Maria LaCalle MS, Abstract Co-Author: Nothing to Disclose

Juan Adan Guzman-De-Villoria MD, PhD, Presenter: Nothing to Disclose

Santiago Reig, Abstract Co-Author: Nothing to Disclose

José María Mateos-Pérez, Abstract Co-Author: Nothing to Disclose

Javier Olazarán, Abstract Co-Author: Nothing to Disclose

Manuel Desco-Menendez MD, PhD, BEng, Abstract Co-Author: Nothing to Disclose

We compare raw cerebral blood volume (CBV) values obtained by MRI with normalized data using four different reference regions: mean cerebellum CBV, mean grey matter (GM) CBV, mean white matter (WM) CBV, and mean centrum semiovale CBV. These four data sets are used to analyze the differences between Alzheimer’s disease patients and controls.

Three groups of subjects with a different level of clinical dementia rating (CDR) were included in the study: AD patients (CDR=1; n=12 mean age=78.2), MCI patients (CDR=0.5; n=19; mean age=70.7) and controls (CDR=0; n=16; mean age=72.3). MRI data were acquired on a 1.5 T system. The imaging protocol included an anatomical scan 3D GE T1-WI (1x1x1.5mm) and parametric maps of CBV using TSE T2-WI (1.8 x 1.8 x 5mm) after the injection of a bolus of gallium chelate. Quantification of CBV maps, we applied a semiautomatic method based on the Talairach proportional grid system that uses the anatomical information of structural MR images to segment CBV maps. This method allowed us to obtain volume and CBV data for the whole brain and for the frontal, and temporal lobes, for each hemisphere. Tissue segmentation was made using SPM2. The main criterion for a valid reference region is to show stable values across subjects and no group differences.  

CBV values in the four regions of normalization considered in our study showed no differences between groups. The pattern of group differences, in CBV means, depends considerably on the choice of the reference region. In frontal and parietal lobes, patients showed hypo-perfusion when normalizing with cerebellum or WM. On the contrary, normalization with GM or centrum semiovale suggests hyper-perfusion. In the temporal lobe, normalized data showed hyper-perfusion in patients, whereas uncorrected data shows hypo-perfusion.  

Description of differences between groups in CBV depends on the choice of the reference region for normalization. The cerebellum and whole brain white matter seem the most valid reference regions for normalization. MCI and AD patients showed hypo-perfusion in frontal and parietal lobes when compared to controls, whereas the opposite pattern is observed in the temporal lobe.  

Perfusion MRI is a promising alternative to nuclear medicine in early detection of dementia. The large physiological variability between subjects makes normalization necessary for comparative analysis

LaCalle, M, Guzman-De-Villoria, J, Reig, S, Mateos-Pérez, J, Olazarán, J, Desco-Menendez, M, Controversy in Differences of Brain Perfusion Data between Alzheimers Disease Patients and Healthy Controls: Could Be Due to the Choice of a Reference Region for Normalization.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034464.html