This is part 1 of a multi-part series which goes through the custom joint replacement to a finger due to rheumatoid arthritis.

Part 1 Scan data to CAD

Part 2 CAD to FEA

Part 3 FEA to Fatigue


Rheumatoid arthritis can affect many joins of the body, however, for this study we will limit the scope to the fingers or phalanges. In a typical hand there are 27 bones. Each finger has proximal intermediate and distal phalanges. The thumbs do not have intermediate phalanges. A schematic of the hand is shown in Figure 1.

Figure 1: Bones of the hand

Each of these joints share many similarities with each other and are predominately different in size. When considering the full population range the differences between an index and middle finger seem minor. For this study we will look at each joint and across the population range as a single design space with similar design variables and inputs along with patient outcome parameters.

Arthritis comes in 2 varieties seen in figure 2. Osteoarthritis is a disorder where the cartilage has been warn away due to repetitive use or trauma. Rheumatoid arthritis is an autoimmune disorder that frequently affects the fingers and results in swollen and/or painful joints.

Figure 2: Arthritis

Advanced cases of rheumatoid arthritis are disabling and disfiguring. Click this link if you want to see an image search but proceed at your own risk! Below in figure 3 is a depiction of a hand with ulnar deviation where the fingers drift away from the thumb. My mom had this condition.

Figure 3: Ulnar Deviation

Total joint replacement surgery is commonly required, in which a silicone elastomer prosthetic is implanted to restore patient hand/finger function and alleviate pain. This design is loosely based on the SWANSON Finger Joint Implant which is a “flexible intramedullary-stemmed one-piece implant developed as an adjunct to resection arthroplasty” shown in figure 4. This study does not reflect the performance of the SWANSON Finger Joint Implant in and way and is simply shown for reference.

Figure 4: SWANSON Finger Joint Replacement


CT Reconstruction (Simpleware)

Clinical Computed Tomography (CT) scan data of the right hand was used for this project.  Clinical CT data is provided in the DICOM format, which is essentially a stack of 2D grey scale images over a 3D volume. An average male cadaver was used. For the current study, the hand was in the neutral position with the fingers extended. Synopsys’ Simpleware™ ScanIP software (Synopsys Inc., Mountain View, CA) was used to process and convert the CT images from voxels, which are 3D pixels or volumetric pixels, to a tessellated surface representation in the STL format (Stereolithography) which is an unstructured triangular mesh of points and connectivity. Below is a rough step by step order of operations.

DICOM CT image data imported into Simpleware in figure 5.

Figure 5: Visible Korean CT scan of hand viewed in Simpleware

The hand was scanned in roughly a neutral hand position similar to the image below of a hand on an ergonomic mouse.

Figure 6: Hand on ergonomic mouse. Photo credit, Rob Stupplebeen’s left hand; hand model, Rob Stupplebeen’s right hand

Bone tissue was separated from the surrounding soft tissue based on gray scale intensity. This method was able to create voxel groups of the bones of the hand shown in figure 7.

Figure 7: Voxel group of the bones of the hand with CT data overlaid in Simpleware

Figure 8: Bones of the hand partitioned from the soft tissue in Simpleware

These bones were exported as an STL.

Computer Aided Design CAD (CATIA)

The individual STLs, from Simpleware, were imported into the 3DExpereince as point clouds with CATIA’s Digitized Shape Preparation.

Figure 9: Imported STLs in CATIA

A parametric CAD model was created for the prosthetic. This geometry is not mean to replicate a commercially available product or imply its performance. Due to the organic nature of the design surface based modelling techniques were used. Care was also given to assure a design which is curvature continuous to reduce any stress risers and to provide smooth geometry for the shape optimization to begin with. The middle metacarpal and proximal phalange or the first joint of the middle finger is the focus of this study.

Figure 10: Middle metacarpal and proximal phalange isolated

The custom implant is to be designed in place on top of the scan data. To begin, a coordinate system needed to be developed along with reference geometry. 3 points were created for the point of contact of the joint and the most distal and proximal points. These created an initial coordinate system and by extruding surface cylinders and viewing the model with reference to this coordinate system. The initial points can be refined iteratively or by using CATIA’s extremum point feature along with the reference coordinate system.

Figure 11: A coordinate system was setup by creating points, planes and cylinders

The implant geometry is a simplified design representative of an actual implant. It consists of an extruded taco shape with 2 tapering stems that will insert into the reamed interamedullary canal.

Figure 12: Implant design with bone overlaid

Figure 12: Implant design

CATIA’s Digitized Shape to Surface app was used to wrap NURBS (Non-uniform rational basis spline) surface geometry around the point cloud. CATIA’s Generative Surface Design was used to design the device and cut the bone geometry similar to how the surgeon would resect the joint and ream the intramedullary canal.

Figure 13: Solid CAD geometry of bones after tissue has been removed

Figure 14: CAD of implanted device

The implant and the 2 prepared bones are exported from the 3DEXPERIENCE Platform by way of a STEP file.

Figure 15: Implant and prepared tissue

Part 2 will continue the journey to the finite element simulation of the prosthetic in Abaqus.

This is part 1 of a multi-part series which goes through the custom joint replacement to a finger due to rheumatoid arthritis.

Part 1 Scan data to CAD

Part 2 CAD to FEA

Part 3 FEA to Fatigue