Prediction of conversion from Mild Cognitive Impairment to Alzheimer's disease

Abstract

Predicting the conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD) and of the time to conversion, remain some of the most important clinical challenges despite having been investigated for many years. To this day the available evidence has not identified any reliable methods that can be applied in clinical settings mostly because of the complexity of the most effective methods. Taking feasibility into account, this thesis aimed to use simple MRI markers such as brain volumes to predict the risk and the time of conversion from MCI to AD. This thesis is built upon five step-by-step studies, which demonstrate that hippocampal volume is a practical, reliable measure for MCI prognosis. The first three studies aimed to develop a novel brain MRI volumetric measure to identify individuals with MCI who progress to AD within five years. The last two studies aimed to explore the contribution of MRI measures in predicting time to conversion and to investigate their interaction with cognitive performance. Data used in this project were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. It was hypothesized that the volumetric ratio of the brain region with the greatest atrophy rate in MCI to that of a region with no or a substantially lower atrophy rate in MCI could be a reliable and sensitive index in the prediction of AD conversion. The first study, a systematic review, revealed that the hippocampus and entorhinal cortex were the brain regions with the greatest atrophy rates in MCI, with atrophy rates about two-fold greater in MCI than cognitively normal (CN) people. The second study revealed that the cerebellum does not shrink faster in MCI than in CN individuals. Based on these findings, the hippocampal volume to cerebellar volume ratio was investigated as a predictor of conversion from MCI to AD within five years. The results revealed that the measure was predictive of conversion, and when combined with a Mini Mental Examination Score (MMSE), could effectively identify individuals at risk of AD from those individuals with MCI who remained stable for at least five years or reverted to CN. Further comprehensive investigation in the last two studies revealed that brain volumes - the whole brain, ventricles, hippocampal and entorhinal cortex volumes - were predictive of time to conversion from MCI to AD. Additionally, although individual cognitive/functional performance was predictive of time to conversion, its predictive values was dependent on hippocampal volume. The same conclusions were drawn from analyses investigating atrophy rates of these regions. That is, the rates of atrophy in whole brain, ventricles, hippocampus, and entorhinal cortex were predictive of time to AD conversion but dependent on their baseline volumes. Moreover, individuals with MCI, who had hippocampal or entorhinal cortex volumes smaller than 5500 mm3 and 2800 mm3 (respectively), progressed to AD more quickly regardless of the ensuing atrophy rate. Taking all these findings into consideration, this thesis suggests that hippocampal volume is a reliable biological marker for the identification of individuals with MCI at demonstrable risk of conversion to AD. Additionally, it is a reliable biomarker of time to conversion from MCI to AD. Indeed, at volumes less than a defined threshold it is highly prognostic of early conversion. Importantly, the prediction accuracy of a simple volumetric measure of the hippocampus is comparable to that of highly complex and sophisticated methods, such as machine learning, but with the advantage of being practical and easier to use in clinical or research settings. In clinical practice, early identification of those at risk can assist with early intervention and lifestyle modification, which subsequently can decrease the burden of the disease on the patients, their caregivers, and the health systems.