Scanning mode effects on relative electron density and effective atomic number accuracy in dual-layer spectral CT: A phantom study.

Abstract

Purpose: To assess the accuracy of relative electron density (RED) and effective atomic number (EAN) measurements using dual-layer spectral CT (DLSCT) in axial and helical scanning modes.

Methods: The CIRS 062M phantom was scanned with three collimation widths (16 × 0.625 mm, 32 × 0.625 mm, 64 × 0.625 mm) in axial mode and combined with two pitches (0.61 and 0.98) in helical mode. The RED and EAN were quantified using the IntelliSpace Portal, with absolute and relative errors calculated for each insert. Error metrics, including mean absolute error (MAE) and mean relative error (MRE), were subsequently derived over all the inserts. High accuracy was defined as the absolute error < 0.2. Statistical analysis was conducted using one-way ANOVA.

Results: In axial mode, the MAE for RED was within 0.01 and the MRE was within 2%; for EAN, the MAE was within 0.25 and the MRE was within 3%. In helical mode, the MAE for RED was within 0.02 and the MRE was within 2.5%; for EAN, the MAE was within 0.5 and the MRE was within 6%. Lung insert showed the highest deviation, while for all other inserts, the MAE was within 0.2 and the MRE was within 2% for EAN across both scanning modes. Smaller collimation widths in axial scanning showed lower relative errors for RED and EAN compared to helical scanning for bone and most of the soft inserts.

Conclusions: DLSCT accurately estimates RED and EAN in both scanning modes, suitable for clinical quantitative tissue analysis, with the exception of lung tissue. Smaller collimation widths in axial scanning are recommended for enhanced measurement accuracy and stability.