Link to Figures Show Figures
	Article Text Figures References
	Send Link Send HTML Article

Double-Wall, Transtibial Prosthetic Socket Fabricated Using Selective Laser Sintering: A Case Study

Bill Rogers, MS
Sean Stephens, MSE
Andrew Gitter, MD
Gordon Bosker, CP, Cped
Richard Crawford, PhD

ABSTRACT

The primary goal of this study was to test the feasibility of using selective laser sintering (SLS) to fabricate a functional transtibial prosthetic socket. Prosthetic socket computer-assisted design and computer-assisted manufacturing techniques were combined with SLS technology to produce a sophisticated, monolithic, transtibial prosthetic socket. The socket combined a rigid outer shell with a variably compliant inner shell and incorporated a fitting for a pylon directly into it. The socket was manufactured for a 65-year-old transtibial amputee using a socket shape identical to his current definitive socket. A prosthesis was then assembled using the same foot as the subject's definitive prosthesis. A comparison of socket performance suggested improved comfort, greater step symmetry, and similar lower extremity joint function.

Key Words: prosthetic socket, amputees, computer-assisted design, solid freeform fabrication

Socket design is the most important aspect of a lower extremity prosthesis. The socket is the interface between the human and the mechanical support system, and its design and fit are what ultimately determine patient acceptance, comfort, suspension, and energy expenditure.¹ Achieving and maintaining a comfortable and tissue-tolerant socket fit remains a clinical problem. As many as 20 to 55% of lower limb amputees report subjective dissatisfaction with socket comfort, residual limb pain, or skin breakdown.^3,4,5 Several approaches to improving the process of fitting these patients have been introduced over the past decade, including the development of silicon gel liners and semiflexible socket designs. A major clinical drawback to the use of flexible-wall socket designs has been the increased cost and complexity of their fabrication, which typically requires a two-part socket design consisting of an outer, rigid structural frame with a flexible, inner-wall socket.

Computer-assisted design (CAD) and computer-assisted manufacturing (CAM) techniques are beginning to be widely used in the design and manufacture of sockets. The residual limb shape is input into the computer with a mechanical digitizer, magnetic digitizer, or noncontact laser scanner.⁹ The prosthetist then uses specialized software to produce a biomechanically correct socket from the limb shape. A computer-controlled milling machine carves the pattern for the socket from plaster or foam, and the socket is made using conventional methods such as vacuum molding or lamination.

Currently available prosthetic CAD systems are limited by the fact that they only design the shape of the inner wall of the socket, which is all that is necessary for conventional fabrication. These systems do not actually design the three-dimensional shape of the socket including the wall thickness, and there are no provisions for incorporating the pylon attachment into the socket.

Solid FreeForm Fabrication (SFF) technologies are a class of technology that can create three-dimensional objects directly from a geometric database without specific tooling or human intervention. Objects are generally built in layers that are fused by a laser or perhaps extruded. Used in conjunction with CAD, advances in SFF technology offer a straightforward pathway to developing flexible-wall socket designs. SFF, which uses a variety of material deposition technologies to create a three-dimensional physical object directly from shape information, has been explored as a tool for eliminating the many time-consuming, manual labor steps used in conventional socket fabrication. Some initial work has been done using SFF methods for socket fabrication.

Stereolithography (SLA), which has been used to make check sockets, is a layer-based manufacturing technique that solidifies liquid photopolymer with an ultraviolet laser. Because the laser does not completely cure the polymer, it is necessary to postcure the part in an ultraviolet oven. Baxter Healthcare and the Northwestern University Prosthetics Research Laboratory produced the first socket of this type. Freeman and Wontorcik² determined that SLA can fabricate test sockets suitable for ambulation; however, because of the brittle nature of available photopolymers, these test sockets are not suitable as definitive sockets.

SquirtShape, which was developed at the Northwestern University Prosthetics Research Laboratory, is an additive method that extrudes a continuous bead of molten plastic to form a socket,⁸ fabricating it in a continuous spiral with a uniform wall thickness. Although useful, this technique can only produce single-layer parts with a uniform wall thickness.

In selective laser sintering (SLS), components are built by material addition rather than by material removal, using a directed laser beam that causes individual particles to adhere in selected regions of space. First, a thin layer of powder is deposited into a container, after which the powder surface is raster-scanned with a laser beam much like the picture is formed on a television screen by a set of horizontal lines. The intensity of the beam is modulated to sinter the powder in areas to be occupied by the socket at that particular cross section. In areas that are not sintered, the powder remains loose and is removed once the part is complete. Successive layers of powder are then deposited and sintered until completion of the entire part.

Preliminary work with SLS socket fabrication involved making scaled-down polycarbonate sockets incorporating an integrated pylon fitting.⁶ Subsequently a full-size, nylon socket was fabricated for an amputee and ambulated upon. In this case, the attachment for the pylon was part of the CAD process and was directly incorporated into the socket.⁷ The socket produced closely resembled a conventional, laminated part in shape and function and served to demonstrate the feasibility of producing a full-size, structurally sound socket. There was no clinical assessment of the socket.

In all previous cases in which SFF has been applied to socket fabrication, there have only been attempts to make sockets equivalent to existing ones. A key part of this work is the monolithic, double-wall design that has not been attempted previously. The rigid outer wall provides support for the pylon, and the variably compliant inner wall provides support for load-bearing areas and pressure relief for sensitive areas.

Methods

The purpose of this study was to further test the feasibility of SLS fabrication by developing a monolithic, flexible-wall socket design and incorporating the socket into a functional prosthesis and to quantitatively compare gait performance with a conventionally fabricated socket.

A 65-year-old, male, transtibial amputee was recruited for test purposes. The amputee had a healthy, mature, residual limb. The amputee subject chosen was representative of a typical elderly, dysvascular transtibial amputee and would be considered an unrestricted indoor and limited community-level ambulator. Informed consent was obtained following local institutional review board protocols.

There are several novel features in the design of the double-wall socket. The inner wall of the socket has a variable wall thickness, with additional thickness added to load-bearing areas. There are cantilever-compliant regions over the fibula head and distal tibia, and there are support structures connecting the inner and outer walls in critical support regions. Each contilever-compliant area consists of a set of triangular cantilever beams composed of radial slots in the inner socket wall. A cantilever beam is supported only on one end, making the beam act as a spring. Finally, there is a fitting for a pylon mount at the distal end of the socket.

This project was split between the Department of Rehabilitation Medicine at the University of Texas Health Science Center at San Antonio (UTHSCSA) and the Department of Mechanical Engineering at the University of Texas at Austin (UT). The UTHSCSA was responsible for the data acquisition and clinical aspects of the project. The UT provided the double-wall socket design, data conversions, and SLS socket fabrication.

A definitive PTS socket was designed for the patient using an Alpha liner (Ohio Willow Wood, Mount Sterling, OH) and standard fabrication techniques. The residual limb of the amputee was cast at the UTHSCSA using a ShapeMate casting sock (Seattle Limb Systems, Poulsbo, WA) over the Alpha liner. The sock hardened 15 minutes after being put in water, then a laser imager (Seattle Limb Systems) was used to measure the outside shape of the ShapeMate sock. The sock had a uniform thickness; by measuring the outside of the sock then subtracting the sock thickness, the correct residual limb shape was calculated. The socket was designed using ShapeMaker prosthetic CAD software (Seattle Limb Systems), and supercondular suspension was incorporated. A socket pattern was then milled out of urethane foam, and the conventional socket was fabricated using carbon fiber. A pylon and a Seattle foot were added to complete the prosthesis. Final alignment of the prosthesis was performed using visual gait analysis and patient feedback.

The urethane foam pattern used to fabricate the definitive socket was measured with the laser imager and served as the template for SLS socket fabrication. Trim lines, load-bearing areas, and pressure-sensitive areas were marked in black on the pattern, with the laser imager capturing these marks. The data (including the marked locations) were saved to an American Academy of Orthotists and Prosthetists (AAOP) data interchange file, which is a nonproprietary file format that is supported by prosthetic CAD/CAM vendors. The AAOP file was e-mailed to UT, where quite a number of steps were required to create the final, three-dimensional socket design from the input data.

A dual-wall socket design was chosen to implement the flexible-wall concept. A rigid outer socket wall provided structural integrity and allowed easy incorporation of an integrated, pylon-mount fitting. The inner wall of the socket was a thin, flexible wall that allowed for regional compliance variability. Regions of maximal flexibility included the distal anterior tibia and the fibular head.

The AAOP data are a single surface representing the inside of the socket. To create a double-walled structure, the inner surface data was offset three times, as shown in Figure 1 . A program was written in C to accept data input in the AAOP format and to create the double-walled socket using a perpendicular offset algorithm. The nominal thickness of the outer wall was 2.921 mm (.115 inches), and the nominal thickness of the inner wall was 1.016 mm (.040 inches), with the nominal separation being 6.508 mm (.256 inches). The program increased the thickness of the inner wall to 3.024 mm (.119 inches) in the support regions, shown in Figure 2 as gray patches. The last function of the program was to export the data in a file format to run in Octave, a free mathematics software package intended for numerical computations that is widely available on the Internet. Octave was used to create a set of fourth-order, b-spline curves to represent the double-wall surface. The b-spline approximation allowed for a large reduction in the amount of data later processes must work through, thus significantly reducing computation time. The b-spline curves were exported to a Scheme format.

Scheme, a computer language very similar to LISP, was used to interact with the ACIS 3D Toolkit (Spatial Technology, Boulder, CO), which is a solid modeling software package that uses the b-spline socket representation to create a geometric solid model. In this case, ACIS was used to add the rest of the features of the socket. These features were integrated using Boolean operations, meaning that three-dimensional solid shapes were added or subtracted from the solid representing the socket design.

Each cantilever-compliant area consisted of a hexagonal set of triangular cantilever beams composed of radial slots in the inner socket wall (Fig. 3 ). These beams were free to deflect over a certain range. The deflection of the end of each beam was designed to be 2 mm at a pressure of 21.8 psi, which is the maximum fibula head pressure during gait in a standard socket as determined by finite element analysis at Northwestern University. Additionally, there were three stiffened regions, including the anterolateral, medial, and popliteal areas. These areas had a wall thickness of about 3 mm. There were additional support struts in these areas that connected the inner and outer walls (Fig. 2 ).

The final design feature was a distal fitting for a nylon pylon adapter. This adapter was a prototype design from Seattle Limb Systems, and the distal fitting was a 1-inch deep cylinder with a one-quarter-inch wall thickness. To accommodate high stresses at the pylon to socket interface, the outer wall of the socket was thickened to allow a smooth transition to the cylindrical fitting (Fig. 4 ). A design compromise was made in deciding not to shape the top of the socket to the socket trim lines. This was done partially to simplify the design process and partially to ensure a proper fit. The trim lines on the model were made based on experience with other materials and other designs. To avoid the possibility of too much material being removed in the design stage, the decision was made to shape the top of the socket manually.

The final step in the design process was to create an STL data file from the ACIS solid model. This file format is the standard for the SFF industry. Unfortunately, there is no export function in ACIS to produce an STL file, so several steps were necessary to achieve STL output. The first step in producing a file was to convert the socket model into a surface composed of triangular facets. There is an ACIS function that accomplishes this, along with a function that writes facets to a file. Next, the UNIX scripting language called AWK (which is widely available on the Internet) was used to write a program converting the facet file to an STL file. Although the resulting file had some holes and ill-defined edges that kept if from being suitable for manufacturing, a program commonly used in the SFF industry called Magics RP (Materialize USA, Ann Arbor, MI) was able to fix all of the problems.

Initially, it was decided to manufacture the part using an early SLS machine at UT. A number of difficulties were encountered with the machine, which made it unsuitable for making the final socket. The DTM Corporation (Austin, TX) very kindly donated both SinterStation 2000 machine time and Duraform polyamide material (a type of nylon) to make the socket. To allow for knee flexion, the socket was trimmed, then the three-piece nylon pylon adapter was pressed into the distal fitting on the socket. The press fit was tight enough that no additional securing was necessary. The pylon was inserted into the fitting and held in place with a setscrew. By moving two oblique, concentric cylinders, the coronal or sagittal angle could be set. The position of the cylinders was also secured with setscrews.

The subject was fitted with the SLS socket prosthesis and aligned using visual gait observation and patient feedback. Quantitative comparison of socket performance used visual analog scale reports of socket comfort and gait ease, along with quantitative biomechanical gait analysis. Reflective markers were placed on the sacrum, bilateral anterior superior iliac spine, knee, ankle, and second metatarsal phalangeal heads (Fig. 5 ). Ground reaction forces were obtained using AMTI force plates (Watertown, MA) embedded into the gait lab floor. Combined-joint kinematic and kinetic data were collected at a sampling rate of 60 cps for three representative strides as the subject walked at his self-selected speed. Using VICON Clinical Manager software (Tustin, CA), basic temporal and spatial features of the gait patterns and lower extremity joint angles were determined. Using inverse dynamics, net joint moment (torque) was calculated. For the prosthetic limb hip and knee joints, this represented the net muscular-generated joint moment developed in response to external forces and loads. For the prosthetic ankle, the joint moment was the resultant internal moment developed at the foot/pylon junction that mirrors the location of the intact limb's ankle joint. The initial test session used the conventional socket prosthesis. After this session, the subject donned the SLS prosthetic limb and was allowed a 10-minute acclimatization period before repeating gait analysis. Immediately after each test session, the subject rated socket comfort and overall ease of ambulation using a 0 to 100 point visual analog scale (100 = greater comfort or ease of ambulation).

Results

Table 1 summarizes the basic temporal, kinematic, and subjective assessments of socket performance. The visual analog scale score (0 to 100) for socket comfort was 10% higher for the SLS socket than for the conventional prosthesis, although overall ease of walking was similar. Walking with the SLS socket resulted in a 3% increase in self-selected speed and improved step-length symmetry between the prosthetic and intact limb. Stance-phase duration and double support were similar, although slightly greater when using the SLS socket design. No statistical analysis was performed.

Quantitative gait analysis of the lower extremities is shown in Figure 6 , Figure 7 , and Figure 8 . Similar patterns of joint motion and resultant muscular moments in the intact limb were present at the ankle, knee, and hip joints. The SLS prosthetic limb was aligned with approximately 3 degrees of ankle dorsiflexion more than was present in the conventional limb (Fig. 8). This alignment was determined by the consensus of the prosthetist in conjunction with patient feedback. The effect of the additional foot dorsiflexion was also reflected by a similar increase in knee flexion during the stance phase (Fig. 7). Net joint moments were similar at the prosthetic ankle and hip. In contrast, at the prosthetic limb knee, a prolonged duration and larger magnitude knee extensor moment was present throughout most of the stance phase with the SLS prosthesis as compared with the conventional prosthesis.

Discussion

The primary goal of this study--to test the feasibility of using SLS to fabricate a functional transtibial prosthetic socket--was achieved. This is the first step toward understanding the role of freeform fabrication and variable compliance sockets in clinical prosthetics practice.

Empirically, the SLS socket design was judged to be acceptable for clinical testing. It weighed 450 grams more than the conventional, laminated, carbon-fiber socket, reflecting the complexity of the design, the heavier nylon material used for fabrication, and an empiric design decision to add material at regions of anticipated stress. However, the subject was unable to detect the part's heavier weight when wearing the limb during clinical testing. The maximal medial-lateral width of the proximal socket was increased by 1.9 cm in the SLS design, resulting in a less cosmetic-appearing socket. The proximal SLS socket brim was stiffer than that of the conventional, laminated socket, which made donning the PTS-style socket more difficult for the subject. Once correctly donned, though, there did not seem to be any significant differences in suspension.

Although the use of a single subject prevents meaningful statistical analysis and limits the interpretation of quantitative gait and function comparison, the results of the subject performance are supportive of a useful role for SLS variable-wall complaint sockets. Small improvements were noted in walking speed and step-length symmetry, both indicators that have previously been shown to suggest improved prosthetic function. As would be expected, quantitative analysis of joint kinematics and moments did not show remarkable differences in joint motion or moment generation, with the probable exception of an increased knee extensor moment when using the SLS prosthesis. This moment was likely related to the alignment differences rather than to a direct socket effect. Using standard visual alignment techniques, the SLS prosthesis was set into 3 to 5 degrees more foot dorsiflexion than the conventional prosthesis. Associated with this alignment difference was the expected increase in generation of knee extensor muscle moments during stance. Although of uncertain clinical value, it is interesting to note that subjective socket comfort was improved, despite the increase in knee extensor moment. This improvement may reflect the enhanced ability of the SLS socket to accommodate for the corresponding higher, distal tibial forces by the cantilever compliance zone incorporated into the socket at this location.

Several areas of design improvement and socket fabrication need to be explored further before this technique could be tested in larger scale clinical studies. First, long-term durability and structural integrity need to be demonstrated. The SLS socket (as designed) included extra material thickness in the proximal brim and distal socket to ensure adequate material strength for clinical testing. Empirically, the socket appeared to have good strength; there was little concern about its use in a supervised setting. Destructive testing of the socket is planned, with a special emphasis on the pylon attachment area. One advantage of SLS over competing freeform fabrication techniques is the greater range of materials available. The prototype socket was fabricated using Duraform. However, since the initial fabrication of this socket, additional materials have become available for use in the part's manufacture. One new material is glass-filled Duraform, which has improved strength characteristics.

A second major concern with freeform fabrication is the cost of socket manufacture. Exclusive of casting and socket modification labor, the approximate cost of manufacturing the SLS socket was $1900. If two were made at the same time, then the cost would be about $1100 each. If sockets could be nested during manufacturing, then the cost would be lower yet. Direct comparison to a conventional socket is difficult when compared with the lightweight, carbon-fiber socket used in this study. The SLS socket is best compared with one that has a soft liner combined with a rigid outer socket, with fenestrations over the fibula head and distal tibia. The cost of a transtibial prosthetic limb of this type is in the range of $2000 to $5000, although it is difficult to say how much of that cost is attributed to socket fabrication. The large number of manual fabrication steps that are eliminated by using SLS has to be balanced against the cost of socket manufacture, which is related to the amount of material sintered, the desired accuracy, and the number of layers of material required to make the socket. Additional socket complexities, such as built-in springs, double walls, and integrated fittings, result in essentially no additional manufacturing cost. This additional level of complexity achieved for little additional cost may allow the fabrication of sockets that are more comfortable and biomechanically improved. Additionally, several sockets may be manufactured at the same time, which could lower cost.

Third, the cost of SLS-socket computer design requires a high level of engineering expertise. Currently cost and time prohibitive, most of the steps in the socket design can be automated with appropriate software development. To be clinically accepted, these design tools should be no more difficult to use than existing software such as TracerCad (Tracer Corporation, Miami, FL) or ShapeMaker. Existing prosthetic CAD packages only design the shape of the inner surface of the socket, whereas software used with SLS needs to design the entire three-dimensional volume shape of the socket. This need poses a much more difficult software development problem.

Ultimately, further clinical testing is needed. Optimally, the subject would have been given a longer period of acclimatization to realistically assess subjective comfort and acceptance. In the absence of data on socket structural integrity with prolonged use, it was deemed that long-term, unsupervised use of the SLS socket exposed the subject to unnecessary risk.

Conclusion

The fabrication of the double-wall socket demonstrated the capability of SLS to manufacture high-quality parts. Furthermore, the socket proved to be clinically acceptable. If there are suitable benefits from the sophisticated socket design, the cost of manufacturing the socket may also prove to be acceptable. Still lacking are software tools that will allow the prosthetist to design sockets in a short amount of time with minimal training. The next step in this process will be a long-term clinical evaluation of the patients involving a number of research subjects. If this evaluation is positive, then an effort will be undertaken to develop design software for clinical use.

References:

1. Foort J, Spiers R, Bannon M. Experimental fittings of sockets for below-knee amputees using computer-aided design and manufacturing techniques. Prosthet Orthot Int. 1985;9:46-47.
2. Freeman D, Wontorik L. Stereolithography and prosthetic test socket manufacture: A cost/benefit analysis. J Prosthet Orthot. 1998;10:17-20.
3. Hoagland FT, Jergesen HE, Wilson CP, et al. Evaluation of problems and needs of veteran lower-limb amputees in the San Francisco Bay Area. J Rehab Res Dev. 1983;20:57-71.
4. Houston VL, Burgess EM, Childress DS, et al. Automated fabrication of mobility aids (AFMA): Below-knee CAD/CAM testing and evaluation program results. J Rehab Res Dev. 1992;29:78-124.
5. Oberg T, Johansson LM, Karsznia T. Clinical evaluation of transtibial prosthesis sockets: A comparison between CAD/CAM and conventionally produced sockets. Prosthet Orthot Int. 1993;17:164-171.
6. Rogers WE, Crawford RH, Beaman JJ, Walsh NE. Fabrication of Prosthetic Sockets by Selective Laser Sintering. In Proceedings of the Solid Freeform Fabrication Symposium, Austin, TX, August 12-14, 1991.
7. Rogers WE, Crawford RH, Faulkner VF, Beaman JJ. Fabrication of an Integrated Prosthetic Socket Using Solid Freeform Fabrication. In Proceedings of the Seventh World Congress of the International Society for Prosthetics and Orthotics, Chicago, June 28-July 3, 1992.
8. Rovick JS. An Additive Fabrication Technique for the Computer-Aided Manufacturing of Sockets. In Proceedings of the Seventh World Congress of the International Society for Prosthetics and Orthotics, Chicago, June 28-July 3, 1992.
9. Walsh NE, Lancaster JL, Faulkner VW, Rogers WE. A Computerized System to Manufacture Prostheses for Amputees in Developing Countries. J Prosthet Orthot. 1989;1:165-181.