Fig 1.
Technical drawing of the 5 mm stenosis-phantom.
The stenosis phantom is shown in its cross section (a) and its longitudinal section (c). For better illustration, the stenosis phantom is also shown as an oblique view (b). It can be seen that the stenosis phantom is a cylinder with a length of 70 mm and a diameter of 20 mm. The lumen (not shaded in (c)) consists of three segments: The two outer segments with a length of 24.5 mm and an inner diameter of 10 mm, which was considered the “normal lumen”. The symmetrical central segment with a length of 21 mm with a 3 mm long cone-shaped narrowing at both its endings, which led to the characteristic stenosis, a 15 mm long cylinder with a diameter of 1–9 mm, in this case 5 mm.
Fig 2.
MPI-Images of the reference phantoms.
Sagittal (a) and axial (b) slices extracted from reconstructed 3D image data of the reference phantoms. Note the anisotropy of the axial images due to the different gradient field strengths, leading to an oval appearance of the phantoms which is the more pronounced, the smaller the lumen is.
Fig 3.
MPI image of the 5 mm stenosis phantom.
Image of the 5 mm stenosis phantom reconstructed as covered by the focus fields in z-/x-plane. One end was not fully covered as it extended beyond the encoded volume. It is thus blurry and would not be used in a quantitative evaluation.
Fig 4.
MPI-Images of the stenosis phantoms.
Sagittal (a) and axial (b) slices extracted from reconstructed 3D image data of the stenosis phantoms. The lumen of the 1 mm stenosis is not discernible. Note the anisotropy of the axial images due to the different gradient field strengths, leading to an oval appearance of the phantoms which is the more pronounced, the smaller the lumen is.
Fig 5.
Scatterplot comparing the calculated and measured degree of stenosis.
The bisectrix indicates the ideal agreement of calculated (x-axis) and measured (y-axis) values. The values of the stenosis phantoms are indicated as black dots and of the reference phantoms as grey triangles. The good agreement of the values can clearly be delineated. Please also see Table 1, Fig 6 and Fig 7.
Table 1.
Comparison of the degree of stenosis.
Fig 6.
Bland-Altman plot of the difference of the calculated and measured lumen loss of the reference phantom.
It can be depicted that the mean absolute difference is 0.9%. Accordingly, the plot shows that there is a slight tendency to underestimate the degree of lumen loss. This effect diminishes with higher degree of lumen loss because of the smaller diameters of the residual lumen.
Fig 7.
Bland-Altman plot of the difference of the calculated and the measured stenosis of the stenosis phantoms.
The mean absolute difference is -2.09%. This indicates that there is a slight overestimation of the degree of stenosis in comparison to the calculated values. This overestimation decreases with increasing degree of stenosis because of the smaller diameters of the residual lumen.
Fig 8.
MPI-Images of the dilution series.
Sagittal (a) and axial (b) slices extracted from reconstructed 3D image data of the 5 mm stenosis-phantom filled with the different Resovist dilutions from 1:100–1:3200 in sagittal (a) and axial (b) reformations with a regularization factor of λ = 1. Additionally, the image reconstruction of the Resovist dilution of 1:3200 with λ = 10 is shown. Note the distinct distortion of the image of the 1:3200 dilution reconstructed with λ = 1, which are smoothed by the higher regularization of λ = 10 due to the reduced noise (detailed data in S4 Table).
Table 2.
Serial dilution measurements.