AbstractThe focus of this thesis is the development and application of personalised computational models of cardiac electromechanics to understand and ultimately inform cardiac resynchronisation therapy (CRT). To achieve this goal, a semi-automatic pipeline for the generation and parameterisation of detailed biophysically based models using clinical data is presented and applied to a cohort of patients. Specifically, an anatomically based finite element model has been developed and applied to simulate cardiac electromechanics through the coupling of the monodomain and large deformation mechanics governing frameworks. Techniques have been implemented for tting high order representations of cardiac anatomy from MRI data, and myocardial conductivity, sti ness, contractility and boundary conditions from endocardial activation recordings and pressure volume loops respectively. Embedding these tting steps within a semi-automatic pipeline, this personalisation work ow has been applied to four CRT patient data sets. Three models were successfully fitted, while the response of a fourth patient to therapy could not be captured with our framework. It is important to note that for this fourth case the recorded response to therapy in this individual was considered, by standard clinical measures, to be an outlier. Seven metrics of cardiac electrophysiology, haemodynamics and energetics were computed from the models of each patient and used to quantify the effects of changes resulting from CRT. Differences between patient cases were analysed, revealing that reductions in total activation time with pacing (correlation between changes across individuals in our virtual patient cohort p= - 0:73) and transmural conduction block were associated with greater acute haemodynamic response (AHR) (97:1% of models in a virtual patient cohort had a greater AHR with block). Patients with a strong AHR also beneffited from increased stroke work ( p = 0:50) and reduced left ventricle (LV) myocardial work ( p = -0:77) , implying an improvement in myocardial efficiency, plus a homogenisation of LV myocardial work ( p = -0:62).
Maps of AHR by LV epicardial pacing site during simultaneous biventricular pacing were generated, and the optimal region for pacing was determined. Optimisation of pacing lead location to maximise AHR was seen to benefifrom improved stroke work and reduced total LV myocardial work and LV work heterogeneity across all patients.
|Date of Award||2015|
|Supervisor||Nicolas Smith (Supervisor) & Steven Niederer (Supervisor)|