RT Journal Article SR Electronic T1 Motion analysis in PET brain studies and the effect of motion on automatic SUVr analysis of β-amyloid investigations JF Journal of Nuclear Medicine JO J Nucl Med FD Society of Nuclear Medicine SP 1926 OP 1926 VO 57 IS supplement 2 A1 Daniel Fakhry-Darian A1 Richard Meades A1 Kuldip Nijran A1 Zarni Win YR 2016 UL http://jnm.snmjournals.org/content/57/supplement_2/1926.abstract AB 1926Objectives BRASS (Hermes Medical Solutions) for florbetapir [18F] provides automatic quantitation in 7 regions of the brain for β-amyloid PET investigations used in the diagnosis of Alzheimer’s disease. The SUV ratio (SUVr) is calculated in each region, relative to the cerebellum, and compared to a template of normal control patients. Several investigations are underway to validate this software for clinical use in our centre, the purpose of this study is to assess the impact of patient motion on the results.Methods 10 patient studies (5 positive and 5 negative) were selected at random from our existing database of clinical studies. All scans were acquired in list mode on a Siemens Biograph 64 PET/CT scanner, data was reconstructed as 20x60 second non-attenuation corrected (NAC) frames in order to check for motion before performing the final clinical reconstruction. The NAC frames were coregistered to the first frame using SPM (v12). Motion was assessed by calculation of the registration matrix between frames (relative to frame 1) and evaluating the straight line distance travelled by a reference point from frame 1 to 20 (Matlab R2015b). The minimal total path length travelled by the reference point over 10 consecutive frames was taken to be the most stable period of the scan. This stable 10 and the manufacturer recommended first 10 minutes were reconstructed as a static image with the following parameters: OSEM2D 4i 14s, 3mm Gaussian filter and 168x168 matrix. Automatic quantitation was performed in BRASS.Results The method used for motion analysis indicates that patient motion is more common at the start of the 20 minute acquisition period, this correlates with previous work done at our centre [1]. The straight line estimate of the path length travelled by a reference point is appropriate for small angle head rotations: for 0°<θ<20°, path length is underestimated by &#8804; -0.5%. The BRASS results from the manufacturer recommended first 10 minutes of the scan were compared to the chosen best 10 minutes. The SUVr in the chosen best 10 minutes is consistently higher than the first 10 minutes, this is significant in 3/10 patients (p<0.05). Each of these three patients shows significant motion within the first 10 minutes of the scan.Conclusions A scan time in excess of 10 minutes is often necessary in order to acquire a stable 10 minute image. The motion analysis method applied here in image space helps to provide insight when visual analysis is difficult. Patient motion has some impact on quantitation in BRASS suggesting florbetapir scans should be checked for motion before reconstruction and automatic quantitation analysis. The results so far indicate that scan times in excess of 10 minutes are necessary if automatic quantitation is to be used accurately. Further study and a larger sample are needed to verify the clinical outcome of these results and the impact on reporting.