Answers are grounded in this paper and cite their sources.
Evaluates effective atomic number and electron density imaging in varied samples, indicating its potential for tissue characterization.
Objective. To develop and evaluate a scanner-specific power law method for voxelwise effective atomic number (Z eff ) and electron density (ρ e ) imaging on a commercial spectral photon-counting CT (SPCCT) system, and to demonstrate its applicability to tissue characterization in heterogeneous samples. Approach. The X-ray linear attenuation coefficient µ(E) was modeled as a power law sum of photoelectric, Compton and empirical correction terms. Model coefficients were fitted to NIST cross sections for Z = 6–21 over the scanner energy range from 7 to 115 keV. A QRM spectral-CT phantom containing soft-tissue-equivalent, hydroxyapatite and iodine inserts was scanned on a MARS Microlab 5×120 SPCCT system in five detector energy bins. For each bin, reconstructed SPCCT µ images were linearly calibrated to NIST µ, and the calibrated µ(E) served as input to four-window and dual-energy models to estimate voxelwise Z eff and ρ e . The same calibration and model were applied, without retuning, to a meat sample with embedded iodine vials. Main results. Using NIST µ as input, the four-window formalism gave mean absolute errors of approximately 1% for both Z eff and ρ e . On calibrated SPCCT images, errors were 4.9% and 2.1% for Z eff and ρ e with the four-window model, and 3.6% and 1.7% with the best dual-energy pair (50–60 / 60–79 keV). The same framework, applied without retuning, separated fat, muscle, bone and iodine in the heterogeneous meat sample. Significance. A scanner-specific power law model of µ(E) combined with bin-wise linear calibration can deliver accurate Z eff and ρ e from a single SPCCT acquisition. Optimized dual-energy windows achieve near-percent ρ e accuracy for clinically relevant materials, providing a quantitative basis for tissue characterization with photon-counting CT.
No takes yet. Share an insight, caveat, or question.
Alzaabi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: