CHD-CXR: a de-identified publicly available dataset of chest x-ray for congenital heart disease

Congenital heart disease is a prevalent birth defect, accounting for approximately one-third of major birth defects. The challenge lies in early detection, especially in underdeveloped medical regions where a shortage of specialized physicians often leads to oversight. While standardized chest x-rays can assist in diagnosis and treatment, their effectiveness is limited by subtle cardiac manifestations. However, the emergence of deep learning in computer vision has paved the way for detecting subtle changes in chest x-rays, such as lung vessel density, enabling the detection of congenital heart disease in children. This highlights the need for further investigation. The lack of expert-annotated, high-quality medical image datasets hinders the progress of medical image artificial intelligence. In response, we have released a dataset containing 828 DICOM chest x-ray files from children with diagnosed congenital heart disease, alongside corresponding cardiac ultrasound reports. This dataset emphasizes complex structural characteristics, facilitating the transition from machine learning to machine teaching in deep learning. To ascertain the dataset's applicability, we trained a preliminary model and achieved an area under the receiver operating characteristic curve (ROC 0.85). We provide detailed introductions and publicly available datasets at: https://www.kaggle.com/competitions/congenital-heart-disease.

The development and validation of an artificial intelligence-based screening method for atrial septal defect in children's chest x-rays

Purpose

For precise diagnosis and effective management of atrial septal defects, it is of utmost significance to conduct elementary screenings on children. The primary aim of this study is to develop and authenticate an objective methodology for detecting atrial septal defects by employing deep learning (DL) on chest x-ray (CXR) examinations.

Methods

This retrospective study encompassed echocardiographs and corresponding Chest x-rays that were consistently gathered at Qingdao Women's and Children's Hospital from 2018 to 2022. Based on a collaborative diagnosis report by two cardiologists with over 10 years of experience in echocardiography, these radiographs were classified as positive or negative for atrial septal defect, and then divided into training and validation datasets. An artificial intelligence model was formulated by utilizing the training dataset and fine-tuned using the validation dataset. To evaluate the efficacy of the model, an assessment of the area under the curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value was conducted employing the validation dataset.

Results

This research encompassed a total of 420 images from individuals. The screening accuracy and recall rate of the model surpass 90%.

Conclusions

One of profound neural network models predicated on chest x-ray radiographs (a traditional, extensively employed, and economically viable examination) proves highly advantageous in the assessment for atrial septal defect.