Enhanced gray-white matter differentiation on non-enhanced CT using a frequency selective non-linear blending

Area of Science:

• Neuroradiology
• Medical Imaging Analysis
• Image Processing Algorithms

Background:

• Non-enhanced CT (NECT) of the brain is a common diagnostic tool.
• Distinguishing gray matter (GM) and white matter (WM) on NECT can be challenging.
• Improving GM-WM contrast can enhance diagnostic accuracy.

Purpose of the Study:

• To evaluate the efficacy of frequency selective non-linear blending in enhancing GM-WM contrast on NECT.
• To quantify the improvement in contrast-to-noise ratio (CNR) using this novel algorithm.
• To determine if the enhancement improves the ability to differentiate between gray and white matter.

Main Methods:

• Retrospective analysis of 30 NECT brain scans.
• Application of a frequency selective non-linear blending algorithm (best contrast, BC) with optimized settings.
• Measurement of GM and WM attenuation values and calculation of CNR for both NECT and BC images.
• Independent review by two radiologists.

Main Results:

• Significant increase in CNR for cortical gray matter (5.91 ± 2.45) and basal ganglia (4.41 ± 1.82) with BC.
• Contrast ratio between cortical GM and WM increased from 1.43 to 1.87 with BC (p < 0.0001).
• Contrast ratio between basal ganglia and WM increased from 1.33 to 1.7 with BC (p < 0.0001).

Conclusions:

• Frequency selective non-linear blending effectively enhances GM-WM contrast on NECT.
• The algorithm improves the discrimination between gray and white matter.
• This technique holds potential for increasing the diagnostic accuracy of brain NECT.