Citations

Abstract Aims An artificial intelligence-enabled electrocardiogram (AI-ECG) is a promising tool to detect patients with aortic stenosis (AS) before developing symptoms. However, functional, structural, or haemodynamic components reflected in AI-ECG responsible for its detection are unknown. Methods and results The AI-ECG model that was developed at Mayo Clinic using a convolutional neural network to identify patients with moderate–severe AS was applied. In patients used as the testing group, the correlation between the AI-ECG probability of AS and echocardiographic parameters was investigated. This study included 102 926 patients (63.0 ± 16.3 years, 52% male), and 28 464 (27.7%) were identified as AS positive by AI-ECG. Older age, atrial fibrillation, hypertension, diabetes, coronary artery disease, and heart failure were more common in the positive AI-ECG group than in the negative group (P < 0.001). The AI-ECG was correlated with aortic valve area (ρ = −0.48, R2 = 0.20), peak velocity (ρ = 0.22, R2 = 0.08), and mean pressure gradient (ρ = 0.35, R2 = 0.08). The AI-ECG also correlated with left ventricular (LV) mass index (ρ = 0.36, R2 = 0.13), E/e′ (ρ = 0.36, R2 = 0.12), and left atrium volume index (ρ = 0.42, R2 = 0.12). Neither LV ejection fraction nor stroke volume index had a significant correlation with the AI-ECG. Age correlated with the AI-ECG (ρ = 0.46, R2 = 0.22) and its correlation with echocardiography parameters was similar to that of the AI-ECG. Conclusion A combination of AS severity, diastolic dysfunction, and LV hypertrophy is reflected in the AI-ECG to detect AS. There seems to be a gradation of the cardiac anatomical/functional features in the model and its identification process of AS is multifactorial.

Abstract Aims Current non-invasive screening methods for cardiac allograft rejection have shown limited discrimination and are yet to be broadly integrated into heart transplant care. Given electrocardiogram (ECG) changes have been reported with severe cardiac allograft rejection, this study aimed to develop a deep-learning model, a form of artificial intelligence, to detect allograft rejection using the 12-lead ECG (AI-ECG). Methods and results Heart transplant recipients were identified across three Mayo Clinic sites between 1998 and 2021. Twelve-lead digital ECG data and endomyocardial biopsy results were extracted from medical records. Allograft rejection was defined as moderate or severe acute cellular rejection (ACR) based on International Society for Heart and Lung Transplantation guidelines. The extracted data (7590 unique ECG-biopsy pairs, belonging to 1427 patients) was partitioned into training (80%), validation (10%), and test sets (10%) such that each patient was included in only one partition. Model performance metrics were based on the test set (n = 140 patients; 758 ECG-biopsy pairs). The AI-ECG detected ACR with an area under the receiver operating curve (AUC) of 0.84 [95% confidence interval (CI): 0.78–0.90] and 95% (19/20; 95% CI: 75–100%) sensitivity. A prospective proof-of-concept screening study (n = 56; 97 ECG-biopsy pairs) showed the AI-ECG detected ACR with AUC = 0.78 (95% CI: 0.61–0.96) and 100% (2/2; 95% CI: 16–100%) sensitivity. Conclusion An AI-ECG model is effective for detection of moderate-to-severe ACR in heart transplant recipients. Our findings could improve transplant care by providing a rapid, non-invasive, and potentially remote screening option for cardiac allograft function.

Abstract Aims Our goal was to evaluate a previously validated artificial intelligence–augmented electrocardiography (AI-ECG) screening tool for left ventricular systolic dysfunction (LVSD) in patients undergoing high-sensitivity-cardiac troponin T (hs-cTnT). Methods and results Retrospective application of AI-ECG for LVSD in emergency department (ED) patients undergoing hs-cTnT. AI-ECG scores (0–1) for probability of LVSD (left ventricular ejection fraction ≤ 35%) were obtained. An AI-ECG score ≥0.256 indicates a positive screen. The primary endpoint was a composite of post-discharge major adverse cardiovascular events (MACEs) at two years follow-up. Among 1977 patients, 248 (13%) had a positive AI-ECG. When compared with patients with a negative AI-ECG, those with a positive AI-ECG had a higher risk for MACE [48 vs. 21%, P < 0.0001, adjusted hazard ratio (HR) 1.39, 95% confidence interval (CI) 1.11–1.75]. This was largely because of a higher rate of deaths (32 vs. 14%, P < 0.0001; adjusted HR 1.26, 95% 0.95–1.66) and heart failure hospitalizations (26 vs. 6.1%, P < 0.001; adjusted HR 1.75, 95% CI 1.25–2.45). Together, hs-cTnT and AI-ECG resulted in the following MACE rates and adjusted HRs: hs-cTnT < 99th percentile and negative AI-ECG: 116/1176 (11%; reference), hs-cTnT < 99th percentile and positive AI-ECG: 28/107 (26%; adjusted HR 1.54, 95% CI 1.01–2.36), hs-cTnT > 99th percentile and negative AI-ECG: 233/553 (42%; adjusted HR 2.12, 95% CI 1.66, 2.70), and hs-cTnT > 99th percentile and positive AI-ECG: 91/141 (65%; adjusted HR 2.83, 95% CI 2.06, 3.87). Conclusion Among ED patients evaluated with hs-cTnT, a positive AI-ECG for LVSD identifies patients at high risk for MACE. The conjoint use of hs-cTnT and AI-ECG facilitates risk stratification.

Abstract Background Artificial intelligence-augmented electrocardiogram (AI-ECG) algorithms have been developed from the standard 12-lead ECG and validated for the recognition of left ventricular systolic dysfunction (LVSD), defined as LV ejection fraction (LVEF)≤35%. Whether AI-ECG facilitates identification of LVSD and is associated with adverse outcomes in emergency department (ED) patients undergoing high-sensitivity cardiac troponin (hs-cTnT) testing is uncertain. Purpose To investigate the diagnostic and prognostic performance of AI-ECG in ED patients undergoing hs-cTnT measurement. Methods Observational US cohort study of ED patients undergoing hs-cTnT measurement. Cases with hs-cTnT increases >sex-specific 99th percentiles were adjudicated following the Fourth Universal Definition of Myocardial Infarction (MI). Post-discharge major adverse cardiac events (MACE) included death, MI, heart failure (HF) hospitalization, stroke or transient ischemic attack, and new onset atrial fibrillation/flutter during 2-years follow-up. The AI-ECG network output, which is a continuous number between 0–1, that provides a probability of LVSD, was obtained for each patient from the first ECG during the index presentation. An AI-ECG threshold of ≥0.256 indicates a positive screen that correlates with a high probability of LVSD. Results Among 1977 patients, 1729 (87%) had a negative AI-ECG screen, while 248 (13%) had a positive AI-ECG screen. Patients with a positive AI-ECG screen were older and had more comorbidities. As compared to patients with hs-cTnT≤99th percentile in whom AI-ECG was positive in 5.8%, those with hs-cTnT>99th percentile had a positive AI-ECG in 22% of cases (p<0.0001). Based on adjudicated diagnoses, the frequency of a positive AI-ECG was 20% in myocardial injury, 38% in type 1 MI, and 20% in type 2 MI. At 2-years follow-up, as compared to patients with a negative AI-ECG, those with a positive AI-ECG had a higher risk for MACE (48% vs. 21%, p<0.0001, adjusted HR 1.39, 95% CI 1.11–1.75) (Figure 1), mainly because of more deaths (43% vs. 30%, p=0.004) and HF hospitalizations (36% vs. 13%, p<0.0001). A positive AI-ECG was associated with a higher risk for MACE (60% vs. 41%, p<0.0001, adjusted HR 1.30, 95% CI 1.02–1.64) in those with hs-cTnT increases >99th percentile, but not in those without hs-cTnT increases. Among patients with an echocardiogram during index presentation or within 30-days (n=452), the diagnostic accuracy of AI-ECG for LVEF ≤35% was 81.4% (95% CI 77.5, 84.9) with a negative predictive value of 96.5% (95% CI 94.0, 98.2). A normal LVEF (>50%) was observed in 87% of those with a negative AI ECG, whereas in those with a positive AI-ECG LVEF was reduced (<50%) in 60%. Conclusions Among ED patients evaluated with hs-cTnT, a positive AI-ECG screen for LVSD identifies patients at high risk of MACE. These findings are largely because of more deaths and HF hospitalizations in those with hs-cTnT increases >sex-specific 99th percentiles. Funding Acknowledgement Type of funding sources: None.

AbstractAimsAn artificial intelligence algorithm detecting age from 12-lead electrocardiogram (ECG) has been suggested to reflect ‘physiological age’. An increased physiological age has been associated with a higher risk of cardiac mortality in the non-transplant population. We aimed to investigate the utility of this algorithm in patients who underwent heart transplantation (HTx).Methods and resultsA total of 540 patients were studied. The average ECG ages within 1 year before and after HTx were used to represent pre- and post-HTx ECG ages. Major adverse cardiovascular event (MACE) was defined as any coronary revascularization, heart failure hospitalization, re-transplantation, and mortality. Recipient pre-transplant ECG age (mean 63 ± 11 years) correlated significantly with recipient chronological age (mean 49 ± 14 years, R = 0.63, P < 0.0001), while post-transplant ECG age (mean 54 ± 10 years) correlated with both the donor (mean 32 ± 13 years, R = 0.45, P < 0.0001) and the recipient ages (R = 0.38, P < 0.0001). During a median follow-up of 8.8 years, 307 patients experienced MACE. Patients with an increase in ECG age post-transplant showed an increased risk of MACE [hazard ratio (HR): 1.58, 95% confidence interval (CI): (1.24, 2.01), P = 0.0002], even after adjusting for potential confounders [HR: 1.58, 95% CI: (1.19, 2.10), P = 0.002].ConclusionElectrocardiogram age-derived cardiac ageing after transplantation is associated with a higher risk of MACE. This study suggests that physiological age change of the heart might be an important determinant of MACE risk post-HTx.

Abstract Artificial intelligence (AI)-based approaches can now use electrocardiograms (ECGs) to provide expert-level performance in detecting heart abnormalities and diagnosing disease. Additionally, patient age predicted from ECGs by AI models has shown great potential as a biomarker for cardiovascular age as recent work has found differences from chronological age ("delta age") to be associated with mortality and co-morbidities. However, the genetic underpinning of delta age is unknown, but crucial for understanding underlying individual risk. By performing a genome-wide association study using UK Biobank data (n=34,432), we identified eight loci associated with delta age (p ≤ 5 × 10−8 ), including genes linked to cardiovascular disease (CVD) (e.g. SCN5A) and (heart) muscle development (e.g. TTN). Our results indicate that the genetic basis of cardiovascular ageing is predominantly determined by genes directly involved with the cardiovascular system rather than those connected to more general mechanisms of ageing. Our insights inform the epidemiology of CVD, with implications for preventative and precision medicine.

Abstract Objective To compare the artificial intelligence‐enabled electrocardiogram (AI‐ECG) atrial fibrillation (AF) prediction model output in patients with migraine with aura (MwA) and migraine without aura (MwoA). Background MwA is associated with an approximately twofold risk of ischemic stroke. Longitudinal cohort studies showed that patients with MwA have a higher incidence of developing AF compared to those with MwoA. The Mayo Clinic Cardiology team developed an AI‐ECG algorithm that calculates the probability of concurrent paroxysmal or impending AF in ECGs with normal sinus rhythm (NSR). Methods Adult patients with an MwA or MwoA diagnosis and at least one NSR ECG within the past 20 years at Mayo Clinic were identified. Patients with an ECG‐confirmed diagnosis of AF were excluded. For each patient, the ECG with the highest AF prediction model output was used as the index ECG. Comparisons between MwA and MwoA were conducted in the overall group (including men and women of all ages), women only, and men only in each age range (18 to <35, 35 to <55, 55 to <75, ≥75 years), and adjusted for age, sex, and six common vascular comorbidities that increase risk for AF. Results The final analysis of our cross‐sectional study included 40,002 patients (17,840 with MwA, 22,162 with MwoA). The mean (SD) age at the index ECG was 48.2 (16.0) years for MwA and 45.9 (15.0) years for MwoA ( p < 0.001). The AF prediction model output was significantly higher in the MwA group compared to MwoA (mean [SD] 7.3% [15.0%] vs. 5.6% [12.4%], mean difference [95% CI] 1.7% [1.5%, 2.0%], p < 0.001). After adjusting for vascular comorbidities, the difference between MwA and MwoA remained significant in the overall group (least square means of difference [95% CI] 0.7% [0.4%, 0.9%], p < 0.001), 18 to <35 (0.4% [0.1%, 0.7%], p = 0.022), and 35 to <55 (0.5% [0.2%, 0.8%], p < 0.001), women of all ages (0.6% [0.3%, 0.8%], p < 0.001), men of all ages (1.0% [0.4%, 1.6%], p = 0.002), women 35 to <55 (0.6% [0.3%, 0.9%], p < 0.001), and men 18 to <35 (1.2% [0.3%, 2.1%], p = 0.008). Conclusions Utilizing a novel AI‐ECG algorithm on a large group of patients, we demonstrated that patients with MwA have a significantly higher AF prediction model output, implying a higher probability of concurrent paroxysmal or impending AF, compared to MwoA in both women and men. Our results suggest that MwA is an independent risk factor for AF, especially in patients <55 years old, and that AF‐mediated cardioembolism may play a role in the migraine–stroke association for some patients.

Abstract Aims Electrocardiogram (ECG)-enabled stethoscope (ECG-Scope) acquires a single-lead ECGs during cardiac auscultation and may facilitate real-time screening for pathologies not routinely identified by cardiac auscultation alone. We previously demonstrated an artificial intelligence (AI) algorithm can identify left ventricular dysfunction (LVSD) [defined as ejection fraction (EF) ≤ 40%] with an area under the curve (AUC) of 0.91 using a 12-lead ECG. Methods and results One hundred patients referred for clinically indicated echocardiography were prospectively recruited. ECG-Scope recordings with the patient supine and sitting were obtained in multiple electrode locations at the time of the echocardiogram. The AI algorithm for the detection of LVSD was retrained using single leads from ECG-12 and validated against ECG-Scope to determine accuracy for low EF detection (≤35%, &lt;40%, or &lt;50%). We evaluated the algorithm with respect to body position and lead location. Amongst 100 patients (aged 61.3 ± 13.8; 61% male, BMI: 30.0 ± 5.4), eight had EF≤40%, and six had EF 40–50%. The best single recording position was V2 with the patient supine [AUC: 0.88 (CI: 0.80–0.97) for EF≤35%, 0.85 (CI: 0.75–0.95) for EF≤40%, and 0.81 (CI: 0.71–0.90) for EF &lt; 50%]. When using an AI model to select the recording automatically, AUC was 0.91 (CI: 0.84–0.97) for EF≤35%, 0.89 (CI: 0.83–0.96) for EF≤40%, and 0.84 (CI: 0.73–0.94) for EF &lt; 50%. Conclusion An AI algorithm applied to an ECG-enabled stethoscope recording in standard auscultation positions reliably detected the presence of a low EF in this prospective study of patients referred for echocardiography. The ability to screen patients with a possible low EF during routine physical examination may facilitate rapid detection of LVSD.

Abstract Aims Some artificial intelligence models applied in medical practice require ongoing retraining, introduce unintended racial bias, or have variable performance among different subgroups of patients. We assessed the real-world performance of the artificial intelligence-enhanced electrocardiogram to detect left ventricular systolic dysfunction with respect to multiple patient and electrocardiogram variables to determine the algorithm’s long-term efficacy and potential bias in the absence of retraining. Methods and results Electrocardiograms acquired in 2019 at Mayo Clinic in Minnesota, Arizona, and Florida with an echocardiogram performed within 14 days were analyzed (n = 44 986 unique patients). The area under the curve (AUC) was calculated to evaluate performance of the algorithm among age groups, racial and ethnic groups, patient encounter location, electrocardiogram features, and over time. The artificial intelligence-enhanced electrocardiogram to detect left ventricular systolic dysfunction had an AUC of 0.903 for the total cohort. Time series analysis of the model validated its temporal stability. Areas under the curve were similar for all racial and ethnic groups (0.90–0.92) with minimal performance difference between sexes. Patients with a ‘normal sinus rhythm’ electrocardiogram (n = 37 047) exhibited an AUC of 0.91. All other electrocardiogram features had areas under the curve between 0.79 and 0.91, with the lowest performance occurring in the left bundle branch block group (0.79). Conclusion The artificial intelligence-enhanced electrocardiogram to detect left ventricular systolic dysfunction is stable over time in the absence of retraining and robust with respect to multiple variables including time, patient race, and electrocardiogram features.