Heterogeneity of neuroanatomical patterns in prodromal Alzheimer's disease: links to cognition, progression and biomarkers.

See Coulthard and Knight (doi:10.1093/aww335) for a scientific commentary on this article.Individuals with mild cognitive impairment and Alzheimer's disease clinical diagnoses can display significant phenotypic heterogeneity. This variability likely reflects underlying genetic, environmental and neuropathological differences. Characterizing this heterogeneity is important for precision diagnostics, personalized predictions, and recruitment of relatively homogeneous sets of patients into clinical trials. In this study, we apply state-of-the-art semi-supervised machine learning methods to the Alzheimer's disease Neuroimaging cohort (ADNI) to elucidate the heterogeneity of neuroanatomical differences between subjects with mild cognitive impairment (n = 530) and Alzheimer's disease (n = 314) and cognitively normal individuals (n = 399), thereby adding to an increasing literature aiming to establish neuroanatomical and neuropathological (e.g. amyloid and tau deposition) dimensions in Alzheimer's disease and its prodromal stages. These dimensional approaches aim to provide surrogate measures of heterogeneous underlying pathologic processes leading to cognitive impairment. We relate these neuroimaging patterns to cerebrospinal fluid biomarkers, white matter hyperintensities, cognitive and clinical measures, and longitudinal trajectories. We identified four such atrophy patterns: (i) individuals with largely normal neuroanatomical profiles, who also turned out to have the least abnormal cognitive and cerebrospinal fluid biomarker profiles and the slowest clinical progression during follow-up; (ii) individuals with classical Alzheimer's disease neuroanatomical, cognitive, cerebrospinal fluid biomarkers and clinical profile, who presented the fastest clinical progression; (iii) individuals with a diffuse pattern of atrophy with relatively less pronounced involvement of the medial temporal lobe, abnormal cerebrospinal fluid amyloid-ß1-42 values, and proportionally greater executive impairment; and (iv) individuals with notably focal involvement of the medial temporal lobe and a slow steady progression, likely representing in early Alzheimer's disease stages. These four atrophy patterns effectively define a 4-dimensional categorization of neuroanatomical alterations in mild cognitive impairment and Alzheimer's disease that can complement existing dimensional approaches for staging Alzheimer's disease using a variety of biomarkers, which offer the potential for enabling precision diagnostics and prognostics, as well as targeted patient recruitment of relatively homogeneous subgroups of subjects for clinical trials.

Neuroanatomical heterogeneity of schizophrenia revealed by semi-supervised machine learning methods.

Schizophrenia is associated with heterogeneous clinical symptoms and neuroanatomical alterations. In this work, we aim to disentangle the patterns of neuroanatomical alterations underlying a heterogeneous population of patients using a semi-supervised clustering method. We apply this strategy to a cohort of patients with schizophrenia of varying extends of disease duration, and we describe the neuroanatomical, demographic and clinical characteristics of the subtypes discovered. METHODS: We analyze the neuroanatomical heterogeneity of 157 patients diagnosed with Schizophrenia, relative to a control population of 169 subjects, using a machine learning method called CHIMERA. CHIMERA clusters the differences between patients and a demographically-matched population of healthy subjects, rather than clustering patients themselves, thereby specifically assessing disease-related neuroanatomical alterations. Voxel-Based Morphometry was conducted to visualize the neuroanatomical patterns associated with each group. The clinical presentation and the demographics of the groups were then investigated. RESULTS: Three subgroups were identified. The first two differed substantially, in that one involved predominantly temporal-thalamic-peri-Sylvian regions, whereas the other involved predominantly frontal regions and the thalamus. Both subtypes included primarily male patients. The third pattern was a mix of these two and presented milder neuroanatomic alterations and comprised a comparable number of men and women. VBM and statistical analyses suggest that these groups could correspond to different neuroanatomical dimensions of schizophrenia. CONCLUSION: Our analysis suggests that schizophrenia presents distinct neuroanatomical variants. This variability points to the need for a dimensional neuroanatomical approach using data-driven, mathematically principled multivariate pattern analysis methods, and should be taken into account in clinical studies.

CHIMERA: Clustering of Heterogeneous Disease Effects via Distribution Matching of Imaging Patterns.

Many brain disorders and diseases exhibit heterogeneous symptoms and imaging characteristics. This heterogeneity is typically not captured by commonly adopted neuroimaging analyses that seek only a main imaging pattern when two groups need to be differentiated (e.g., patients and controls, or clinical progressors and non-progressors). We propose a novel probabilistic clustering approach, CHIMERA, modeling the pathological process by a combination of multiple regularized transformations from normal/control population to the patient population, thereby seeking to identify multiple imaging patterns that relate to disease effects and to better characterize disease heterogeneity. In our framework, normal and patient populations are considered as point distributions that are matched by a variant of the coherent point drift algorithm. We explain how the posterior probabilities produced during the MAP optimization of CHIMERA can be used for clustering the patients into groups and identifying disease subtypes. CHIMERA was first validated on a synthetic dataset and then on a clinical dataset mixing 317 control subjects and patients suffering from Alzheimer's Disease (AD) and Parkison's Disease (PD). CHIMERA produced better clustering results compared to two standard clustering approaches. We further analyzed 390 T1 MRI scans from Alzheimer's patients. We discovered two main and reproducible AD subtypes displaying significant differences in cognitive performance.

Regularized tensor factorization for multi-modality medical image classification.

This paper presents a general discriminative dimensionality reduction framework for multi-modal image-based classification in medical imaging datasets. The major goal is to use all modalities simultaneously to transform very high dimensional image to a lower dimensional representation in a discriminative way. In addition to being discriminative, the proposed approach has the advantage of being clinically interpretable. We propose a framework based on regularized tensor decomposition. We show that different variants of tensor factorization imply various hypothesis about data. Inspired by the idea of multi-view dimensionality reduction in machine learning community, two different kinds of tensor decomposition and their implications are presented. We have validated our method on a multi-modal longitudinal brain imaging study. We compared this method with a publically available classification software based on SVM that has shown state-of-the-art classification rate in number of publications.