Characterization of a New MR Compatible Small Animal PET Scanner Using Monte-Carlo Simulations

Abstract

We are currently designing a small animal PET insert for use in an MRI with a bore size that constrains the insert inner diameter to be no larger than 66 mm while leaving 25 mm for ring thickness. The insert will be made from 10 mm thick DOI-capable Dual Layer Offset LYSO blocks coupled to MR-compatible SiPMs. The block is made from a 9 × 9 array of 1.345 × 4 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> crystals in the front layer and a 10 × 10 array of 1.345 × 1.345 × 6 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> crystals in the back layer (crystal pitch = 1.422 mm). A ring of blocks is made by repeating a block around a ring with inner diameter of 64.776 mm 16 times. Here, GATE simulations have been made to estimate mousenoise-equivalent count rate (NECR), mouse-scatter fraction (SF), peak sensitivity (S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> ) resolution, and resolution uniformity to evaluate the design of our PET insert. These simulations make use of hardware performance estimates measured from a prototype block. For the one, three, and six ring tomographs, NECR curves, SF, and fígures were produced for the best and worst expected hardware performance. Simulations of a point source in a one-ring tomograph were made to estimate resolution across the field of view (FOV). For a six-ring tomograph with a 250-750 keV energy window and best expected hardware performance, the peak NECR, peak NECR activity, and S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> were 1273 kcps, 96 MBq and 10.0%. With three rings, these figures were 389 kcps at 95 MBq, and 5.9%. And with one-ring, they were 43 kcps, 85 MBq, and 2.0%. SF was ~ 16% in these three cases. Spatial resolution in the radial direction was found to change from 1.0 to 1.9 mm FWHM moving from the center of the FOV to a 15 mm offset. These results indicate that our scanner design is highly suited for high-resolution preclinical mouse imaging.