Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.687129
Title: Computational modelling of diastole for human ventricle
Author: Palit, Arnab
ISNI:       0000 0004 5922 0913
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2015
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Thesis embargoed until 15 Mar 2018
Access through Institution:
Abstract:
Diastolic heart failure (DHF) with normal systolic pump function has been typically observed in the majority of HF patients. DHF changes regular diastolic behaviour of left-ventricle (LV), and increases the ventricular wall stress. Therefore, normalisation of increased LV wall stress is the cornerstone of many existing and new therapeutic treatments. However, information regarding such regional stress-strain distribution for human LV is extremely limited in the literature. Thus, the study aimed at estimating the normal range and regional variation of diastolic stress-strain field in healthy human LVs, and exploring the infl uence of fibre structure, geometrical heterogeneity and material properties on passive infl ation of LV. It is envisaged that such information could be used as targets for future in-silico studies to design optimised HF treatments. FE modelling of passive diastolic mechanics was carried out using personalised ventricular geometry, that was constructed from magnetic resonance imaging (MRI), and structure-based orthotropic constitutive law. Laplace-Dirichlet-Region growing-Finite element (LDRF) algorithm was developed in order to assign the myocardium fibre map on ventricular geometry. The effect of right ventricle (RV) deformation, that has not been taken into account by the majority of researchers due to modelling simplification, was investigated for the first time by comparing the results predicted by bi-ventricle (BV) and single LV models, constructed from the aforementioned MRI data. In addition, personalised in-vivo measurement of fibre structure, that might be different in individual subjects and diseased conditions, is still an open question. Therefore, the sensitivity of LV diastolic mechanics to the details of the fibre structure was accomplished for the first time using eight different fibre orientations. In-vivo passive orthotropic myocardium properties for healthy human myocardium, indispensable for personalised LV wall stress estimation, was identified, and subsequently, the regional variations of LV wall stress-strain were investigated by incorporating geometrical heterogeneity, personalised myocardium properties and LV base movements in the FE models. RV deformation increased average fibre and sheet stress-strain in LV wall during diastole, and therefore, the effect should always be included in cardiac biomechanics study. Any pathological remodelling, that increased the amount of transmural fibre angle, led to an additional LV infl ation. The study indicates that a change in fibre orientation may contribute to the heart failure with preserved ejection fraction (HFpEF) development. Future therapeutic intervention should consider the effect of altered fibre orientation for better outcome. Due to the ill-posed nature of the inverse optimisation problem, the average myocardial stiffness was extracted by identifying the normal ranges of the parameters. A novel method was developed by combining FE modelling, response surface method (RSM) and genetic algorithm (GA) to identify the passive orthotropic myocardium properties for healthy human myocardium using routinely used clinical data. These myocardium properties can directly be utilised in future computational studies. Although the regional stress-strain distribution of the LV wall was highly heterogeneous amongst the individuals, it was observed that the inner wall of the LV experienced higher fibre stress compared to the outer wall. The LV wall near the base and the lateral region received greater stress-strain compared to the other regions. The incorporation of LV base movement (not addressed in the literature) improved the FE model predictions, and therefore, it is recommended to be considered in later studies. In addition, normal ranges of various stress-strain components in different regions of LV wall were reported for five healthy human ventricles considering RV deformation, LV base movement, and subject-specific myocardium properties. This information could be used as a reference map for future studies. The study revealed that the FE modelling can be employed to analyse the effect of geometry, fibre-structure and material properties on normal ventricular mechanics, and therefore, can provide a greater insight into the underlying mechanics of failing heart and plan for optimised surgical intervention. Hence, the research has impacts on computational cardiac biomechanics as well as clinical cardiac physiology fields.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.687129  DOI: Not available
Keywords: QA76 Electronic computers. Computer science. Computer software ; RC Internal medicine
Share: