Use this URL to cite or link to this record in EThOS:
Title: Shape and function in congenital heart disease : a translational study using image, statistical and computational analyses
Author: Bruse, J. L.
ISNI:       0000 0004 7224 4071
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
While medical image analysis techniques are becoming technically more advanced, analysis of shape and structure in clinical practice is often limited to two-dimensional morphometry, neglecting potentially crucial three-dimensional (3D) anatomical information provided by the original images. This thesis aims at closing this gap by combining state-of-the-art medical image analysis, engineering and data analysis tools to elucidate relationships between 3D shape features and clinically relevant functional outcomes. In particular, patient cohorts affected by congenital heart disease were studied since shape and structure of the heart and its components are crucial for diagnosis, therapy and management of those patients. At first, a statistical shape model was coupled with partial least squares regression to extract anatomical 3D shape biomarkers related to clinical parameters from cardiovascular magnetic resonance image data. After establishing a step-by-step protocol to guide the user with respect to parameter selection, results were shown to be in accordance with traditional morphometry as well as with clinical expert opinion. Novel aortic arch shape biomarkers relating to cardiac functional parameters were found in a cohort of patients post aortic coarctation repair (CoA). By combining statistical shape modelling results with computational fluid dynamics simulations, a mechanistic basis for the observed results was provided. Methods were then extended towards a hierarchical shape clustering framework, which achieved good unsupervised classification performance in a population of healthy and pathological aortic arch shapes. Applied to a cohort of CoA patients, previously unknown anatomical patterns were discovered. This thesis demonstrates that combining medical image analysis and engineering tools with data mining and statistics provides a powerful platform to detect novel shape biomarkers and patient sub-groups. Results may ultimately improve risk-stratification, treatment-planning and medical device development, thereby promoting translation of advanced computational analysis techniques into clinical practice.
Supervisor: Schievano, S. S. ; Hsia, T. Y. H. ; Taylor, A. M. T. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available