Example PhD

Simulation and management of surgical intervention for craniofacial surgery

Supervisor: Professor A.D. Marshall

Keywords: Image Processing, Computer Vision.

The aim is to create an individualised fully functioning biomechanical head model that can be used to plan and simulate facial surgery

Personalised healthcare is currently an international priority and has been acknowledged worldwide as one of the "grand challenges" facing healthcare in the 21st Century. Cardiff University has an established inter-disciplinary team of world leading researchers utilising state of the art scanners (3D Video, Diffusion MRI, Computerised Tomography (CT) to provide a unique person-specific facial model having a vast number of potential applications. Craniofacial surgery is commonly undertaken for craniofacial anomalies, orofacial cancer, facial disharmony and trauma. Traditionally, surgical planning has involved two-dimensional radiographs and photographs and validity of their use has been questioned. However, recently the 3D acquisition of hard and soft tissues using low radiation dose (compared with traditional CT) cone beam CT and non-invasive surface scanning has enabled more realistic and detailed 3D computerised models to be routinely constructed. The facial structures consist of a skeletal framework to which muscles and tendons are attached, enveloped by a complex soft tissue matrix through which nerves and blood vessels run and covered with skin (dermis and epidermis). However, anatomical and constitutive data is scarce restricting the building of an accurate biomechanical model for the various facial actions. Futhermore, current complete functional biomechanical only offer a generic model of a hman head. For detailed surgical planning and other applications it is essential that a patient specific head model is available

CT scans can readily give skeletal scan and 3D surface scanning/video can yield dynamic exterior facial structure and motion. What is missing is detailed interior sturcture such a mussle structure. Current techniques for acquiring this involves anatomical dissection of muscles indicates the multi-directional nature of muscle fibres.

For the planning of surgery, detail in the muscle fibre structure is important and this can be potentially be acquired using Magnetic Resonance Imaging and diffusion MRI-based methods (e.g. diffusion tensor MRI) in particular. A pilot study has already shown that we can track facial movement and transfer this information to a mechanical model built from full colour facial dynamic 10-second video sequences and cone beam CT data. These data enable the assessment of the temporo-mandibular joint in function. Currently we are world leaders in these areas (Modelling facial dynamics, 3D data processing and diffusion MRI) and have made considerable advances in surgical simulations. Detailed characterisation of muscle fibre volume and orientation of the facial muscles using DTI, new segmentation techniques and skeletal tracking will allow the team to develop precise individualised surgical management which will keep the team as world leaders in their respective disciplines and collectively world beaters in craniofacial surgical planning. This work builds on established research links between the schools of Computer Science, Psychology and Dentistry at Cardiff University.

Key Skills/Background: Strong computing and mathematical skills required

Contact: Professor A.D. Marshall to discuss this research topic.