Optimization of Low-Contrast Detectability in Abdominal Imaging: A Comparative Analysis of PCCT, DECT, and SECT...

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• Clear anatomical representation is crucial for diagnosing pathologies in computed tomography (CT).
• Photon-counting CT (PCCT) and advanced iterative reconstruction (IR) algorithms offer potential for enhanced low-contrast detectability.
• Task-based assessments are vital for optimizing CT protocols and comparing image quality across different CT technologies.

Purpose of the Study:

• To evaluate low-contrast detectability in abdominal CT protocols using PCCT, DECT, and SECT.
• To compare the performance of PCCT, DECT, and SECT in abdominal imaging scenarios.
• To utilize both model and human observers for a comprehensive assessment of image quality.

Main Methods:

• Acquired data using standard resolution scan modes on PCCT and DECT/SECT scanners.
• Investigated detectability using a low-contrast module phantom simulating abdominal conditions.
• Adjusted various imaging parameters including tube potentials, virtual monoenergetic image (VMI) energy levels, radiation doses, and iterative reconstruction settings.
• Quantified noise power spectrum, task transfer function, and detectability index using model and human observers.

Main Results:

• Image noise was minimized at specific VMI energy levels for PCCT and DECT.
• PCCT with VMI demonstrated superior low-contrast detectability compared to DECT VMI and DECT mixed imaging.
• Low-contrast detectability in PCCT could be maintained at reduced doses by increasing quantum iterative reconstruction (QIR) strength.

Conclusions:

• Optimizing VMI energy levels in PCCT and DECT is essential for maximizing low-contrast detectability by minimizing image noise.
• PCCT shows potential for improved low-contrast detectability and possible dose reduction in abdominal protocols compared to DECT and SECT.
• The non-prewhitening model observer with internal noise provided a more accurate quantification of low-contrast detectability.