Total ankle replacement (TAR) is the main surgical option in case of severe joint osteoarthritis. The high failure rate of current TAR is often associated to inappropriate prosthetic articulating surfaces designed according to old biomechanical concepts such the fixed axis of rotation, thus resulting in non-physiological joint motion. A recent image-based 3D morphological study of the normal ankle (Siegler et al. 2014) has demonstrated that the ankle joint surfaces can be approximated by a saddle-shaped cone with its apex located laterally (SSCL). We aimed at comparing the kinematic effects of this original solution both with the intact joint and with the traditional prosthetic articulating surfaces via in-silico models and in-vitro measurements.

Native 3D morphology of ten normal cadaver ankle specimens was reconstructed via MRI and CT images. Three custom-fit ankle joint models were then developed, according to the most common TAR designs: cylindrical, symmetrically-truncated medial apex cone (as in Inman's pioneering measures), and the novel lateral apex cone, i.e. SSCL. Bone-to-bone motion, surface-to-surface distance maps, and ligament forces and deformations were evaluated via computer simulation. Prototypes of corresponding prosthesis components were designed and manufactured via 3D-printing, both in polymer-like-carbon and in cobalt-chromium-molybdenum powders, for in-vitro tests on the cadaver specimens. A custom testing rig was used for application of external moments to the ankle joint in the three anatomical planes; a motion tracking system with trackers pinned into the bone was used to measure tibial, talar and calcaneal motion (Franci et al. 2009), represented then as tibiotalar, subtalar and ankle complex 3D joint rotations. Each ankle specimen was tested in the intact joint configuration and after replacement of the articulating surfaces according with the three joint models: cylindrical, medial apex cone and SSCL.

Results. Small intra-specimen data variability in cycle-to-cycle joint kinematics was found in all cadaver ankles, the maximum standard deviation of all rotation patterns being smaller than 2.0 deg. In-silico ligament strain/stress analysis and in-vitro joint kinematic and load transfer measurements revealed that the novel SSCL surfaces reproduce more natural joint patterns than those with the most common surfaces used in current TAR.

TAR based on a saddle-shaped skewed truncated cone with lateral apex is expected to restore more normal joint function. Additional tests are undergoing for further biomechanical validation. The present study has also demonstrated the feasibility and the quality of the full process of custom TAR design and production for any specific subject. This implies a thorough procedure, from medical imaging to the production of artificial surfaces via 3D printing, which is allowing for personalised implants to become the future standard in total joint replacement.