A case of progressive right ventricular failure with ventricular arrhythmia and aortic insufficiency after implantable left ventricular assist device implantation.

Right ventricular failure (RVF) is a serious complication after left ventricular assist device (LVAD) implantation. In this report, a case of RVF that developed over two years after LVAD implantation is presented. The patient was a 12-year-old male with dilated phase of hypertrophic cardiomyopathy. He had no risk factors for early or late-onset RVF. However, his right ventricular function worsened after he developed ventricular arrhythmia (VA), and right ventricular dysfunction became exacerbated with an increasing frequency of VAs. He also developed moderate aortic insufficiency (AI), which became severe. Two years after implantation, he was admitted for treatment of recurrent ventricular tachycardia and became inotropic-dependent during hospitalization. Finally, he underwent successful heart transplantation 2 years and 9 months after LVAD implantation. This case suggests that vicious cycle of RV dysfunction, recurrent VAs and severe AI could lead to RVF in patients without known risk factors for RVF, even long after LVAD implantation. Learning objective: This report shows a progressive right ventricular failure (RVF) two years after left ventricular assist device (LVAD) implantation. Although the patient had no known risk factor, vicious circle of RV dysfunction, ventricular arrhythmias (VAs) and aortic insufficiency (AI) lead to RVF. Patients with LVAD as destination therapy will increase and require long-term LVAD management. We should recognize that these patients could develop RVF even years after LVAD implantation in association with VAs and AI.

Shared and Reciprocal Mechanisms Between Heart Failure and Cancer　- An Emerging Concept of Heart-Cancer Axis.

Epidemiological evidence of increased risks of cancer in heart failure (HF) patients and HF in cancer patients has suggested close relationships between the pathogenesis of both diseases. Indeed, HF and cancer share common risk factors, including aging and unhealthy lifestyles, and underlying mechanisms, including activation of the sympathetic nervous system and renin-angiotensin-aldosterone system, chronic inflammation, and clonal hematopoiesis of indeterminate potential. Mechanistically, HF accelerates cancer development and progression via secreted factors, so-called cardiokines, and epigenetic remodeling of bone marrow cells into an immunosuppressive phenotype. Reciprocally, cancer promotes HF via cachexia-related wasting and metabolic remodeling in the heart, and possibly via cancer-derived extracellular vesicles influencing myocardial structure and function. The novel concept of the "heart-cancer axis" will help in our understanding of the shared and reciprocal relationships between HF and cancer, and provide innovative diagnostic and therapeutic approaches for both diseases.

Deep Learning Models for Predicting Left Heart Abnormalities From Single-Lead Electrocardiogram for the Development of Wearable Devices.

BACKGROUND: Left heart abnormalities are risk factors for heart failure. However, echocardiography is not always available. Electrocardiograms (ECGs), which are now available from wearable devices, have the potential to detect these abnormalities. Nevertheless, whether a model can detect left heart abnormalities from single Lead I ECG data remains unclear.Methods and Results: We developed Lead I ECG models to detect low ejection fraction (EF), wall motion abnormality, left ventricular hypertrophy (LVH), left ventricular dilatation, and left atrial dilatation. We used a dataset comprising 229,439 paired sets of ECG and echocardiography data from 8 facilities, and validated the model using external verification with data from 2 facilities. The area under the receiver operating characteristic curves of our model was 0.913 for low EF, 0.832 for wall motion abnormality, 0.797 for LVH, 0.838 for left ventricular dilatation, and 0.802 for left atrial dilatation. In interpretation tests with 12 cardiologists, the accuracy of the model was 78.3% for low EF and 68.3% for LVH. Compared with cardiologists who read the 12-lead ECGs, the model's performance was superior for LVH and similar for low EF. CONCLUSIONS: From a multicenter study dataset, we developed models to predict left heart abnormalities using Lead I on the ECG. The Lead I ECG models show superior or equivalent performance to cardiologists using 12-lead ECGs.

PGC-1α-mediated angiogenesis prevents pulmonary hypertension in mice.

Pulmonary hypertension (PH) is a life-threatening disease characterized by a progressive narrowing of pulmonary arterioles. Although VEGF is highly expressed in lung of patients with PH and in animal PH models, the involvement of angiogenesis remains elusive. To clarify the pathophysiological function of angiogenesis in PH, we compared the angiogenic response in hypoxia (Hx) and SU5416 (a VEGFR2 inhibitor) plus Hx (SuHx) mouse PH models using 3D imaging. The 3D imaging analysis revealed an angiogenic response in the lung of the Hx-PH, but not of the severer SuHx-PH model. Selective VEGFR2 inhibition with cabozantinib plus Hx in mice also suppressed angiogenic response and exacerbated Hx-PH to the same extent as SuHx. Expression of endothelial proliferator-activated receptor γ coactivator 1α (PGC-1α) increased along with angiogenesis in lung of Hx-PH but not SuHx mice. In pulmonary endothelial cell-specific Ppargc1a-KO mice, the Hx-induced angiogenesis was suppressed, and PH was exacerbated along with increased oxidative stress, cellular senescence, and DNA damage. By contrast, treatment with baicalin, a flavonoid enhancing PGC-1α activity in endothelial cells, ameliorated Hx-PH with increased Vegfa expression and angiogenesis. Pulmonary endothelial PGC-1α-mediated angiogenesis is essential for adaptive responses to Hx and might represent a potential therapeutic target for PH.

Coronary Periarteritis and Pericardial Thickening Could Be Predictors for Coronary Artery Events Complicated by IgG4-Related Disease.

Background: IgG4-related disease (IgG4-RD) is a systemic disease characterized by serum IgG4 upregulation, massive infiltration of IgG4-positive plasma cells, and storiform fibrosis, which results in nodules or thickening of the involved organs. Cardiologists have recently recognized that IgG4-RD can be complicated by coronary artery events (CAEs); however, the mechanisms and clinical characteristics of this phenomenon are unknown. We evaluated the clinical signs of patients with coronary periarteritis (CP), aortic periarteritis (AP), and pericardial thickening, which are complications of IgG4-RD, to determine the contributing factors. Methods: We retrospectively examined 19 patients with IgG4-RD who attended or consulted a cardiologist in our department at the University of Tokyo Hospital between January 1, 2004, and December 31, 2021. Results: The frequency of CAEs was significantly higher in the CP group than in the non-CP group. Furthermore, the CP group had significantly lower event-free survival than the non-CP group (log-rank test, P = 0.008). However, the frequency of incidents and event-free survival for CAEs after IgG4-RD diagnosis did not differ significantly between the AP and non-AP groups. Although no statistically significant difference was present between the frequency of CAEs for those with vs without pericardial thickening, the group with pericardial thickening had significantly worse event-free survival than the group without pericardial thickening (log-rank test, P = 0.017). Conclusions: The incidence and clinical course of CAEs complicated by IgG4-RD could be predicted by identifying CP and pericardial thickening in IgG4-RD but not AP.

Contexte: La maladie liée aux immunoglobulines de type G4 (ML-IgG4) est une maladie généralisée caractérisée par une augmentation du taux sérique d'IgG4, par une infiltration massive de plasmocytes exprimant les IgG4 et par une fibrose storiforme, qui produit des nodules ou un épaississement des organes touchés. Les cardiologues ont récemment reconnu que la ML-IgG4 peut être compliquée par des événements coronariens; les mécanismes et caractéristiques cliniques de ce phénomène demeurent cependant inconnus. Nous avons évalué les signes cliniques chez des patients atteints de périartérite coronarienne (PC), de périaortite (PA) et d'épaississement du péricarde, des complications de la ML-IgG4, pour tenter d'établir les facteurs contributifs. Méthodologie: Nous avons examiné de manière rétrospective les dossiers de 19 patients atteints de ML-IgG4 qui ont été admis à notre service de l'Hôpital de l'Université de Tokyo ou qui ont consulté un cardiologue du service entre le 1er janvier 2004 et le 31 décembre 2021. Résultats: La fréquence des événements coronariens était significativement plus élevée dans le groupe PC que dans les autres groupes. Par ailleurs, le groupe PC avait une survie sans événement significativement plus courte que les autres groupes (test logarithmique par rangs; p = 0,008). En outre, la fréquence des événements coronariens et la survie sans événement coronarien après un diagnostic de ML-IgG4 ne variaient pas de manière significative entre le groupe PA et les autres groupes. Bien qu'aucune différence statistiquement significative n'ait été constatée quant à la fréquence des événements coronariens entre les patients présentant un épaississement du péricarde et les autres patients, le premier groupe affichait une survie sans événement significativement plus courte que l'autre (test logarithmique par rangs; p = 0,017). Conclusions: L'incidence et le déroulement clinique des événements coronariens compliqués par la ML-IgG4 pouvaient être anticipés dans les cas de ML-IgG4 en présence de PC et d'un épaississement du péricarde, mais pas de PA.

An explainable artificial intelligence-enabled electrocardiogram analysis model for the classification of reduced left ventricular function.

Aims: The black box nature of artificial intelligence (AI) hinders the development of interpretable AI models that are applicable in clinical practice. We aimed to develop an AI model for classifying patients of reduced left ventricular ejection fraction (LVEF) from 12-lead electrocardiograms (ECG) with the decision-interpretability. Methods and results: We acquired paired ECG and echocardiography datasets from the central and co-operative institutions. For the central institution dataset, a random forest model was trained to identify patients with reduced LVEF among 29 907 ECGs. Shapley additive explanations were applied to 7196 ECGs. To extract the model's decision criteria, the calculated Shapley additive explanations values were clustered for 192 non-paced rhythm patients in which reduced LVEF was predicted. Although the extracted criteria were different for each cluster, these criteria generally comprised a combination of six ECG findings: negative T-wave inversion in I/V5-6 leads, low voltage in I/II/V4-6 leads, Q wave in V3-6 leads, ventricular activation time prolongation in I/V5-6 leads, S-wave prolongation in V2-3 leads, and corrected QT interval prolongation. Similarly, for the co-operative institution dataset, the extracted criteria comprised a combination of the same six ECG findings. Furthermore, the accuracy of seven cardiologists' ECG readings improved significantly after watching a video explaining the interpretation of these criteria (before, 62.9% ± 3.9% vs. after, 73.9% ± 2.4%; P = 0.02). Conclusion: We visually interpreted the model's decision criteria to evaluate its validity, thereby developing a model that provided the decision-interpretability required for clinical application.

Pericardial Grafting of Cardiac Progenitor Cells in Self-Assembling Peptide Scaffold Improves Cardiac Function After Myocardial Infarction.

Many studies have explored cardiac progenitor cell (CPC) therapy for heart disease. However, optimal scaffolds are needed to ensure the engraftment of transplanted cells. We produced a three-dimensional hydrogel scaffold (CPC-PRGmx) in which high-viability CPCs were cultured for up to 8 weeks. CPC-PRGmx contained an RGD peptide-conjugated self-assembling peptide with insulin-like growth factor-1 (IGF-1). Immediately after creating myocardial infarction (MI), we transplanted CPC-PRGmx into the pericardial space on to the surface of the MI area. Four weeks after transplantation, red fluorescent protein-expressing CPCs and in situ hybridization analysis in sex-mismatched transplantations revealed the engraftment of CPCs in the transplanted scaffold (which was cellularized with host cells). The average scar area of the CPC-PRGmx-treated group was significantly smaller than that of the non-treated group (CPC-PRGmx-treated group = 46 ± 5.1%, non-treated MI group = 59 ± 4.5%; p < 0.05). Echocardiography showed that the transplantation of CPC-PRGmx improved cardiac function and attenuated cardiac remodeling after MI. The transplantation of CPCs-PRGmx promoted angiogenesis and inhibited apoptosis, compared to the untreated MI group. CPCs-PRGmx secreted more vascular endothelial growth factor than CPCs cultured on two-dimensional dishes. Genetic fate mapping revealed that CPC-PRGmx-treated mice had more regenerated cardiomyocytes than non-treated mice in the MI area (CPC-PRGmx-treated group = 0.98 ± 0.25%, non-treated MI group = 0.25 ± 0.04%; p < 0.05). Our findings reveal the therapeutic potential of epicardial-transplanted CPC-PRGmx. Its beneficial effects may be mediated by sustainable cell viability, paracrine function, and the enhancement of de novo cardiomyogenesis.

Novel AGL variants in a patient with glycogen storage disease type IIIb and pulmonary hypertension caused by pulmonary veno-occlusive disease: A case report.

Glycogen storage disease type III (GSD-III) is an autosomal recessive metabolic disorder caused by mutations in the AGL gene, and may develop various types of pulmonary hypertension (PH). Here, we report a case of 24-year-old man with GSD-IIIb with two novel null variants in AGL (c.2308 + 2T>C and c.3045_3048dupTACC). He developed multi-drug-resistant pulmonary veno-occlusive disease (PVOD) and was registered as a candidate for lung transplantation. No pathogenic variants were detected in previously known causative genes for pulmonary hypertension and the underlying mechanism of coincidence of two rare diseases was unknown. We discuss the association of the loss of glycogen-debranching enzyme with incident PVOD.

Cross-ancestry genome-wide analysis of atrial fibrillation unveils disease biology and enables cardioembolic risk prediction.

Atrial fibrillation (AF) is a common cardiac arrhythmia resulting in increased risk of stroke. Despite highly heritable etiology, our understanding of the genetic architecture of AF remains incomplete. Here we performed a genome-wide association study in the Japanese population comprising 9,826 cases among 150,272 individuals and identified East Asian-specific rare variants associated with AF. A cross-ancestry meta-analysis of >1 million individuals, including 77,690 cases, identified 35 new susceptibility loci. Transcriptome-wide association analysis identified IL6R as a putative causal gene, suggesting the involvement of immune responses. Integrative analysis with ChIP-seq data and functional assessment using human induced pluripotent stem cell-derived cardiomyocytes demonstrated ERRg as having a key role in the transcriptional regulation of AF-associated genes. A polygenic risk score derived from the cross-ancestry meta-analysis predicted increased risks of cardiovascular and stroke mortalities and segregated individuals with cardioembolic stroke in undiagnosed AF patients. Our results provide new biological and clinical insights into AF genetics and suggest their potential for clinical applications.

Deep learning-based age estimation from chest X-rays indicates cardiovascular prognosis.

BACKGROUND: In recent years, there has been considerable research on the use of artificial intelligence to estimate age and disease status from medical images. However, age estimation from chest X-ray (CXR) images has not been well studied and the clinical significance of estimated age has not been fully determined. METHODS: To address this, we trained a deep neural network (DNN) model using more than 100,000 CXRs to estimate the patients' age solely from CXRs. We applied our DNN to CXRs of 1562 consecutive hospitalized heart failure patients, and 3586 patients admitted to the intensive care unit with cardiovascular disease. RESULTS: The DNN's estimated age (X-ray age) showed a strong significant correlation with chronological age on the hold-out test data and independent test data. Elevated X-ray age is associated with worse clinical outcomes (heart failure readmission and all-cause death) for heart failure. Additionally, elevated X-ray age was associated with a worse prognosis in 3586 patients admitted to the intensive care unit with cardiovascular disease. CONCLUSIONS: Our results suggest that X-ray age can serve as a useful indicator of cardiovascular abnormalities, which will help clinicians to predict, prevent and manage cardiovascular diseases.

Chest X-ray is one of the most widely used medical imaging tests worldwide to diagnose and manage heart and lung diseases. In this study, we developed a computer-based tool to predict patients' age from chest X-rays. The tool precisely estimated patients' age from chest X-rays. Furthermore, in patients with heart failure and those admitted to the intensive care unit for cardiovascular disease, elevated X-ray age estimated by our tool was associated with poor clinical outcomes, including readmission for heart failure or death from any cause. With further testing, our tool may help clinicians to predict outcomes in patients with heart disease based on a simple chest X-ray.

Age prediction from coronary angiography using a deep neural network: Age as a potential label to extract prognosis-related imaging features.

Coronary angiography (CAG) is still considered the reference standard for coronary artery assessment, especially in the treatment of acute coronary syndrome (ACS). Although aging causes changes in coronary arteries, the age-related imaging features on CAG and their prognostic relevance have not been fully characterized. We hypothesized that a deep neural network (DNN) model could be trained to estimate vascular age only using CAG and that this age prediction from CAG could show significant associations with clinical outcomes of ACS. A DNN was trained to estimate vascular age using ten separate frames from each of 5,923 CAG videos from 572 patients. It was then tested on 1,437 CAG videos from 144 patients. Subsequently, 298 ACS patients who underwent percutaneous coronary intervention (PCI) were analysed to assess whether predicted age by DNN was associated with clinical outcomes. Age predicted as a continuous variable showed mean absolute error of 4 years with R squared of 0.72 (r = 0.856). Among the ACS patients stratified by predicted age from CAG images before PCI, major adverse cardiovascular events (MACE) were more frequently observed in the older vascular age group than in the younger vascular age group (p = 0.017). Furthermore, after controlling for actual age, gender, peak creatine kinase, and history of heart failure, the older vascular age group independently suffered from more MACE (hazard ratio 2.14, 95% CI 1.07 to 4.29, p = 0.032). The vascular age estimated based on CAG imaging by DNN showed high predictive value. The age predicted from CAG images by DNN could have significant associations with clinical outcomes in patients with ACS.

Automatic Detection of Left Ventricular Dilatation and Hypertrophy from Electrocardiograms Using Deep Learning.

Left ventricular dilatation (LVD) and left ventricular hypertrophy (LVH) are risk factors for heart failure, and their detection improves heart failure screening. This study aimed to investigate the ability of deep learning to detect LVD and LVH from a 12-lead electrocardiogram (ECG). Using ECG and echocardiographic data, we developed deep learning and machine learning models to detect LVD and LVH. We also examined conventional ECG criteria for the diagnosis of LVH. We calculated the area under the receiver operating characteristic (AUROC) curve, sensitivity, specificity, and accuracy of each model and compared the performance of the models. We analyzed data for 18,954 patients (mean age (standard deviation): 64.2 (16.5) years, men: 56.7%). For the detection of LVD, the value (95% confidence interval) of the AUROC was 0.810 (0.801-0.819) for the deep learning model, and this was significantly higher than that of the logistic regression and random forest methods (P < 0.001). The AUROCs for the logistic regression and random forest methods (machine learning models) were 0.770 (0.761-0.779) and 0.757 (0.747-0.767), respectively. For the detection of LVH, the AUROC was 0.784 (0.777-0.791) for the deep learning model, and this was significantly higher than that of the logistic regression and random forest methods and conventional ECG criteria (P < 0.001). The AUROCs for the logistic regression and random forest methods were 0.758 (0.751-0.765) and 0.716 (0.708-0.724), respectively. This study suggests that deep learning is a useful method to detect LVD and LVH from 12-lead ECGs.

Nonsyndromic arteriopathy and aortopathy and vascular Ehlers-Danlos syndrome causing COL3A1 variants.

Vascular Ehlers-Danlos syndrome (vEDS) is an autosomal dominant genetic disorder characterized by soft connective tissue vulnerability due to dysfunction of Type III collagen and caused by the pathogenic variants in COL3A1 gene. In the era of next-generation sequencing, multiple genes including COL3A1 can be simultaneously analyzed, and among patients suffering from aortopathy even without any other clinical features suggestive of vEDS, pathogenic COL3A1 variants have been increasingly identified. Here, we briefly summarize the characteristics of 12 Japanese patients from 11 families with arteriopathy and pathogenic or likely pathogenic COL3A1 variants in our hospital. Five patients did not have any extra-arterial clinical features, however, the multigene panel testing for hereditary thoracic aortic aneurysm and dissection unexpectedly revealed that two had glycine substitutions in the triple-helical region and three had haploinsufficient type variants in the COL3A1 gene, whose pathogenicities were all classified as pathogenic or likely pathogenic. Further genetic screening and identification of pathogenic variants in patients with nonsyndromic arteriopathy and aortopathy will enable us to develop risk-stratification and management based on the genetic diagnosis.

Antihypertensive Drugs and Cancer Risk.

Hypertension is the most prevalent comorbidity in cancer patients. Consequently, many cancer patients are prescribed antihypertensive drugs before cancer diagnosis or during cancer treatment. However, whether antihypertensive drugs affect the incidence, treatment efficacy, or prognosis of cancer remains unanswered. For instance, renin-angiotensin and ß-adrenergic signaling may be involved not only in blood pressure elevation but also in cell proliferation, angiogenesis, and tissue invasion. Therefore, the inhibition of these pathways may have beneficial effects on cancer prevention or treatment. In this article, we reviewed several studies regarding antihypertensive drugs and cancer. In particular, we focused on the results of clinical trials to evaluate whether the use of antihypertensive drugs affects future cancer risk and prognosis. Unfortunately, the results are somewhat inconsistent, and evidence demonstrating the effect of antihypertensive drugs remains limited. We indicate that the heterogeneity in the study designs makes it difficult to clarify the causal relationship between antihypertensive drugs and cancer. We also propose that additional experimental studies, including research with induced pluripotent cells derived from cancer patients, single-cell analyses of cancer cell clusters, and clinical studies using artificial intelligence electronic health record systems, might be helpful to reveal the precise association between antihypertensive drugs and cancer risk.

Oral Administration of Euglena Gracilis Z Alleviates Constipation and Cardiac Dysfunction in a Mouse Model of Isoproterenol-Induced Heart Failure.

Background: Patients with heart failure (HF) often experience gastrointestinal problems such as constipation, diarrhea, and disturbances to drug absorption. In HF, hypoperfusion and congestion cause structural and functional changes in the gut, which, in turn, lead to impaired cardiac function. Euglena gracilis Z (hereafter "Euglena"), called Midorimushi in Japanese, is a microalga that is used as a food or nutritional supplement. It is unclear whether Euglena is beneficial for bowel habitus and cardiac function in subjects with HF. Methods and Results: We injected C57BL/6 male mice subcutaneously with isoproterenol (ISO) (20 mg/kg/day) for 7 days to examine bowel movement in HF. Euglena was orally administered to mice on an ad libitum-feeding to a normal chow containing 2% dietary mixture. ISO induced a decrease in bowel movement and an increase in fecal retention in the cecum, as well as a decrease in left ventricular (LV) contraction. Euglena accelerated intestinal transit, relieved fecal retention, and prevented the alterations in gut pathology in ISO-treated mice. Euglena also suppressed ISO-induced decreases in LV contraction, although it had no significant effect on LV hypertrophy. Conclusions: The results suggested that oral administration of Euglena alleviated constipation and cardiac dysfunction in a mouse model of ISO-induced HF, and highlight the potential clinical benefit of Euglena in patients with HF in preventing constipation and contractile deterioration.

Oxidized LDL but not angiotensin II induces cardiomyocyte hypertrophic responses through the interaction between LOX-1 and AT1 receptors.

It is well known that lectin-like oxidized low-density lipoprotein (ox-LDL) and its receptor LOX-1, angiotensin II (AngII) and its type 1 receptor (AT1-R) play an important role in the development of cardiac hypertrophy. However, the molecular mechanism is not clear. In this study, we found that ox-LDL-induced cardiac hypertrophy was suppressed by inhibition of LOX-1 or AT1-R but not by AngII inhibition. These results suggest that the receptors LOX-1 and AT1-R, rather than AngII, play a key role in the role of ox-LDL. The same results were obtained in mice lacking endogenous AngII and their isolated cardiomyocytes. Ox-LDL but not AngII could induce the binding of LOX-1 and AT1-R; inhibition of LOX-1 or AT1-R but not AngII could abolish the binding of these two receptors. Overexpression of wild type LOX-1 with AT1-R enhanced ox-LDL-induced binding of two receptors and phosphorylation of ERKs, however, transfection of LOX-1 dominant negative mutant (lys266ala / lys267ala) or an AT1-R mutant (glu257ala) not only reduced the binding of two receptors but also inhibited the ERKs phosphorylation. Phosphorylation of ERKs induced by ox-LDL in LOX-1 and AT1-R-overexpression cells was abrogated by an inhibitor of Gq protein rather than Jak2, Rac1 or RhoA. Genetically, an AT1-R mutant lacking Gq protein coupling ability inhibited ox-LDL induced ERKs phosphorylation. Furthermore, through bimolecular fluorescence complementation analysis, we confirmed that ox-LDL rather than AngII stimulation induced the direct binding of LOX-1 and AT1-R. We conclude that direct binding of LOX-1 and AT1-R and the activation of downstream Gq protein are important mechanisms of ox-LDL-induced cardiomyocyte hypertrophy.

Deep learning model to detect significant aortic regurgitation using electrocardiography.

BACKGROUND: Aortic regurgitation (AR) is a common heart disease, with a relatively high prevalence of 4.9% in the Framingham Heart Study. Because the prevalence increases with advancing age, an upward shift in the age distribution may increase the burden of AR. To provide an effective screening method for AR, we developed a deep learning-based artificial intelligence algorithm for the diagnosis of significant AR using electrocardiography (ECG). METHODS: Our dataset comprised 29,859 paired data of ECG and echocardiography, including 412 AR cases, from January 2015 to December 2019. This dataset was divided into training, validation, and test datasets. We developed a multi-input neural network model, which comprised a two-dimensional convolutional neural network (2D-CNN) using raw ECG data and a fully connected deep neural network (FC-DNN) using ECG features, and compared its performance with the performances of a 2D-CNN model and other machine learning models. In addition, we used gradient-weighted class activation mapping (Grad-CAM) to identify which parts of ECG waveforms had the most effect on algorithm decision making. RESULTS: The area under the receiver operating characteristic curve of the multi-input model (0.802; 95% CI, 0.762-0.837) was significantly greater than that of the 2D-CNN model alone (0.734; 95% CI, 0.679-0.783; p<0.001) and those of other machine learning models. Grad-CAM demonstrated that the multi-input model tended to focus on the QRS complex in leads I and aVL when detecting AR. CONCLUSIONS: The multi-input deep learning model using 12-lead ECG data could detect significant AR with modest predictive value.

Prospects for cardiovascular medicine using artificial intelligence.

As the importance of artificial intelligence (AI) in the clinical setting increases, the need for clinicians to understand AI is also increasing. This review focuses on the fundamental principles of AI and the current state of cardiovascular AI. Various types of cardiovascular AI have been developed for evaluating examinations such as X-rays, electrocardiogram, echocardiography, computed tomography, and magnetic resonance imaging. Cardiovascular AI achieves high accuracy in diagnostic support and prognosis prediction. Furthermore, it can even detect abnormalities that were previously difficult for cardiologists to detect. Randomized controlled trials begin to be reported to verify the usefulness of cardiovascular AI. The day is approaching when cardiovascular AI will be commonly used in clinical practice. Various types of medical AI will be used for cardiovascular care; however, it will not replace medical doctors. We need to understand the strengths and weaknesses of medical AI so that cardiologists can effectively use AI to improve the medical care of patients.

The Effectiveness of a Deep Learning Model to Detect Left Ventricular Systolic Dysfunction from Electrocardiograms.

Deep learning models can be applied to electrocardiograms (ECGs) to detect left ventricular (LV) dysfunction. We hypothesized that applying a deep learning model may improve the diagnostic accuracy of cardiologists in predicting LV dysfunction from ECGs. We acquired 37,103 paired ECG and echocardiography data records of patients who underwent echocardiography between January 2015 and December 2019. We trained a convolutional neural network to identify the data records of patients with LV dysfunction (ejection fraction < 40%) using a dataset of 23,801 ECGs. When tested on an independent set of 7,196 ECGs, we found the area under the receiver operating characteristic curve was 0.945 (95% confidence interval: 0.936-0.954). When 7 cardiologists interpreted 50 randomly selected ECGs from the test dataset of 7,196 ECGs, their accuracy for predicting LV dysfunction was 78.0% ± 6.0%. By referring to the model's output, the cardiologist accuracy improved to 88.0% ± 3.7%, which indicates that model support significantly improved the cardiologist diagnostic accuracy (P = 0.02). A sensitivity map demonstrated that the model focused on the QRS complex when detecting LV dysfunction on ECGs. We developed a deep learning model that can detect LV dysfunction on ECGs with high accuracy. Furthermore, we demonstrated that support from a deep learning model can help cardiologists to identify LV dysfunction on ECGs.