Skip to Main Content

An official website of the United States government

Analysis of Chest X-rays Based on Deep Learning for Health Risk Assessment

Principal Investigator
Name
Yu Fei Huang
Degrees
M.E.
Institution
China Medical University
Position Title
postgraduate
Email
About this CDAS Project
Study
NLST (Learn more about this study)
Project ID
NLST-1351
Initial CDAS Request Approval
Nov 5, 2024
Title
Analysis of Chest X-rays Based on Deep Learning for Health Risk Assessment
Summary
This project leverages deep learning techniques to analyze chest X-rays (CXR) for health risk assessment and biological age prediction. By automatically segmenting and extracting features from key anatomical structures in chest images, the study aims to quantify changes in structures such as the lungs, heart, and blood vessels, which are indicative of the biological aging process. Additionally, the research will establish a multi-task learning framework that integrates anatomical feature extraction and disease risk screening for high-risk conditions. The goal of this project is to advance the use of medical imaging in proactive health management, providing novel scientific insights for preventive health assessments at an individual level.
Aims

Develop a Deep Learning-Based CXR Image Analysis Model: Construct and optimize a deep learning model focused on analyzing anatomical structures in chest X-rays (CXR) to predict individual biological age, providing technical support for personalized health assessment in medical imaging.
Health Risk Assessment and Early Screening: Extract features from CXR images to assess potential health risks associated with lung diseases. Explore non-invasive, image-based early screening methods to improve disease prevention in high-risk populations.
Multi-Task Learning Framework Construction: Develop an integrated multi-task learning framework that combines anatomical labeling and lesion detection to enhance model accuracy and robustness, thereby supporting multifunctional applications in image analysis.
Clinical Translation and Application of the Model: Validate the feasibility and effectiveness of the developed model in real clinical settings, further exploring its broad application potential in medical imaging, and advancing the use of artificial intelligence in medical diagnostics and health management.

Collaborators

1.Professor Jiang Xiran, China Medical University
2.Department of Radiology, The First Affiliated Hospital of China Medical University