Artificial intelligence for ventricular arrhythmia capability using ambulatory electrocardiograms.

Aims: European and American clinical guidelines for implantable cardioverter defibrillators are insufficiently accurate for ventricular arrhythmia (VA) risk stratification, leading to significant morbidity and mortality. Artificial intelligence offers a novel risk stratification lens through which VA capability can be determined from the electrocardiogram (ECG) in normal cardiac rhythm. The aim of this study was to develop and test a deep neural network for VA risk stratification using routinely collected ambulatory ECGs. Methods and results: A multicentre case-control study was undertaken to assess VA-ResNet-50, our open source ResNet-50-based deep neural network. VA-ResNet-50 was designed to read pyramid samples of three-lead 24 h ambulatory ECGs to decide whether a heart is capable of VA based on the ECG alone. Consecutive adults with VA from East Midlands, UK, who had ambulatory ECGs as part of their NHS care between 2014 and 2022 were recruited and compared with all comer ambulatory electrograms without VA. Of 270 patients, 159 heterogeneous patients had a composite VA outcome. The mean time difference between the ECG and VA was 1.6 years (⅓ ambulatory ECG before VA). The deep neural network was able to classify ECGs for VA capability with an accuracy of 0.76 (95% confidence interval 0.66-0.87), F1 score of 0.79 (0.67-0.90), area under the receiver operator curve of 0.8 (0.67-0.91), and relative risk of 2.87 (1.41-5.81). Conclusion: Ambulatory ECGs confer risk signals for VA risk stratification when analysed using VA-ResNet-50. Pyramid sampling from the ambulatory ECGs is hypothesized to capture autonomic activity. We encourage groups to build on this open-source model. Question: Can artificial intelligence (AI) be used to predict whether a person is at risk of a lethal heart rhythm, based solely on an electrocardiogram (an electrical heart tracing)? Findings: In a study of 270 adults (of which 159 had lethal arrhythmias), the AI was correct in 4 out of every 5 cases. If the AI said a person was at risk, the risk of lethal event was three times higher than normal adults. Meaning: In this study, the AI performed better than current medical guidelines. The AI was able to accurately determine the risk of lethal arrhythmia from standard heart tracings for 80% of cases over a year away-a conceptual shift in what an AI model can see and predict. This method shows promise in better allocating implantable shock box pacemakers (implantable cardioverter defibrillators) that save lives.

Machine learning in sudden cardiac death risk prediction: a systematic review.

AIMS: Most patients who receive implantable cardioverter defibrillators (ICDs) for primary prevention do not receive therapy during the lifespan of the ICD, whilst up to 50% of sudden cardiac death (SCD) occur in individuals who are considered low risk by conventional criteria. Machine learning offers a novel approach to risk stratification for ICD assignment. METHODS AND RESULTS: Systematic search was performed in MEDLINE, Embase, Emcare, CINAHL, Cochrane Library, OpenGrey, MedrXiv, arXiv, Scopus, and Web of Science. Studies modelling SCD risk prediction within days to years using machine learning were eligible for inclusion. Transparency and quality of reporting (TRIPOD) and risk of bias (PROBAST) were assessed. A total of 4356 studies were screened with 11 meeting the inclusion criteria with heterogeneous populations, methods, and outcome measures preventing meta-analysis. The study size ranged from 122 to 124 097 participants. Input data sources included demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data ranging from 4 to 72 variables per model. The most common outcome metric reported was the area under the receiver operator characteristic (n = 7) ranging between 0.71 and 0.96. In six studies comparing machine learning models and regression, machine learning improved performance in five. No studies adhered to a reporting standard. Five of the papers were at high risk of bias. CONCLUSION: Machine learning for SCD prediction has been under-applied and incorrectly implemented but is ripe for future investigation. It may have some incremental utility in predicting SCD over traditional models. The development of reporting standards for machine learning is required to improve the quality of evidence reporting in the field.

Effects of Exercise on Nutritional Status in People with Cystic Fibrosis: A Systematic Review.

BACKGROUND: Physical exercise is an important part of regular care for people with cystic fibrosis (CF). It is unknown whether such exercise has beneficial or detrimental effects on nutritional status (body composition). Thus, the objective of this review was to evaluate the effect of exercise on measures of nutritional status in children and adults with CF. METHODS: Standardized reporting guidelines for systematic reviews were followed and the protocol was prospectively registered. Multiple databases were utilized (e.g., PubMed, Scopus, and CINHAL). Two reviewers independently reviewed titles/abstracts and then the full text for selected studies. RESULTS: In total, 924 articles were originally identified; data were extracted from 4 eligible studies. These four studies included only children; pulmonary function ranged from severe to normal, and the majority of participants were at or below their recommended weight. Exercise training did not worsen nutritional status in any study; two studies that included resistance exercise reported an increase in fat-free mass. Three of the four studies also reported increased aerobic capacity and/or muscle strength. CONCLUSIONS: Exercise training can produce positive physiologic changes in children with CF without impairing their nutritional status. In fact, resistance exercise can help improve body mass. Much less is known about how exercise may affect adults or those who are overweight.

Application of two novel electrical restitution-based ECG markers of ventricular arrhythmia to patients with nonischemic cardiomyopathy.

INTRODUCTION: Sudden cardiac death (SCD) risk assessment is limited, particularly in patients with nonischemic cardiomyopathies. This is the first application, in patients with cardiomyopathies, of two novel risk markers, regional restitution instability index (R2I2) and peak electrocardiogram restitution slope (PERS), which have been shown to be predictive of ventricular arrhythmias (VA) or death in ischemic heart disease patients. METHODS: Blinded retrospective study of 50 patients: 33 dilated cardiomyopathy and 17 other; undergoing electrophysiological study (EPS) for SCD risk stratification, and 29 controls with structurally normal hearts undergoing EPS. R2I2 was calculated from an EPS using electrocardiogram surrogates for action potential duration and diastolic interval. Cut-offs for high and low R2I2/PERS were predefined. RESULTS: R2I2 was significantly higher in study than control patients (0.99 ± 0.05 vs. 0.63 ± 0.04, p < .001). PERS showed a trend to higher values in the study group (1.18[0.63] vs. 1.09[0.54], p = .07). During median follow up of 5.6 years [interquartile range 1.9 years], nine study patients reached the endpoint of VA/death. Patients who experienced VA/death showed trends to higher mean R2I2 (1.14 ± 0.07 vs.0.95 ± 0.05, p = .12) and PERS (1.46[0.49] vs. 1.13[0.62], p = .22). A Cox proportional hazards model using grouped markers: R2I2 < 1.03 + PERS < 1.21/either R2I2 ≥ 1.03 or PERS ≥ 1.21/R2I2 ≥ 1.03 + PERS ≥ 1.21; significantly predicted VA/death (p = .02) with a hazard ratio per positive component of 3.2 (95% confidence interval 1.2-8.8). CONCLUSION: R2I2≥ 1.03 + PERS ≥ 1.21 may predict VA/death in patients with cardiomyopathies. R2I2 ≥ 1.03 + PERS ≥ 1.21 have the potential to play an important role in SCD risk stratification in cardiomyopathies but their validity should be confirmed in a larger study.

A case report: upgrade to cardiac resynchronization therapy with a blocked persistent left-sided superior vena cava.

BACKGROUND: Pacemaker-induced cardiomyopathy (PICM) can occur in up to 9% of patients having a pacemaker. Pacemaker-induced cardiomyopathy can be treated by upgrade to a biventricular pacemaker with a left ventricular (LV) lead implantation. The procedure can be technically challenging in patients with persistent left-sided superior vena cava (PLSVC). CASE SUMMARY: We report the case of a 72-year-old gentleman with a PLSVC, who had a dual-chamber pacemaker implanted 15 years ago for complete heart block. After 12 years of good health, the gentleman developed breathlessness due to PICM. At upgrade to biventricular pacemaker, his coronary sinus was found to be occluded and a collateral branch was used to successfully position an LV lead. Marked clinical improvement was seen before representation with syncope after 2 years due to simultaneous failure of both LV and right ventricular leads. Subsequently, a right-sided de novo biventricular pacemaker was implanted. In this instance, the PLSVC was beneficial because it isolated the existing leads from the new implant, thereby reducing the risk of SVC obstruction. DISCUSSION: Although implantation of pacemaker leads through a PLSVC constitutes a challenging procedure due to manoeuvring difficulties of the pacing leads into the cardiac chambers, in this particular case, the presence of PLSVC was beneficial because it meant that no leads were present in the true SVC, reducing the risk of occlusion and avoiding the need for lead extraction.

Investigation of the relationship between two novel electrocardiogram-based sudden cardiac death risk markers and autonomic function.

BACKGROUND: Regional Restitution Instability Index (R2I2) and Peak ECG Restitution Slope (PERS) are promising sudden cardiac death (SCD) risk markers. R2I2 and PERS use the standard 12­lead ECG to measure properties of electrical restitution implicated in ventricular arrhythmogenesis. We investigated the relationship between R2I2, PERS and autonomic function to inform future application of these risk markers. METHODS: Blinded, prospective, observational study of 44 patients with ischaemic cardiomyopathy undergoing risk stratification for an ICD. Patients underwent an electrophysiological study for determination of R2I2 and PERS. 24-hour ambulatory ECG monitoring was carried out for determination of time-domain heart rate variability (HRV). RESULTS: During median follow up of 22 months, 11 patients experienced ventricular arrhythmia (VA)/SCD. Weak inverse correlation was seen between R2I2 and HRV-i (rho: -0.36, p = 0.02). R2I2 and PERS were significantly higher in patients experiencing VA/SCD than those not (mean ±â€¯SEM:1.14 ±â€¯0.11 vs 0.84 ±â€¯0.05, p = 0.01) and (1.73 ±â€¯0.27 vs 1.07 ±â€¯0.08, p = 0.002) respectively. Patients with low HRV-i and high PERS had an incidence rate ratio for VA/SCD 14.5 times that of patients with high HRV-i and low PERS (p = 0.02). CONCLUSION: This small study suggests that there is minimal correlation between R2I2, PERS and autonomic function as measured by HRV. Combining PERS with HRV identified patients at particularly high risk of ventricular arrhythmia/SCD. A combined PERS+HRV risk marker may improve SCD risk stratification in patients with ischaemic cardiomyopathy.

LifeMap: towards the development of a new technology in sudden cardiac death risk stratification for clinical use.

Sudden cardiac death (SCD) is a major cause of mortality presenting a significant unmet clinical need. Patients at risk of SCD are implanted with implantable cardioverter-defibrillators (ICDs) according to international guidelines based on clinical trial evidence. Implantable cardioverter-defibrillators are not inexpensive and not without problem in terms of inappropriate shocks and infection risk. Also, only a minority of patients implanted with the ICD ever use the device during its battery lifetime highlighting the fact that methods used for SCD risk stratification are inadequate. Better ways of predicting who is at risk of SCD are needed. In addition, there is no effective prevention due to the lack of understanding of the electrical mechanisms underlying SCD. Our group has been investigating the electrophysiological basis of ventricular fibrillation and have successfully applied our preclinical findings to translational studies in patients with ischaemic cardiomyopathy. We have developed two ECG markers which have been shown to be strong predictors of ventricular arrhythmias and SCD. Ongoing clinical studies are being carried out including a multicentre UK study to consolidate the evidence base. They are being incorporated into the technology, LifeMap, with the aim to develop a successful clinical tool for the assessment of SCD risk. We hereby present the scientific data leading to the technology and the development to date. The information provided here was presented at the European Heart Rhythm Association (EHRA) Europace/Cardiostim conference at which LifeMap won the EHRA Inventors Award 2016.

Recurrent valve obstruction in a patient with a pure carbon bileaflet metallic mitral valve: a case report.

BACKGROUND: Despite overcoming the morbidity from severe native valve disease, prosthetic metallic valve replacement is not without its inherent morbidity, in particular from prosthetic valve thrombosis (PVT). The contemporary pure carbon bileaflet metallic valve confers reduced thrombogenicity. CASE SUMMARY: We describe the case of a 45-year-old woman with a pure carbon bileaflet metallic mitral valve replacement (27/29 mm On-X) 6 months previously for severe rheumatic mitral stenosis, who presented with a rapid onset of dyspnoea, paroxysmal nocturnal dyspnoea, and haemoptysis. This was preceded by an interruption in therapeutic anticoagulation. On admission the patient was in cardiogenic shock. Transthoracic and transoesophageal (TOE) echocardiograms revealed increased transmitral gradients with disc hypomobility, suggestive of PVT, unexpected given the favourable safety profile of the On-X valve. Fluoroscopy confirmed the findings. The patient was thrombolysed successfully with alteplase, with restoration of normal transmitral gradients. A target international normalized ratio of 3.5-4.5 was chosen, in addition to aspirin 75 mg, to minimize thrombotic sequalae. Repeat TOE 6 weeks later revealed disc hypomobilty with a large adherent clot. Due to the high risks from thrombolysis, emergency redo-mitral bioprosthetic valve surgery was performed, to negate the need for long-term anticoagulation. DISCUSSION: Subtherapeutic anticoagulation and the rapid development of dyspnoea, should prompt the clinician to suspect PVT. Thorough clinical examination and immediate bedside echocardiography are critical for assessing prosthetic valve patients in cardiogenic shock. The treatment of PVT is complex, with considerable risks to the patient, irrespective of the strategy (thrombolysis/emergency valve replacement), necessitating the expertise of cardiologists and cardiac surgeons.

Prospective evaluation of two novel ECG-based restitution biomarkers for prediction of sudden cardiac death risk in ischaemic cardiomyopathy.

OBJECTIVE: To improve prediction of sudden cardiac death (SCD) in patients with ischaemic cardiomyopathy (ICM). Electrical heterogeneity is known to contribute to risk of SCD. We have previously developed Regional Restitution Instability Index (R2I2), an ECG-based biomarker, which quantifies cardiac electrical instability by measuring heterogeneity in electrical restitution, and demonstrated its potential utility for risk stratification in a retrospective analysis of patients with ICM. Here, we examined R2I2 in a prospective ICM cohort and also tested the predictive value of another ECG-based biomarker, Peak ECG Restitution Slope (PERS). METHODS: Prospective, blinded, observational study of 60 patients with ICM undergoing implantable cardioverter defibrillator risk stratification. R2I2 was calculated from an electrophysiological study (EPS) using ECG surrogates for action potential duration and diastolic interval. R2I2 quantifies inter-lead electrical restitution heterogeneity. PERS was the peak restitution curve slope taken as a mean across the 12 ECG leads. Endpoints were ventricular arrhythmia (VA)/SCD. RESULTS: Over median follow-up of 22 months, 16 (26.6%) patients achieved endpoint. R2I2 was significantly higher in these patients compared with those without an event (mean ± SEM: 1.11 ± 0.09 vs 0.84 ± 0.04, p=0.003) as was PERS (median(IQR): 1.35(0.60) vs 1.08(0.52), p=0.014). R2I2≥1.03, the cut-off used in our previous study, identified patients with a significantly higher risk of VA/SCD independent of EPS result, LVEF or QRS duration with a relative risk of 6.5 (p=0.008). Patients positive for R2I2 and PERS had a relative risk of VA/SCD 21.6 times that of those negative for R2I2 and PERS (p<0.0001). CONCLUSIONS: R2I2 and PERS each independently and in combination, identify patients with ICM that are at high risk of developing ventricular arrhythmias (VA). R2I2/PERS represent promising risk markers for SCD discrimination. TRIAL REGISTRATION NUMBER: ClinicalTrials.gov Identifier: NCT01944514.

New approach for T-wave peak detection and T-wave end location in 12-lead paced ECG signals based on a mathematical model.

This paper presents an innovative approach for T-wave peak detection and subsequent T-wave end location in 12-lead paced ECG signals based on a mathematical model of a skewed Gaussian function. Following the stage of QRS segmentation, we establish search windows using a number of the earliest intervals between each QRS offset and subsequent QRS onset. Then, we compute a template based on a Gaussian-function, modified by a mathematical procedure to insert asymmetry, which models the T-wave. Cross-correlation and an approach based on the computation of Trapezium's area are used to locate, respectively, the peak and end point of each T-wave throughout the whole raw ECG signal. For evaluating purposes, we used a database of high resolution 12-lead paced ECG signals, recorded from patients with ischaemic cardiomyopathy (ICM) in the University Hospitals of Leicester NHS Trust, UK, and the well-known QT database. The average T-wave detection rates, sensitivity and positive predictivity, were both equal to 99.12%, for the first database, and, respectively, equal to 99.32% and 99.47%, for QT database. The average time errors computed for T-wave peak and T-wave end locations were, respectively, -0.38±7.12 ms and -3.70±15.46 ms, for the first database, and 1.40±8.99 ms and 2.83±15.27 ms, for QT database. The results demonstrate the accuracy, consistency and robustness of the proposed method for a wide variety of T-wave morphologies studied.

A novel surface electrocardiogram-based marker of ventricular arrhythmia risk in patients with ischemic cardiomyopathy.

BACKGROUND: Better sudden cardiac death risk markers are needed in ischemic cardiomyopathy (ICM). Increased heterogeneity of electrical restitution is an important mechanism underlying the risk of ventricular arrhythmia (VA). Our aim was to develop and test a novel quantitative surface electrocardiogram-based measure of VA risk in patients with ICM: the Regional Restitution Instability Index (R2I2). METHODS AND RESULTS: R2I2, the mean of the standard deviation of residuals from the mean gradient for each ECG lead at a range of diastolic intervals, was measured retrospectively from high-resolution 12-lead ECGs recorded during an electrophysiology study. Patient groups were as follows: Study group, 26 patients with ICM being assessed for implantable defibrillator; Control group, 29 patients with supraventricular tachycardia undergoing electrophysiology study; and Replication group, 40 further patients with ICM. R2I2 was significantly higher in the Study patients than in Controls (mean ± standard error of the mean: 1.09±0.06 versus 0.63±0.04, P<0.001). Over a median follow-up period of 23 months, 6 of 26 Study group patients had VA or death. R2I2 predicted VA or death independently of demographic factors, electrophysiology study result, left ventricular ejection fraction, or QRS duration (Cox model, P=0.029). R2I2 correlated with peri-infarct zone as assessed by cardiac magnetic resonance imaging (r=0.51, P=0.024). The findings were replicated in the Replication group: R2I2 was significantly higher in 11 of 40 Replication patients experiencing VA (1.18±0.10 versus 0.92±0.05, P=0.019). In combined analysis of ICM cohorts, R2I2 ≥1.03 identified subjects with significantly higher risk of VA or death (43%) compared with R2I2 <1.03 (11%) (P=0.004). CONCLUSIONS: In this pilot study, we have developed a novel VA risk marker, R2I2, and have shown that it correlated with a structural measure of arrhythmic risk and predicted risk of VA or death in patients with ICM. R2I2 may improve risk stratification and merits further evaluation. (J Am Heart Assoc. 2012;1:e001552 doi: 10.1161/JAHA.112.001552.).