Identifying pediatric heart murmurs and distinguishing innocent from pathologic using deep learning.

OBJECTIVE: To develop a deep learning algorithm to perform multi-class classification of normal pediatric heart sounds, innocent murmurs, and pathologic murmurs. METHODS: We prospectively enrolled children under age 18 being evaluated by the Division of Pediatric Cardiology. Parents provided consent for a deidentified recording of their child's heart sounds with a digital stethoscope. Innocent murmurs were validated by a pediatric cardiologist and pathologic murmurs were validated by echocardiogram. To augment our collection of normal heart sounds, we utilized a public database of pediatric heart sound recordings (Oliveira, 2022). We propose two novel approaches for this audio classification task. We train a vision transformer on either Markov transition field or Gramian angular field image representations of the frequency spectrum. We benchmark our results against a ResNet-50 CNN trained on spectrogram images. RESULTS: Our final dataset consisted of 366 normal heart sounds, 175 innocent murmurs, and 216 pathologic murmurs. Innocent murmurs collected include Still's murmur, venous hum, and flow murmurs. Pathologic murmurs included ventricular septal defect, tetralogy of Fallot, aortic regurgitation, aortic stenosis, pulmonary stenosis, mitral regurgitation and stenosis, and tricuspid regurgitation. We find that the Vision Transformer consistently outperforms the ResNet-50 on all three image representations, and that the Gramian angular field is the superior image representation for pediatric heart sounds. We calculated a one-vs-rest multi-class ROC curve for each of the three classes. Our best model achieves an area under the curve (AUC) value of 0.92 ± 0.05, 0.83 ± 0.04, and 0.88 ± 0.04 for identifying normal heart sounds, innocent murmurs, and pathologic murmurs, respectively. CONCLUSION: We present two novel methods for pediatric heart sound classification, which outperforms the current standard of using a convolutional neural network trained on spectrogram images. To our knowledge, we are the first to demonstrate multi-class classification of pediatric murmurs. Multiclass output affords a more explainable and interpretable model, which can facilitate further model improvement in the downstream model development cycle and enhance clinician trust and therefore adoption.

Chylous pericardial effusion with cardiac tamponade in a child treated with imatinib.

Chylous pericardial effusions are extremely rare outside of thoracic and cardiac surgery patients. We report the case of an 8-year-old girl with history of recurrent benign giant cell granulomas who developed a large chylous pericardial effusion with cardiac tamponade soon after beginning therapy with imatinib. In this article, we discuss the presentation, diagnosis, and management and review the published literature of this rarely reported side effect of this medication.

Cyanosis in a patient after Fontan palliation due to unrecognised hepatic vein to coronary sinus communication.

A 6-year-old male with heterotaxia, abnormal systemic and pulmonary venous drainage, and a history of Fontan completion presented with desaturations and was found by cardiac catheterisation to have a hepatic vein to coronary sinus connection. This was successfully occluded using an Amplatzer Muscular Ventricular Septal Defect Occluder.

Extubation in the operating room after congenital heart surgery in children.

OBJECTIVE: Early extubation in the operating room after surgery for congenital heart disease has been described; however, postoperative mechanical ventilation in the intensive care unit remains common practice in many institutions. The goal of this study was to identify perioperative factors associated with not proceeding with planned operating room extubation. METHODS: We performed a retrospective chart review of 224 patients (aged 1 month to 18 years, median 20 months) undergoing surgery for congenital heart defects requiring cardiopulmonary bypass. Patients mechanically ventilated preoperatively were excluded. A stepwise logistic regression model was used to test for the independent influence of various perioperative factors on extubation in the operating room. RESULTS: Overall, 79% of patients were extubated in the operating room. Younger age and longer cardiopulmonary bypass time were the strongest predictors for not extubating. Each step down to a younger age group (<2, 2-4, 4-6, 6-12, >12 months) reduced the chance of extubation in the operating room by 56%. Cardiopulmonary bypass time for more than 150 minutes was associated with an 11.8-fold increased risk of not being extubated. Male gender and high inotrope requirement after cardiopulmonary bypass were also significantly associated with fewer children being extubated. CONCLUSION: Extubation in the operating room after surgery for congenital heart disease was successful in the majority of patients. The strongest independent risk factors for failure of this strategy included younger age and longer cardiopulmonary bypass time.

Frequency of aberrant subclavian artery, arch laterality, and associated intracardiac anomalies detected by echocardiography.

Tetralogy of Fallot is generally considered to be the most common congenital heart defect associated with an aberrant subclavian artery (ASA), but the prevalence of ASA in patients with other cardiac anomalies is not well described. The pediatric echocardiography database, with 15,871 initial echocardiograms, was searched for all patients with ASA. Arch laterality and associated intracardiac anomalies were documented for each patient. ASA was found in 226 patients, of whom 171 had a left aortic arch (LAA) and 55 had a right aortic arch (RAA). The occurrence of ASA was 1% in patients with LAA (171 of 15,650) and 25% in patients with RAA (55 of 221; p = 0.001). Intracardiac anatomy was normal in 32% of patients with ASA/LAA and 25% with ASA/RAA. Conotruncal anomalies occurred more frequently with ASA/RAA than ASA/LAA (36% vs 18%; p = 0.01). Atrioventricular canal defects accounted for 10% and left-sided cardiac obstructive lesions accounted for 11% of subjects with ASA/LAA. ASA was rarely associated with d-transposition of the great arteries (1 of 226) and double-outlet right ventricle (5 of 226). The prevalence of ASA was highest in patients with interrupted aortic arch (11 of 38; 29%). In patients with tetralogy of Fallot, the overall prevalence of ASA was 8% (34 of 447), but was higher with RAA (16 of 103; 16%). The highest prevalence of ASA occurred in the subgroup of patients with tetralogy of Fallot with pulmonary atresia and RAA (6 of 25; 24%). In conclusion, ASA was more common in patients with RAA, especially with conotruncal anomalies. In patients with LAA, hypoplastic left heart syndrome, aortic coarctation, and atrioventricular canal defects were commonly associated with ASA.

Acute myopericarditis after multiple vaccinations in an adolescent: case report and review of the literature.

We report a case of postvaccination acute myopericarditis in an adolescent. The patient presented with acute chest pain, diffuse ST-segment elevation, and elevated cardiac enzyme levels. Cardiac MRI was consistent with acute myocarditis. He recovered within a few days with nonsteroidal antiinflammatory treatment and remains clinically stable, with improvement of MRI findings at the 10-week follow-up. Postvaccination cases of myopericarditis reported in the pediatric literature are also reviewed.