3D stent recovery from one X-ray projection

Area of Science:

• Medical Imaging
• Computer-Aided Surgery
• Biomedical Engineering

Background:

• Current endovascular abdominal aortic repair (EVAR) relies on 2D imaging, lacking depth information.
• This limitation makes precise stent graft placement challenging, particularly for junior surgeons.
• Advanced 3D visualization is crucial for improving EVAR accuracy and safety.

Purpose of the Study:

• To develop a novel algorithm for automatic 3D stent graft model registration to intraoperative 2D images.
• To enhance the accuracy and robustness of stent graft placement during EVAR procedures.
• To reduce reliance on user interaction for device positioning.

Main Methods:

• A novel algorithm for automatic matching of 3D stent graft models to intraoperative 2D images.
• Utilized automatic preprocessing and a global-to-local registration approach.
• Employed a semi-simultaneous optimization strategy with geometric constraints to reduce complexity.

Main Results:

• The algorithm demonstrated accurate matching of the 3D stent graft model to 2D image data.
• Successful validation was achieved using synthetic, phantom, and real interventional data.
• The method achieved robustness without requiring user interaction.

Conclusions:

• The presented algorithm offers an accurate and robust solution for 3D stent graft registration in EVAR.
• This technology has the potential to improve surgical outcomes and reduce complications.
• Automated 3D visualization aids in precise stent graft deployment, benefiting surgeons of all experience levels.