The prediction of rapid conversion to Alzheimer's disease in mild cognitive impairment using regional cerebral blood flow SPECT

Abstract

Mild cognitive impairment (MCI) comprises a heterogeneous group with a variety of clinical outcomes and they are at risk for developing Alzheimer's disease (AD). The prediction of conversion from MCI to AD using the initial neuroimaging studies is an important research topic. We investigated the initial regional cerebral blood flow (rCBF) measurements using single photon emission computed tomography (SPECT) in individuals with 76 amnesic MCI (52 subjects converted to AD and 24 subjects did not convert to AD at 3-year follow-up) and 57 age- and gender-matched controls. We sought functional profiles associated with conversion to AD, then evaluated the predictive value of the initial rCBF SPECT. As compared with controls, AD converters demonstrated reduced blood flow in the bilateral parahippocampal gyri, precunei, posterior cingulate cortices, bilateral parietal association areas, and the right middle temporal gyrus. Non-converters also demonstrated significant reduction of rCBF in the posterior cingulated cortices and the right caudate nucleus when compared to controls. As compared with non-converters, converters showed reductions of rCBF in the bilateral temporo-parietal areas and the precunei. The logistic regression model revealed that reduced rCBF in the inferior parietal lobule, angular gyrus, and precunei has high predictive value and discriminative ability. Although a cross-validation study is needed to conclude the usefulness of rCBF SPECT for the prediction of AD conversion in individuals with MCI, our data suggest that the initial rCBF SPECT studies of individuals with MCI may be useful in predicting who will convert to AD in the near future.

Introduction

Mild cognitive impairment (MCI) is an operational diagnostic term developed to describe the preclinical stage of Alzheimer's disease (AD) (Petersen et al., 2001a). The risk for conversion to AD is higher in individuals with MCI than in the general aged population, as annual conversion rate of 6%–25% from MCI to AD (Petersen et al., 2001b). Furthermore, a recent study suggested that progression from MCI to AD is time-dependent. According to Palmer's study, people with MCI have a high risk of progressing to dementia over the next 3 years, and the risk starts to decrease after this point (Palmer et al., 2003). The early detection of MCI individuals who will later convert to AD is an important issue for both clinical and research interests.

The recent advance of computer-assisted statistical image analyses revealed that subjects with very mild AD typically show abnormal metabolic and regional cerebral blood flow (rCBF) patterns, even at the preclinical stage. Using glucose metabolism positron emission tomography (PET) with a voxel-by-voxel statistical analysis, Minoshima et al. reported that the earliest changes observed in very mild AD were in the posterior cingulate cortex (PCC) (Minoshima et al., 1997). This unexpected finding has been replicated by other groups using both glucose metabolism measurements with PET and even less sophisticated measurement techniques such as regional cerebral blood flow (rCBF) measurements with single photon emission computed tomography (rCBF SPECT). Our previous rCBF SPECT study demonstrated significantly decreased rCBF in the posterior cingulate gyri and precunei bilaterally in MCI subjects as compared with controls at least 2 years before they met a clinical diagnosis of AD (Kogure et al., 2000). We also reported a diagnostic value of reduced rCBF in the posterior cingulate cortex (PCC) to assist in discriminating between normal subjects and MCI subjects who later developed AD (Imabayashi et al., 2004). Furthermore, a PET study demonstrated hypometabolism of the PCC in young subjects with a high genetic risk of developing AD (Reiman et al., 2004). These results suggest that functional neuroimaging techniques such as PET and SPECT may be promising techniques for the preclinical diagnosis of AD.

However, MCI is a heterogeneous diagnostic category comprised of individuals with a variety of clinical outcomes (Petersen et al., 2001). As such, only a longitudinal study comparing MCI subjects who convert to AD at follow-up (converters) with MCI subjects who do not convert at follow-up (non-converters) is appropriate to determine the predictive value of initial neuroimaging for progression of MCI to AD. Only a few longitudinal studies have been published so far (Celsis et al., 1997, Arnaiz et al., 2001, Huang et al., 2002, Chetelat et al., 2003, Drzezga et al., 2003, Mosconi et al., 2004). These studies have suggested that reduced glucose metabolism in the right temporo-parietal cortex or reduced rCBF and glucose metabolism in the PCC might be good predictors of progression to AD.

On the other hand, morphological magnetic resonance imaging (MRI) studies have demonstrated that higher atrophy rates in the medial temporal regions such as the entorhinal cortex and hippocampus may be good predictors of conversion to AD (Jack et al., 1999, Mungas et al., 2002, Nestor et al., 2004). However, such serial MR studies require at least a 1-year follow-up period to predict AD conversion. As with functional imaging studies, the predictive value of morphological MR studies has not been yet clarified.

The present retrospective cohort study assessed initial rCBF SPECT images in a group of amnesic MCI subjects consisting of AD converters and non-converters who were followed clinically for 3 years. The aim of the present study was to find highly specific and sensitive rCBF changes capable of discriminating between MCI subjects who eventually develop AD from those who do not convert to AD, as early as possible. We also demonstrated the predictive value of the initial rCBF SPECT studies in MCI subjects.