Introduction
To investigate the rate of acetylcholinesterase (AChE) in Alzheimer's disease, N-[11C]-Methylpiperidin-4-yl-Proprionate and Positron Emission Tomography are used. In a clinical environment, it is important to simplify the acquisition as much as possible by avoiding arterial blood sampling and/or to have fast methods available.
Methods
In 18 subjects (10 normals and 8 MCI patients; mean age: 67 yr in both groups), data were acquired dynamically for one hour (4 × 15 s, 4 × 60 s, 2 × 150 s, 10 × 300 s) on a HR+ after the injection of 300 MBq [11C]-PMP. Arterial blood samples were drawn to measure the plasma input function and to determine the metabolite fraction. Dynamic images were reconstructed using 3D-FBP with randoms, scatter and attenuation correction. An adjustment for small movements was applied between all the time frames and the data were coregistered to the subject's T1 weighted MRI. Segmentation of the MRI was done according to the optimized VBM approach and the dynamic data were transformed into the same MNI space. A partial volume correction was then applied to each frame, according to the method of Müller-Gartner 1 . Kinetic modelling was performed using three different approaches: (1) a “standard” irreversible two-tissue compartment model with non-linear fitting of the parameters, (2) a fast linearized model 2 and (3) a ‘non-invasive’ model which does not require arterial blood sampling 3 with the cerebellum, characterized by a high AChE uptake, as reference region. A correlation analysis between the three models was performed taking model (1) a priori as the most accurate model. Correlation coefficients were calculated in the cerebellum, the cortical (temporal, frontal, parietal and occipital) and subcortical regions.
Results
Comparison of the linearized model with model (1) showed a good agreement (median correlation value 0.95) in all cortical regions (range over all subjects: 0.86 to 0.99) except for one MCI patient in which the correlation in the parietal region was very low (r = 13). Lower correlations were found in the subcortical regions and the cerebellum (range over all subjects: 0.74 to 0.94 respectively 0.64 to 0.89), except for another MCI patient with a correlation of r = 0.13 in the subcortical regions. Comparing the non-invasive model with model (1) showed a similar behaviour: good correlation in the cortical regions (median correlation in the temporal lobe: 0.92, occipital lobe: 0.98, parietal lobe: 0.94 and frontal lobe: 0.78) but low correlations were seen in the subcortical regions and the cerebellum. Overall, a systemic underestimation of AChE was found for the non-invasive model in cortex. The correlation between the models did not differ between normals and MCI patients.
Conclusion
A simplified tissue reference model and the (fast) linearized model can be used to study AChE in cortical regions but not in subcortical or cerebellar regions.