Using flow information to support 3D vessel reconstruction from rotational angiography

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Cerebrovascular disease assessment requires 3D morphologic and hemodynamic data.
• Rotational angiography provides 3D vascular geometry but standard reconstruction assumes uniform contrast filling, leading to artifacts.
• Existing methods struggle with non-uniform contrast enhancement during acquisition.

Purpose of the Study:

• To develop and validate a method using blood flow information to improve 3D vessel reconstruction from rotational angiography.
• To address artifacts caused by non-uniform contrast agent distribution.
• To enable accurate quantitative blood flow assessment.

Main Methods:

• Utilized the fast marching algorithm guided by contrast agent arrival times and enhancement characteristics.
• Estimated bolus arrival time and enhancement intensity/duration from rotational time intensity curves (R-TICs).
• Reconstructed 3D vessel centerlines via backtracking and 3D radii from 2D estimates.

Main Results:

• Achieved mean absolute errors of 0.42 mm (centerline) and 0.25 mm (radius) in simulations.
• Obtained a mean absolute error of 0.45 mm for 3D centerline reconstruction in experimental datasets.
• Demonstrated successful 3D reconstruction under pulsatile and nonpulsatile flow conditions.

Conclusions:

• Flow information significantly improves 3D vessel reconstruction from rotational angiography, even with contrast inflow/outflow.
• X-ray rotation range (≥135°) is critical for accurate centerline and radius reconstruction.
• The method enables more reliable quantitative blood flow analysis from rotational angiography.