Related Experiment Video
Updated: Aug 26, 2025
06:18
Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions
Published on: April 5, 2024
1.2K
Self-supervised learning enables 3D digital subtraction angiography reconstruction from ultra-sparse 2D projection
Huangxuan Zhao1, Zhenghong Zhou2, Feihong Wu1
1Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430022, China; Hubei Province Key Laboratory of Molecular Imaging, Wuhan 430022, China.
Cell Reports. Medicine
|October 8, 2022
Summary
This study introduces a novel self-supervised learning method for 3D digital subtraction angiography (DSA) reconstruction using minimal X-ray projections. This approach significantly reduces radiation exposure while maintaining diagnostic accuracy for intracranial aneurysms.
Area of Science:
- Medical Imaging
- Radiology
- Artificial Intelligence
Background:
- 3D digital subtraction angiography (DSA) is crucial for diagnosing and treating intracranial aneurysms (IAs).
- Current gold standard 3D DSA reconstruction requires numerous projection views, leading to high radiation dosage.
- A need exists for methods reducing radiation exposure without compromising reconstruction quality.
Purpose of the Study:
- To develop and validate a self-supervised learning method for 3D DSA reconstruction using ultra-sparse 2D projections.
- To significantly reduce radiation dosage in 3D DSA while maintaining diagnostic accuracy.
Main Methods:
- A self-supervised learning framework was employed for 3D DSA reconstruction.
- The method utilized ultra-sparse 2D projection data.
- Validation was performed on 202 patient cases from multiple hospitals.
Main Results:
- The proposed method achieved high-quality 3D DSA reconstruction using only eight projections.
- Radiologists identified all 82 lesions with high diagnostic confidence.
- Radiation dosage was reduced by approximately 16.7 times compared to the gold standard.
Conclusions:
- Self-supervised learning enables revolutionary 3D DSA reconstruction from ultra-sparse projections.
- This method offers a significant reduction in radiation dosage for IA diagnosis and treatment.
- The technique shows promise for clinical application in neurovascular imaging.
