Salvaging low contrast abdominal CT studies using noise-optimised virtual monoenergetic image reconstruction

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• Dual-energy CT (DECT) is crucial for abdominal cancer staging.
• Impaired portal-venous contrast can compromise diagnostic accuracy.
• Virtual monoenergetic imaging (VMI) offers potential for image quality improvement.

Purpose of the Study:

• To evaluate the impact of noise-optimised VMI+ on image quality.
• To assess diagnostic performance in abdominal DECT with suboptimal contrast.
• To compare VMI+ reconstructions with standard imaging techniques.

Main Methods:

• Retrospective analysis of 11,746 abdominal DECT scans.
• Identification of 262 patients with poor portal-venous contrast.
• Reconstruction of standard linearly-blended and VMI+ (40, 50, 60 keV) images.
• Objective (SNR, CNR) and subjective (Likert scale) image quality assessment.

Main Results:

• VMI+ at 40 keV yielded the highest contrast-to-noise ratio (CNR) for organs and vessels.
• Subjective analysis showed peak image contrast at 40 keV VMI+.
• Optimal overall image quality was achieved with VMI+ at 50 keV (p < 0.001).

Conclusions:

• VMI+ reconstruction at 50 keV significantly improves image contrast and quality.
• The VMI+ algorithm effectively reduces non-diagnostic scans in poorly opacified abdominal DECT.
• This study validates VMI+ utility in a large cohort of challenging DECT studies.