Marcella Cline BS, Presenter: Nothing to Disclose

Satoshi Minoshima MD, PhD, Abstract Co-Author: License agreement, General Electric Company Research Grant, Koninklijke Philips Electronics NV Research Grant, Hitachi, Ltd Research Consultant, Hamamatsu Photonics KK Grant, Nihon Medi-Physics Co, Ltd Research Grant, Astellas Group Research Grant, Seattle Genetics, Inc

Donna Jean Cross PhD, Abstract Co-Author: Research Grant, Hitachi, Ltd Research Grant, Astellas Group

Alexander Drzezga MD, Abstract Co-Author: Consultant, Siemens AG Consultant, Eli Lilly and Company Consultant, General Electric Company Consultant, Piramal Enterprises Limited Speaker, Siemens AG Speaker, Eli Lilly and Company Speaker, General Electric Company Speaker, Piramal Enterprises Limited

Daniel S. Hippe MS, Abstract Co-Author: Research Grant, Koninklijke Philips NV Research Grant, General Electric Company

Barbara L. Lewellen BS, Abstract Co-Author: Nothing to Disclose

Greg Garwin, Abstract Co-Author: Nothing to Disclose

The functional integrity of neural activity via circuitries/pathways is thought to be reflected on regional intercorrelation of neuronal activity ('functional connectivity'). This study investigates the feasibility of such parametric mapping using individual FDG-PET imaging and compared to standard static images.

Ten non-human primates underwent dynamic brain PET imaging under sevoflurane anesthesia. Following a slow-bolus injection of 3 mCi [F-18]FDG, 120 30-second dynamic frames were obtained over 60 min. Following frame-to-frame image coregistration, stereotactic transformation, and global normalization, voxel-wise principal component analysis (PCA) with matrix transposition was applied to the individual data sets, followed by Varimax rotation of initial components. Individual quantitative Metabolic Coherence (MC) maps were created by averaging absolute component loadings and compared to conventional static FDG maps.

In all subjects, the first 2 components represented large variances (76% +/-11 SD to total variance) resulting from general blood flow and tissue FDG uptake that were eliminated by exclusion of the initial vascular phase in the dynamic data. Individual MC maps elucidated cerebral structures involved in the default mode network with high composite correlation coefficients: posterior cingulate cortex (0.070+/-0.006); frontal (0.070+/-0.005), parietal (0.069+/-0.006), and temporal (0.069+/-0.005) association cortices. MC values were modest in the striatum (0.059+/-0.007) and low in the visual cortex (0.039+/-0.005, presumably due to anesthesia) and cerebellum (0.035+/-0.007). In contrast, conventional static FDG maps from the same subjects showed high metabolic values (normalized to global activity 100) in the striatum (148+/-9.4); posterior cingulate cortex (136+/-6.5); parietal (134+/-6.9) and frontal (124+/-7.7) association cortices.

While static FDG maps represent regional neuronal activity, MC maps potentially provide unique supplementary information concerning regional functional integration via intercorrelation across regions within the brain. Further validation and optimization are underway.

New parametric analysis of dynamic FDG-PET depicts regional neuronal activity and functional integrity that can supplement conventional static image interpretation and shed light on disease processes.

Cline, M, Minoshima, S, Cross, D, Drzezga, A, Hippe, D, Lewellen, B, Garwin, G, Metabolic Coherence Mapping of the Brain to Elucidate Regional Neuronal Activity and Functional Integration: Multivariate Correlational Analysis Using Dynamic FDG PET.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14006763.html