BENCHMARKING OF CT FOR PATIENT EXPOSURE OPTIMISATION.

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Diagnostic Imaging Technology

Background:

• Patient dose optimization in computed tomography (CT) is crucial for diagnostic imaging.
• Assessing image quality using clinically relevant tasks is essential for effective dose reduction strategies.
• Standardized methods are needed to evaluate low-contrast detectability across diverse CT systems.

Purpose of the Study:

• To assess low-contrast detectability in CT imaging for average patient morphology.
• To evaluate the performance of 56 different CT units using a consistent image quality assessment method.
• To determine the impact of CT dose index (CTDIvol) on image quality metrics.

Main Methods:

• Utilized an anthropomorphic phantom with spheres to simulate average patient morphology.
• Employed a channelized Hotelling observer (CHO) with dense difference of Gaussian channels for image analysis.
• Performed receiver operating characteristics (ROC) analysis, using the area under the curve (AUC) as the primary figure of merit.

Main Results:

• A small disparity in image quality was observed across CT units at a CTDIvol of 15 mGy.
• For 8-mm targets, AUC values ranged from 0.999 ± 0.018 at 20 HU to 0.927 ± 0.054 at 10 HU.
• For 5-mm targets, AUC values were 0.947 ± 0.059 at 20 HU and 0.702 ± 0.068 at 10 HU.

Conclusions:

• The channelized Hotelling observer (CHO) provides a robust method for assessing CT image quality.
• The findings support the use of CHO for benchmarking and optimizing CT protocols.
• Clinically relevant image quality assessments are vital for effective patient dose optimization in CT.