Abstract
Recent developments in arthroplasty have demonstrated limitation in traditional assumptions surrounding joint reconstruction, such as population wide safe zones for acetabular cup version. Instead surgeons are increasingly targeting a patient-specific joint reconstruction in which patient-specific anatomy informs both surgical technique and joint reconstruction.The first technical section of this work addresses limitations identified in computer geometry and machine vision tools, namely point correspondence and statistical shape modelling workflows, which facilitate study of patient-specific morphology. This work produced a set of tools that improved generalisation of correspondence algorithms across geometries, whilst minimising the need for specialised user knowledge, via full automated adaptive workflows. This has the potential to democratise useful, but currently specialised, tools.
The second technical section of this work takes these developed tools and applies them to patient-specific hip arthroplasty. Firstly, projection error in pre-operative radiographs is addressed with a tool developed to correct erroneous femoral offset measurements and the impact of projection error on the appearance of the internal femoral geometry evaluated. Further a tool was implemented to predict radiographic stem version in post-operative radiographs. Combined, these tools immediately improve surgeons ability to implement a patient-specific hip reconstruction.
Thesis is embargoed until 31st December 2028.
| Date of Award | Dec 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Sponsors | Trauma and Orthopaedics Research Charity |
| Supervisor | Alex Lennon (Supervisor) & David Beverland (Supervisor) |
Keywords
- arthroplasty
- biomechanics
- medical imaging
- statistical shape modelling
- shape correspondence
- total hip arthroplasty