The topographical description and quantification of an amputee's residual limb is a major problem in prosthetic design and manufacture.6 This procedure prolongs the rehabilitation process. The problem is partly due to the unavailability of a device which can rapidly and accurately provide a reproducible topographical description of a residual limb. Development of a low cost, reliable device to provide comparative data would facilitate future attempts to determine the natural development of the mature amputated residual limb. The device would also provide a data base for quantification of optimal limb/socket designs. Ultimately, it would be the basis for the initial imaging mechanism necessary for long-term, computer-aided design (CAD) and computer-aided manufacture (CAM) of prostheses.1,8,11
Fernie3 describes a "shape sensing" machine that uses a series of video cameras, and Klasson6 reports on similar work in Sweden. The major problem with this procedure is the inability of video cameras to see below the skin surface to image bone surfaces.
As part of our ongoing research project designed to devise a computerized system for prostheses manufacture, we decided to investigate computerized tomography as a possible imaging method. Computerized tomography (CT) scanning provides an excellent two-dimensional picture of human anatomy.5 Recently, work has been done in sev eral areas of the country concerning evaluation and surgical planning of human anatomy using computerized tomography scanning.4,7,12
The Rehabilitation Engineering Lab (REL) at The University of Texas Health Science Center at San Antonio recently acquired a physician's imaging computer console developed by Contour Medical Systems of Mountain View California (Figure 1) . This computer system, the CEMAX 1000, has the capability of reconstructing three-dimensional images of tissue and bone from an analysis of computed tomography data. Using this system, it is potentially possible to design a custom fitted prosthetic socket without touching the amputee.
A CT Scan was taken of an amputee's below-knee residual limb, starting approximately 5cm above the knee joint space and ending at the distal end of the residual limb. This information was stored on a magnetic tape acquired from a GE 9800 scanner, and delivered to the REL. The CEMAX 1000 computer system was used to read the tape, and then produce a three-dimensional image of the residual limb.
The CEMAX 1000 has software that allows the operator to view the residual limb in its entirety (Figure 2) , or one slice at a time (Figure 3) . This software allows the operator to view the residual limb from any angle (Figure 4) . It can be rotated, translated, scaled, and modified one slice at a time (Figure 5) .
Using standard software designed for the CEMAX 1000, the three dimensional image was reconfigured using normal prosthetic biomechanical considerations, one slice at a time (Figure 6) . This was done to relieve pressure on pressure sensitive areas, and to apply pressure to pressure tolerant areas of the residual limb (Figure 7) . This information was then recorded on a cassette tape that was delivered to Contour Medical Systems in California where a computer controlled milling machine (Figure 8) was used to carve positive model (Figure 9) and socket (Figure 10) replicas in wax. This was done as a "proof of concept" purely to assess the possibility of using the technique to produce functional sockets, not to produce functional sockets.
The currently preferred methods for obtaining a complete image of the residual limb (including the bone) are either computerized axial tomography (CAT) scans or magnetic resonance imaging (MRI). Excellent image quality can be obtained; however, it takes a long time to produce an image, involves significant movement artifact, and the cost is high. In addition, a CAT scan involves ionizing radiation, which should be avoided whenever possible. Most companies that have developed this technology hold their systems as proprietary, and will not allow program alterations. Finally, these systems are much more expensive, sophisticated, and complex than is required to acquire the necessary measurements of the residual limb. For these reasons, it was ultimately decided that the use of these techniques do not lend themselves to potential prosthetic CAD/CAM applications.
In an attempt to overcome these objections, the REL is currently investigating ultrasound as an imaging medium.2 We are now installing a second generation ultrasound shape sensing device which appears to be promising. This device will allow us to scan skin and bone surfaces.
Photos are by Baltazar Farias, Photo Tech III, Department of Radiology, University of Texas Health Science Center at San Antonio.
Virgil W. Faulkner, C.P.O., is Associate Professor and Director of the Rehabilitation Engineering Lab, Department of Physical Medicine and Rehabilitation, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7799.
Nicolas Walsh, M.D., is Associate Professor, Department of Physical Medicine and Rehabilitation, University of Texas Health Science Center at San Antonio.