A comprehensive dual energy method for CBCT metal artifact reduction.

Area of Science:

• Medical Imaging
• Radiology
• Cone Beam CT (CBCT)

Background:

• Metal artifacts are a significant limitation in CBCT imaging, degrading image quality and obscuring anatomical details.
• Existing metal artifact reduction methods often struggle to balance artifact removal with preservation of surrounding tissue information.

Purpose of the Study:

• To develop and validate a novel dual-energy based method for effective metal artifact reduction in CBCT.
• To improve the diagnostic accuracy of CBCT in the presence of metallic implants or high-attenuation materials.

Main Methods:

• Generation of virtual monoenergetic (VM) projections from high- and low-energy datasets to mitigate beam hardening effects.
• Application of the normalized metal artifact reduction (NMAR) technique on VM projections, followed by metal reintegration.
• Iterative reconstruction using VM projections and the NMAR CBCT as initial input for final image generation.

Main Results:

• The proposed dual-energy method demonstrated superior performance in artifact reduction and preservation of anatomical details compared to the frequency split metal artifact reduction (FSMAR) method.
• Quantitative analysis showed an improvement in Structural Similarity Index Measurement (SSIM) from 99.48% to 99.86% and a reduction in Root Mean Square Error (RMSE) from 93.62 HU to 70.75 HU.
• The method achieves efficient processing, with implementation times under two minutes using GPU acceleration.

Conclusions:

• The developed dual-energy based metal artifact correction method effectively reduces artifacts while preserving critical details in the metal-affected region.
• This technique integrates virtual monoenergetic imaging, projection interpolation, and iterative reconstruction for enhanced CBCT image quality.
• The method shows strong potential for clinical implementation due to its superior performance, efficiency, and preservation of anatomical structures.