Single-pass metal artifact reduction using a dual-layer flat panel detector.

Main Methods:

• Metal objects were directly detected in projections using DE imaging, generating material-specific images where metal is clearly delineated.
• Metal detection was achieved through thresholding and entropy filtration to minimize false positives.
• Scatter correction was applied to DE raw projections to enhance material decomposition accuracy.
• Image reconstruction was performed using filtered backprojection after correcting identified metal regions via interpolation.
• The method was evaluated using phantoms mimicking liver biopsy and a cadaver head with dental fillings and external metal tags.

Main Results:

• The proposed DE-based MAR method demonstrated robustness to FOV truncation, outperforming standard two-pass reconstruction in such cases.
• Metal segmentation accuracy was superior to Markov Random Field (MRF) and single-energy methods, which were prone to false-positive errors.
• For a liver biopsy phantom, spatial nonuniformity was reduced from 0.127 to 0.077.
• In a cadaver head scan, average standard deviation in soft tissue regions decreased from 209.1 HU to 46.8 HU.
• Processing time was reduced by 31% compared to the two-pass method.

Conclusions:

• A novel, single-pass MAR method using DE imaging has been successfully developed.
• The method is robust to truncation artifacts and offers superior performance compared to single-energy imaging techniques.
• This approach significantly reduces processing time, offering a more efficient solution for metal artifact reduction in CBCT.