Published Quarterly by The American Society of Mechanical Engineers November 1993
`
`.•·.
`.--1 t..
`
`~~-"'?
`
`•.
`
`;
`~:-
`/
`'~,..
`
`'
`· .~
`_; .;; . •fin
`· "'l.y~
`
`~ Special Issue:
`
`' I
`
`I
`
`' 1 '
`
`' 1 '- ' )
`
`l ) I
`
`• ant
`• ng
`
`20th Anniversary
`Biomechanics Symposium
`
`i
`1 J~~~~1~~!!~m1
`l Jl~li~UI~l
`t rlllllllllllllllllllllllllllll
`t ISSUE CODE 9304
`
`Mako Exhibit 1007 Page 1
`
`

`

`E. Y. S. Chao
`Orthopaedic Biomechanics Laboratory,
`Johns Hopkins University,
`Baltimore, Maryland 21205-2196
`
`J. D. Lynch
`
`M. J. Vanderploeg
`
`Department of Mechanical Engineerig,
`Iowa State University,
`Ames, Iowa 50010
`
`Simulation and Animation of
`Musculoskeletal Joint System
`
`This paper describes the development of computer-based software for three-dimen(cid:173)
`sional geometric data base of the human musculoskeletal system. Using a computer
`graphics workstation, a user of the software will interactively display detailed in(cid:173)
`formation about the muscles, tendons, ligaments, bone, and joint anatomy. This
`software will enable a wide range of health care workers to viualize complex phys(cid:173)
`iological data. In addition to geometric and visual realism, this software will include
`kinematic relationships which allow the calculation and display of the motion and
`forces of the joints, muscles, and tendons. This will permit a user to interactively
`move joints or tendons and display the resulting motion of the surrounding tissues,
`as well as internal reactive forces and joint pressure distribution. A two-dimensional
`version of this software is currently being used for knee and hip osteotomy pre(cid:173)
`operative planning, total joint replacement prosthesis design and dimensional se(cid:173)
`lection, and osteochondral allograft sizing and reconstruction using radiographic
`data.
`
`Introduction
`Simulation and animation of musculoskeletal system is a
`relatively young field in biomechanics although joint mechan(cid:173)
`ics and multi-body dynamic simulation have had a long history
`of development. The advantage of high power work stations
`with advanced graphic capability greatly facilitated the com(cid:173)
`putational speed of the dynamic problem and at the same time,
`allowed visualization of the system performance using pre(cid:173)
`stored anatomical database and graphic display models. This
`new development represents a long and systematic evolution
`of progress in musculoskeletal joint mechanics. As we are
`commemorating the twentieth anniversary of biomedical en(cid:173)
`gineering of ASME, this paper not only explores a new horizon
`of biomechanics but also bring together the knowledge we have
`accumulated during the last two decades in this exciting field.
`The multi-body simulation programs, ADAMS [56] and
`DADS [57] were developed in the mid 70's which provided the
`ability to formulate and solve the equations of motion of
`complex dynamic systems. The physical properties of con(cid:173)
`necting bodies and joints were modeled while numerical so(cid:173)
`lutions were obtained after long computation in large digital
`computer main frames. Without the aid of computer graphics
`software permitting 3-D system visualization, it was difficult
`to interpret dynamics solutions and their implication to the
`system performance. Although computer graphics has been
`incorporated into these simulation programs, no provision was
`given to allow interactive interface in the software that enables
`system set-up, alteration of system configuration and display
`of dynamic results during solution process.
`Chao and Rim [13] was first to introduce the concept of
`
`Contributed by the Bioengineering Division of THE AMERICAN SOCIETY OF
`MECHANICAL ENGINEERS and presented at the 1993 ASME/ A!ChE/ ASCE Sum(cid:173)
`mer Bioengineering Conference, Forum on the 20th Anniversary of ASME
`Biomechanics Symposium, Breckenridge, CO, June 25-29, 1993. Revised man(cid:173)
`uscript received by the Bioengineering Division August 10, 1993.
`
`Inverse Dynamics Problems in biomechanical systems simu(cid:173)
`lation. Statically indeterminate problem was later formulated
`for the hand and elbow [12, 5]. Serig and Arvikar [58] worked
`on a general package to calculate forces in lower extremity
`joints during gait. These were some of the early simulation
`studies involving human musculoskeletal joint systems. Still
`images of line drawings were the only graphic display capa(cid:173)
`bilities.
`Buford et a!. [59] developed a kinematic model of the hand
`that utilized interactive 3-D line drawings to facilitate tendon
`placement in finger digit control. The users of the software
`were allowed to interactively control the joint angle and viewing
`the spatial position of the hand. Delp et a!. [20, 21] created
`an interactive computer graphic package to analyze the dy(cid:173)
`namics of the lower extremity. The software enabled modeling
`of different musculoskeletal joint systems using 3-D shaded
`computer graphic display and allowed system parameters to
`be altered using graphic interface. No software package permits
`a total interactive environment to alter system parameters,
`model configuration, external forces for both kinematic and
`dynamic analyses while able to display analysis results graph(cid:173)
`ically.
`The present musculoskeletal system simulation/visualization
`software is totally interactive and its graphic presentations are
`entirely different from the commonly referred to as medical
`imaging, i.e., to post-process medical measurements from a
`particular patient for use in diagnosis and treatment of that
`patient [40, 41, 45, 46, 50]. Rather, this simulation/visualiz(cid:173)
`ation package deals with the creation of very detailed generic
`models which will facilitate the understanding of complicated
`information through graphic interrogation of the anatomic
`system [1, 2, 9, 10, 23, 24, 29, 31, 37, 39, 42, 43, 44, 47, 53].
`The technology presented here goes beyond medical illus(cid:173)
`tration because images can be manipulated interactively [60].
`
`562/ Vol. 115, NOVEMBER 1993
`
`Transactions of the ASME
`
`Mako Exhibit 1007 Page 2
`
`

`

`Viewpoints can be changed, anatomical parts can be removed
`or made transparent and anatomical function can be simulated.
`Image storage and display .using the voxel based techniques
`are cumbersome and require a huge amount of computer core
`storage. In contrast, the present system uses geometric mod(cid:173)
`eling technique, generally known as descriptive geometry, a
`technique developed primarily by the CAD/CAM research
`community [31, 39, 42].
`The use of CAD/CAM-based descriptive geometry is an
`innovative approach for the modeling of biological systems.
`This brings to bear advanced mathematical modeling, includ(cid:173)
`ing sculptured surfaces, solid modeling, and computational
`geometry to model the systems of the human body. These
`models have several significant advantages over data typically
`obtained from CT or MRI scans because they require much
`less data to model a complex object, such as a tendon or bone,
`than does scan data. As an example, huge grid points resulting
`from CT scan can be indicated by a mathematical function
`defining surface that fits closely through the contour indicated
`by the data. The advantage of the surface is that its description
`requires much less stored information than was used for the
`data points from aCT scan. Furthermore, the surface provides
`a continuous model so texture and color can be added to create
`more realistic images not possible with CT or MRI scans. In
`addition, since the surfaces of models depend upon a relatively
`small amount of data, it becomes feasible to display images
`in real time [20, 21, 27, 36, 38, 49, 51]. However, surface curve
`fits introduce error into the resulting image while further re(cid:173)
`finement patch density would increase computational time
`greatly. Modern data compression schemes can significantly
`reduce the storage space of bit-mapped images obtained from
`CT /MRI directly. The real advantage of surface representation
`is the ability to visualize a 3-D object on a graphics terminal
`which is very important in the analysis and animation of mus(cid:173)
`culoskeletal systems.
`The objectives of this developmental effort are to:
`
`1. Collect anatomical data in a uniform data base using
`geometric models developed from the raw data.
`2. Provide a general human body motion and force analysis
`software program.
`3. Develop an interactive user-friendly graphics interface
`that enables efficient use and interpretation of human
`body function.
`
`The resulting software and data base will provide a powerful
`new tool for the medical research and education communities.
`Based on the current rate of increasing computer performance
`and reduced cost, this capability will become cost effective to
`a large segment of the health care industry over the next few
`years. This marks a significant advance since the early effort
`biomechanical modeling and analysis of human musculoske(cid:173)
`letal joint system using manual derivation, lengthy calculation
`and numerical presentation of results practiced more than
`twenty years ago.
`
`Data Acquisition and Model Development
`For bone structural and geometric data, a GE CT IT 9800
`Scanner with a bone optimization technique (120 MA, 2 S,
`and matrix 512) was used to obtain the whole-body data. The
`whole-body data came from a middle-aged adult unembalmed
`cadaveric body selected from donors who have died of disease
`or trauma unrelated to the musculoskeletal system. The ca(cid:173)
`daveric body was positioned in neutral spine posture in a
`wooden box immobilized in polyurethane foam with imbedded
`lead marker wires serving as guides to construct the global
`reference coordinate system. The whole-body was scanned at
`1.5 mm intervals from the skull to the feet in several sessions
`to permit data block management and machine cooling [40,
`41, 45, 46, 50]. The scan data were stored in uncompresses
`
`Fig 1 Top
`
`10·24 X 1.25 LG
`NYLON MTG. SCREW
`(3 REO'D)
`
`SENSOR
`
`~~J"l ~X
`
`NYLON MTG. SCREW
`(2 REO'D)
`
`z
`
`Fig 1 Bottom
`Fig. 1 The 3SPACETM lsotrack device used to obtain three·dimensional
`human body segment relative motion
`
`format on digital tape for later processing purposes. The
`wooden box was custom made according to the size and geo(cid:173)
`metric shape of the cadaver so that the entire system could be
`stored and moved for scanning without changing the relative
`position of the cadaveric body and its limb segments in ref(cid:173)
`erence to the platform. Standard limb positions were main(cid:173)
`tained and stabilized using nonmetallic fasteners, and the body
`was frozen before each scanning session.
`Custom-made calibration phantoms containing different
`mineral density solutions was used to assist differentiating bone
`structure and morphology versus other soft connective tissue.
`The calibration phantoms also enabled selection of the proper
`gray-level threshold window for the CT number in order to
`obtain clear and accurate bone cortex margins to reconstruct
`the periosteal and intramedullary boundaries. The cancellous
`bone morphology was revealed in the volumetric model to
`facilitate cross-sectional views and projection images to em(cid:173)
`ulate regular radiographs.
`For muscle, tendon, ligament and cartilage geometric data,
`a 2.5T GE MR axial Scanner using a spin echo technique with
`repetition time of 700 ms, echo time of 20 ms, field of view
`(FOV) of 16 em in a 256 x 256 Matrix was used [50]. The scan
`
`Journal of Biomechanical Engineering
`
`NOVEMBER 1993, Vol. 115/563
`
`Mako Exhibit 1007 Page 3
`
`

`

`slices were at 3 mm intervals. The increased scan slice thickness
`as compared to CT was mainly due to the equipment resolution
`~imit~tion. No morphologic or structural data was necessary
`m this phase of study since only the soft tissues were required
`for volumetric and surface reconstruction. The centroidalline
`of each soft tissue unit was used to provide the line of action
`in. the biomechanical model required to perform force analysis.
`Different surface textures and colors were used to differentiate
`tissue type and functional unit so they can be displayed together
`with the underlying bony structure. Integrated projections,
`stereo or transparent images, shaded surfaces and volume
`rendering were available for visualization and' measurement
`purposes. This scan preceded the CT scan since the cadaver
`body must not be in a frozen state. The potential swelling
`e~fect ~ue to freezing was ignored which may produce certain
`dimensiOnal error. At the present time, no information is avail(cid:173)
`able to estimate such error.
`
`Motion Data Acquisition. Experimental techniques were
`used to measure joint motion under prescribed activities. Triax(cid:173)
`ial electrogoniometers, stereo-optical video system, 3Space™
`lsotrack device utilizing an electromagnetic field source and
`sensors (Fig. 1) were adopted to obtain three-dimensional hu(cid:173)
`man body segment motion using rigid body assumption [15,
`16]. Adequate data smoothing techniques and cubic spine func(cid:173)
`tions were utilized to express limb segment relative and absolute
`motion in closed form analytical expressions [55]. When small
`bones with irregular geometry were involved such as in the
`hand, wrist, spine and foot, stereoradiographic method was
`applied marking individual bones using radio-opaque markers
`[3, 7, 35, 48, 52, 54]. Published joint kinematic data in the
`literature was also utilized to provide joint motion simulation
`for visualization purpose. Different techniques have inherent
`advantages as well as disadvantages depending upon the an(cid:173)
`atomic region involved and the type of kinematic data to be
`analyzed. [61]. Discussion of the appropriate selection of each
`technique for motion data acquisition in beyond the scope of
`this paper.
`
`Geometric Model Development. The many forms of tissue
`in the human body will require a number of different geometric
`~o?els to best represent the anatomy, and, at the same time,
`limit the amount of data needed to define the model. A number
`of different geometric models include several parametric sur(cid:173)
`face models, hyperpatch models, general solid primitives and
`several parametric curves [20, 21, 27, 36, 51]. These modules
`will each be tied to specific data record in the data base. As
`entities in the data base are displayed, the appropriate module
`will be called to display the entities. In having a number of
`these display modules one can define the appropriate models
`and develop very efficient display modules for each model
`type.
`Scan data typically takes form of a three-dimensional array
`of values representing, for example, density value in the case
`of CT scans. Each array point corresponds to a XYZ location
`and possesses a density value. The task is to identify boundaries
`between different tissue types and create mathematical surfaces
`for these boundaries. There is software available which fits
`curves through constant value data points from CT slice data.
`Software menu provides combinations of automated and in(cid:173)
`teractive methods to develop the geometric curves defining the
`bone boundaries.
`It was necessary to define the interconnectivity and joint
`kenematics for the entire skeletal system. This was then im(cid:173)
`plemented in the data base to provide the rigid body motion
`for the skeleton. The bones of the upper extremity wer modeled
`us~ng. rational B-sp~ine surfaces [10, 11, 18, 19, 22, 26]. Using
`this simple geometnc data base for the bones and the kinematic
`motion protocol developed, a user interface was created that
`allows the user to select any view point and move the position
`
`of any joint in the limb [60]. To develop relationships between
`the position of geometric models of the skeleton, the soft tissue
`must follow the motion of the skeleton. To deal with the motion
`of deformable surfaces, several algorithms were developed for
`the addition of geometric models to be used in the data base.
`
`Musculoskeletal Model Development. The surface model
`combined with volumetric information and the orientation of
`constraining soft tissue centroidallines could be used for joint
`resultant force analysis under static and dynamic conditions.
`External loads obtained from experimental measurements (load
`cells or force plates) and the estimated gravitational force of
`each limb segment were used to formulate the equilibrium
`equations of a free-body diagram involving a specific anatomic
`region containing the joint of interest. The resulting analytical
`problem c~uld then be solved, since joint contact forces, lig(cid:173)
`ament tens10ns, and the muscle contractions were lumped into
`resultant forces. In addition, the joint moment caused by ex(cid:173)
`ternally applied forces and gravity was calculated with respect
`to reference axes passing through the estimated joint center
`[5, 6, 12, 14, 17, 33, 34, 60].
`In order to solve the complete static or dynamic problem
`including joint contact force, ligamentous tension, and muscle
`contr~ctions, a complete free-body diagram including all key
`soft tissue elements was established. Joint contact force di(cid:173)
`rection and point of application were assumed known (normal
`to the contact surface and at the centroid of the area of contact)
`s.o that the transverse shear forces would be borne by the
`ligamentous structure. Due to the multiple number of con(cid:173)
`straining ligaments and active muscles, the static analysis prob(cid:173)
`lem became indeterminate (more unknowns than the number
`of equilibrium equations). Currently available theories and
`techniques were used to obtain estimated solutions [5, 21, 33,
`30, 60]. Several available dynamic analysis softwares for multi(cid:173)
`body systems could be used to obtain the resultant forces and
`moments at each joint [56, 57, 60].
`. ~tatic and quasi-static analyses provided the loading con(cid:173)
`dition of any musculoskeletal joint system at one time instance.
`This loading condition combined with the joint orientation
`and articulating surface contact area estimation were combined
`to determine joint contact pressure distribution using a Discrete
`EleJ?ent Analysis (DEA) technique described as the Rigid Body
`Sprm.g Model (RBSM) of Kawai and others [4, 25, 28, 30, 32].
`In this technique all bony segments are assumed to be rigid
`bodies, while ligaments and articular cartilage are modeled as
`distributed tensile or compressive springs (Fig. 2). The redun(cid:173)
`da.nc~ of the system will be handled using minimum energy
`pnnctple. The model will contain all spring element forces as
`unknown variables, while the equilibrium condition can be
`determined based on an iterative scheme (Fig. 3). When com(cid:173)
`pressive spring elements are subject to tensile forces they will
`be eliminated in the model and the equilibrium equations will
`be solved again with a new stiffness matrix of varying spring
`elements. The final rigid body equilibrium state provides de(cid:173)
`formation of the spring elements and thus, determines the joint
`contact pressure and ligament tension. The 2-D models were
`developed mainly for clinical application. A true 3-D model
`was also formulated for uni-condylar or by-condylar joints
`
`.. -...._,___....,...-Series of
`elastic springs
`normal to the
`articulating
`surface
`
`Spring constant Kd (N/mm/spring)
`
`Fig. 2 Rigid Body Spring Model (RBSM) or the discrete element model
`(OEM) of articulating joint, lor contact pressure
`
`564/ Vol. 115, NOVEMBER 1993
`
`Transactions of the ASME
`
`Mako Exhibit 1007 Page 4
`
`

`

`Total Hip Replacement Simulation Model
`RIGID BODY SPRING MODELING (RBSMJ TECHIIOUE
`Body weight
`
`Springs m
`tension
`(eliminated)
`
`~u·· Virtual displacement
`Iterative process for RBSM analysis for joint contact pressure
`Fig. 3
`determination under load F
`
`MAIN
`
`Fig. 4 Flow chart diagram for the Visual Interactive Multi·body Software
`(VIMS) for musculoskeletal joint system simulation/animation
`
`but the method of identifying surface of contact geometry and
`size non-invasively was difficult. At the present time, only
`assumed 3-D contact surfaces were used for demonstration
`purpose in research development. The entire simulation and
`animation process were incorporated in the Visual Interactive
`Multi-body Software (VIMS) described in the flow chart (Fig.
`60).
`
`Clinical Applications
`The dependence on computers in medical research and clin(cid:173)
`ical patient care has gained widespread and enthusiastic ac(cid:173)
`ceptance, but the degree of technical sophistication in such
`applications still lags far behind modern computer capabilities.
`Such utilizations can be grossly categorized as computer-aided
`diagnostic techniques, computer-controlled treatment proce(cid:173)
`dures, analysis, design and fabrication of artificial joint pros(cid:173)
`thetic devices, computer simulation and optimization of surgical
`operations, and computer-assisted modeling and analysis of
`musculoskeletal rehabilitation. Although these developments
`and applications could be applied to all fields of medicine and
`surgery, orthopedic surgery appears to take a leading position
`in this exciting development. This review paper summarizes
`the applications of the current technology in computer-aided
`joint implant design and selection; simulation and optimization
`of surgical procedures; and in musculoskeletal rehabilitation.
`The present simulation/visualization software and data base
`were utilized in several exciting applications. Some areas have
`already been benefited by this analysis for preoperative plan(cid:173)
`ning using simplified two-dimensional models. When the cur(cid:173)
`rent three-dimensional CT and MRI based model are completely
`
`Bone
`cement
`
`Normal
`spring
`
`Shear Bone
`spring
`
`Fig. 5 Using RBSM to study stem bone interface stress. If bone cement
`is used, a small gap between bone cement and bone must be assumed.
`
`developed, it is conceivable that future surgery can be simulated
`and planned on a graphic workstation, especially in difficult
`cases. Certain areas are particularly attractive and they are
`briefly described. Although the models are based on idealized
`and individual cadaveric data, proper scaling method and as(cid:173)
`sumed abnormal conditions may allow a close simulation of
`individual patients with specific pathological conditions. How(cid:173)
`ever, it should be emphasized here that linear or non~linear
`scaling techniques have significant limitation in extrapolation
`of the normal database collected from only a finite number
`of cadavers to each individual patient's abnormal conditions.
`Extensive anthropometric studies must be performed in order
`to establish such capability.
`
`Selection and Planning in Total Joint Replacement.
`In cer(cid:173)
`tain failed total joint replacement cases or when radical bone
`resection due to tumor is required, special prosthetic devices
`must be custom-designed according to the anatomical defect
`involved. Computer-aided design, analysis and manufacturer
`are required in order to ensure optimal clinical and functional
`outcome. A similar approach has also been studied in the
`selection and fixation of massive allografts to reconstruct os(cid:173)
`teochondral defects. Pathological lesion and anticipated sur(cid:173)
`gical resection margins can be displayed in 3-D color graphics
`by computer using CT or MRI scan information. A properly
`sized allograft from an existing bone bank can be selected and
`prepared for the surgical reconstruction. Joint pressure and
`motion were obtained through model simulation. Hence, com(cid:173)
`plex surgical procedures could be analyzed on computer to
`seek the ideal condition to improve functional results and min(cid:173)
`imize complications. Utilizing the same RBSM technique, bone/
`prosthetic joint implant interface normal and shear stresses
`were determined without the use of three-dimensional finite
`element analysis. The bone and prosthetic component were
`treated as rigid bodies interfaced with normal and shear springs
`to simulate the implant fixation condition, with or without
`bone cement (Fig. 5). The muscle and ligament forces were
`represented as a series of tensile springs. The elastic properties
`of these springs were determined experimentally using cadav(cid:173)
`eric models. The same equilibrium analysis based on the it(cid:173)
`erative scheme used for joint surface contact pressure analysis
`was used.
`Different prosthetic designs and placement conditions were
`simulated and analyzed, and the minimization of interface
`stress distribution and rigid body displacement of the implant
`component relative to the bone were used as selection criteria
`
`Journal of Biomechanical Engineering
`
`NOVEMBER 1993, Vol. 115/565
`
`Mako Exhibit 1007 Page 5
`
`

`

`since implant loosening was assumed to be associated with
`these biomechanical parameters. The advantage of using this
`analysis scheme is to save modeling and computational time,
`while the results of each parametric condition can be obtained
`in real time. Both two-dimensional and three-dimensional anal(cid:173)
`yses can be conveniently performed. This technique is incor(cid:173)
`porated as one of the feature analysis algorithms of the data
`base-software to provide clinical applications in pre-surgical
`planning for total joint replacement procedures.
`
`Joint Osteotomy Preoperative Planning. A computer(cid:173)
`based system using long-standing X-rays to analyze knee joint
`axial alignment and pressure distribution was developed uti(cid:173)
`lizing the DEA or RBSM theory. This system is now being
`applied in three clinical areas involving reconstructive surgery
`of the knee, hip and the wrist joints. The first application is
`to determine the optimal wedge angle and location for knee
`(femoral/tibial) osteotomy.
`A reference data base on the normal range of knee axial
`alignment and joint pressure in 120 subjects was established.
`This program was used to study a group of patients retro(cid:173)
`spectively after knee osteotomy. Based on an assumed statis(cid:173)
`tical power, the biomechanical parameters used in the computer
`analysis were correlated with clinical and functional results in
`a prospective study involving 120 patients using the principal
`component analysis technique. These patients were followed
`for a minimum of one year and their clinical, radiographic
`and functional gait data were compred between two groups:
`1) Those following the recommended wedge angle (within ±3
`deg range), and 2) those falling outside the range of ±3 deg.
`In group 1, pre and one-year postoperative knee performance
`index based on gait analysis improved from a mean of 45 to
`58, while 50 is the separation point between normals and ab(cid:173)
`normals. There was no significant change of knee performance
`index in group 2 patients, with both pre and postoperative
`mean values to be 45. Guidelines and computer software (OA(cid:173)
`SIS) have been developed for both varus and valgus knee
`deformities using either the anatomic axes or the load bearing
`axis of the involved lower extremity (Fig. 6). Even if the com(cid:173)
`puter software is not available, simple application of these
`guidelines can be utilized to assist the practicing orthopedic
`surgeon in deciding the osteotomy wedge and the site of cor(cid:173)
`rection for each patients. This technique is being evaluated
`further with longer follow-up to justify its clinical relevance.
`Radiocarpal joint realignment is a common orthopedic pro(cid:173)
`cedure. However, the precise surgical techniques lack bio(cid:173)
`mechanical justification on the amount of correction to be
`used to order to ensure consistent and lasting good results.
`Using full-length radiographs or CT scan, wrist and intercarpal
`joint contact stresses can be determined based on the preop(cid:173)
`erative abnormal anatomic condition (Fig. 7). One can then
`proceed to simulate various amounts and locations of surgical
`correction on computer so that predictable clinical and func(cid:173)
`tional improvements can be accomplished.
`The same technique is now extended to provide preoperative
`planning in total knee replacement patients for optimal selec(cid:173)
`tion of implant type, dimension, and surgical placement. The
`effects of the lower extremity axial alignment through the knee
`joint and the placement and orientation of the prosthetic com(cid:173)
`ponents on the implant/bone interface stresses were also in(cid:173)
`vestigated these results were used to suggest surgical technique
`improvement and instrumentation design. The same analysis
`has also been applied to the proximal femur and the pelvis for
`prosthetic replacement or osteotomy procedures.
`
`Simulation
`and Rehabilita(cid:173)
`Joint Function
`of
`tion. Objective functional evaluation of the musculoskeletal
`system is of great importance in determining the efficacy of
`treatment modalities, comparing implant design and surgical
`techniques, establishing rehabilitation regimens, and accu-
`
`Fig. 6 Knee joint model based on RBSM technique to predict frontal
`plane pressure distribution as a function of joint alignment. This model
`can be used for knee osteotomy preoperative planning.
`
`Fig. 7 The same RBSM technique as applied to the wrist to study the
`effect of limited intercarpal fusion on carpal pressure distribution
`
`rately assessing the degree and extent of disabilities. Different
`electromechanical devices are interfaced with the computer so
`that the entire evaluation process, data acquisition, results
`archiving and report preparation can be automated. More re(cid:173)
`liable disability ratings of patients with varying degrees of
`functional impairment can now be recommended by comparing
`their results with the large data base of a matched population.
`Using the current model and data base, internal joint and
`soft tissue forces during rehabilitation can be monitored in
`real time during exercise. Therefore, the efficieny of exercise
`can be analyzed and optimized for the planned therapeutic
`treatment. More importantly, adverse effect of exercise when
`performed incorrectly can be predicted and thus provide proper
`warning to the patient and the therapist so that long-term
`detrimental effects can be accurately prevented (Fig. 8). The
`concept of computer -aided rehabilitation can become a reality,
`while safe and more efficient equipment, plus optimal exercise
`protocols can be developed.
`
`Summary
`Although current computer technological capabilities are not
`
`566/ Vol. 115, NOVEMBER 1993
`
`Transactions of the ASME
`
`Mako Exhibit 1007 Page 6
`
`

`

`cation. With available computer technologies still not yet ex(cid:173)
`plored, the future of orthopedic research and the related clinical
`advancement is indeed unlimited. The proposed computer soft(cid:173)
`ware presented here will undoubtedly open up the most prom(cid:173)
`ising application of engineering technology in medicine and
`surgery. It is, indeed, most exciting that biomedical engineering
`has finally emerged into the forefront of health care manage(cid:173)
`ment and delivery as a mainstay in medicine and medical sci(cid:173)
`ences. Since the early development in biomechanical analysis
`of bone and joint system twenty years ago, simulation/ani(cid:173)
`mation of dynamic performance of human skeletal structure
`can now be visualized with unprecendented accuracy, speed
`and reality using attractive graphics and interactive computer
`capabilities (Fig. 9). It is very possible that a computer-aided
`treatment program for patients with musculoskeletal diseases
`and disabilities can be established, which will be a natural
`outgrowth of imaging techniques combined with simulation/
`animation and artificial intelligence.
`
`Acknowledgment
`The computer software and data base are being developed
`through an Advanced Technology Program grant awarded to
`the Engineering Animation, Inc. (EAI), in Ames, Iowa, by the
`United States Department of Commerce. Additional support
`for this research was also provided by grant CA 40583 funded
`by NCI, NIH, DHHS.
`
`References
`1 Abi-Ezzi, S., "The Graphical Processing of B-Splines in a Highly Dynamic
`Environment," Ph.D. thesis, RPI, 1989.
`2 Akima, H., "A New Method of Interpolation and Smooth Curve Fitting
`Based on Local Procedures," J. ACM, 17(4):589-602, 1970.
`3 An, K. N., Chao, E. Y., Cooney, W. P., III, and Linscheid, R. L.,
`"Normative Model of Human Hand for Biomechanical Analysis," J. Biomech.,
`12:775-788, 1979.
`4 An, K. N., Himeno,