Automatic thresholding for frame-repositioning using external tracking in PET brain imaging

Motion correction (MC) in positron emission tomography (PET) brain imaging become of higher importance with increasing scanner resolution. Several motion correction methods have been suggested and so far the Polaris Vicra tracking system has been the preferred one for motion registration. We present an automated algorithm for dividing PET acquisitions into subframes based on the registered head motion to correct for intra-frame motion with the frame repositioning MC method. The method is tested on real patient data (five
¹¹
C-SB studies and five
¹¹
C-PIB studies) and compared with an image based registration method (AIR). Quantitative evaluation was done using a correlation measure. The study shows that MC improves the correlation of the PET images and that AIR performed slightly better than the Polaris Vicra. We found significant intra-frame motion of 1-5 mm in 9 frames but the correlation was not significantly improved using intra-frame MC.