Truncation effect reduction for fast iterative reconstruction in cone-beam CT

Area of Science:

• Medical Imaging
• Computational Imaging

Background:

• Cone-beam computed tomography (CBCT) utilizes iterative reconstruction to enhance image quality and reduce radiation dose.
• Truncated data in CBCT, caused by projections exceeding detector size, degrades image quality within the field of view (FOV).

Purpose of the Study:

• To propose and evaluate a method for reducing truncation effects in fast iterative reconstruction for CBCT imaging.
• To improve image quality, particularly for soft-tissue visualization, in the presence of truncated projections.

Main Methods:

• Extrapolated projection data to a suitable size and reconstructed using fast iterative reconstruction.
• Modified a smoothing parameter for noise regularization to minimize accumulated errors during processing.
• Evaluated image quality using a head phantom and compared results with conventional filtered backprojection (FBP).

Main Results:

• Reconstructed images from the head phantom demonstrated improved image quality.
• Fast iterative reconstruction maintained consistent performance over multiple iterations.
• Enhanced contrast-to-noise ratio and visualization of soft tissues and ventricles compared to FBP at 5 mGy dose index.

Conclusions:

• The proposed method effectively reduces truncation effects in CBCT, leading to superior soft-tissue image quality within the FOV.
• Fast iterative reconstruction exhibited stable convergence characteristics over numerous iterations.