The CAPOD System-A Scandinavian CAD/CAM System for Prosthetic
Sockets
Kurt Oberg, Ph.D.
Jonathan Kofman, M.Sc.A., P.Eng.
Arne Karisson
Billy Lindstrom
Goran Sigblad, C.P.O.
Introduction
Current applications of Computer-Aided-Design / Computer-Aided-Manufacturing (CAD/CAM) to amputee socket design and manufacturing aim to achieve
some of the following: better and more uniform quality, shorter production time, and automated shape and process recording. The processes using CAD/CAM may
include some of the following steps: determination of the residual limb shape by means of a measuring system, socket design based on these measurements using
some type of custom- made CAD computer program, and machining a positive mold for the socket. The prosthetic socket can then be made by conventional
methods including plastic vacuum forming.
At present, various methods exist of measuring the residual limb defining initial socket shape prior to CAD modifications. Simple measurements of the residual limb's
diameter and circumference using caliper and tape have been used to define a stylized primitive socket selected from a set of stored socket shapes.4 These can then be
scaled, tapered, and modified in relief and compression areas using a CAD program. In another method, a more complete yet time consuming recording of the residual
limb shape is obtained by first taking a plaster cast of the limb in the conventional manner and then digitizing points along the inner surface of the cast. This is followed
by a CAD modification of
an average rectification pattern based upon information obtained from several below-knee amputee residual limbs.
Non-contacting methods of determining residual limb shape are used to eliminate the unrepeatability associated with applying different pressures in taking caliper
measurements or in forming a plaster cast. These techniques make use of optical systems which employ several cameras and laser-light planes2 or silhouette images of
the limb as developed at UCL.5 Although using several cameras enables quick scanning of the complete limb, a high cost is implied. The silhouette device is compact
and can scan quickly; however, it is not able to detect concavities in a horizontal cross section.
This paper presents a CAD/CAM system which includes residual limb shape sensing, socket design, and positive mold machining for below-knee prostheses. In our
current system, we aimed at achieving full sensing of the residual limb in a relatively simple, low-cost, and compact device and have, therefore, employed a single camera
and single laser plane in a scanner. Our custom-written CAD program enables us to design the socket directly from the limb's shape according to the prosthetist's empirical
experience of socket-stump biomechanics as conventionally done with plaster cast and mold forming.7
The system is a joint Scandinavian effort in cooperation with several orthopedic cen
ters where clinical evaluation of the system will soon begin on a broad scale. Our intention is to eventually use the system for both prostheses and orthoses. Therefore,
the system is called The Computer-Aided-Prosthetic and Orthotic-Design (CAPOD) System. Our general design criteria have been described previously.6
The CAD/CAM System
The main components of the CA POD system are illustrated in Figure 1
. They include a laser-scanner for limb measurement; a CAD/CAM workstation to control
the scanning of the limb, perform socket design modifications, and control the socket-mold machining; and a custom-designed numerically-controlled (NC) milling
machine to cut the socket mold.
Stump Measuring Apparatus
Measurement of the residual limb is carried out using the scanner shown in Figure 2
.
It consists of a CCD video camera and a laser-light source mounted to a rotatable frame, and a cover to completely contain the frame except where the amputee's limb
is located in the center of the frame. A plane of laser light is directed toward the coplaner central axis of the rotatable frame and intersects with the limb to form a line
contour on the limb's surface. To view this intersection as a contour rather than as a straight line, the camera is positioned at a 45° angle to the plane of light. In order to
view the full length of the limb while maintaining a compact device, the camera is positioned to obtain a reflected image of the limb's contour via two plane mirrors. This
results in an optical axis with a virtual angle of 56° in respect to the rotation axis.
Position the patient in respect to the measuring apparatus by angling and displacing the device horizontally and by means of a special patient supporting stand as
shown in Figure 3
. This stand holds up the patient with either an adjustable seat or with a foot stand which supports the sound limb of a unilateral amputee and allows
the residual limb to remain free for placement in the scanner. The patient can be elevated or lowered simply by the stand's motorized drive.
Stump Preparation
Prior to measuring the limb, place a thin nylon stocking with a round piece of Velcro attached at the distal end over the residual limb as shown in Figure 4
. This
provides a smooth and reflective surface to increase the sharpness and intensity of the laser-light con-tour which would otherwise be formed on the skin. A hemispherical
marker, lOmm in diameter, is placed at the mid-patellar tendon for later use as a reference point during CAD socket design. Markers may be placed at other sights such
as bony prominences if desired. After the amputee is lowered in place by the motorized stand, a cup supported on a vertically adjustable rod to hold the stump in place
is raised. The Velcro® on the stocking fastens to the clothlike Velcro which is glued to the inside of the cup to aid in obtaining a moderate fixation. The amputee's thigh
can be held in place by an adjustable diaphragm in the top cover of the measuring apparatus. This gives sufficient fixation at the proximal end.
Scanning the Residual Limb
To measure the limb's shape, the frame is rotated containing the video camera and a laser light source around the limb under the control of an IBM AT compatible
computer with an Intel 8386 processor and a frame-grabber. The frame-grabber is used to obtain 100 images of the stump-shape dependent line contour at 3.6° intervals.
Raw planar coordinates of the leading edge of the contour image are stored for processing. For the below-knee residual limb, this edge is represented by approximately
350 points, while the system permits up to 512 points per contour or 51,200 points for complete shape representation. The planar coordinates are then filtered and
transformed into three-dimensional (3-D) position coordinates to obtain be complete shape of the limb.
CAD Program
The CAD program, run on the same computer, is then used to design a socket (Figure 5)
to correctly fit the residual limb and to incorporate the usual modifications
made to a plaster negative and positive. The socket can be modified by stretching or shrinking any partial socket length; by modifying the radius at any cross section
by absolute addition or subtraction; or by proportional scaling of the length and radius simultaneously. A rectangular region of the socket surface can also be raised or
compressed to form a relief or load-bearing area. These can be in the form of a small point or a long ridge (Figure 6)
. For every modification, the new socket volume
is calculated and compared to the volume of the limb. The user may view a 3-D wire-frame image of the limb or socket in separate images, as well as profiles or cross
sections of the limb and socket in the same image to examine the fit. In the sectional view, the circumference of both the limb and socket are displayed (Figure 7)
. Diameter
and length measurements may be made roughly using a moveable ruler on the screen (Figure 8)
, or more precisely between any two points marked by the cursor.
The socket shape may be saved or recalled at any stage in the design process. Since several socket shapes may be stored, a prosthetist can try various modification
approaches without having to repeat steps.
A typical CAD sequence is outlined as follows: First, the shape of the residual limb is copied to a socket shape, which will be subsequently modified. Then the image
of the reference marker is found in a sectional
view. The socket and residual limb are then rotated to view the market in a profile view. The cup at the distal end of the socket is cut off and a special function is used
to complete the end with a parabolic closure, which fits smoothly with the remaining socket end.
The socket can then be stretched and raised, for example, in the distal tibial region or at bony prominences to provide relief, or compressed to form a flare in the
popliteal area. These modifications are done while observing the limb-socket volume difference. The socket volume is finally adjusted as desired by radial increment
or decrement along the whole or partial length.
Mold Cutting
Once the final socket shape is determined, cylindrical coordinates describing its shape are sent to an NC milling machine to cut the positive mold. The mold blank
is currently made from equal parts by volume of plaster
and microballoons. The NC machine was specially designed and built for the CAPOD system as commercial machines generally had a higher specification than required
and were highly expensive. Machining is carried out with the mold plug stationary. The cutter itself is stepped radially, while its carriage is stepped longitudinally
corresponding to each line contour scanned, as shown in Figure 8
. Between each longitudinal cut, the blank is rotated 3.6°. This method is used as opposed to the
helical cutting with constant cutter travel speed to enable faster cutter movements over flat areas.3 A finished mold, shown in Figure 9
, can be produced in approximately
20 minutes.
Clinical Experience and Results
Preliminary clinical testing of the system has been carried out with 21 below-knee amputees who are elderly with amputation for vascular problems. While the trials
were a learning experience for both the prosthetists and the system designers, we generally had very good results.
Various types of stockings were tried before arriving at a suitable thin nylon stocking made by Teufel, which would not form excessive folds or creases, and which
would not deform soft tissues.
We have experienced some problems in finding an appropriate sized reference marker to place on the stocking. A marker that was too large would obstruct the
camera's view of the laser-contour at several points behind the marker, while too small a marker would be difficult to locate when using the CAD program.
We found that when placed in lower bony regions, the markers produced this problem, but when placed in the mid-patellar region, the problem was almost entirely
eliminated. This was probably due to the marker appearing smaller to the camera in the latter position. To enable placement and detection of the markers in any region,
the measuring program was slightly modified to accept slightly longer strings of hidden points.
For complete limb-shape measurement, the limb must intersect the rotation axis with no discontinuities along its entire length and throughout the entire duration of
the camera-laser rotation. Patients with flexion contractures at the knee presented some difficulty in aligning the scanner rotation axis with the limb. It was necessary to adjust the entire scanner orientation with respect to the amputee by raising one side.
The scanning time required for measurement of the limb is 10 seconds. The complete measurement period, including preparation of the residual limb with a stocking and marker and positioning the amputee in the scanner, requires
approximately 25 minutes for the most difficult case, and as short as 5 minutes for the least difficult case.
Concerning the CAD program, prosthe-tists generally found the system quite easy to learn and use. They were able to work comfortably and independently with the system after using it with only two to four patients. The complete
CAD design takes 15 to 25 minutes, while a second CAD session to perform modifications, can take two to 10 minutes.
The cutting of the socket molds has presented no problems. Although there is a slight roughness in the machined surface, there have been no adverse patient reactions directly related to it, and the socket can therefore be formed
directly on the mold.
In the ongoing preliminary clinical trials, complete fitting information has been obtained for 15 of the 21 below-knee amputees. Ten subjects were fitted satisfactorily with sockets made from CAD/CAM molds. Eight of these subjects
are currently using their
new prostheses, while the other two, although satisfied, did not wear new prostheses as they had only participated in the trial for testing purposes and did not require a new prosthesis. Of the ten satisfactory prostheses, eight were
satisfactory after one CAD session, while two required modifications in a second CAD session. Of the 15 subjects-fittings, three were not satisfactory after a first CAD session and a modification session has not taken place. No
conclusion can be made whether a second CAD session to perform modifications to these would result in a satisfactory fitting, although we have no reason to believe that they wouldn't. For one subject, the fitting was not satisfactory
and a re-scan was suggested. For one last case, the fitting was originally described by the prosthetist and subject as very good. However, after some use of the prosthesis, the patient decided that some modification was necessary and
discontinued wearing it.
Discussion
After preliminary clinical testing, we are generally quite satisfied with the performance of the CAPOD system. Furthermore, the quality of the fittings has been good. Considering that the prosthetists in the trial began with no
experience in using the system, and that some only fit between one and three amputees, the overall present status is favorable.
For all but one of the subjects who had an unsatisfactory prosthesis, there remains the possibility that a CAD modification session would result in a good prosthesis.
The performance of the laser scanner has also been acceptable. Some changes will be made to facilitate adjustment of the complete scanner orientation for amputees
with flexion contractures. As well, various types of markers will be tested to facilitate and improve the measurement procedure.
After using the system, the prosthetists had quite a positive attitude toward using CAD/CAM. They found the program easy to learn and use, and enjoyed being able
to work directly from the shape of the limb in a similar manner as they had previously done with plaster casts. However, they were required to adapt to the computer
system by thinking in terms of numbers when palpating the residual limb.
With the present system, the prosthetist can fully control the changes made to the socket beginning from the limb shape. We are considering including decision-making
aids in the form of suggested socket modifications for the purpose of improving fit quality. Also, we intend to include some automatic routines that will be used solely
for time-saving purposes, such as locating the markers and rotating the limb to the appropriate start-up profile.
Other improvements in the CAD program will be made by upgrading computer hardware and software to include 3-D solid modeling. This feature allows shading
of solid surfaces to offer the prosthetist a more realistic view of the socket and residual limb. Upgrading is also expected to reduce CAD session time.
Concerning the cutting of the socket mold, modifications to replace the current stepper motors in the milling machine with servo motors is expected to reduce the
cutting time to 10 minutes.
Finally, future research aims at adaptation and testing of the system for higher level amputees.
Summary
The application of CAD/CAM to amputee socket design and manufacturing aims at achieving better and more uniform quality,
shorter production time, and automated shape and process recording. This paper presents a low-cost CAD/CAM system to design and manufacture below-knee prosthetic
socket molds. The system consists of the residual limb measuring apparatus, a milling machine to cut the socket mold, and a computer used to control these devices and
make socket-design modifications. Preliminary clinical evaluation has been carried out with 21 below-knee amputees with very good results.
Acknowledgments
This work was carried out with the gratefully acknowledged financial support and cooperation in clinical testing of the councils of Jönköping, ölvsborgs, and örebro
counties, and Gothenburg City, as well as Sahva BS, Copenhagen, Denmark and the Prosthetic Foundation, Helsinki, Finland.
Kurt öberg, Ph.D., is department director
Jonathan Kofman, M.Sc.A., P.Eng., is research engineer
Arne Karlsson is computer-system engineering technician
Billy Lindström is design engineer at the Department
of Biomechanics, College of Health and Care (Hälsohögsko Ian), Box 1038, S-551 11 Jönköping, Sweden, tel. (46)-(36)- 16-50-10.
Göran Sigblad, C.P.O., is director of the Department of Prosthetics and Orthotics, Västra Klinikerna, Jönköping County Hospital, S-552 85, Jönköping, Sweden.
References:
- Dewar, M., P. Jarman, D. Reynolds, H.
Crawford, J. MacCoughlan and A. Crew, "Computer Aided Socket Design (CASD):UCL System Based on Full Shape Sensing," Bioengineermg Center, University
College London, Report 1985, 1985, pp.19-30.
- Fernie, G.R., G. Griggs, S. Bartlett and K. Lunau, "Shape Sensing for Computer Aided Below-Knee Prosthetic Socket Design," Prosthet. Orthot. Int., 9, 1985,
pp.12-16.
- Lawrence, R.B., W. Knox and H.V. Crawford, "Prosthetic Shape Replication Using A Computer Controlled Carving Technique," Prosthet. Orthot. Int., 9,
1985, pp.23-26.
- Saunders, C.G., J. Foort, M. Bannon, D. Lean and L. Panych, "Computer Aided Design of Prosthetic Sockets for Below-Knee Amputees," Prosthet. Orthot.
Int., 9, 1985, pp.17-22.
- Smith, D.M., A. Crew and A. Hankin, "Silhouette Shape Sensor," Bioengineering Center, University College London, Report 1986, 1986, pp.41-42.
- öberg, K., "Swedish Attempts in Using CAD/CAM Principles for Prosthetics and Orthotics," Clinical Prosthetics and Orthotics 9:2, 1985, pp.19-23.
- öberg, K. and A. Karisson, "The experience of CAD/CAM in Scandinavia." X Congresso International de la Union Internacional de Teenicos Ortopedicos,
Barcelona, Italy, June, 1987.
|
|
