Figure 5
From: Quantifying multiple stain distributions in bioimaging by hyperspectral X-ray tomography
Design and spectral analysis of chemical phantoms. (A) Illustration of the phantom layout for each contrast agent (left), with chemical concentrations labelled in units of mg/ml. The I2KI phantom is labelled as iodine to highlight that the concentrations represent aqueous I3−. Reconstructed image slices for each of the respective chemical phantoms (right), with orientation matched to the concentrations labelled. The BaSO4 phantoms appear heterogeneous due to partial sinking and separation of the barium sulphate crystals from the agarose base. Images are measured in terms of 'standardised intensity' value, with the same overall scaling range. (B) Reconstructed slices of each chemical phantom, shown in the axial plane, for two energy channels, taken just before (left column) and just after (right column) the relevant K-edge position. ROIs covering each chemical phase are highlighted and numbered, from which spectral plots were extracted. Minor ring artefacts appear on the PTA reconstructed phantom. (C) Voxel spectra showing attenuation as a function of energy for the ROIs highlighted in (B). The attenuation values were averaged over the full ROI. Known K-edge positions of the phantom’s heavy element are overlaid. (D) Fitted relationship for each chemical phantom, based on the average K-edge step change over six vertical slices through the sample depth. Error bars measure the standard deviation across the slices analysed. A line of best fit is applied for each phantom following linear interpolation. (E) Fitted relationship between chemical concentration and 'standardised intensity’ value in each phantom phase following conventional XCT imaging. Values were calculated for the central slice reconstruction. Phantom datasets were scanned at 80 kV and 100 µA with a Zeiss Xradia MicroXCT-400. Standardised intensity values are based on 2% agarose and air, as the barium phantom uses an agarose base.
