Fig 1.
Photographs of the construction process for the anthropomorphic phantom.
The white matter shell was 3D printed (A), and then a silicone holder was created to facilitate the agarose gel assembly (B). The top of the white matter shell was removed (C) for filling. Immediately prior to filling with the agarose gel white matter mimic, the nano-iron oxide disc inclusions were created from agarose and iron-oxide nanoparticles. Food coloring was used to make positioning within the white matter easier (D). The nano-iron oxide discs were made in a silicone mold (E). The nano-iron oxide discs were removed from the disc mold (F). The white matter shell was filled with the white matter agarose gel mimic, and the nano-iron oxide discs (marked with green arrows) were carefully positioned (G). Then the white matter gel was allowed to solidify. Next, the 3D printed ABS shell was dissolved using acetone (H). The white matter was positioned within the 3D printed skull (I) and filled with the gray matter agarose gel mimic. Finally, the top of the skull was fixed to the bottom part of the skull, and the remaining space was filled with gray matter agarose gel mimic (J). The complete anthropomorphic phantom, within the vacuum sealed bag, is shown in (K).
Table 1.
Chemical compositions for the tissue mimic materials.
Table 2.
3 T imaging sequences (TE = echo time; TR = repetition time; BW = bandwidth).
Table 3.
64 mT imaging sequences.
Fig 2.
Representative axial T1 and T2 maps acquired at 3 T (A, B) and 64 mT at construction (C, D); 10 weeks (E, F); 22 weeks (G, H); and 31 weeks (I, J) post-construction.
Table 4.
Measured relaxation times for white matter, gray matter and CSF mimics at 3 T, 19.8 °C +/- 0.5 °C.
The experimental measurements are reported as the mean value over six image slices with the standard deviation in parentheses.
Table 5.
Measured relaxation times for white matter, gray matter and CSF mimics at 64 mT, 21.5 °C +/- 0.5 °C at construction and again at 10, 22, and 31 weeks after construction.
The experimental measurements are reported as the mean value over six image slices with the standard deviation in parentheses.
Fig 3.
Detection of microbleed mimics.
From left to right, a susceptibility weighted image, a single gradient echo acquired (TE = 16 ms), and the final echo from a multi-echo gradient echo (TE = 21 ms). The top row is magnitude images. The bottom row is unwrapped phase images. Phase images were automatically masked in the ROMEO pipeline. In the multi-echo phase image, the masking did not remove all of the 3D printed skull (white arrows). While there are many artifacts in the images, none of these correspond to the location of the nano-iron oxide discs. For example, the red arrow indicates ABS remaining from incomplete removal of the white matter shell.
Fig 4.
Sagittal MR images of the completed phantom at 3 T (MPRAGE) and 64 mT (FLAIR).
In both images the blue arrow shows the water tube; the red arrow points to incomplete removal of the 3D printed white matter shell; the yellow arrow points to remaining silicone from the white matter holder; and the white arrow shows water contained in the 3D printed skull.
Fig 5.
Additional details of the construction flaws.
(A) and (B) show the anomalies in more detail on a 1 mm isotropic volume scan (spoiled gradient echo) acquired at 3 T. (A) the red arrow points to leftover 3D printed ABS (dark in the image) and the yellow arrow highlights silicone mold material (bright in the image), and the blue arrows indicate the water tube. (B) water from the assembly process remains in the 3D printed skull shell (white arrow).
Fig 6.
Photo of the anthropomorphic brain cross-section following deconstruction at 31 weeks.
A red arrow points to leftover 3D printed ABS, and the yellow arrow highlights silicone mold material. Additionally, the location of the samples for NMR assessment are shown with the black circles for white matter edge (A) and gray matter (B).