Technical performance of a dual-energy CT system with a novel deep-learning based reconstruction process: Evaluation...

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• A novel tube voltage-switching dual-energy (DE) CT system with deep learning reconstruction has been developed.
• Characterizing its performance is crucial for understanding its clinical utility.

Purpose of the Study:

• To evaluate the technical performance of a new DECT system.
• Compare it against standard single-kilovolt (kV) CT and a rotate/rotate DECT system for abdominal imaging.

Main Methods:

• Acquired phantom images using kV-switching DECT, rotate/rotate DECT, and single-kV CT.
• Reconstructed virtual monochromatic images (VMIs) and iodine maps.
• Assessed image quality using task transfer functions (TTFs), noise power spectrum (NPS), and contrast-to-noise ratios (CNRs).
• Evaluated low-contrast visibility, beam-hardening artifacts, and radiation dose (CTDIvol).

Main Results:

• kV-switching DECT demonstrated higher TTFs than single-kV and rotate/rotate DECT.
• 70 keV VMIs exhibited lower noise (12.0 HU) compared to single-kV (18.3 HU).
• Iodine quantification accuracy was within 1 mg I/ml, with CNRs up to 2.5 times higher than single-kV.
• Radiation dose (CTDIvol) was higher for DECT at manufacturer-recommended settings.

Conclusions:

• The novel DECT system produces sharper, lower-noise VMIs with superior iodine CNR compared to single-kV CT.
• Despite potentially higher radiation dose, the enhanced image quality offers significant advantages for abdominal imaging.