Abstract
Recent polls of collegiate and professional sports have shown large occurrence rates of Lisfranc injuries affecting 1.58/100 and 1.8/100 of these athletes, respectively. Currently, no experimental study has demonstrated a consistent mechanism for Lisfranc injuries, nor found a tolerance for the injury. Furthermore, it is unknown what influence geometric and material variables could have on this injury, and no computational models exist that focus on this region of the foot to explore these questions. The goal of this thesis was to gain knowledge about Lisfranc injuries in order to create tools to aid the prevention of these injuries and the future development of injury countermeasures. In order to achieve this goal, five aims were defined: 1) Develop a lower extremity finite element model. 2) Reproduce Lisfranc injuries in a human cadaveric model. 3) Confirm an injury mechanism and create an injury tolerance for Lisfranc injuries. 4) Create a framework for specimen-specific finite element lower extremity models. 5) Evaluate the influence of model parameters on model prediction. Sixteen large male cadaveric lower extremities were tested in quasi-static compression of the first metatarsal with two different boundary conditions; the calcaneus was positioned medial or lateral of the first metatarsal to bias gross bending direction of the foot (either medial or lateral). Post-testing, bony kinematic data was evaluated to determine injury mechanism. Furthermore, experimental injury timing was determined via analysis of abrupt changes in time-synchronized traces of force, moment, bone kinematics, acoustic sensors, and strain gages. Once injury timing was determined, kinetic and kinematic responses were evaluated for accuracy of injury prediction through the development of injury risk functions using survival analysis assuming an underlying Weibull distribution. Development of a methodology to produce specimen-specific models was achieved using a template lower extremity finite element model and target CT scan. Several numerical methods from literature and commercial software were combined in a custom Matlab script to generate specimen-specific models. This custom code included the following techniques: iterative closest point, single value decomposition, elastic registration, radial basis functions, inverse distance weighting, Delaunay triangulation, ray/triangle intersection, HyperMesh, and TCL scripts. Following the creation of specimen-specific models, simulations were performed to match the experimental boundary conditions for five of the experimental tests and were evaluated against their respective injury metrics. Error analysis between experimental and simulated response was completed to evaluate the legitimacy of the model. Finally, the sensitivity of model parameters (bone geometry, initial position, and ligament properties) was evaluated to determine the influence of each of these parameters on finite element model injury prediction. Experimental testing showed that both boundary conditions demonstrated a consistent injury mechanism for Lisfranc injuries. Three separate injury risk functions were produced that can be implemented into countermeasure design. A framework was created that used a template finite element model to make additional models using the CT data of new specimens. A total of 150 simulations were created representing five geometries with 30 different variations in model parameters. The predictions of these models and results from experimental tests identified the injury prediction metric that was best predicted by the model were kinematic based injury tolerances. The results of these simulations also determined specimen-specific geometry, initial position, and ligament failure strain had a significant influence on the model’s predictions. Lisfranc injury was found to be caused by hyper-plantarflexion in combination with either adduction or abduction of the forefoot relative to the hindfoot. An angle metric (relative angle of the 1st metatarsal to the calcaneus) is recommended for use for future Lisfranc injury prediction, as it is a direct measurement of the mechanism identified that caused these injuries. A robust framework for production of specimen-specific lower extremity finite element models was generated that can be extended to examine further questions pertaining to foot and ankle biomechanics and was successfully applied to model a group of the Lisfranc experiments. Finally, sensitivity analysis demonstrated that injury metric prediction of the specimen-specific models was influenced the most by changes in bony geometry, initial position, and ligament failure strain. This suggests that the developed methodology is able to address two of the three most sensitive model parameters for Lisfranc injury prediction.
