Fast shading correction for cone-beam CT via partitioned tissue classification.

Area of Science:

• Medical Imaging
• Radiation Oncology
• Image Processing

Background:

• Cone beam computed tomography (CBCT) is vital in radiation therapy but suffers from shading artifacts, limiting its quantitative use.
• Existing correction methods often require planning CT (pCT) images, reducing clinical practicality.

Purpose of the Study:

• To develop an effective CBCT shading artifact correction method that does not require prior CT images.
• To enhance the clinical applicability and image quality of CBCT for radiation therapy.

Main Methods:

• A novel method utilizing partitioned tissue classification to generate a 'shading-free' template image from CBCT data.
• Employing sparse sampling of shading artifacts and local filtration (a Fourier transform-based algorithm) for efficient correction.
• Validation using an anthropomorphic pelvis phantom and 6 pelvis patient datasets.

Main Results:

• The proposed method significantly reduces signal non-uniformity (SNU) and maximum CT number errors in CBCT images of phantoms and patients.
• On average, SNU was reduced from 9.22% to 1.06% (fat) and 11.41% to 1.67% (muscle) in patient data.
• Maximum CT number errors were reduced from 95 HU to 9 HU (fat) and 88 HU to 8 HU (muscle) in patient data.
• Processing time for one CBCT dataset is approximately 45 seconds on a standard PC.

Conclusions:

• The developed method effectively corrects CBCT shading artifacts without prior CT information, achieving image quality comparable to pCT.
• This approach enhances the quantitative accuracy and clinical utility of CBCT in radiation therapy.
• The method is efficient and practical for routine clinical application.