TY - JOUR T1 - PiB-PET Imaging of Post-Operative Delirium JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1073 LP - 1073 VL - 62 IS - supplement 1 AU - Maribel Torres-Velazquez AU - Margaret Parker AU - Amber Bo AU - Marissa White AU - Robert Pearce AU - Richard Lennertz AU - Barbara Bendlin AU - Sterling Johnson AU - Vivek Prabhakaran AU - Alan McMillan AU - Robert Sanders Y1 - 2021/05/01 UR - http://jnm.snmjournals.org/content/62/supplement_1/1073.abstract N2 - 1073Introduction: Strong epidemiological associations between delirium and dementia have been previously observed1. Studies suggest that dementia both predisposes to, and is a consequence of, delirium. However, associations between Amyloid-β pathology (observed in Alzheimer’s disease and related dementias2) and delirium have been inconsistent. Therefore, in this pilot study we investigated the relationship between amyloid-β deposition, as defined by relative standardized uptake value (SUVR) measures, and delirium severity (DRS). Methods: Study participants included patients with a postoperative diagnosis of delirium (n = 6; male = 5; female = 1; age = 73.0 ± 3.3 years) and controls with no diagnosis of delirium (n = 5; male = 5; age = 78.0 ± 5.1). Delirium severity was assessed with the validated Delirium Rating Scale-98 (DRS)3. Subjects were imaged 90 days postoperatively on PET/MR (GE Signa PET/MR) with the beta amyloid tracer Pittsburgh Compound B (PiB). High density EEG was also obtained postoperatively4. Imaging parameters are shown in Fig. 1A. The SPM5-based image processing pipeline shown in Fig. 1B included the coregistration of PiB-PET to T1 images, segmentation of T1 images, spatial normalization, and background+skull extraction. SUVR was calculated using whole cerebellum (WCB) as the reference region. Voxel-wise analysis was performed using the Statistical NonParametric Mapping (SnPM) toolbox6, to determine clusters within the SUVR images that are monotonically related with DRS. ROIs were obtained from the AAL7 and Hammers8 atlases and eroded subcortical white matter (esWM) derived from SPM segmentation. A two-tailed Spearman’s rank correlation coefficient test was used to compare SUVR vs. DRS, SUVR vs. clinical variables (age, operative time, intraoperative blood loss), and SUVR vs. EEG (Oz delta power) for each ROI. Results: The ROI analysis revealed a statistically significant positive correlation between SUVR and DRS within the occipital lobe (Fig. 2). While not statistically significant, a moderate positive correlation was observed at the frontal lobe and a negative correlation with esWM. Age was not correlated to SUVR in any ROI. Operative time had a statistically significant positive correlation with SUVR within the frontal lobe, occipital lobe, and esWM. Intraoperative blood loss had a statistically significant positive correlation with SUVR within the occipital lobe. EEG had a statistically significant positive correlation with SUVR in the occipital lobe and a moderate negative correlation with esWM. Voxel-wise analysis uncovered statistically significant clusters of focal negative correlation between SUVR and DRS located within the white matter and cerebellum (Fig. 3). Conclusion: This pilot study demonstrates that delirium severity may correlate with increased occipital Amyloid-β deposition, which is consistent with correlations with the EEG data. However, interpretation of this is complex due to associations with various operative factors which also associate with delirium. Decreased PiB uptake in subcortical white matter also appears to be correlated with DRS which may reflect white matter integrity9 (that is also associated with delirium4). These results suggest that the links between dementia, delirium, and amyloid-β deposition should be further investigated. In particular, the use of white matter as a reference region with amyloid-β PET tracers in studies of delirium appears imprudent. Acknowledgements: This work is supported by NIH R01 AG063849-01 and a pilot grant from the Department of Radiology, University of Wisconsin-Madison. Figure 1. (A) Image reconstruction and acquisition parameters. (B) SPM-based image processing pipeline. Figure 2. Results of Spearman’s rank correlation analysis comparing SUVR to study variables. Bold entries are statistically significant. Figure 3. Voxel-wise analysis, negative correlation with DRS, p-value < 0.05, FWE corrected. ER -