Personalised Model Based Optimal Lead Guidance in Cardiac Resynchronisation Therapy

With each heart beat a wave of electrical activation sweeps across the heart stimulating the muscles to contract. In the healthy heart the wave is initiated from many locations across the wall and rapidly activates the whole heart leading to a synchronous, efficient and effective pumping of blood around the body.

In patients suffering dyssynchronous heart failure the activation wave starts on the right hand side of the heart and slowly progresses to the left hand side of the heart. This asynchronous activation pattern causes an asynchronous, inefficient and ineffective pumping of blood. To treat these patients a pacing device is implanted with leads attached to the left and right hand side of the heart. By activating the left and right side of the heart from these two leads the patient's activation pattern can be resynchronised leading to a synchronous and effective contraction. This treatment is referred to as cardiac resynchronisation therapy or CRT.

CRT is an effective treatment in most patients but 30-50% of patients fail to improve or respond to treatment. Due to the invasive nature and cost of the procedure it is undesirable to treat patients who will not respond. Identifying the patients who cannot respond is currently obfuscated by the inability to guarantee optimal treatment in all cases. Hence it is not possible to differentiate from patients that did not respond as they did not receive the optimal treatment from those that were unable to benefit from CRT under any conditions. At present guidelines suggest a "one size fits all" approach to the location of the leads on the patient's heart despite significant evidence that the location of the leads plays a critical role in determining outcome. This indicates that some patients may respond to CRT but only if they receive optimal lead placement.

The aim of this project is to determine the best location to place the pacing lead on the left side of the heart in each individual patient receiving CRT, based on the physiology and pathology of the specific patient's heart. To achieve this aim we propose to use advanced high fidelity and resolution imaging techniques to characterise the shape of the patient's heart, the potential pacing locations, and the location of any dead non-conducting tissue in the heart. We will combine this anatomical information with measurements of electrical activation time to create a biophysical model of the electrical properties of the individual patient's heart.

Using the model we will be able to simulate the activation patterns in the patient's heart for each potential pacing location. In a training data set we will compare the activation patterns at each pacing location with measured pump function, in response to pacing, to identify the activation pattern that best predicts the optimal pacing location.
A prospective clinical study will then be performed where patient specific models will be created for each patient prior to procedure and the optimal pacing site identified. The predictive capacity of the model will then be evaluated when the device is implanted by testing if the model has correctly predicted the optimal pacing location.
The project represents a significant advance for patient specific models - moving from a technique for analysing patient data to a tool for guiding patient treatment. Improving outcomes for CRT patients will reduce morbidity and hospitalisation rates, decrease the financial burden of non-responding patients on the NHS and improve our ability to identify what characteristics determine if a patient will respond to treatment.

The UK has a long history in developing cardiac electrophysiology models and this proposed research aims to continue this trajectory, moving computational models of cardiac electrophysiology into clinical applications. This project falls within the Clinical Technologies research area and in line with EPSRC guidance is focused heavily on translation to clinical impact and engaging with clinical end users. In light of this, the primary goal of this project is:

To move computational models of the heart from data integration and analysis techniques to predictive tools that directly informs and impacts patient care.

This transformative project will bring computational biophysical models from a position of a novel analysis tool to a clinical application over 4 years. At the end of the project we will have demonstrated the ability of personalised models to guide cardiac resynchronisation therapy (CRT) lead placement. The prospective clinical study, included in this project, will provide the necessary pilot data to underpin the application for funding for the first biophysical patient specific model-guided CRT clinical trial.

Options for delivering CRT are growing with multipolar lead, multi lead pacing, endocardial pacing and combinations of these approaches. Conventional CRT required hundreds of patients to identify the best location to place the lead on average. The identification of activation pattern phenotypes that leads to effective cardiac function will provide an important tool for optimising these novel treatment options.

The ability to predict where to optimally place leads from non-invasive pre-procedural data is expected to improve response to CRT. Increasing the efficacy of CRT will improve patient quality of life and life expectancy while reducing the costs to the NHS of unnecessary implants and increased hospitalisation due to sub-optimal lead positions. Improving response will further isolate the patients that are unable to respond, greatly facilitating the identification of patient phenotypes that prohibit them benefiting from CRT aiding in improved patient selection.

The framework for creating validated patient specific models of cardiac electrophysiology is not limited to informing CRT treatment. Adapting this approach has the potential to be used for modelling the atria to inform atrial arrhythmias, for simulating activation patterns to guide ventricular tachycardia ablations or for interpreting complex activation patterns in congenital heart defect patients.

A by-product of developing patent specific models is the creation of a growing virtual cohort of patients with validated models of activation. We have previously used a model of a single patient case to evaluate pacing strategies and efficacy in the presence of scar for a St. Jude multipolar lead (Niederer et al., 2012) and are currently working with Boston Scientific to evaluate the electrode spacing in their leads. The development of a larger cohort would provide a resource for performing further commercial or clinical studies to evaluate novel lead technologies and pacing strategies characterising both the mean and the variation in the expected response.