Predicting fracture risk in metastatic bone disease: A combined finite element modelling and machine learning approach

The femur (or thighbone) is a common site for the growth of tumours due to different types of cancers that originate elsewhere but spread to the bone. Such a bone cancer, called the metastatic bone disease, is referred to orthopaedic surgeons for surgical management as it poses a risk of fracture. Surgery to prevent fracture presents significant risks and decisions on whether to operate are based on a scoring system which is based on the site, size, nature and the level of pain caused by the lesion. Fracture is caused by the mechanical inability of the bone to carry loads, so an objective decision support system based on sound principles of mechanics is required to predict fracture risk.

The project will employ finite element analysis to simulate the mechanical behaviour and the risk of fracture in the first phase; femurs with varying geometries and bone quality, lesion size, shape and location subjected to loadings expected from physiological activities will be analysed using the finite element method. While finite element models can be successfully applied for diagnosis of individual patients, the computational cost of modelling is high, which limits their clinical adoption. The project will, therefore, in the second phase develop a novel machine-learning based emulator which can be trained by the above database of finite element analyses. Considerable data of patients with metastatic bone disease is being collated at the University of Edinburgh as part of another project. This data will be available for finite element analysis and validation of the developed machine learning algorithms.

This project will be co-supervised by Ms Chloe Scott, Dr Sohan Seth and Professor Hamish SImpson. 

 

Further Information: 

The selected candidate will be jointly supervised by engineers and clinicians and will be part of the Edinburgh Computational Biomechanics Group: https://ecbm.eng.ed.ac.uk/home

Closing Date: 

Friday, December 18, 2020

Principal Supervisor: 

Eligibility: 

Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree. Further information on English language requirements for EU/Overseas applicants.

The candidate should have an undergraduate or a master’s degree in Engineering or Physics. They should have had undertaken courses in solid mechanics and finite element modelling. Prior experience in machine learning is not required but the candidate should be skilled in advanced programming languages such as Python, Matlab, R, C/C++ or Java and interested in machine learning and its application in biomedical engineering. 

Funding: 

Applications are welcomed from those who have secured their own funding through scholarship, sponsorship, or similar. No funding is provided with this project.

Further information and other funding options.

 

Informal Enquiries: