Virtual Twins and Tools for Personalised Clinical Care- the Heart Failure With Reduced Ejection Fraction Sub-study
Virtual Twins and Tools for Personalised Clinical Care- the Heart Failure With Reduced Ejection Fraction Sub-study
Heart failure affects over 1 million people in the United Kingdom. Approximately 50% have heart failure with reduced ejection fraction (HFrEF), a condition where the heart muscle does not contract properly. If the right and left chambers (called 'ventricles') of the heart are not pumping at the same time, devices called cardiac resynchronisation therapy (CRT) can be inserted to encourage the ventricles to pump together, reducing heart failure symptoms and helping people to live longer.
CRT devices are made of wires which are placed into the ventricles through the blood vessels that carry blood to and from the heart. However, one third of patients do not show any improvement following CRT insertion, and another third show only a partial improvement. Doctors cannot accurately predict who will respond well, meaning patients may undergo a procedure with no benefit, being exposed to risks including bleeding and infection.
Virtual models of a patient's heart and blood vessels, known as a 'digital twin', can be generated using artificial intelligence. These models can be used to predict a patient's response to a procedure without them having to undergo the procedure first, meaning patients can avoid being exposed to risks.
30 patients with HFrEF who have been selected to have CRT implanted by their consultant cardiologist will be recruited. Before having their procedure patients will attend University College London (UCL) for tests including a blood test, urine test, magnetic resonance imaging (MRI) scan of the heart, ultrasound of the heart, a 5-minute heart recording and a 6- minute walking/ stepping test. After the CRT is implanted, the patient's response to the device will be assessed with a blood test, a 6-minute walking/ stepping test and a heart ultrasound scan 3 and 6 months after the procedure. The virtual models will be used to determine if a patient's response to CRT can be accurately predicted.
Heart failure affects over 1 million people in the United Kingdom. Approximately 50% of these people have heart failure with reduced ejection fraction (HFrEF), a condition where the muscle of the heart does not contract properly. Without treatment, patients are at risk of developing fluid around the lungs and legs (called oedema), abnormal heart rhythms (known as arrhythmias), and death. Treatment involves a combination of tablets, but some patients show no improvement with medications. When the left and right chambers of the heart are not pumping at the same time, devices called cardiac resynchronisation therapy (CRT) can be implanted to encourage the chambers to pump together, reducing symptoms and improving the patient's quality of life.
CRT devices consist of wires which are placed into the ventricles through the blood vessels which lead to the heart. However, despite an initially successful procedure, over one third of patients do not show any improvement following CRT implantation. Risks associated with the procedure include bleeding, infection and damage to the underlying lung or heart muscle (which can be life-threatening), meaning a significant proportion of patients undertake these risks with no benefit afterwards. Moreover, the infection risk is lifelong, and patients may develop device-related infective endocarditis at any stage, which is associated with a high mortality.
Virtual models of a patient's anatomy and blood flow, known as a 'digital twin', can be generated using artificial intelligence. These models can be used to predict a patient's response to a treatment or procedure. Digital twins have been studied in research projects looking at coronavirus disease 2019 (COVID-19), asthma and cancer. Within cardiology, digital twins have been studied in patients who are undergo valve surgery. So far, no models have been developed to predict patients' responses to heart failure treatments.
Working closely with bioengineers, physiologists and computer scientists, this study aims to develop a validated, multi-scale, multi-organ modelling platform than can create an individualised virtual twin, incorporating cardiovascular anatomy as well as complex physiological processes including the inter-connection of organ systems, the autonomic nervous system activity and hormonal actions.
A minimum of 30 patients who are undergoing a clinically-indicated implantation of CRT will be recruited. Before the procedure, each patient will attend University College London (UCL) for a range of tests including measurement of height, weight and blood pressure, an echocardiogram, a cardiac magnetic resonance imaging (MRI) scan, a 12-lead electrocardiogram (ECG), a 256-lead electrocardiographic imaging recording (a 5-minute detailed recording of the heartbeat), a 6-minute walking/ stepping test and blood and urine sample collection. A subset of patients will also be provided with a wearable heart monitor which will be worn during their UCL appointment. These data will be integrated into the modelling framework to create a personalised virtual 'twin' for each patient.
Following implantation of the CRT, each patient will return for follow-up visits at 3 months and 6 months when they will undergo a further echocardiogram, blood test and 6 minute step/walk test.
A clinically-validated virtual platform would have the potential to transform how patients are selected to undergo device implantation in the following ways:
Patients who are predicted to be 'non-responders' can be directed to other treatments and avoid the serious potential risks associated with device implantation.
Device selection can be optimised. There are 2 forms of CRT:
CSP is considered to better mimic the heart's own physiology, but is more technically challenging to achieve, often leading to a longer and higher risk procedure. Patients referred for both forms of CRT will be recruited to this study. A validated modelling platform may be able to predict which form of pacing would benefit the patient most, meaning clinicians will be able to select the optimal form of pacing without exposing a patient to unnecessary complications.
Inclusion Criteria:
Exclusion criteria:
gabriella.captur@ucl.ac.uk
debbie.falconer@ucl.ac.uk07984548291