In recent years, significant advances in clinical imaging have provided a wealth of detailed information on the internal function and anatomy of the human body. Fundamental questions on the progression of diseases and on their treatment can be addressed using this information. A highly promising avenue for the exploitation of this potential is the creation of biophysical computer models of the organs based on these state-of-the-art data. These models combine the available clinical information in a consistent mathematical framework that can be tailored to the specific characteristics of the individual patient. They can be used to predict the complex internal dynamics of a working organ, as well as its diseases' mechanisms, providing a powerful tool to personalise treatment.

This project will create a toolbox to generate personalised models of flow and temperature in the heart. Specifically, this platform will be designed to simulate atrial fibrillation, a disease that commonly affects one chamber of the heart (the left atrium), and its treatment, known as catheter ablation. When the left atrium is in fibrillation, the wall stops contracting and starts quivering. This behavior is triggered by abnormal electrical impulses at a specific site, called the driver site. As a result, the motion of the blood flow in the atrium is weakened. This abnormal behavior can reduce the supply of blood to the body. Catheter ablation consists in burning the driver site in the atrial wall by applying heat via a catheter, in order to suppress the abnormal electrical impulses and restore contraction. However, patient outcomes are suboptimal and reoccurrence of fibrillation after a single procedure is high. This is due to the strong dependence of this treatment effectiveness on patient-specific factors that are difficult to quantify from the imaging data alone.

The proposed research will focus on improving outcomes using personalised computer models. Previous pilot work has proven the potential of this type of approach in predicting the size of the lesion caused by ablation. The overall goal of this project is to create physiologically accurate, personalised models to inform the choice of the ablation parameters such as the catheter voltage and the ablation time. This modelling toolbox will then be applied to a cohort of patients with the aim of increasing significantly the clinical impact of the approach. The research will be undertaken at St Thomas' Hospital, one of the biggest UK referral centres for atrial fibrillation. Given the multi-disciplinary aspect of the project, internal collaborations have been set up with groups within different areas of expertise. Clinical guidance and patient datasets will be provided by Dr M. O'Neill and his team, the Cardiac Arrhythmia Research Group, while Dr O. Aslanidi and Dr D. Nordsletten will provide expertise in scar and blood flow modelling, respectively. This work will therefore generate impact in different areas, from clinical research to mathematical modelling. Ultimately, however, the beneficiaries of the proposed project are the patients themselves, who will benefit from a personalised and more efficient approach to catheter ablation.

Potential Impact:
The main goal of the proposed project is to create and apply an integrated imaging-modelling toolbox to improve ablation therapy in atrial arrhythmias. In the first part of the project, the focus will be on developing and testing personalised models of the thermal and fluid dynamics in the left atrium. In the second stage, the validated models will be applied to a cohort of patients to improve our understanding of the pathophysiology of a common atrial arrhythmia, i.e. atrial fibrillation, and of its treatment. In particular, we will simulate the effect of radiofrequency catheter ablation therapy for each individual patient in their personal clinical context. We envisage that this work will have a strong potential to benefit end-users such as clinicians, patients and the NHS, as well as the research community.
For the clinician, placing the focus on a patient-specific approach is a fundamental advance in the clinical decision-making process: it allows to improve significantly the diagnosis potential in diseases like atrial fibrillation, where the heterogeneity of patient outcomes has been demonstrated to be highly dependent on patient-specific factors. The proposed research will inform the choice of ablation strategy by providing insight into the effect of settings such as catheter voltage and contact time on the treatment outcomes.
For the patient, a customised treatment strategy will primarily contribute to enhance the quality and quantity of life by addressing specific needs. These include reducing the incidence of re-interventions from inaccurate patient assessments and thus the re-hospitalisation incidence. Developing more effective ablation stretegies will also put less strain on the patient. Further, improvements in home monitoring and care have been demonstrated to help increasing the event-free survival rates: a better awareness of the disease's dynamics, and of its potentially fatal consequences will therefore benefit a wider public, including the patients themselves and their families, and has the potential to make a social impact.
For the NHS, faster procedures will mean more efficient use of time and infrastructures, enabling improved scheduling and effectiveness of service. Reduced re-intervention rates will also have a direct economic impact. Further, a more complete understanding of the diseases' mechanisms will facilitate and speed up new clinical research and, consequently, improve the level of specialists training.
The proposed research will also benefit the research community by placing emphasis on the synergy between different scientific areas, from computing to imaging sciences. Computer modelling relies on the quality of the clinical imaging datasets: to achieve impact it is essential that the patient data are collected following standardised procedures and organised in databases that can be easily accessed by both the modelling and the clinical community. The models and the imaging protocols used throughout this project will be made publicly available via the King's College database (http://amdb.isd.kcl.ac.uk) to raise clinical awareness of this matter, paving the way for future exploitation of the data and progression of research.