Baseline and longitudinal patterns of brain atrophy in MCI patients, and their use in prediction of short-term conversion to AD: Results from ADNI

Abstract

High-dimensional pattern classification was applied to baseline and multiple follow-up MRI scans of the Alzheimer's Disease Neuroimaging Initiative (ADNI) participants with mild cognitive impairment (MCI), in order to investigate the potential of predicting short-term conversion to Alzheimer's Disease (AD) on an individual basis. MCI participants that converted to AD (average follow-up 15 months) displayed significantly lower volumes in a number of grey matter (GM) regions, as well as in the white matter (WM). They also displayed more pronounced periventricular small-vessel pathology, as well as an increased rate of increase of such pathology. Individual person analysis was performed using a pattern classifier previously constructed from AD patients and cognitively normal (CN) individuals to yield an abnormality score that is positive for AD-like brains and negative otherwise. The abnormality scores measured from MCI non-converters (MCI-NC) followed a bimodal distribution, reflecting the heterogeneity of this group, whereas they were positive in almost all MCI converters (MCI-C), indicating extensive patterns of AD-like brain atrophy in almost all MCI-C. Both MCI subgroups had similar MMSE scores at baseline. A more specialized classifier constructed to differentiate converters from non-converters based on their baseline scans provided good classification accuracy reaching 81.5%, evaluated via cross-validation. These pattern classification schemes, which distill spatial patterns of atrophy to a single abnormality score, offer promise as biomarkers of AD and as predictors of subsequent clinical progression, on an individual patient basis.

Introduction

The most common cause of dementia is Alzheimer's disease, with incidence doubling approximately every 5 years after the age of 65. As life expectancy increases, AD is becoming an important health problem in the elderly, and a significant societal and financial burden. As many treatments are being developed and evaluated, it is becoming important to develop diagnostic and prognostic biomarkers that can predict which individuals are relatively more likely to progress clinically. This is especially important in individuals with MCI, who present a conversion rate of approximately 15% per year.

MCI has attracted a lot of attention in the recent years, in part because it offers opportunities for relatively early diagnosis of AD, and in part a number of pharmachological interventions typically target MCI patients. A number of studies, including both region-of-interest (ROI) and voxel-based analyses, have reported relatively reduced brain volumes in the hippocampus, parahippocampal gyrus, cingulate, and other brain regions in both MCI and AD patients (Jack et al., 1999, Jack et al., 2008, Karas et al., 2004, Fox and Schott, 2004, Chetelat et al., 2002, Convit et al., 2000, Dickerson et al., 2001, Kaye et al., 1997, Killiany et al., 2000, Stoub et al., 2005, Visser et al., 2002, De Leon et al., 2006). The spatial pattern of brain atrophy in MCI is complex and highly variable, and it evolves in time as the disease progresses. Therefore, capturing such spatio-temporal patterns of structural brain change requires sophisticated image analysis methods, including high-dimensional image warping often used to quantify the regional distribution of brain tissue. Even more challenging is the development of biomarkers that classify individuals, as opposed to characterizing group differences, as inter-individual variations and statistical overlap among groups renders it difficult to characterize individuals with sufficient sensitivity and specificity.

In order to be able to achieve individual classification, we leverage upon work on high-dimensional pattern classification that has shown great promise in the past 5 years as a means to measure subtle and spatially complex imaging patterns that have diagnostic value (Davatzikos et al., 2005, Davatzikos et al., 2008b, Liu et al., 2004, Lao et al., 2003, Lao et al., 2004, Li et al., 2007, Adeli et al., 2005, Tandon et al., 2006, Davatzikos et al., 2005, Kloppel et al., 2008a, Vemuri et al., 2008). This approach aims to provide computational tools that classify individuals, based on their imaging measurements, rather than determining statistical group differences. The current study builds upon previous work in (Lao et al., 2004, Davatzikos et al., 2008a, Fan et al., 2008a, Duchesne et al., 2008), by investigating longitudinal images from a relatively large sample of MCI individuals from ADNI. The primary goal of the current study is to utilize high-dimensional pattern classification of baseline scans to determine predictors of short-term conversion from MCI to AD. The secondary goal of this study is to measure the spatial distribution of brain atrophy, as well as its longitudinal change, in MCI converters (MCI-C) and in MCI non-converters (MCI-NC) in the ADNI cohort, and to evaluate differences between these two groups. The hypothesis was that advanced pattern analysis and classification techniques applied to baseline and longitudinal images of the regional distribution of brain tissues would allow us to predict future conversion from MCI to AD.