Related Experiment Video
Updated: May 22, 2025
Retrospective Cardiac Gating with A Prototype Small-Animal X-ray Computed Tomograph
Published on: February 21, 2025
Comparing two deep learning spectral reconstruction levels for abdominal evaluation using a rapid-kVp-switching
Hakki Serdar Sagdic1,2, Mohammadreza Hosseini-Siyanaki3,4, Abheek Raviprasad3,4
1Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, Gainesville, FL, USA. drsagdic@gmail.com.
Strong-level Deep Learning Spectral Reconstruction (DLSR) significantly enhances dual-energy CT (DECT) abdominal scan image quality. This advanced DLSR improves signal-to-noise and contrast-to-noise ratios for better diagnostic evaluation.
Area of Science:
- Medical Imaging
- Radiology
- Computational Imaging
Background:
- Dual-energy CT (DECT) provides material decomposition capabilities.
- Deep Learning Spectral Reconstruction (DLSR) is an emerging technique to enhance DECT image quality.
- Limited research exists on DLSR for human abdominal DECT scans.
Purpose of the Study:
- To evaluate and compare image quality between standard and strong DLSR levels in abdominal DECT.
- To identify optimal virtual monochromatic image (VMI) energy levels for abdominal DECT using DLSR.
- To quantitatively and qualitatively assess the impact of DLSR on DECT image quality.
Main Methods:
- Retrospective analysis of 51 abdominal/pelvic DECT scans.
- Reconstruction of VMIs from 35 to 200 keV using standard and strong DLSR.
- Quantitative assessment of signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) in abdominal structures.
- Qualitative assessment of VMIs at specific low energy levels (35-65 keV).
Main Results:
- Strong DLSR significantly outperformed standard DLSR in SNR and CNR (p < 0.0001).
- Optimal SNR was achieved at 70 keV, and optimal CNR at 65 keV (p < 0.0001).
- Significant differences in qualitative scores between DLSR levels were noted at 45, 55, and 65 keV (p < 0.0001).
Conclusions:
- A strong DLSR level significantly improves abdominal DECT image quality over standard DLSR.
- Enhanced SNR and CNR with strong DLSR may improve diagnostic accuracy in abdominal imaging.
- 65 keV VMIs demonstrated the highest CNR with strong DLSR, differing from traditional DECT reconstruction methods.
