Publications by authors named "W Border"
Article Synopsis
- Researchers investigated how machine learning could improve predicting cardiomyopathy risk in childhood cancer survivors using echocardiograms from multiple centers.
- They designed deep convolutional neural networks (DCNNs) to analyze different sets of echocardiographic data and gauge model performance using standard metrics like F1-score and AUPRC.
- Results showed that the DCNN effectively distinguished between survivors who developed cardiomyopathy and those who didn't, suggesting a potential tool for early prediction in at-risk patients.
Objective: Risk stratification of fetuses diagnosed with congenital heart disease (CHD) helps provide a delivery plan and prepare families and medical teams on expected course in the delivery room. Our aim was to assess the accuracy of echocardiographically determined risk-stratification assignments in predicting postnatal cardiac outcomes beyond the delivery room.
Study Design: This was a retrospective study at a single center evaluating all fetuses with CHD who were risk-stratified by echocardiographically determined level of care (LOC) assignment, ranging from 1a (lowest risk) to 4 (highest risk).
The Fate of the Left Ventricular Outflow Tract Following Interrupted Aortic Arch Repair.
World J Pediatr Congenit Heart Surg
September 2024
To examine the probability of left ventricular outflow tract (LVOT) reintervention following interrupted aortic arch (IAA) repair in neonates with LVOT obstruction (LVOTO) risk. This retrospective multicenter study included 150 neonates who underwent IAA repair (2003-2017); 100 of 150 (67%) had isolated IAA repair (with ventricular septal defect closure) and 50 of 150 (33%) had concomitant LVOT intervention: conal muscle resection (n = 16), Ross-Konno (n = 7), and Yasui operation (n = 27: single-stage n = 8, staged n = 19). Demographic and morphologic characteristics were reviewed.
View Article and Find Full Text PDFBackground: A previous multicenter study showed that longitudinal changes in standard cardiac functional parameters were associated with the development of cardiomyopathy in childhood cancer survivors (CCS). Evaluation of the relationship between global longitudinal strain (GLS) changes and cardiomyopathy risk was limited, largely due to lack of quality apical 2- and 3-chamber views in addition to 4-chamber view. We sought to determine whether apical 4-chamber longitudinal strain (A4LS) alone can serve as a suitable surrogate for GLS in this population.
View Article and Find Full Text PDF