Fig 1.
Illustration of the experimental phase-contrast computed tomography system.
Three gratings with micrometer-sized structures are installed at a rotating anode X-ray tube (Nonius FR 591). The spatial resolution of the photon-counting X-ray detector (Pilatus II) is (172 × 172) μm2. Two gratings G1 and G2 together resolve the very small refraction angles induced by an object in the beam. Grating G0 (placed directly behind the source) guarantees proper functioning of the grating interferometer when operated under standard laboratory X-ray conditions. The investigated samples were mounted in front of G1 and submerged in a water bath during measurements to avoid image artefacts. The water was cooled to 4°C to decelerate the decay process of the soft tissues.
Fig 2.
Representative phase-contrast imaging results of non-fixated and fixated tissue samples.
(a) Photograph of porcine fat and rind samples on which the effects of formalin fixation were investigated (note the presence of skin, muscle and adipose tissues). Samples were immersed in physiological saline or fixated with formaldehyde solutions of varying concentrations. (b) Phase-contrast axial slice through a non-fixated porcine sample. Skin, muscle and adipose tissues are clearly differentiated by their signal intensities. (c) Phase-contrast image of a sample fixated in typical preservation solution (containing 3.7% formaldehyde). (d) Phase-contrast image of a sample fixated in 18.5% formaldehyde solution. The signal intensities in b) are visually indistinguishable from those of (c).
Fig 3.
Graphical representation of the quantitative analysis of non-fixated and fixated tissue samples.
(a) Comparison of phase-contrast Hounsfield units evaluated for five non-fixated porcine samples, and five samples fixed in 3.7% formaldehyde solution. Formalin fixation does not affect adipose tissue but increases the Hounsfield units of muscle and skin by approximately 10%, most likely by tissue shrinkage. (b) Phase-contrast Hounsfield units of the three investigated tissue types fixated at different formaldehyde concentrations. Again, adipose tissue is insensitive to formaldehyde concentration, but the HUp of muscle and skin follow the same trend as the actual formaldehyde solutions. This effect is due to replacement of water by the formaldehyde solution within the tissues.
Table 1.
HUp values of five non-fixated (fresh) and five fixated (3.7% formaldehyde solution) tissue samples containing adipose tissue, muscle tissue and skin.
The numbers in regular typeface represent the mean values and standard deviations (round brackets) obtained from ten regions-of-interest that were analysed for each of the three tissue types and each individual sample. The bold numbers finally give the mean values and standard deviations (square brackets) of these individual results and reflect the differences in the tissues’ HUp in non-fixated and fixated state.
Table 2.
Effect of formaldehyde concentration on the quantitative phase-contrast Hounsfield unit of adipose tissue, muscle tissue and skin.
For higher formaldehyde concentrations, the increase in the Hounsfield unit of muscle tissue and skin correlates with the Hounsfield unit of the formaldehyde solution itself.
Fig 4.
Representative phase-contrast imaging results of non-fixated human tissue samples.
(a) Sample comprising tendon (bright, top), skin (bright, bottom), muscle (medium signal intensity) and adipose tissue (dark structures). (b) Heart muscle tissue with portions of fatty tissue. (c) Renal and adipose tissue. (d) Brain tissue. (e) Spleen tissue. (f) Piece of liver tissue. In each image, the high-intensity circular areas are the PMMA rods used to calibrate the quantitative phase-contrast Hounsfield units.
Table 3.
Phase-contrast Hounsfield units of various human soft tissue types evaluated from measurements and calculated from tabulated literature data.
Ten regions-of-interest were analysed for each tissue type and measurement. The listed numbers are the mean values and standard deviations (round brackets) of the regions’ averages.