Correcting Spatial Positioning Errors in Pre-defined Hardware Attenuation Correction Maps for PET/MR

A spatially accurate µ-map map of all materials traversed by acquired PET lines of response (LORs) is required for attenuation correction. Materials include the patient, patient table, and for PET/MR systems, local RF coils that cannot be neglected due to their design. Patient attenuation coefficients are usually derived from MR images, however the remaining hardware is typically corrected by referencing pre-defined CT derived µ-maps. Spatial positioning errors in the hardware µ-map, for example vertical table displacement resulting from patient weight, can produce attenuation correction errors. A simple approach that corrects spatial positioning error in the pre-defined hardware µ-maps was developed. Using the principle of PET data consistency, the rigid-body spatial transform that minimizes azimuthal variations in attenuation corrected PET emission data is determined. To improve efficiency, only LORs which are localized to approximately 100 coordinates are considered. Phantom experiments were performed with a head-neck coil, a spine coil, and a flexible body coil. Only the spine coil showed poor alignment accuracy in the x and y directions, with remaining degrees of freedom and all other coils resulting in good positioning accuracy. For the head-neck coil, simulated y-displacement errors were all accurately corrected with 0 mm residual error. The alignment method was shown to be sensitive to both phantom µ-map noise and PET count rate reduction which, when combined, increased the alignment error up to 14.9 mm.