Fast three-material modeling with triple arch projection for electronic cleansing in CTC

Area of Science:

• Medical Imaging
• Radiology
• Computational Imaging

Background:

• Computed tomographic colonography (CTC) is a crucial tool for colorectal cancer screening.
• Artifacts, particularly at T-junctions, can obscure colonic structures, impacting diagnostic accuracy.
• Existing electronic cleansing methods face challenges in speed and preservation of subtle colonic features.

Purpose of the Study:

• To develop and evaluate a fast three-material modeling approach for electronic cleansing in CTC.
• To effectively remove ridge-shaped artifacts at T-junctions where air, soft-tissue, and tagged residues converge.
• To improve the speed and quality of colon surface reconstruction while preserving submerged folds.

Main Methods:

• Segmentation of colonic components including air, tagged residues (TRs), and transitional layers.
• Estimation of three-material fractions in T-junction regions using a novel triple arch projection model.
• Updating CT density values with a fold-preserving reconstruction model.
• Comparison with a previous two-material transition model.

Main Results:

• The proposed three-material modeling successfully removed T-junction artifacts in clinical datasets.
• The method achieved clear reconstruction of the entire colon surface, preserving submerged folds.
• A five-fold increase in processing speed was observed compared to the previous model.
• Similar cleansing quality in T-junction regions was maintained.

Conclusions:

• The fast three-material modeling offers a significant advancement in electronic cleansing for CTC.
• This technique enhances diagnostic confidence by improving visualization of the colon surface.
• The balance of speed, artifact removal, and fold preservation makes it a valuable tool for clinical application.