We review some of the latest approaches to analysing cardiac electrophysiology data using machine learning and predictive modelling. Cardiac arrhythmias, particularly atrial fibrillation, are a major global healthcare challenge. Treatment is often through catheter ablation, which involves the targeted localised destruction of regions of the myocardium responsible for initiating or perpetuating the arrhythmia. Ablation targets are either anatomically defined, or identified based on their functional properties as determined through the analysis of contact intracardiac electrograms acquired with increasing spatial density by modern electroanatomic mapping systems. While numerous quantitative approaches have been investigated over the past decades for identifying these critical curative sites, few have provided a reliable and reproducible advance in success rates. Machine learning techniques, including recent deep-learning approaches, offer a potential route to gaining new insight from this wealth of highly complex spatio-temporal information that existing methods struggle to analyse. Coupled with predictive modelling, these techniques offer exciting opportunities to advance the field and produce more accurate diagnoses and robust personalised treatment. We outline some of these methods and illustrate their use in making predictions from the contact electrogram and augmenting predictive modelling tools, both by more rapidly predicting future states of the system and by inferring the parameters of these models from experimental observations.

The contact cardiac electrogram is derived from the extracellular manifestation of cellular action potentials and cell-to-cell communication. It is used to guide catheter based clinical procedures. Theoretically, the contact electrogram and the cellular action potential are directly related, and should change in conjunction with each other during arrhythmogenesis, however there is currently no methodology by which to concurrently record both electrograms and action potentials in the same preparation for direct validation of their relationships and their direct mechanistic links. We report a novel dual modality apparatus for concurrent electrogram and cellular action potential recording at a single cell level within multicellular preparations. We further demonstrate the capabilities of this system to validate the direct link between these two modalities of voltage recordings.

Fibrillation is the most common arrhythmia observed in clinical practice. Understanding of the mechanisms underlying its initiation and maintenance remains incomplete. Functional re-entries are potential drivers of the arrhythmia. Two main concepts are still debated, the “leading circle” and the “spiral wave or rotor” theories. The homogeneous subclone of the HL1 atrial-derived cardiomyocyte cell line, HL1-6, spontaneously exhibits re-entry on a microscopic scale due to its slow conduction velocity and the presence of triggers, making it possible to examine re-entry at the cellular level.We therefore investigated the re-entry cores in cell monolayers through the use of fluorescence optical mapping at high spatiotemporal resolution in order to obtain insights into the mechanisms of re-entry.Re-entries in HL1-6 myocytes required at least two triggers and a minimum colony area to initiate (3.5 to 6.4 mm2). After electrical activity was completely stopped and re-started by varying the extracellular K+ concentration, re-entries never returned to the same location while 35% of triggers re-appeared at the same position. A conduction delay algorithm also allows visualisation of the core of the re-entries. This work has revealed that the core of re-entries is conduction blocks constituted by lines and/or groups of cells rather than the round area assumed by the other concepts of functional re-entry. This highlights the importance of experimentation at the microscopic level in the study of re-entry mechanisms.

The SERCA-LVAD trial was a phase 2a trial assessing the safety and feasibility of delivering an adeno-associated vector 1 carrying the cardiac isoform of the sarcoplasmic reticulum calcium ATPase (AAV1/SERCA2a) to adult chronic heart failure patients implanted with a left ventricular assist device. The SERCA-LVAD trial was one of a program of AAV1/SERCA2a cardiac gene therapy trials including CUPID1, CUPID 2 and AGENT trials. Enroled subjects were randomised to receive a single intracoronary infusion of 1 × 10 13 DNase-resistant AAV1/SERCA2a particles or a placebo solution in a doubleblinded design, stratified by presence of neutralising antibodies to AAV. Elective endomyocardial biopsy was performed at 6 months unless the subject had undergone cardiac transplantation, with myocardial samples assessed for the presence of exogenous viral DNA from the treatment vector. Safety assessments including ELISPOT were serially performed. Although designed as a 24 subject trial, recruitment was stopped after five subjects had been randomised and received infusion due to the neutral result from the CUPID 2 trial. Here we describe the results from the 5 patients at 3 years follow up, which confirmed that viral DNA was delivered to the failing human heart in 2 patients receiving gene therapy with vector detectable at follow up endomyocardial biopsy or cardiac transplantation. Absolute levels of detectable transgene DNA were low, and no functional benefit was observed. There were no safety concerns in this small cohort. This trial identified some of the challenges of performing gene therapy trials in this LVAD patient cohort which may help guide future trial design.

No abstract