Estimation of heart inclination angle using posteroanterior chest radiograph and comparison with cardiac axis obtained from synthesized vectorcardiogram

ABSTRACT In medical practice, it is common to perform electrocardiography exams and by mathematical transformations to obtain the vectorcardiogram. The vectorcardiogram provides important information for medical diagnosis, such as the angle of inclination of the heart. This article aims to present a methodology for estimating the QRS vector-related angle of the heart using a posteroanterior chest radiograph image. We used an open source image processing software (Icy software version 2.3.0.0, Institut Pasteur, France, 2021) to perform a manual measurement of the target angle by analyzing relevant morphological structures from the x-ray images and using some functions to help the user to measure it. 18 radiographic images were selected to measure the angle of the heart by two independent individuals. The measured angles were compared using the mean absolute error (MAE). We then computed the QRS peak elevation angles of the vectorcardiogram (VCG) of the 57 patients collected at Dante Pazzanese Institute of Cardiology. In addition, an individual was randomly selected to measure a set of 57 radiographic images of these same patients. We performed the statistical treatments and the results suggested that the proposed manual method may be an alternative, viable and fast approach to estimating the anatomical heart axis for the purpose of aiding in medical diagnosis. However, further comparisons with more data and information are needed to determine its validity and possible method improvements.

Leveraging advanced ai algorithms for predicting outcomes in non-acute coronary syndrome chest pain patients

INTRODUCTION: Chest pain is often encountered in emergency rooms and the detection of acute coronary syndrome (ACS) is a major focus. However, a notable percentage of patients present with a diverse range of nonACS conditions. Accurately identifying the causes and outcomes of these cases can prevent unnecessary interventions, reduce healthcare costs, and optimize resource allocation. This study aims to explore how advanced AI algorithms can enhance risk assessment, refine classification, and predict outcomes in nonACS chest pain patients using conventional ECG analysis. METHODS: We studied 3458 nonACS patients referred to the Emergency Room at Instituto Dante Pazzanese de Cardiologia with chest pain. A total of 185 features, including sex, height, ECG diagnostic statements, and measures, were used. The predicted outcome was defined as hospitalization within 14 days and/or death (1 or 0). We employed the AutoGluon framework for feature engineering and early model selection. XGBoost, a tree-based model, was chosen as the architecture. Training and k-fold stratified cross-validation were performed using an oversampled balanced dataset, and evaluation metrics such as AUROC, specificity, and sensitivity were measured using the original data. RESULTS: In this study, 18.2% (630 patients) had a positive outcome. The sex distribution was comparable between outcome groups, with men accounting for 57-58% and women for 42-43%. Significant differences (p<0.01) were observed in ECG intervals (QRS, corrected QT, RR interval, PSP) between the groups. The AI model identified important diagnostic statements, including normal ECG (19.4), atrial fibrillation (7.4), left ventricular hypertrophy (7.1), Ischemic T-wave inversion in inferior leads (6.7), T-wave changes in inferior leads (5.9), and first-degree atrioventricular block (5.8). The AI model performed exceptionally well, with a sensitivity of 97.93%, specificity of 96.08%, and an AUROC of 0.97. CONCLUSIONS: The AI model demonstrated its ability to predict outcomes in patients with acute chest pain without ACS, making it an appealing tool for effective risk stratification. The early identification provided by the AI model presents an opportunity for timely intervention to mitigate adverse outcomes.

An automated machine learning method to efficientlyclassify the 12-lead ECG signal acquisition quality

ABSTRACT: Given their low cost and non-invasive nature, ElectroCardioGram(ECG) signals have been widely used as a useful tool for diagnosing heart dis-eases. However, acquisition issues such as electrode interchange and oscilla-tion noise may negatively impact expert exam interpretation and even automaticclassification tasks. Here we propose an automated machine learning methodto efficiently classify the 12-lead ECG signal acquisition quality. It consists ofa two-stage classification process. Firstly, the ECG signals are processed andsegmented aiming to classify them as noisy or acceptable signals. Then, thesecond classification stage yields the binary classification correct acquisition orlimb electrodes interchange. Concerning the electrode positioning, the RandomForest technique presented interesting results (precision of 97%, recall of 89%, and F1-Score of 93%). Concerning noise detection, Random Forest presented ageneral accuracy of 85%, a recall of 57%, and a precision of 91%. All the ob-tained results yield to consider the proposed framework for application within areal telemedicine environment.

Impacto da pandemia de Covid-19 nos exames cardiológicos no Brasil: uma análise do SUS e da medicina suplementar / Impact of the Covid-19 pandemic on cardiological exams in Brazil: an analysis of the SUS and supplementary medicine

INTRODUÇÃO: a pandemia de COVID-19 ocasionou grande impacto no sistema de saúde brasileiro. Além da sobrecarga de pacientes infectados pelo virus, o risco de contaminação obrigou múltiplos serviços de saúde a interromper ou reduzir drasticamente a oferta de exames diagnósticos em cardiologia. O objetivo desse estudo foi avaliar o impacto da pandemia sobre a realização de exames cardiológicos no Sistema Único de Saúde (SUS) e na saúde suplementar. MÉTODOS: coletamos dados referentes aos exames de teste ergométrico, ecocardiograma (repouso e estresse; transtorácico e transesofágico), cintilografia (perfusão de estresse e repouso; avaliação de necrose; viabilidade), cineangiocoronariografia e ressonância magnética cardíaca realizados nos anos de 2019 e 2020 no SUS e no sistema suplementar (mesmos métodos acrescidos de tomografia para escore de cálcio e angiotomografia coronariana) em pacientes internados e ambulatoriais. Os dados foram coletados online a partir do DATASUS e da plataforma da Agência Nacional de Saúde (ANS) e os anos de 2019 e 2020 foram comparados. RESULTADOS: No ano de 2019, foram realizados no Brasil 10.647.379 exames diagnósticos cardiológicos no SUS e no sistema suplementar no Brasil; já no ano 2020, foram apenas 8.315.215 - redução de 2.332.164 exames ou queda de 21,9% em relação ao ano anterior. No SUS, a redução no número de exames foi de 781.529 (24,4%) no comparativo entre os dois anos; já na saúde suplementar, a queda foi de 1.550.635 exames (20,8%). Na análise nacional por método diagnóstico, observamos comportamento heterogêneo ao longo do biênio 2019/2020: teste ergométrico apresentou queda relativa de 28,1%, ecocardiograma 19%, cintilografia miocárdica 26,7% e cineangiocoronariografia 16,5%; já a ressonância magnética cardíaca apresentou aumento de 5% no período. Tendências heterogêneas também foram observadas na comparação entre os exames realizados no SUS e na saúde suplementar ao longo do período de análise. Dados de tomografia cardíaca (disponíveis somente na saúde suplementar) mostraram aumento de 36,7 mil exames em 2019 para 47,1 mil exames em 202 (aumento de 28,4%). CONCLUSÕES: a pandemia de COVID-19 ocasionou alteração importante na estratégia de investigação diagnóstica entre pacientes cardiológicos no Brasil no ano de 2020, com redução significativa no número de exames no SUS e na saúde suplementar e impacto heterogêneo conforme a modalidade diagnóstica. Esses resultados podem contribuir para o planejamento de ações no âmbito público e privado voltadas para a melhoria na investigação cardiológica nos próximos anos.

Afterload increase challenge unmasks systolic abnormalities in heart failure with preserved ejection fraction.

BACKGROUND: Provocative maneuvers have the potential to overcome the low sensitivity of resting echocardiography and biomarkers in the detection of heart failure with preserved ejection fraction (HFpEF). We investigate the mechanical response of the left ventricle to an afterload challenge in patients with preclinical and early-stage HFpEF (es-HFpEF). METHODS: Three groups of patients (non-HFpEF - n = 42, pre-HFpEF - n = 43, and es-HFpEF - n = 39) underwent echocardiography at rest and during an afterload challenge induced by handgrip maneuver combined with pneumatic constriction of limbs. RESULTS: Patients in the non-HF group displayed a median ΔLPSS = -4% (IQR: -10%, +2%), LPSS rest<16% in 3/42(7%) and LPSS stress<16% in 6/43(14%). Subjects in the pre-HFpEF group displayed median ΔLPSS = -3% (IQR: -10%, +5%) LPSS rest<16% in 13/43(30%) and LPSS stress<16% in 19/43 (44%). 11/43 (25%) subjects in this group increased at least one absolute point in LPSS during stress. Patients in es-HFpEF group displayed a median ΔLPSS = -10% (IQR: -18%, -1%), LPSS rest<16% in 15/39(38%) and LPSS stress<16% in 25/39(64%). Changes in LPSS (ΔLPSS) were significantly greater in es-HFpEF than pre-HFpEF (p = 0.022). In multivariate analysis, this group effect was maintained after adjustment of the LPSS for systolic blood pressure, use of ß-blockers, LV mass, RWT, age, and sex. CONCLUSION: Our data suggest that patients with HFpEF have a marked decrease in peak strain during acute pressure overload. Longitudinal studies are needed to test and compare the clinical impact of each pattern in early and long-term follow-ups.

Afterload increase challenge unmasks systolic abnormalities in heart failure with preserved ejection fraction

BACKGROUND: Provocative maneuvers have the potential to overcome the low sensitivity of resting echocardiography and biomarkers in the detection of heart failure with preserved ejection fraction (HFpEF). We investigate the mechanical response of the left ventricle to an afterload challenge in patients with preclinical and early-stage HFpEF (es-HFpEF). METHODS: Three groups of patients (non-HFpEF - n = 42, pre-HFpEF - n = 43, and es-HFpEF - n = 39) underwent echocardiography at rest and during an afterload challenge induced by handgrip maneuver combined with pneumatic constriction of limbs. RESULTS: Patients in the non-HF group displayed a median ΔLPSS = -4% (IQR: -10%, +2%), LPSS rest<16% in 3/42(7%) and LPSS stress<16% in 6/43(14%). Subjects in the pre-HFpEF group displayed median ΔLPSS = -3% (IQR: -10%, +5%) LPSS rest<16% in 13/43(30%) and LPSS stress<16% in 19/43 (44%). 11/43 (25%) subjects in this group increased at least one absolute point in LPSS during stress. Patients in es-HFpEF group displayed a median ΔLPSS = -10% (IQR: -18%, -1%), LPSS rest<16% in 15/39(38%) and LPSS stress<16% in 25/39(64%). Changes in LPSS (ΔLPSS) were significantly greater in es-HFpEF than pre-HFpEF (p = 0.022). In multivariate analysis, this group effect was maintained after adjustment of the LPSS for systolic blood pressure, use of ß-blockers, LV mass, RWT, age, and sex. CONCLUSION: Our data suggest that patients with HFpEF have a marked decrease in peak strain during acute pressure overload. Longitudinal studies are needed to test and compare the clinical impact of each pattern in early and long-term follow-ups.