Image quality in CT: From physical measurements to model observers

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Diagnostic Imaging Technology

Background:

• Computed Tomography (CT) image quality (IQ) evaluation is crucial for accurate diagnostics and patient radiation dose optimization.
• Routine quality control of medical X-ray devices includes assessing individual IQ aspects.
• Evolving CT technology necessitates advanced IQ characterization methods.

Purpose of the Study:

• To review diverse methods for characterizing CT image quality.
• To cover objective physical measurements, human observer approaches, and task-based model observer (MO) methods.
• To propose updated figures of merit (FOM) for modern CT scanners, including iterative reconstruction.

Main Methods:

• Review of objective physical IQ measurements.
• Analysis of subjective human observer performance studies.
• Evaluation of model observer (MO) approaches for task-based IQ assessment.
• Integration of IQ metrics with dose indicators for dose efficiency analysis.

Main Results:

• Various IQ assessment methods exist, ranging from physical measurements to observer-based approaches.
• Model observer (MO) methods offer objective IQ assessment for clinically relevant tasks.
• Updated figures of merit (FOM) are needed to reflect technological advancements in CT.
• Combining IQ and dose metrics enables generalized dose efficiency indices.

Conclusions:

• The model observer (MO) approach shows significant promise for objectively assessing CT image quality relevant to clinical tasks.
• MO methods align with radiologist sensitivity performance, making them highly relevant for clinical environments.
• Optimizing CT scanners requires updated figures of merit (FOM) that incorporate advanced reconstruction algorithms and dose efficiency.