is the most common file type used with RP machines. Once the human anatomy has been recorded and a digital model has been created, the produced STL file instructs the RP machine about how to manufacture the intended medical model [4, 5].
Rapid Prototyping for Medical Modeling and Rehabilitation
RP has been extensively used in medicine . Depending on the anatomy that is being modeled and the application of interest, different types of RP machines may be most appropriate.
The most broadly used RP technique for surgical planning and training is stereolithography (SLA) . An SLA machine uses a laser beam to sequentially trace the cross sectional slices of an object in a liquid photopolymer resin. The area of photopolymer that is hit by the laser partially cures into a thin sheet. The platform upon which this sheet sits is then lowered by one layer's thickness (resolution on the order of 0.05 mm) and the laser traces a new cross section on top of the first layer. These sheets continue to be built one on top of another to create the final three-dimensional shape. Some of the advantages of SLA are its high accuracy, the ability to build clear models for examination, and - with some materials - sterilization for biocompatibility.
Another RP technique known to the medical field is selective laser sintering (SLS) [e.g. 8]. This technology is similar to SLA since it relies upon a laser to sketch out the region to be built on a substrate. In this process, however, the laser binds a powder substrate rather than curing a liquid. This powder is typically rolled over the layer built before it by precision rollers, and each layer is dropped down exposing an area for a second layer to be applied. This technology can utilize stainless-steel, titanium, or nylon powders as fabrication materials.
In rehabilitation, RP has been used for the fabrication of prosthetic sockets [9, 10]. It has been also proposed as a way to optimize the design of customized rehabilitation tools . Research on the development of custom-fit orthoses using RP has been very limited. A 3D scanner in conjunction with SLS was used by Milusheva et al. [12, 13] to develop 3D models of customized AFO's. However, the SLS prototype of the customized AFO was used only for design evaluation purposes and not as the functional prototype. Another customized AFO manufactured using SLS was presented by Faustini et al. . The geometry of these AFOs was captured by Computed-Tomography (CT) scanning of an AFO built using a conventional technique rather than generating the surface model directly from the subject's anatomy.
It is clear that although some important pioneering research has already been performed in the area of RP patient-specific orthoses, several aspects of the implementation of the technique to manufacture AFOs using RP need to be addressed including: a) demonstrating the full design/manufacturing cycle starting from obtaining scans of the human anatomy to fabricating the customized orthosis; and b) performing gait analysis experiments to evaluate the mechanical effect of orthoses manufactured using RP and compare their performance with that achieved using orthoses fabricated by means of conventional techniques.
Current Methodology to Develop Custom-Fit AFOs
To show that the proposed technique can lead to manufacturing an AFO comparable to a prefabricated one, we chose a posterior leaf spring AFO (Type C-90 Superior Posterior Leaf Spring, AliMed, Inc., Dedham, MA) as an exemplary orthotic device to be matched by using the proposed RP-based technique . The RP implementation of the posterior leaf spring AFO used a 3D FaceCam 500 from Technest Inc.  for acquiring the data of the human's anatomy and a Viper Si2 SLA machine from 3D Systems Inc. for layered manufacturing .
The process began with removing redundant data points (Figure 4). This includes data from the parts of the leg that were not needed as well as mismatching surfaces and data from the floor or background for each captured view. The points within each cloud were then connected to each other with three-sided polygons to create a surface mesh. The individual surface meshes were aligned and merged to create one complete surface model of the ankle-foot complex. The polygon surface curvature was smoothened and edges then trimmed with a boundary curve. This surface was then offset to prevent the fabricated AFO from over-compressing the subject's leg. The offset surface was extruded to a thickness of 3 mm as typically done for fitting of standard AFOs . Once completed, the model was exported from Rapidform as a STL file.
The model was manufactured using the 3D Systems Viper Si2 SLA machine . This system uses a solid state Nd YVO4 laser to cure a liquid resin. STL files were prepared with 3D Lightyear for part and platform settings, and Buildstation to optimize the machine's configuration.
The effectiveness of using RP for the application at hand is largely dependent on material properties. The prefabricated AFO selected for the study (i.e. the one we attempted to match using the proposed methodology based on RP) was the Type C-90 Superior Posterior Leaf Spring from AliMed . This AFO comes in a pre-determined range of sizes of injection molded polypropylene.
AFO material properties:
Accura SI 40
Somos® 9120 UV
Tensile Strength (MPa)
31 - 37.2
57.2 - 58.7
7 - 13
4.8 - 5.1
15 - 25%
Young's Modulus (GPa)
1.1 - 1.5
2.6 - 3.3
1.2 - 1.4
Flexural Strength (MPa)
41 - 55
93.4 - 96.1
41 - 46
Flexural Modulus (MPa)
1172 - 1724
2836 - 3044
1310 - 1455
Gait studies were conducted at Spaulding Rehabilitation Hospital, Boston, MA using a motion capture system. We collected data from a healthy subject (the one for which scans were taken in order to manufacture the AFO) walking without an AFO, walking with the above-mentioned standard, prefabricated AFO, and walking with each of the AFOs manufactured using the proposed RP-based technique. The subject wore the AFOs on the right side. Four different conditions were therefore tested: 1) with sneakers and no AFO (No AFO); 2) with the standard, prefabricated polypropylene AFO (Standard AFO); 3) with the rigid AFO made with the Accura 40 resin (Rigid RP AFO), and 4) with the flexible AFO made from the Somos 9120 resin (Flexible RP AFO).
Gait parameters derived from the walking trials included spatio-temporal parameters and kinematics and kinetics of the hip, knee and ankle of each leg in the sagittal plane. Kinematics (joint angles) and kinetics (joint moments and powers) were estimated using a standard model (Vicon Plug-in-Gait, Vicon Peak, Oxford, UK).
Testing and Validation
Mean (± SD) spatiotemporal gait parameters of the right side for the 4 testing conditions
Flexible RP AFO
Rigid RP AFO
Walking speed (m/s)
1.49 ± 0.05
1.46 ± 0.02
1.44 ± 0.05
1.50 ± 0.06
Step length (m)
0.79 ± 0.02
0.79 ± 0.01
0.79 ± 0.03
0.82 ± 0.03
Double support time (s)
0.22 ± 0.02
0.24 ± 0.01
0.24 ± 0.01
0.23 ± 0.01
Figure 8B shows that the ankle is slightly more plantarflexed at initial contact when wearing no AFO compared to wearing an AFO, and that for each of the AFO conditions initial contact was made with the ankle-foot complex in a more neutral position. This is likely due to the AFOs being made from castings and scans, respectively, of the subject's foot set in a neutral position. During controlled plantarflexion (CD) the ankle showed a similar range of motion (RoM) for each of the AFOs with the standard, prefabricated AFO allowing slightly more plantarflexion compared to the RP AFOs (Figure 8C). This may be due to greater compliance of the polypropylene material from which the standard AFO was made.
During the phase of controlled dorsiflexion (CD), the standard AFO allowed more RoM compared to the two RP AFOs, which performed similarly (Figure 8D). This greater RoM was due to a combination of greater plantarflexion during the CP phase and also greater dorsiflexion during the CD phase.
The ankle showed the greatest RoM during the power plantarflexion (PP) phase at push-off when the subject was wearing no brace since the movement of the ankle was not restricted by an AFO. When wearing the AFOs, the amount of plantarflexion was substantially decreased (Figure 8B) while the RoM during the PP phase was slightly greater for the standard AFO compared to the two RP AFOs (Figure 8E).
In the final phase of dorsiflexion during swing (SD), the ankle showed the greatest RoM when it was not restricted by an AFO, while the three AFO testing conditions showed lower but similar ranges of motion (Figure 8F). This was partly due to the reduced amount of plantarflexion achieved during the PP phase. Importantly the two RP AFOs enabled a similar amount of ankle dorsiflexion at the end of swing as that allowed by the standard AFO (Figure 8B).
Overall, when comparing the three AFOs, it was clear that they performed similarly in terms of controlling ankle kinematics and kinetics during the gait cycle.
The flexible RP AFO performed almost identically to the standard AFO. Both required less ankle power than normal (i.e. with no AFO). The rigid AFO results showed that this testing condition was associated with high ankle power; most likely because the rigid AFO provided resistance to bending that the subject had to overcome. Despite differences among AFO's, it was noted that the change in ankle power was still relatively small, and that increased material flexibility would have been likely to help improving performance.
In this paper, we presented a process to combine state of the art 3D scanning hardware and software technologies for human surface anatomy with advanced RP techniques so that novel custom made orthoses and rehabilitation devices can be rapidly produced. Two custom-fit AFOs were rapidly prototyped to demonstrate the proposed process. Preliminary biomechanical data from gait analyses of one subject wearing the AFOs indicated that the RP AFOs can match the performance of the standard, prefabricated, polypropylene design. This new platform technology for developing custom-fit RP orthoses has the potential to provide increased freedom with geometric features, cost efficiencies and improved practice service capacities while maintaining very high quality-of-service standards. In the long run, this technology aims at bringing the manufacturing of orthoses from the current manual labor/expert craftsman's skills to a 21st century computerized design process. The proposed technology has the potential for increasing the numbers of patients serviced per year per orthotist while reducing overall the orthosis fabrication cost and time.
- Chelule K, Coole T, Chesire D: Fabrication of medical models from scan data via rapid prototyping techniques. In Proceedings of the 2000 Conference on Time Compression Technologies. Cardiff International Arena, UK; 2000.Google Scholar
- CC Kai, Meng CS, Ching LS, Teik LS, Aung SC: Facial prosthetic model fabrication using rapid prototyping tools. Integrated Manufacturing Systems 2000,11(1):42-53. 10.1108/09576060010303668View ArticleGoogle Scholar
- Hieu LC, Zlatov N, Sloten JV, Bohez E, Khanh L, Binh PH, Oris P, Toshev Y: Medical rapid prototyping: applications and methods. Assembly Automation 2005,25(4):284-292. 10.1108/01445150510626415View ArticleGoogle Scholar
- Zollikofer CPE, Ponce de Leon MS: Tools for rapid prototyping in the biosciences. IEEE Computer Graphics and Applications 1995,15(6):48-55. 10.1109/38.469515View ArticleGoogle Scholar
- Noorani R: Rapid Prototyping: Principles and Applications. John Wiley & Sons Inc. Hoboken;; 2006.Google Scholar
- Webb PA: A review of rapid prototyping (RP) techniques in the medical and biomedical sector. Journal of Medical Engineering & Technology 2000,24(4):149-153.View ArticleGoogle Scholar
- Sinn DP, Cillo JE, Miles BA: Stereolithography for craniofacial surgery. The Journal of Craniofacial Surgery 2006,17(5):869-875. 10.1097/01.scs.0000230618.95012.1dView ArticlePubMedGoogle Scholar
- Chua CK, Leong KF, KH Tan, Wiria FE, Cheah CM: Development of tissue scaffolds using selective laser sintering of polyvinyl alcohol/hydroxyapatite biocomposite for craniofacial and joint defects. Journal of Material Science: Materials in Medicine 2004, 15: 1113-1121. 10.1023/B:JMSM.0000046393.81449.a5Google Scholar
- Herbert N, Simpson D, Spence WD, W Ion: A preliminary investigation into the development of 3-D printing of prosthetic sockets. Journal of Rehabilitation Research & Development 2005, 42: 141-146.View ArticleGoogle Scholar
- Rogers B, Bosker GW, Faustini MF, Walden G, Neptune RR, Crawford RH: Variably Compliant Transtibial Prosthetic Socket Fabricated Using Solid Freeform 'a case study'. Journal of Prosthetics and Orthotics 2008,20(1):1-7. 10.1097/JPO.0b013e31815ea839View ArticleGoogle Scholar
- Kumar V, Bajcsy R, Harwin W, Harker P: Rapid design and prototyping of customized rehabilitation aids. Communications of the ACM 1996,39(2):55-61. 10.1145/230798.230804View ArticleGoogle Scholar
- Milusheva S, Tochev D, Stefanova Y, Y Toshev Y: Virtual models and prototype of individual ankle foot orthosis. In Proceedings of ISB XXth Congress - ASB 29th Annual Meeting. Cleveland, Ohio; 2005:227.Google Scholar
- Milusheva S, Tosheva E, Tochev D, Toshev Y: Personalized ankle foot orthosis with exchangeable elastic elements. Journal of Biomechanics 2007,40(S2):S592. 10.1016/S0021-9290(07)70580-8View ArticleGoogle Scholar
- Faustini MC, Neptune RR, Crawford RH, Stanhope SJ: Manufacture of passive dynamic ankle-foot orthoses using selective laser-sintering. IEEE Transactions of Biomedical Engineering 2008,55(2):784-790. 10.1109/TBME.2007.912638View ArticleGoogle Scholar
- Alimed Inc.:Type C-90 Superior Posterior Leaf Spring. 2009. [http://www.alimed.com]Google Scholar
- Technest Holdings, Inc.:3D Imaging Products. [http://www.genextech.com/pdfs/3DImaging/Genex_3DSolutions_web.pdf]
- 3D Systems Inc.:Products - SLA Systems - Viper Si2. 2009. [http://www.3dsystems.com/products/sla/viper/datasheet.asp]Google Scholar
- Rapidform Inc.:The Standard Software for 3D Scanning. 2009. [http://www.rapidform.com/]Google Scholar
- 3D Systems Inc.:Products - Accura 40 SLA Resin. 2009. [http://www.3dsystems.com/products/datafiles/accura/datasheets/DS-Accura_25_SL_material.pdf]Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.