رفع ابهام پیری طبیعی از بیماری آلزایمر در تصاویر رزونانس مغناطیسی از ساختار

The morphology observed in the brains of patients affected by Alzheimer's disease (AD) is a combination of different biological processes, such as normal aging and the pathological matter loss specific to AD. The ability to differentiate between these biological factors is fundamental to reliably evaluate pathological AD-related structural changes, especially in the earliest phase of the disease, at prodromal and preclinical stages. Here we propose a method based on non-linear image registration to estimate and analyze from observed brain morphologies the relative contributions from aging and pathology. In particular, we first define a longitudinal model of the brain's normal aging process from serial T1-weight magnetic resonance imaging scans of 65 healthy participants. The longitudinal model is then used as a reference for the cross-sectional analysis. Given a new brain image, we then estimate its anatomical age relative to the aging model; this is defined as a morphological age shift with respect to the average age of the healthy population at baseline. Finally, we define the specific morphological process as the remainder of the observed anatomy after the removal of the estimated normal aging process. Experimental results from 105 healthy participants, 110 subjects with mild cognitive impairment (MCI), 86 with MCI converted to AD, and 134 AD patients provide a novel description of the anatomical changes observed across the AD time span: normal aging, normal aging at risk, conversion to MCI, and the latest stages of AD. More advanced AD stages are associated with an increased morphological age shift in the brain and with strong disease-specific morphological changes affecting mainly ventricles, temporal poles, the entorhinal cortex, and hippocampi. Our model shows that AD is characterized by localized disease-specific brain changes as well as by an accelerated global aging process. This method may thus represent a more precise instrument to identify potential clinical outcomes in clinical trials for disease modifying drugs.

The objective of computational anatomy when applied to neurodegenerative diseases, such as Alzheimer's disease (AD), is to understand the pathological changes affecting brain morphology (Frisoni et al., 2010 and Scahill et al., 2002). However, the morphology of the brain affected by AD is not completely related to the disease, especially in asymptomatic and prodromal stages, because the brain structure is also the result of patient phenotype and clinical history. In a brain affected by AD, we can identify 2 major processes contributing to morphological changes: normal aging and AD pathology itself. • Age-related anatomical changes. It is known that aging is related to progressive impairment of neural mechanisms ( Burke and Barnes, 2006), to chemical alterations in the brain, and to changes in cognition and behaviour ( Hof and Mobbs, 1984). It has been observed that morphological changes in the aging brain are heterogeneous and primarily lead to gray matter loss in frontal, temporal, and parietal areas ( Long et al., 2012 and Sowell et al., 2003). • Disease-related anatomical changes. AD is a neurodegenerative disease characterized by the cooccurrence of different phenomena. It starts with the deposition of amyloid plaques and tau proteins in neurofibrillary tangles, which is followed by the development of function brain loss, and finally by widespread structural atrophy ( Jack et al., 2010). The typical pattern of brain tissue loss seen in AD mirrors tau deposition ( Thompson et al., 2003) and involves primarily hippocampi, the entorhinal cortex, the posterior cingulate, and secondarily the temporal, parietal, and frontal cortices ( Frisoni et al., 2010). Aging is the primary risk-factor in AD and leads to patterns of structural loss overlapping the pathological ones. However, the magnitude of brain atrophy caused by AD is generally striking compared with normal aging. As claimed in previous studies, AD is more likely to be a pathological state concurrent to aging, identified by specific biochemical and structural hallmarks ( Barnes, 2011 and Nelson et al., 2011). Being able to separately model healthy aging and AD would allow us to describe a given anatomy as being composed of distinct and concurrent factors. Such a decomposition would be extremely interesting not only to improve the understanding of the disease but also for clinical purposes, such as for early diagnosis and for the development of drugs targeting the atrophy specific to the pathology. It is important to notice that, although brought on completely different biological mechanisms, aging and AD often map to common areas, and the correct identification of the respective contributions can be difficult, especially in morphometric studies. Moreover, it is plausible that these phenomena are not completely independent and may overlap to create a positive “feedback” process. Thus, the onset of pathological changes may lead to accelerated global aging in the long term (Fjell et al., 2012), and vice versa. A reliable estimate of the aging component is also important for modeling the evolution of the disease and for subsequent statistical analysis. When comparing the longitudinal observations from different clinical groups, at different aging stages, it is crucial to correctly position the observations on the time axis. This is not straightforward because the disease appears at different ages and chronologically older brains may have greater structural integrity than younger ones affected by the pathology. Therefore, it might be of practical interest to compute an index of age shift “relative” to a reference anatomical model. The idea of modeling the time course of AD with respect to clinical and demographic factors was proposed in previous statistical studies (Ito et al., 2012, Samtani et al., 2012 and Yang et al., 2011). However, these works were limited to scalar observations such as clinical scores and demographics and thus do not provide an explicit model which relates structural changes in the entire brain to the disease and aging. Moreover, the disease progression was identified by clinical measures and was not therefore explicitly associated with a temporal time course. Although imaging-based surrogate measures of aging have been provided by different methodological studies (Davatzikos et al., 2009, Franke et al., 2010 and Konukoglu et al., 2013), the idea of separately investigating aging and residual morphological changes has not been proposed before. The objective of this work is to introduce a framework to identify and disentangle the brain anatomical changes related to normal aging from those related to other biological processes, such as AD. In particular, our framework is based on the hypothesis that relates the development of AD to the abnormal accumulation of beta-amyloid (Aβ) peptide in the brain (Jack et al., 2010). We thus define “normal aging” as the morphological brain evolution which is not caused by Aβ. This evolution is modeled by nonlinear registration and is used as a reference to characterize observed anatomy as a contribution from normal morphological aging (normal aging process) plus a specific morphological process that encodes the subject’s specific variability such as pathological traits. We test our framework on healthy participants positive to the cerebrospinal fluid (CSF) Aβ42 marker, in participants affected by mild cognitive impairment (MCI) and in AD patients. The method is based on diffeomorphic nonlinear registration and is detailed in Section 2. In Section 3, we show that such a framework provides a meaningful and accurate description of anatomical brain changes across the stages of AD, characterized by increased morphological aging plus specific and local atrophy features.

We proposed a method to describe brain anatomy as contributions of 2 independent processes: morphological aging and a specific component. These components identify different clinical stages, and are compatible with the hypothesis that points to the abnormal levels of CSF Aβ42 as a presymptomatic marker of AD in the early stages. We showed that more advanced AD stages (from Aβ+ to MCI converters, and finally to AD) are associated with both “virtually older” brains, and with increased specific morphological changes not related to the normal aging process. Thus, according to our model, AD might have an influence on the overall aging of the brain acting as an acceleration factor. 4.1. Relationship with classical morphometric studies Compared with standard analysis approaches of group-wise structural changes such as classical voxel/tensor-based morphometry (Good et al., 2002 and Hua et al., 2008), or discriminative analysis (Chincarini et al., 2011 and Cuingnet et al., 2011), our method has the advantage of providing complementary information representing morphological aging and specific deformation parameters that carry relevant biological meaning. For this reason, the proposed method provides a novel way of interpreting morphometric results. For instance, it may be of great interest to investigate the relationship between the morphological age shift and specific changes in characterizing brain structural and clinical reserve in preclinical/prodromal stages. 4.2. Combining cross-sectional and longitudinal analysis Our method proposes cross-sectional comparison of brain images by means of a previously defined longitudinal model of morphological changes. The statistical modeling of the intersubject variability in computational anatomy is challenging, and the interpretation of group-wise comparison results is always bound to the statistical assumptions about the measured features (for instance, concerning the statistical distribution of the Jacobian/divergence values). In this study we showed that by removing the aging component we increase the ability in detecting specific group-wise differences. For this reason, the proposed method represents a novel and promising approach to the appropriate modeling and interpretation of group-wise anatomical variation. We note in fact that the voxel-by-voxel analysis of the divergence maps in Section 3.6 is compatible with standard voxel/tensor-based morphometry approaches and leads to very similar results to those provided by classical group-wise comparison (Good et al., 2002 and Hua et al., 2008). It is, however, providing a slightly greater effect size when comparing MCIs. This indicates that the removal of the normal aging process could enhance the estimation of pathological volume changes. When comparing AD with healthy controls, the analysis of the specific morphological process provides a slightly larger effect size in the temporal horns but generally lower in the white matter. This latter result can be explained by noting that the removal of the aging process aims to decorrelate the morphological changes explained by the aging model. In this way we reduce, for instance, the effect of ventricles expansion, detectable in tensor based morphometry as an apparent contraction in the white matter. 4.3. Interpretation of aging and specific processes Concerning the modeling of the specific deformation parameter, we note that by definition this component is highly heterogeneous across the population because it includes normal anatomical variability as well as pathological features. In this study we have shown that despite this high variability, the specific deformation is able to accurately describe anatomical features specific to AD. The discriminative analysis performed in Section 3.6. showed that the specific deformation parameter includes specific pathological traits which characterize the whole disease time span. Future studies based on more sophisticated machine learning techniques, as proposed for instance in Chincarini et al., (2011) and Cuingnet et al., (2011), may lead to improved classification results. In our model, the morphological age shift is based on the whole brain average of the projection on the normal aging model. Here, we make a precise assumption about the aging process, which is in fact defined globally. Therefore, accelerated aging is constrained with respect to the whole brain normal aging model, and any local departure from it (for instance in some specific regions), is interpreted as a specific morphological change, independent from aging. Different MRI-based indices of brain aging have been proposed in the past ( Davatzikos et al., 2009 and Franke et al., 2010). Our model integrates these approaches into a more general description of the AD process. We note that the morphological age shift for AD patients is lower than the aging score estimated in Franke et al. (2010) (4 vs. 10 years). In fact we have shown that AD is characterized by a more pronounced specific and concurrent pattern of atrophy. In this case, the present result motivates and provides clinical outcomes for the design of disease-specific modifying drugs that do not have an impact on normal aging. We observed a weak but significant positive correlation between morphological age shift and years of education. The correlation is significant and stronger when considered in healthy participants only (Supplementary Table S1). This latter finding could indicate that morphologically older participants with high education are more likely to appear cognitively healthy. Therefore, the morphological age shift might represent a measure of cognitive and structural reserve in normal aging ( Coffey et al., 1999). The proposed average model of aging assumes that the baseline acquisition time is unique for the healthy cohort, as already proposed in previous longitudinal studies of brain volume changes in AD, based on mixed-effects modeling (Ridha et al., 2006). We showed in the experimental section that under these assumptions we obtained similar results to those provided by classical univariate LME models, by correctly describing the temporal variability of brain changes in the healthy cohort. However, it will be of interest to explicitly account for the inter-subject baseline differences in future studies. This amounts to modifying the assumption about the constant evolution in time of the healthy aging process. Finally, the proposed model could be extended in future work to account for different evolution trajectories, and to explicitly model different neurodegenerative pathologies, and AD subtypes. To conclude, our approach provides new insights that may help the understanding of AD dynamics and could thus represent a more precise instrument to identify outcomes in clinical trials for disease modifying drugs. 4.4. Limitations Some methodological limitations should be considered in this study. The proposed model of aging progression is based on nonlinear registration and is therefore estimated from image data only. This means that no biological/biomechanical information was accounted in the definition of the average trajectory. Nonlinear registration is driven by image intensities, thus it only models apparent observable anatomical changes. Moreover, nonlinear registration results are dependent on the registration model and may potentially provide different results depending on the choice of parameters, similarity measure and regularization scheme. However, we have already shown in previous work that LCC-Demons nonlinear registration provides reliable and meaningful results when applied to brain image registration problems, especially for the longitudinal analysis of atrophy (Lorenzi et al., 2013). Finally, only 3-year follow-up imaging data are available for the healthy cohort, and therefore the group-wise evolution was limited to a linear model in time for the SVF due to the lack of sufficient longitudinal observations. Disclosure statement The authors have actual or potential conflicts of interest. Acknowledgements This work was partially funded by the European Research Council (ERC advanced Grant MedYMA 2011-291080), ANR blanc Karametria and the EU project Care4Me. Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI; National Institutes of Health Grant U01 AG024904). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: Alzheimer's Association; Alzheimer Drug Discovery Foundation; BioClinica, Inc.; Biogen Idec Inc.; Bristol-Myers Squibb Company; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; GE Healthcare; Innogenetics, N.V.; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Medpace, Inc.; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Synarc Inc.; and Takeda Pharmaceutical Company. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer's Disease Cooperative Study at the University of California, San Diego. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. The authors thank Sheila and John Stark, and Kristin McLeod for their valuable help in proofreading the manuscript.