Dynamic-dual-energy spectral CT for improving multi-material decomposition in image-domain

Area of Science:

• Medical Imaging
• Computed Tomography
• Image Reconstruction

Background:

• Dual-energy CT (DECT) shows promise for material decomposition but is limited to two energy bins, hindering multi-material analysis.
• Conventional spectral CT (sCT) offers more energy bins but suffers from high statistical noise due to low photon counts, compromising material decomposition accuracy.
• Existing methods struggle with noise and artifacts, limiting the quality of multi-material decomposition in CT imaging.

Purpose of the Study:

• To propose and validate a dynamic-dual-energy spectral CT (DDECT) method for accurate multi-material decomposition.
• To improve image reconstruction and reduce noise and artifacts in spectral CT.
• To demonstrate the feasibility of DDECT for high-quality material decomposition applications.

Main Methods:

• Developed a dynamic-dual-energy detector where the energy threshold randomly changes during scanning to capture spectral photon information.
• Implemented a statistical noise-weighted tPRISM algorithm for multi-energy image reconstruction from DDECT data.
• Performed multi-material decomposition on the reconstructed images.

Main Results:

• DDECT achieved more accurate multi-energy image reconstruction and multi-material decomposition compared to conventional static methods with the same number of energy bins.
• The proposed method significantly reduced noise levels in the reconstructed images.
• Ring artifacts, prevalent in conventional sCT, were substantially mitigated using the DDECT approach in both simulations and experiments.

Conclusions:

• The dynamic-dual-energy spectral CT method is highly feasible for improving multi-material decomposition.
• This technique offers a significant potential for high-quality spectral CT imaging and material analysis.
• DDECT provides superior accuracy and reduced noise/artifacts compared to traditional spectral CT methods.