Publications by authors named "Jim Segala"
Pediatr Cardiol
April 2024
Article Synopsis
- A new technique using Total Cardiac Volume (TCV) and CT scans aims to improve pediatric heart transplant outcomes by allowing better size matching of donor and recipient hearts, but current methods require extensive manual work.
- This study explores a Deep Learning approach, specifically a 3D Convolutional Neural Network (3D-CNN), to automatically measure TCV, demonstrating high accuracy in estimating heart size quickly.
- With a strong validation performance (average Dice coefficient of 0.94 and mean absolute percent error of 5.5%), the study emphasizes the need for future multicenter trials to enhance the model's applicability across different patient demographics and heart conditions.
Article Synopsis
- Total Cardiac Volume (TCV) size matching using CT scans can help match donor and recipient hearts for pediatric transplants, but current manual methods are time-consuming and require specialized training.
- This study investigates the effectiveness of a Deep Learning method using 3D Convolutional Neural Networks (3D-CNN) to quickly and accurately measure TCV, aiming to improve transplant matching across various centers.
- Results show that the deep learning model achieved a high accuracy (Dice coefficient of 0.94) and a low average error (5.5%) in estimating TCV, making it a promising tool for future pediatric heart transplants.
