Subject-specific image analysis and computational modelling of the human Achilles tendon

Abstract: Achilles tendon rupture is becoming more prevalent with an increasing population of middle aged people participating in recreational sports. A torn Achilles tendon will likely not regain its original function, and it is unknown what treatment options are optimal. The Achilles tendon structure and how the tendon heals after a rupture varies between humans. A potentially useful tool in understanding Achilles tendon healing is the use of computational models. Data from an ongoing unpublished study (by Pernilla Eliasson) with 41 patients with ruptured Achilles tendon was made available for this project, five of these were selected for analysis. In the study, all patients did early motion training and a test group had early load training as well. The study included computer tomography (CT) images of the Achilles tendon, along with tensile loading data. The access to this patient data resulted in two aims for this project. One aim was to develop a pipeline for modelling subject-specific healing human Achilles tendons. Another aim of this project was to perform image analysis on healing human Achilles tendons from computed tomography images of these five patients, to study dimensions and tissue density throughout healing. From CT images taken at 7, 19 and 52 weeks post rupture, the healing Achilles tendon was segmented. The free tendon length was also measured at the three time points. The segmented tendon was analysed for cross sectional area and tissue density. Segmentations of tendons were also prepared for finite element modelling. Using subject-specific experimentally measured tensile load data that had been obtained from roentgen stereophotogammic analysis (RSA), two material models were applied to the geometry, one linear elastic and one viscoelastic. The subject-specific tendon models were used for simulating tensile load and creep. The results from the image analysis suggest that the cross sectional area and the tendon length increased within 7 weeks post rupture. The tissue density distribution was similar throughout the tendon. The simulations predicted maximum stress and strain at the smallest cross sectional area and at the boundary conditions. In order to get more accurate and in vivo like results from the simulations more experimental data is required. One year post rupture the tissue density had gone back to normal, but an increase in cross sectional area remained, suggesting that the tendon had not fully recovered. Due to the small sample size of five patients, no conclusions could be drawn about the two treatments protocols.