RESUMO
Objectives: To evaluate the performance of convolutional neural network (CNN) architectures to distinguish eyes with glaucoma from normal eyes. Materials and Methods: A total of 9,950 fundus photographs of 5,388 patients from the database of Eskisehir Osmangazi University Faculty of Medicine Ophthalmology Clinic were labelled as glaucoma, glaucoma suspect, or normal by three different experienced ophthalmologists. The categorized fundus photographs were evaluated using a state-of-the-art two-dimensional CNN and compared with deep residual networks (ResNet) and very deep neural networks (VGG). The accuracy, sensitivity, and specificity of glaucoma detection with the different algorithms were evaluated using a dataset of 238 normal and 320 glaucomatous fundus photographs. For the detection of suspected glaucoma, ResNet-101 architectures were tested with a data set of 170 normal, 170 glaucoma, and 167 glaucoma-suspect fundus photographs. Results: Accuracy, sensitivity, and specificity in detecting glaucoma were 96.2%, 99.5%, and 93.7% with ResNet-50; 97.4%, 97.8%, and 97.1% with ResNet-101; 98.9%, 100%, and 98.1% with VGG-19, and 99.4%, 100%, and 99% with the 2D CNN, respectively. Accuracy, sensitivity, and specificity values in distinguishing glaucoma suspects from normal eyes were 62%, 68%, and 56% and those for differentiating glaucoma from suspected glaucoma were 92%, 81%, and 97%, respectively. While 55 photographs could be evaluated in 2 seconds with CNN, a clinician spent an average of 24.2 seconds to evaluate a single photograph. Conclusion: An appropriately designed and trained CNN was able to distinguish glaucoma with high accuracy even with a small number of fundus photographs.
Assuntos
Aprendizado Profundo , Glaucoma , Fundo de Olho , Glaucoma/diagnóstico , Humanos , Redes Neurais de Computação , FotografaçãoRESUMO
OBJECTIVE: ECG recordings often suffer from a set of artifacts with varying types, severities, and durations, and this makes an accurate diagnosis by machines or medical doctors difficult and unreliable. Numerous studies have proposed ECG denoising; however, they naturally fail to restore the actual ECG signal corrupted with such artifacts due to their simple and naive noise model. In this pilot study, we propose a novel approach for blind ECG restoration using cycle-consistent generative adversarial networks (Cycle-GANs) where the quality of the signal can be improved to a clinical level ECG regardless of the type and severity of the artifacts corrupting the signal. METHODS: To further boost the restoration performance, we propose 1D operational Cycle-GANs with the generative neuron model. RESULTS: The proposed approach has been evaluated extensively using one of the largest benchmark ECG datasets from the China Physiological Signal Challenge (CPSC-2020) with more than one million beats. Besides the quantitative and qualitative evaluations, a group of cardiologists performed medical evaluations to validate the quality and usability of the restored ECG, especially for an accurate arrhythmia diagnosis. SIGNIFICANCE: As a pioneer study in ECG restoration, the corrupted ECG signals can be restored to clinical level quality. CONCLUSION: By means of the proposed ECG restoration, the ECG diagnosis accuracy and performance can significantly improve.
Assuntos
Algoritmos , Eletrocardiografia , Humanos , Projetos Piloto , Artefatos , Arritmias Cardíacas/diagnóstico , Processamento de Sinais Assistido por ComputadorRESUMO
OBJECTIVE: Despitethe proliferation of numerous deep learning methods proposed for generic ECG classification and arrhythmia detection, compact systems with the real-time ability and high accuracy for classifying patient-specific ECG are still few. Particularly, the scarcity of patient-specific data poses an ultimate challenge to any classifier. Recently, compact 1D Convolutional Neural Networks (CNNs) have achieved the state-of-the-art performance level for the accurate classification of ventricular and supraventricular ectopic beats. However, several studies have demonstrated the fact that the learning performance of the conventional CNNs is limited because they are homogenous networks with a basic (linear) neuron model. In order to address this deficiency and further boost the patient-specific ECG classification performance, in this study, we propose 1D Self-organized Operational Neural Networks (1D Self-ONNs). METHODS: Due to its self-organization capability, Self-ONNs have the utmost advantage and superiority over conventional ONNs where the prior operator search within the operator set library to find the best possible set of operators is entirely avoided. RESULTS: Under AAMI recommendations and with minimal common training data used, over the entire MIT-BIH dataset 1D Self-ONNs have achieved 98% and 99.04% average accuracies, 76.6% and 93.7% average F1 scores on supra-ventricular and ventricular ectopic beat (VEB) classifications, respectively, which is the highest performance level ever reported. CONCLUSION: As the first study where 1D Self-ONNs are ever proposed for a classification task, our results over the MIT-BIH arrhythmia benchmark database demonstrate that 1D Self-ONNs can surpass 1D CNNs with a significant margin while having a similar computational complexity.
Assuntos
Eletrocardiografia , Complexos Ventriculares Prematuros , Algoritmos , Bases de Dados Factuais , Eletrocardiografia/métodos , Frequência Cardíaca , Humanos , Redes Neurais de Computação , Neurônios , Processamento de Sinais Assistido por ComputadorRESUMO
OBJECTIVE: Noise and low quality of ECG signals acquired from Holter or wearable devices deteriorate the accuracy and robustness of R-peak detection algorithms. This paper presents a generic and robust system for R-peak detection in Holter ECG signals. While many proposed algorithms have successfully addressed the problem of ECG R-peak detection, there is still a notable gap in the performance of these detectors on such low-quality ECG records. METHODS: In this study, a novel implementation of the 1D Convolutional Neural Network (CNN) is used integrated with a verification model to reduce the number of false alarms. This CNN architecture consists of an encoder block and a corresponding decoder block followed by a sample-wise classification layer to construct the 1D segmentation map of R-peaks from the input ECG signal. Once the proposed model has been trained, it can solely be used to detect R-peaks possibly in a single channel ECG data stream quickly and accurately, or alternatively, such a solution can be conveniently employed for real-time monitoring on a lightweight portable device. RESULTS: The model is tested on two open-access ECG databases: The China Physiological Signal Challenge (2020) database (CPSC-DB) with more than one million beats, and the commonly used MIT-BIH Arrhythmia Database (MIT-DB). Experimental results demonstrate that the proposed systematic approach achieves 99.30% F1-score, 99.69% recall, and 98.91% precision in CPSC-DB, which is the best R-peak detection performance ever achieved. Results also demonstrate similar or better performance than most competing algorithms on MIT-DB with 99.83% F1-score, 99.85% recall, and 99.82% precision. SIGNIFICANCE: Compared to all competing methods, the proposed approach can reduce the false-positives and false-negatives in Holter ECG signals by more than 54% and 82%, respectively. CONCLUSION: Finally, the simple and invariant nature of the parameters leads to a highly generic system and therefore applicable to any ECG dataset.
Assuntos
Eletrocardiografia , Processamento de Sinais Assistido por Computador , Algoritmos , Arritmias Cardíacas , Eletrocardiografia Ambulatorial , Humanos , Redes Neurais de ComputaçãoRESUMO
Medications and treatments are said to have a palliative effect if they relieve symptoms without having a curative effect on the underlying disease such as atherosclerosis or cancer. Some authors speculated that atherosclerotic coronary artery disease (CAD) could be considered a "cancer of the coronary arterial wall". Although the percutaneous coronary intervention (PCI) has proven to be effective in decreasing mortality rates among patients with acute coronary syndromes, the previous meta-analyses of PCI versus optimal medical therapy for stable CAD have not been able to demonstrate a reduction in major adverse cardiac outcomes. However, few cardiologists discussed the evidence-based benefits of angiogram and PCI for stable CAD, and some implicitly or explicitly overstated the benefits. Recently, the precision medicine is defined as an evidence-based approach that uses innovative tools and biological and data science to customize disease prevention, detection, and treatment, and improve the effectiveness and quality of patient care. Providing patients with accurate and complete information appears to be an effective way to combat the reliance on the oculostenotic reflex. The foundation of precision medicine is the ability to tailor therapy based upon the expected risks and benefits of treatment for each individual patient. As said by Doctor William Osler, "The good physician treats the disease; the great physician treats the patient who has the disease."
Assuntos
Síndrome Coronariana Aguda/cirurgia , Aterosclerose/complicações , Vasos Coronários/cirurgia , Stents Farmacológicos , Cuidados Paliativos/métodos , Intervenção Coronária Percutânea , Medicina de Precisão/métodos , Síndrome Coronariana Aguda/etiologia , Aterosclerose/cirurgia , HumanosAssuntos
Prótese Vascular , Ponte de Artéria Coronária/métodos , Doença da Artéria Coronariana/diagnóstico , Vasos Coronários/diagnóstico por imagem , Anastomose Cirúrgica/métodos , Angiografia Coronária , Doença da Artéria Coronariana/cirurgia , Vasos Coronários/cirurgia , Humanos , Desenho de PróteseRESUMO
It has been rarely encountered some patients in clinical practice with coronary artery chronic total occlusion (CTO) on angiography but no any clinical history or electrocardiographic, echocardiographic, or left ventriculographic evidence of previous myocardial infarction. These noninfarct-related artery CTO (non-IRA CTO) lesions may be used as a clinical role model of endogenous cardioprotective mechanisms in addition to continuing the process of atherosclerosis. The objective of this study was to characterize the clinical characteristics of patients with non-IRA CTO patients and compared them to those with infarct-related CTO (IRA-CTO). We reviewed our invasive cardiology database searching for the CTO of any major coronary arteries, and assessed whether or not they have the clinical history or electrocardiographic, echocardiographic, and left ventriculographic evidence of previous myocardial infarction. Interestingly, we detected that all these patients with non-IRA CTO had diabetes mellitus, and the clinical and demographic features of these non-IRA CTO patients were compared with age- and sex-matched diabetic IRA-CTO patients with regard to conventional coronary risk factors and the angiographic collateral grading system. There were total 99 CTO patients (49 patients with non-IRA CTO and 50 patients with IRA-CTO).All patients with non-IRA CTO had better collateral circulation (96 vs. 40% p < 0.001) compared with those having IRA-CTO. The only significant difference between the groups was the status of current smoking (4 vs. 88%; p < 0.001). The present study showed that the non-IRA CTOs were associated with diabetes mellitus and better collateral development compared with IRA-CTO. In diabetic patients, the concomitant smoking use may be harmful by preventing endogenous cardioprotective mechanisms.
Assuntos
Aneurisma/etiologia , Ponte de Artéria Coronária/efeitos adversos , Veia Safena/transplante , Idoso , Aneurisma/diagnóstico por imagem , Aneurisma/terapia , Angiografia Coronária/métodos , Progressão da Doença , Humanos , Masculino , Flebografia/métodos , Valor Preditivo dos Testes , Veia Safena/diagnóstico por imagem , Fatores de Tempo , Tomografia Computadorizada por Raios XRESUMO
Multiple endocrine neoplasia 2 (MEN 2) is a hereditary syndrome associated with medullary thyroid carcinoma, pheochromocytoma (PCC), and hyperparathyroidism. PCCs in patients with MEN 2 are usually found in the adrenals after the manifestation of medullary thyroid cancer and are commonly bilateral and hormonally active. Unfortunately, a diagnosis of MEN 2 or PCC often is delayed until after the patient has developed an advanced MEN 2-related tumor. We present unusual electrocardiographic changes on exercise testing in MEN 2 syndrome. Transient peaked T waves and shortening QT during exercise stress testing may provide an early clue for undiagnosed PCC.
RESUMO
BACKGROUND: Both obstructive sleep apnea (OSA) and coronary slow-flow phenomenon (CSFP) are known to share similar etiopathogenic mechanisms, such as chronic sympathetic activation, upregulation of inflammatory pathways, oxidative stress and, finally, endothelial dysfunction. OBJECTIVE: We evaluated whether there is an association between OSA and coronary flow rates. METHOD: We retrospectively reviewed medical records of all patients who underwent diagnostic nocturnal polysomnography for suspected OSA. Those who had coronary angiography performed within the same year of polysomnography were divided into two main groups: those with (group 1) and without (group 2) OSA; also, angiographic coronary TIMI (thrombolysis in myocardial infarction) frame counts (TFC) were compared between the groups. Patients with coronary arterial stenosis and angiograms with inadequate filling of the coronary arteries or visualization of the distal landmarks for frame counting were excluded from the study. RESULTS: There was a statistically significant difference between the groups regarding TFCs. We found a significant positive correlation between mean TFC and apnea-hypopnea index (r=0.611, P<0.001). CONCLUSION: The current study demonstrated that sleep apnea impairs coronary flow rates and is associated with CSFP.
Assuntos
Fenômeno de não Refluxo/fisiopatologia , Apneia Obstrutiva do Sono/fisiopatologia , Angiografia Coronária , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Polissonografia , Estudos RetrospectivosRESUMO
PURPOSE: Sleep and sleep position have a significant impact on physical, cardiac and mental health, and have been evaluated in numerous studies particularly in terms of lateral sleeping positions and their association with diseases. We retrospectively examined the relationship between the sleeping position and position-specific apnea-hypopnea index (AHI) in obstructive sleep apnea-hypopnea (OSA) patients. METHODS: We assessed the sleeping body position and the body position-specific AHI score in patients who were referred for suspected OSA and underwent diagnostic nocturnal polysomnography. In order to eliminate inter-individual differences, only those who had a similar percentage of time spent in the LSSP and RSSP for each patient were enrolled. To provide this validity, only subjects that had a similar percentage of left and right lateral sleep time (±10%) were included in the analysis. RESULTS: A total of 864 patients had nocturnal diagnostic PSG. Of them, 131 patients met the inclusion criteria. The percent rate spent in the supine sleeping position (SSP) was 31.3 ± 18.7%, in the LSSP was 31.8 ± 10% and in the RSSP was 32.6 ± 10.8%. Whereas the SSP-specific AHI score was the highest with 60.4 ± 36.2/h among all the sleeping position-specific AHI scores (p < 0.001), the LSSP-specific AHI score was statistically higher than that for RSSP (30.2 ± 32.6/h vs. 23.6 ± 30.1/h; p < 0.001). When comparing individuals sub-grouped based on OSA severity, there was a statistically significant difference between the LSSP-specific AHI score and RSSP-specific AHI score in patients with severe (p = 0.002) and moderate (p = 0.026), but not mild (p = 0.130) OSA. CONCLUSION: We found that the sleeping position had a significant influence on apneic events and RSSP decreased the frequency of obstructive respiratory events in patients with moderate and severe disease.
Assuntos
Postura , Apneia Obstrutiva do Sono/diagnóstico , Apneia Obstrutiva do Sono/etiologia , Adulto , Distúrbios do Sono por Sonolência Excessiva/diagnóstico , Distúrbios do Sono por Sonolência Excessiva/etiologia , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Polissonografia , Estudos Retrospectivos , Fatores de RiscoRESUMO
OBJECTIVE: Both heart rate (HR) and blood pressure parameters provide important information on the pathophysiology of the cardiovascular regulatory mechanisms, and are mainly affected by the autonomic nervous system. We sought to clarify whether the severity of obstructive sleep apnea (OSA) affects nocturnal HRs and whether there is a relationship between nocturnal HRs and the presence of hypertension. METHODS: We retrospectively reviewed medical records of all patients who performed nocturnal polysomnography with monitoring of HRs, and examined whether there is a relationship among the nocturnal HRs, the severity of OSA and the presence of hypertension. RESULTS: A total of 540 patients were included in the study. Nocturnal mean and maximal HRs were significantly higher in severe OSA group than in moderate (p=0.002 and p>0.05 in females; p<0.049 and p=0.044, in males, respectively) and mild OSA groups (p<0.001 and p=0.003, respectively in females, p<0.001 and p=0.004, respectively in males); and there was a positive correlation between the nocturnal mean HR and apnea-hypopnea index (Pearson's p=0.504, p<0.001 in female group; Pearson's p=0.254, p<0.001 in male group) and again the nocturnal mean HR and the presence of HT (Spearman's p=0.090, p=0.394 in female group; Spearman's p=0.272, p<0.001 in male group) in both gender groups. CONCLUSION: We found that nocturnal mean and maximal HRs to be associated with severity of OSA and the presence of hypertension. We speculated that increased nocturnal mean and maximal HRs caused by sympathetic nervous system activation in OSA might be one of the mechanisms in explaining the hypertension and OSA association.
Assuntos
Hipertensão/fisiopatologia , Apneia Obstrutiva do Sono/fisiopatologia , Pressão Sanguínea , Ritmo Circadiano , Feminino , Frequência Cardíaca , Humanos , Hipertensão/etiologia , Masculino , Prontuários Médicos , Pessoa de Meia-Idade , Polissonografia , Estudos Retrospectivos , Índice de Gravidade de Doença , Apneia Obstrutiva do Sono/complicaçõesRESUMO
Obesity has reached global pandemic that threatens the health of millions of people and is associated with numerous morbidities such as hypertension, type II diabetes mellitus, dyslipidemia, cor pulmonale, gallbladder disease, obstructive sleep apnea (OSA), certain cancers, osteoarthritis, increased surgical risk and postoperative complications, lower extremity venous and/or lymphatic problems, pulmonary embolism, stroke/cerebrovascular diseases and coronary arterial disease. Despite all these adverse associations, numerous studies and meta-analyses have documented an "obesity paradox" in which overweight and obese population with established cardiovascular disease have a better prognosis than do their lean counterparts. There are potential and plausible explanations offered by literature for these puzzling data; however, it still remains uncertain whether this phenomenon is attributable to a real protective effect of high body fat mass. In recent years, the survival advantage of patients with OSA, combined with the potential cardioprotective effects of chronic intermittent hypoxia, raise the possibility that apneas during sleep may activate preconditioning-like cardioprotective effect. Chronic intermittent hypoxia, one of the physiological markers of OSA, is characterized by transient periods of oxygen desaturation followed by reoxygenation, and is a major cause of its systemic harmful (oxidative stress, inflammation, sympathetic activity, vasculature remodelling and endothelial dysfunction) and/or protective (preconditioning-like cardioprotective) effects. Since many OSA subjects are obese, and obesity is an independent risk factor for many comorbidities associated with OSA; and also most OSA has never been diagnosed in obese patients, we hypothesed that the chronic intermittent hypoxia caused by OSA in obese patients may be one of the underlying mechanisms in morbi-mortality paradox of obesity.