A comparative study of artifact reduction techniques in metal-implanted CT scans

Area of Science:

• Medical Imaging
• Radiology
• Image Processing

Background:

• X-ray computed tomography (CT) is crucial in clinical applications but suffers from metal artifacts caused by implants.
• These artifacts, appearing as lines, degrade image quality and hinder diagnostic accuracy and radiotherapy planning.

Purpose of the Study:

• To evaluate the effectiveness of five metal artifact reduction (MAR) algorithms.
• To compare their performance in reducing artifacts from metallic implants in CT images.

Main Methods:

• Assessed five MAR algorithms: linear interpolation (LI-MAR), normalization metal artifact reduction (NMAR), metal removal (MDT), orthopedic implant MAR (OMAR), and iteration-based (MAP).
• Evaluated algorithms using simulation and clinical studies with diverse metal implants and body regions.
• Quantitatively assessed image quality using normalized root mean square error (NRMSE).

Main Results:

• The normalization metal artifact reduction (NMAR) algorithm demonstrated superior efficiency in reducing metal artifacts across most cases.
• NMAR also showed favorable processing time, indicating clinical applicability.
• NRMSE analysis provided quantitative evidence of artifact reduction effectiveness.

Conclusions:

• NMAR is a highly effective algorithm for reducing metal artifacts in CT images.
• The NMAR algorithm offers significant clinical value due to its performance and efficiency.