Regional covariance of muscarinic acetylcholine receptors in Alzheimer’s disease using (R, R) [ 123 I]-QNB SPECT

Alzheimer’s disease (AD) is characterised by deficits in cholinergic neurotransmission and subsequent receptor changes. We investigated 123 I-iodo-quinuclidinyl-benzilate (QNB) SPECT images using spatial covariance analysis (SCA), to detect an M1/M4 receptor spatial covariance pattern (SCP) that distinguished AD from controls. Furthermore, a corresponding regional cerebral blood flow (rCBF) SCP was also derived. Thirty-nine subjects (15 AD and 24 healthy elderly controls) underwent 123 I-QNB and 99m Tc-exametazime SPECT. Voxel SCA was simultaneously applied to the set of smoothed/registered scans, generating a series of eigenimages representing common intercorrelated voxels across subjects. Linear regression identified individual M1/M4 and rCBF SCPs that discriminated AD from controls. The M1/M4 SCP showed concomitant decreased uptake in medial temporal, inferior frontal, basal forebrain and cingulate relative to concomitant increased uptake in frontal poles, occipital, pre-post central and precuneus/superior parietal regions ( F 1,37 = 85.7, p < 0.001). A largely different perfusion SCP was obtained showing concomitant decreased rCBF in medial and superior temporal, precuneus, inferior parietal and cingulate relative to concomitant increased rCBF in cerebellum, pre-post central, putamen, fusiform and brain stem/midbrain regions ( F 1,37 = 77.5, p < 0.001). The M1/M4 SCP expression correlated with the duration of cognitive symptoms ( r = 0.90, p < 0.001), whereas the rCBF SCP expression negatively correlated with MMSE, CAMCOG and CAMCOG memory ( r ≥ |0.63|, p ≤ 0.006). 123 I-QNB SPECT revealed an M1/M4 basocortical covariance pattern, distinct from rCBF, reflecting the duration of disease rather than current clinical symptoms. This approach could be more sensitive than univariate methods in characterising the cholinergic/rCBF changes that underpin the clinical phenotype of AD.