Brain glucose metabolism and cognitive function in patients with amnestic Mild Cognitive Impairment (MCI)

Aim: The need of early diagnosis of Alzheimer’s disease before the onset of dementia comes from the several ongoing phase 3 pharmacological studies that will hopefully lead to effective pathogenetic treatment of the disease in the next few years. In this frame, both brain PET and neuropsychological assessment are under extensive evaluation.
Methods: Twenty-one consecutive patients (9 males; mean age:76.5±5.9; mean MMSE score:27.8±1.5) with amnestic MCI (criteria of the European Alzheimer’s Disease Consortium, JNNP 2006, in press), underwent 18FDG-PET (GE Advance), 2D acquisition, OSEM reconstruction (16 subset, 2 iterations). Ten normal subjects (2 males; mean age:70.4±7.3; mean MMSE score:28.7±1.3) were recruited as control group (CTR). Neuropsychological evaluation included tests for ideomotor speed and shifting attention (trailmaking A and B), categorial verbal fluency, logic reasoning (Raven’s PM38), episodic verbal memory (Buschke), spatial memory (digit symbol), visuospatial abilities (constructional apraxia and clock drawing), cognitive flexibility (Stroop colour-word). PET data were analysed by a 3-D computerized brain atlas and 27 Brodmann areas (BA) bilaterally were submitted to ANCOVA. Step-wise discriminant analysis assessed the concordance between PET data and clinical diagnosis.
Results: Overall PET analysis showed significant group (p<0.05), VOI*group (p<0.001) and VOI*hemisphere*group (p<.025) effects. Post-hoc analyses showed at MCI versus CTR comparison significant 18FDG uptake decreases in prefrontal cortex (BA 46), superior parietal lobule and precuneus (BA 7), posterior cingulate (BA 31), both lateral and mesial occipital cortex (BA 19), lateral temporal cortex (BA 21) and temporal pole (BA 38). Tracer uptake was relatively increased in primary motor (BA 4) and somato-sensory (BA 1-3) cortex in patients. Significant group differences were found for Buschke (p<0.01) and Stroop (p<0.05) tests. Discriminant analysis using all significant VOI identified seven useful BAs (either in the left or in the right hemisphere) reaching 95% sensitivity and 100% specificity. Discriminant analysis with neuropsychological tests identified Buschke and clock drawing together as having 89% sensitivity and 70% specificity.
Conclusion: VOI-based analysis of 18FDG-PET has showed the involvement of several associative cortical areas in amnestic MCI, including the prefrontal, lateral temporal and occipital cortex, besides the classical regions affected in the disease represented by the posterior cingulate and the parietal precuneus. Both brain 18FDG-PET and neuropsychological assessment share similarly high sensitivity in amnestic MCI, but specificity is better with 18FDG-PET.