Quantitative accuracy of CT numbers: Theoretical analyses and experimental studies

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Diagnostic Imaging Technology

Background:

• Accurate CT number values are vital for reliable clinical diagnosis.
• Controlling CT number bias is essential when implementing strategies for radiation dose reduction.
• Understanding the relationship between radiation dose, image contrast, and CT number accuracy is critical for optimizing CT protocols.

Purpose of the Study:

• To investigate the relationship between CT number bias and radiation dose levels in Computed Tomography (CT).
• To examine how image contrast (CT number difference between ROI and background) influences CT number bias.
• To provide insights into maintaining CT number accuracy under reduced radiation exposure.

Main Methods:

• A theoretical lesion-background model was developed to analyze CT number bias variations with radiation exposure and contrast.
• Experimental validation was performed using benchtop and clinical MDCT scanners with photon-counting and energy-integrating detectors.
• Phantoms (Catphan and anthropomorphic) were scanned at varying mAs levels to evaluate bias across different contrast scenarios.

Main Results:

• CT number bias demonstrated an inverse proportionality to radiation exposure levels.
• A linear dependence of CT number bias on the CT number contrast between the region of interest (ROI) and the background was observed.
• Bias was found to increase as the milliampere-second (mAs) setting decreased, with both positive and negative biases occurring.

Conclusions:

• Quantitative accuracy of CT numbers requires careful consideration, especially during radiation dose reduction.
• The filtered backprojection (FBP) reconstruction algorithm's bias is sensitive to changes in radiation dose (mAs).
• The observed bias is influenced by the contrast difference between the target ROI and surrounding tissues, highlighting the need for protocol optimization.