Deep Learning for Automated Measurement of Total Cardiac Volume for Heart Transplantation Size Matching

Nicholas A. Szugye,Neeraja Mahalingam,Elanchezhian Somasundaram,Chet Villa,Jim Segala,Michael Segala,Farhan Zafar,David L. S. Morales,Ryan A. Moore

DOI: https://doi.org/10.1007/s00246-024-03470-4

2024-04-05

Pediatric Cardiology

Abstract:Total Cardiac Volume (TCV)-based size matching using Computed Tomography (CT) is a novel technique to compare donor and recipient heart size in pediatric heart transplant that may increase overall utilization of available grafts. TCV requires manual segmentation, which limits its widespread use due to time and specialized software and training needed for segmentation. This study aims to determine the accuracy of a Deep Learning (DL) approach using 3-dimensional Convolutional Neural Networks (3D-CNN) to calculate TCV, with the clinical aim of enabling fast and accurate TCV use at all transplant centers. Ground truth TCV was segmented on CT scans of subjects aged 0–30 years, identified retrospectively. Ground truth segmentation masks were used to train and test a custom 3D-CNN model consisting of a DenseNet architecture in combination with residual blocks of ResNet architecture. The model was trained on a cohort of 270 subjects and a validation cohort of 44 subjects (36 normal, 8 heart disease retained for model testing). The average Dice similarity coefficient of the validation cohort was 0.94 ± 0.03 (range 0.84–0.97). The mean absolute percent error of TCV estimation was 5.5%. There is no significant association between model accuracy and subject age, weight, or height. DL-TCV was on average more accurate for normal hearts than those listed for transplant (mean absolute percent error 4.5 ± 3.9 vs. 10.5 ± 8.5, p = 0.08). A deep learning-based 3D-CNN model can provide accurate automatic measurement of TCV from CT images. This initial study is limited as a single-center study, though future multicenter studies may enable generalizable and more accurate TCV measurement by inclusion of more diverse cardiac pathology and increasing the training data.

pediatrics,cardiac & cardiovascular systems

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to automatically measure the Total Cardiac Volume (TCV) by using deep - learning techniques in pediatric heart transplantation, so as to achieve an accurate match of the heart sizes between donors and recipients. Traditionally, the size matching in heart transplantation is accomplished by comparing the body weights of donors and recipients, but the accuracy of this method has been shown to be insufficient in recent years. The TCV method can match the heart sizes more accurately by directly comparing the heart volumes of donors and recipients, which may increase the number of available donors. However, a major drawback of the TCV method is that it requires manual segmentation by trained imaging experts, which limits its wide application, because manual segmentation is time - consuming and requires specialized software and training, and these resources are not available in most transplant centers. For this reason, the research team has developed a deep - learning model based on 3D - Convolutional Neural Network (3D - CNN), aiming to automatically and accurately measure TCV from CT images. The training data set of this model contains 270 subjects, and the validation data set contains 44 subjects (among which 36 have normal hearts and 8 are heart disease patients). The main objective of the study is to create a model that can derive TCV from CT scan data of the pediatric population, so as to reduce the time of manual segmentation, decrease the inter - rater variability, and enable all transplant centers to use the directly measured TCV.

Deep Learning for Automated Measurement of Total Cardiac Volume for Heart Transplantation Size Matching

Deep learning enables the fully automated quantification of high-resolution cardiac CT images: human in loop for model development

Segmentation and Measurements of Major Blood Vessels from CT Pulmonary Angiography Using Deep Learning

Fully automatic cardiac four chamber and great vessel segmentation on CT pulmonary angiography using deep learning

Evaluation of deep learning estimation of whole heart anatomy from automated cardiovascular magnetic resonance short- and long-axis analyses in UK Biobank

Deep Learning in Cardiothoracic Ratio Calculation and Cardiomegaly Detection

Deep 3D attention CLSTM U-Net based automated liver segmentation and volumetry for the liver transplantation in abdominal CT volumes

Self-adaptive deep learning-based segmentation for universal and functional clinical and preclinical CT image analysis

Applying a Deep Learning Model for Total Kidney Volume Measurement in Autosomal Dominant Polycystic Kidney Disease

Improving Reproducibility of Volumetric Evaluation Using Computed Tomography in Pediatric Patients with Congenital Heart Disease

Automated biventricular quantification in patients with repaired tetralogy of Fallot using a 3D deep learning segmentation model

“Horses for courses” computed tomography or predicted total lung capacity for size matching in lung transplantation

Automatic quantification of the LV function and mass: a deep learning approach for cardiovascular MRI

Segmentation-free Estimation of Aortic Diameters from MRI Using Deep Learning

Automated segmentation of biventricular contours in tissue phase mapping using deep learning

A CT-based deep learning system for automatic assessment of aortic root morphology for TAVI planning

A Three-Dimensional Deep Convolutional Neural Network for Automatic Segmentation and Diameter Measurement of Type B Aortic Dissection

Real-time Cardiovascular MR with Spatio-temporal Artifact Suppression using Deep Learning - Proof of Concept in Congenital Heart Disease

Precision measurement of cardiac structure and function in cardiovascular magnetic resonance using machine learning

Left Heart Chamber Volumetric Assessment by Automated Three-Dimensional Echocardiography in Heart Transplant Recipients

AI-based, automated chamber volumetry from gated, non-contrast CT