Deformable respiratory motion correction for hepatic rotational angiography.

Area of Science:

• Medical Imaging
• Image Reconstruction
• Computational Anatomy

Background:

• Cone-beam rotational angiography provides 3D hepatic vasculature imaging crucial for transcatheter arterial chemoembolization.
• Respiratory motion during acquisition introduces artifacts like blurring and streaking, degrading image quality.
• Existing reconstruction algorithms struggle with intra-scan motion, necessitating motion correction strategies.

Purpose of the Study:

• To develop and evaluate an automated method for respiratory motion estimation and compensation in cone-beam rotational angiography.
• To improve the quality of 3D hepatic vasculature imaging for enhanced guidance of chemoembolization procedures.
• To reduce artifacts caused by respiratory motion in rotational angiography reconstructions.

Main Methods:

• An automated method was developed using registration of a 3D arterial model to vesselness-enhanced 2D projection images.
• Arterial tree centerlines were modeled as time-dependent B-splines, with control points optimized via graph cuts (α-expansion).
• The method estimates both 3D rigid translations and non-rigid deformations caused by respiratory motion.

Main Results:

• The proposed method demonstrated notable reductions in reprojection error in both pre-clinical and clinical datasets.
• Significant increases in vessel sharpness were observed, with reduced blurring and streaking artifacts.
• The motion-compensated reconstructions showed improved vessel contrast compared to uncompensated data.

Conclusions:

• The automated motion estimation and compensation method effectively reduces artifacts in cone-beam rotational angiography.
• This technique enhances the visualization of hepatic vasculature, offering superior guidance for chemoembolization.
• The generic nature of the method suggests potential applicability to other rotational angiography protocols, such as coronary angiography.