Image filtering for improved dose resolution in CT polymer gel dosimetry

Area of Science:

• Medical Physics
• Radiological Imaging
• Image Processing

Background:

• X-ray computed tomography (CT) is used for dosimetry with polyacrylamide gels (PAGs).
• Irradiation causes density changes in PAGs, creating contrast in CT images.
• Low dose resolution currently limits clinical use.

Purpose of the Study:

• To investigate image filtering techniques for reducing noise and enhancing dose resolution in PAG CT images.
• To evaluate the impact of various filters on dose resolution and spatial integrity.

Main Methods:

• Analyzed CT image noise characteristics (Gaussian, independent of contrast).
• Applied several spatial noise reduction filters (mean, median, midpoint, adaptive mean, alpha-trimmed mean, sigma mean, SUSAN) with varying mask sizes (3x3, 5x5, 7x7).
• Calculated dose resolution (D(delta)95%) and assessed spatial integrity of the dose distribution for each filtered image.

Main Results:

• Filters vary significantly in their ability to improve dose resolution and maintain spatial integrity.
• Larger mask sizes generally improve dose resolution but degrade spatial information.
• Mean filter offers the best dose resolution improvement but with significant spatial distortion.
• SUSAN, adaptive mean, and alpha-trimmed mean filters provide comparable dose resolution with better spatial integrity.
• SUSAN and adaptive filters are identified as the best performing for this application.

Conclusions:

• Postprocessing PAG CT images with spatial noise reduction filters can significantly improve dose resolution for CT dosimetry.
• The choice of filter is critical, as performance varies in dose resolution improvement and spatial integrity preservation.
• SUSAN and adaptive filters are recommended for balancing dose accuracy and spatial information in clinical CT PAG dosimetry.