Citations

Introduction: Accurate pre-procedure localization of the premature ventricular contractions (PVCs) is critical in preparing for the ablation. So far, no tools exist for artificial intelligence guided PVC-localization. Hypothesis: Trained deep learning model will be able to predict the site of origin (SOO) of PVCs using 12 lead electrocardiogram (ECG). Methods: We analyzed 707 ECG-ablation pairs from 2403 patients who had at least one ECG flagged for PVCs in the 6 months prior to their procedure. First, we used a cohort of 100,000 ECGs without PVCs and 10,000 ECGs with PVCs to train a convolutional neural network (CNN) to isolate the ECG segments with PVCs in them from a 12-lead ECG (Figure 1). We then trained two models, one utilizing a 10-second 12-lead ECG, and the other with just the isolated PVC beats as input. The models were trained using the Keras framework with TensorFlow (Google). Results: The mean age of the patients in the cohort was 61.2 ± 15.0 years. 66.3 percent of patients underwent left ventricle ablation, 54.6 percent had outflow ablation, and 55.3 percent had ablation of sites with leftward axis (septal location for midline and left ventricle PVCs). The model trained on 10-second ECGs yielded validation and test AUCs of 0.941 and 0.931 for outflow, 0.898 and 0.891 for ventricle, and 0.863 and 0.846 for leftward axis (Figure 1A). The model trained on isolated segments containing PVC beats yielded validation and test AUCs of 0.966 and 0.957 for outflow, 0.879 and 0.907 for ventricle, and 0.846 and 0.829 for SOO for leftward axis (Figure 1B). Conclusions: We have trained an accurate convolutional neural network that can predict SOO using the surface ECG. Further fine-tuning would be needed to refine the model and develop prediction of specific sites.

Background: The impact of computer ECG interpretation (CEI) on medical professionals’ performance is understudied. This study aims to assess the interpretation proficiency of diverse medical professionals with and without access to CEI reports. Methods: Medical professionals from diverse backgrounds interpreted 60 12-lead ECGs with urgent and non-urgent findings. The interpretation process involved two interpretation phases: (i) 30 ECGs with clinical statements and (ii) the same 30 ECGs and clinical statements in random order along with a CEI report. Diagnostic performance was evaluated based on interpretation accuracy, time per ECG, and self-reported confidence (rated 0 [not confident], 1 [somewhat confident], or 2 [confident]). Results: Among 892 participants, there were 44 (4.9%) primary care physicians, 123 (13.8%) cardiology fellows-in-training, 259 (29.0%) resident physicians, 137 (15.4%) medical students, 56 (6.3%) advanced practice providers, 82 (9.2%) nurses, and 191 (21.4%) allied health professionals. The inclusion of the CEI significantly improved interpretation accuracy by 15.1% (95% CI, 14.3 to 16.0; P<0.001), reduced interpretation time by 52 seconds (95% CI, -56 to -48; P<0.001), and increased confidence by 0.06 (95% CI, 0.03 to 0.09; P=0.003) (Table 1) . Conclusions: CEI integration improves ECG interpretation accuracy, efficiency, and confidence across medical professionals. Nevertheless, even with CEI, significant gaps in ECG interpretation proficiency remain.

Background: Accurate ECG interpretation is vital, but variations in skills exist among healthcare professionals. This study aims to identify factors contributing to ECG interpretation proficiency. Methods: Survey data and ECG test scores from participants in the EDUCATE Trial were analyzed. The dependent variable was the test score for interpreting 30 12-lead ECGs. Independent variables included non-modifiable factors (physician status, clinical experience, patient care impact) and modifiable factors (weekly interpretations, training hours, expert supervision frequency). Bivariate and multivariate analyses generated the Comprehensive Model (all factors) and Actionable Model (modifiable factors only). Results: Among 1206 participants, there were 72 (6.0%) primary care physicians, 146 (12.1%) cardiology fellows-in-training, 353 (29.3%) resident physicians, 182 (15.1%) medical students, 84 (7.0%) advanced practice providers, 120 (9.9%) nurses, and 249 (20.7%) allied health professionals. Physicians accounted for 571 (47.3%) of participants, while 453 (37.6%) were non-physicians. Bivariate analysis showed associations between test scores and multiple variables (Table 1) . In the Comprehensive Model, scores were independently associated with weekly interpretations (9.9 score increase; 95% CI 7.9-11.8; P<0.001), physician status (9.0 score increase; 95% CI 7.2-10.8; P<0.001), and training hours (5.7 score increase; 95% CI 3.7-7.6; P<0.001). In the Actionable Model, scores were independently associated with weekly interpretations (12.0 score increase; 95% CI, 10.0-14.0; P<0.001) and training hours (4.7 score increase; 95% CI 2.6-6.7; P<0.001). The Comprehensive and Actionable Models accounted for 18.7% and 12.3% of the variance in test scores, respectively. Conclusions: Being a physician is a non-modifiable predictor, while training and regular practice are modifiable predictors of ECG interpretation performance.

Background: Gaps in ECG interpretation competency among medical professionals exist, with a need for practical, evidence-based, and accessible learning solutions. We conducted an international, prospective, randomized controlled trial to assess the effectiveness of web-based, self-directed ECG learning. Methods: A total of 1206 diverse medical learners and professionals underwent an initial test, followed by random assignment to one of four groups: (i) online question-bank (questions), (ii) online lectures (lectures), (iii) online, question-bank and lectures (hybrid), or (iv) no resources (control). After 4 months, a post-test was administered to evaluate the overall change in performance. Results: Of all participants, 863 (72%) completed the trial. After the 4-month follow-up, all three online, self-directed intervention groups (questions, lectures, and hybrid) showed significant improvement in overall performance (P<0.001), while the control group did not (P=0.544). The questions, lectures, and hybrid interventions demonstrated score improvements of +11.4% (95% CI, 9.1-13.7; P<0.001), +9.8% (95% CI, 7.8-11.9; P<0.001), and +11.0% (95% CI, 9.2-12.9; P<0.001) respectively, compared to +0.8% (95% CI, -1.2-2.8; P=0.544) for the control group. These improvements were consistent across medical professional groups (Figure 1) . Conclusions: Web-based, self-directed learning resources effectively improved ECG interpretation proficiency for diverse medical learners and professionals.

Background: ECG interpretation is crucial in medical practice, but professionals struggle with achieving and maintaining competency. Identifying proficiency gaps aids in designing educational interventions. Methods: Medical professionals from various disciplines interpreted 30 12-lead ECGs containing commonly taught urgent and non-urgent findings. Performance metrics evaluated were overall adjusted score (% of correctly identified findings using a point-value score based on clinical relevance), unadjusted score (% of correctly identified findings), interpretation time per ECG, and self-reported confidence (rated on an ordinal scale of 0 [not confident], 1 [somewhat confident], or 2 [confident]). Results: Among 1206 participants, there were 72 (6%) primary care physicians (PCPs), 146 (12%) cardiology fellows-in-training (FIT), 353 (29%) resident physicians, 182 (15%) medical students, 84 (7%) advanced practice providers (APPs), 120 (10%) nurses, 249 (21%) allied health professionals, 571 (47%) physicians, and 453 (38%) non-physicians. Participants had a mean adjusted score of 46.9% (± 15.9%), unadjusted score of 56.4% (± 17.2%), interpretation time of 142 seconds (± 67 seconds), and confidence of 0.83 (± 0.53) (Table 1) . Performance varied across groups; cardiology FIT had superior performance in all metrics. Physicians generally outperformed non-physicians, with higher overall adjusted score (52% vs. 43%; p<0.01), unadjusted score (62% vs. 52%; p<0.01), and confidence (0.91 vs. 0.80; p<0.01). Unadjusted score for PCPs was higher than nurses and APPs (58% vs. 47% and 51%; p<0.01) but lower than resident physicians (58% vs. 60%; p<0.01). Allied health professionals outperformed nurses and APPs and closely matched resident physicians and PCPs. Conclusions: Significant gaps in ECG interpretation proficiency exist, emphasizing the need for comprehensive, scalable, and accessible educational tools.

High dose chemotherapy followed by autologous hematopoietic cell transplantation (auto HCT) is a standard therapy for patients with plasma cell disorders (PCD). A barrier to improving candidate selection for auto HCT is appraisal of risk for transplant related complications. Electrocardiography (ECG) is a standard pre-HCT test to evaluate the potential for arrhythmia and structural heart disease. We hypothesized that an artificial intelligence (AI) based interpretation of ECG waveform data (AI-ECG) could enhance risk prediction for cardiac-specific and other outcomes following auto HCT for patients with PCD. To test this hypothesis, we used several AI-ECG tools developed using a deep-learning based approach that were trained to predict comorbidity (Attia et al, Lancet 2019 & Grogan et al, Mayo Clin Proc 2021). Model outputs include a scaled prediction value indicating increasing risk for atrial fibrillation/flutter (AF), reduced ejection fraction (EF), or cardiac amyloidosis. We evaluated this tool in 9,225 ECGs obtained from 849 patients (median 6 ECGs/patient, inter-quartile range: 3-12) who underwent first or subsequent auto HCT at Memorial Sloan Kettering Cancer Center between 2016-2022 and who had a diagnosis of multiple myeloma (N = 810), amyloidosis (N = 37), or other PCD (N = 2). The endpoints were to determine whether baseline AI-ECG prediction of AF risk correlated with peri-transplant AF events, and to explore whether model outputs correlated with survival and other cardiac clinical outcomes. The mean age of patients was 61.0 years (range 23.7 - 81.2) and median follow up was 2.7 years post auto HCT. Contributing cardiovascular risk factors included a prior history of AF (N = 101, 11.9%), coronary or cerebrovascular disease (N = 95, 11.2%), and diabetes mellitus (N = 98, 11.5%). Reduced dose melphalan (&lt; 200 mg/m 2) was administered in 26%. The HCT-comorbidity index (HCT-CI) was high risk (≥3) in 52.7%. The one-year overall survival and progression-free survival in the whole cohort was 94% (95% confidence interval (95% CI): 92 - 95%) and 86% (84 - 89%), respectively. Among patients not in AF at pre HCT evaluation, the 90-day cumulative incidence of AF was 5.6% (4.2 - 7.3%). In univariable competing risk regression, baseline pre HCT AI-ECG appraisal of AF risk was strongly correlated with incidence AF in patients who were in sinus rhythm before HCT (hazard ratio (HR): 25.1, 95% CI: 11.7 - 53.7, p &lt; 0.001). In a multivariable model adjusting for patient age and HCT-CI, AI-ECG AF predicted risk remained associated with incidence of post HCT AF (HR: 13.4, 95% CI: 5.9 - 30.4, p &lt; 0.001). Considering a decision threshold identified during original model training, AI-ECG predicted AF events with an accuracy of 75% (Mcnemar's p &lt; 2 x 10 -16). The precision, recall, and F1-score were 0.94, 0.77, and 0.85, respectively. AI-ECG predicted age correlated highly with actual patient age (Spearman rho = 0.59, p &lt; 2 x 10 -16). Amyloid risk scores were highly correlated with a known clinical history of amyloidosis (median 0.45 versus 0.05, Wilcoxon rank-sum p &lt; 0.001). Patients with EF &lt; 50% had greater prediction scores for cardiomyopathy compared to patients with normal EF (median 0.23 versus 0.01, Wilcoxon rank-sum p &lt;0.001). In univariable analysis, increasing AI-ECG predicted age (HR 1.02, CI: 1.003 - 1.04, p = 0.02) and predicted AF risk score (HR: 3.2, CI: 1.4 - 7.2, p = 0.006) were associated with risk for death, whereas predicted amyloid score and predicted low EF were not associated with decreased survival. In multivariable analysis of all patients that considered patient age, hematopoietic comorbidity index, and remission status at HCT, increasing predicted AF score trended towards association with increased risk for death (HR: 2.3, 95%CI: 1.0 - 5.5, p = 0.059). Predicted AF score correlated highly with the HCT-CI and patient age. Finally, we examined whether risk prediction for low EF correlated with peri-HCT fluid overload (FO) defined by absolute weight gain. The incidence of ≥ 10% weight gain (grade 2 FO) and ≥ 20% weight gain (grade 3 FO) at 180 days post HCT was 7.4% (95% CI: 5.8%, 9.3%) and 1.0% (95% CI: 0.5%, 1.8%), respectively. Predicted low EF was not associated with development of either grade 2 or grade 3 FO in this population. In summary, AI-ECG predicts episodes of AF in patients undergoing auto HCT for PCD and may augment transplant risk appraisal when considered in conjunction with other measures of comorbidity.

Abstract Background Atrial functional mitral regurgitation (AFMR) is an increasingly recognized cause of functional MR thought to result from left atrial enlargement and myopathy in the setting of preserved left ventricular geometry and function. AFMR mostly occurs in patients with atrial fibrillation (AF) and/or heart failure with preserved ejection fraction. Purpose We investigated whether the use of an artificial intelligence-based electrocardiogram (AI-ECG) can identify patients with new-onset AF and sinus rhythm (SR) who are at highest risk of developing AFMR. Methods Adults with a new diagnosis of AF between 2010 and 2021 who had a transthoracic echocardiogram (TTE) within 1 month after AF diagnosis were identified. Similarly, adults in SR who never had AF diagnosis and had a TTE during the same period were identified. Exclusion criteria from both groups included &gt;mild MR at baseline, primary mitral valve disease, cardiomyopathy, ≥moderate aortic valve disease, current or previous low ejection fraction, enlarged left ventricle, cardiac devices, previous cardiac surgeries, and the absence of follow-up TTE ≥6 months from baseline. Only patients who had ECG within 10 days of the TTE were included. When multiple ECGs were available, the closest was considered. A previously trained AI-ECG algorithm to detect the probability of AF was utilized. Incident AFMR was defined as &gt;mild functional MR in the absence of &gt;mild left ventricle enlargement or dysfunction. Diastolic dysfunction was defined according to ASE/EACVI guidelines. Results Overall, 1,570 patients with new-onset AF (median age 67 years; 34% females) and 16,802 patients in SR (median age 59 years; 46% females) were included. The median AI ECG probability for AF was 43% (IQR 26%-59%) in the new-onset AF group and 2% (0.5%-7%) in the SR group, Figure 1. Incident AFMR occurred in 152 patients in the AF group over median 3.06 (IQR 1.5-5.2) years of follow-up and in 482 patients in the SR group over median 3.0 (1.5-5.6) years of follow-up. The probability of silent AF by the AI-ECG was associated with incident AFMR in the new-onset AF group [hazard ratio 1.19 (95% CI 1.11-1.28) per 10%, p &lt;.001] and the SR group [hazard ratio 1.19 (95% CI 1.12-1.25) per 10%, p &lt;.001]. The AI ECG-derived AF probability remained as an independent risk factor for AFMR in multivariable analysis in both AF and SR, Table 1. Conclusion The AI ECG for AF might be able to reflect the degree of left atrial myopathy and thus the risk of incident AFMR in patients with AF and SR.Distribution of AI ECG probability of AFMultivariable risk factors for AFMR

Abstract Background Cardiomyopathy is a leading cause of pregnancy-related mortality and represents the number one cause of death in the late postpartum period. Purpose The goal of this study was to prospectively evaluate the effectiveness of an artificial intelligence-based electrocardiogram (AI-ECG) and a point-of-care artificial intelligence-enhanced digital stethoscope for cardiomyopathy screening among pregnant and postpartum women. Methods We conducted a single-arm prospective study of pregnant and postpartum women (up to 12 months following delivery) enrolled at three sites between October 2021 and October 2022. Study participants completed a standard 12-lead ECG, ECG + phonocardiogram recordings with a digital stethoscope, a demographic questionnaire, and a transthoracic echocardiogram (TTE) within a 24-hour period. Diagnostic performance of the AI-ECG and digital stethoscope for cardiomyopathy detection, defined as left ventricular ejection fraction (LVEF) &lt; 45%, was evaluated using the area under the curve (AUC) and standard measures of diagnostic performance. Results We included 100 consecutive pregnant and postpartum women who provided informed consent and had at least one ECG recorded, completed the questionnaire, and had an echocardiogram performed. All participants self-identified as female, and 41% were recruited from a federally qualified health center. The median age of participants was 31 years (Q1: 27, Q3: 34). Thirty-eight percent identified as non-Hispanic White, 32% as non-Hispanic Black, 21% as Hispanic or Latino, 6% as Asian, and 3% as Other (1 Native Hawaiian and 2 multiracial). Five percent (5/100) had LVEF &lt;45%. The 12-lead AI-ECG model demonstrated close to perfect discrimination (accuracy 99%; AUC 1.00, sensitivity 100%) for detection of cardiomyopathy with 1 false positive screen (LVEF 49%), Figure 1A. The digital stethoscope (angled position) had an AUC of 0.99 (95% CI: 0.97, 1.00) for detection of LVEF &lt;45%. Sensitivity, specificity, positive predictive value, and negative predictive value for the digital stethoscope were 60%, 98%, 60% and 98%, respectively for detection of LVEF &lt;45% (Figure 1B & 1C). Conclusions In this small prospective study, we found that 5% of pregnant and postpartum women had cardiomyopathy, a potentially life-threatening condition, and that a rapid, inexpensive, ubiquitous, massively scalable point-of-care test - the ECG - showed promising results, identifying it with an AUC &gt; 0.98 with the application of artificial intelligence.Figure 1

Abstract Background Premature ventricular contractions (PVCs) are the most common arrhythmia affecting the ventricle. Frequent PVCs are associated with the development of PVC-cardiomyopathy (CM). It is known that the PVC burden has a direct correlation to the severity of left ventricular (LV) dysfunction. However, only 5% of patients with frequent PVCs (&gt;10,000 per day) have associated PVC-CM. Thus, selecting patients for therapy options with drugs or catheter ablation is challenging, especially, when patients are asymptomatic with the PVCs. Purpose To train a deep learning model to predict the probability of low LV ejection fraction (LVEF) in patients with PVCs on a 12-lead electrocardiogram (ECG). Methods Using 69,440 paired 12-lead ECGs (containing at least one PVC), and echocardiogram data, including the left ventricular ejection fraction at Mayo Clinic, we trained a convolutional neural network to identify patients with ventricular dysfunction, defined as ejection fraction ≤35%, using the ECG data alone. ECGs were acquired at a sampling rate of 500 Hz using a Marquette ECG machine (GE Healthcare) and stored using the MUSE data management system for later retrieval. Comprehensive 2-D or 3-D and Doppler echocardiography were available for all patients. We trained a convolutional neural network using the Keras framework with TensorFlow (Google). The ECG-echocardiogram pairs were divided into a 60:20:20 training:validation:test split. Results The mean age of the patients used to train the model was 69.9 ± 13.0 years, and 65.2 percent of patients were male. The mean LVEF for this population was 52 ± 14.9 %, and the proportion of patients with low LVEF was 16.8 percent. The model yielded values for area under the curve, sensitivity, specificity, and accuracy of 0.897, 0.830, 0.813, and 0.816, respectively (Figure 1). A representative example of model prediction of normal LVEF and low LVEF is shown in Figure 2. Conclusion Using large-scale data, we have trained a highly accurate convolutional neural network with excellent predictive accuracy for LV systolic dysfunction using the surface ECG. Such a tool would be invaluable for cardiologists when risk-stratifying patients with frequent PVCs, especially those that are asymptomatic. Patients at risk will benefit from early referral to electrophysiologists for drug therapy and catheter ablation to mitigate the risk of PVC-CM.Figure 1Figure 2