Random forest-based similarity measures for multi-modal classification of Alzheimer's disease

• Published in: NeuroImage (Orlando, Fla.) Vol. 65; pp. 167 - 175
• Main Authors: Gray, Katherine R., Aljabar, Paul, Heckemann, Rolf A., Hammers, Alexander, Rueckert, Daniel
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Inc 15-01-2013; Elsevier; Elsevier Limited
• Subjects: Accuracy; Adult and adolescent clinical studies; Aged; Alzheimer Disease - classification; Alzheimer Disease - diagnosis; Alzheimer Disease - physiopathology; Alzheimer's disease; Artificial Intelligence; Biological and medical sciences; Biomarkers; Biomarkers - analysis; Cerebrospinal fluid biomarkers; Classification; Degenerative and inherited degenerative diseases of the nervous system. Leukodystrophies. Prion diseases; Female; Genetics; Glucose; Humans; Magnetic Resonance Imaging; Male; Medical imaging; Medical sciences; Metabolism; Methods; Mild cognitive impairment; Models, Theoretical; Multi-modality classification; Neurology; NMR; Nuclear magnetic resonance; Organic mental disorders. Neuropsychology; Positron emission tomography; Psychology. Psychoanalysis. Psychiatry; Psychopathology. Psychiatry; Studies; Magnetic resonance imaging; Mild cognitive impairment; Multi-modality classification; Genetics; Cerebrospinal fluid biomarkers; Alzheimer's disease; Positron emission tomography; Nervous system diseases; Alzheimer disease; Biological marker; Cerebrospinal fluid; Nuclear magnetic resonance imaging; Cerebral disorder; Central nervous system disease; Degenerative disease; mild cognitive impairment; Emission tomography
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1016/j.neuroimage.2012.09.065

Neurodegenerative disorders, such as Alzheimer's disease, are associated with changes in multiple neuroimaging and biological measures. These may provide complementary information for diagnosis and prognosis. We present a multi-modality classification framework in which manifolds are constructed based on pairwise similarity measures derived from random forest classifiers. Similarities from multiple modalities are combined to generate an embedding that simultaneously encodes information about all the available features. Multi-modality classification is then performed using coordinates from this joint embedding. We evaluate the proposed framework by application to neuroimaging and biological data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Features include regional MRI volumes, voxel-based FDG-PET signal intensities, CSF biomarker measures, and categorical genetic information. Classification based on the joint embedding constructed using information from all four modalities out-performs the classification based on any individual modality for comparisons between Alzheimer's disease patients and healthy controls, as well as between mild cognitive impairment patients and healthy controls. Based on the joint embedding, we achieve classification accuracies of 89% between Alzheimer's disease patients and healthy controls, and 75% between mild cognitive impairment patients and healthy controls. These results are comparable with those reported in other recent studies using multi-kernel learning. Random forests provide consistent pairwise similarity measures for multiple modalities, thus facilitating the combination of different types of feature data. We demonstrate this by application to data in which the number of features differs by several orders of magnitude between modalities. Random forest classifiers extend naturally to multi-class problems, and the framework described here could be applied to distinguish between multiple patient groups in the future.
► Multi-modal classification based on pairwise similarities derived from random forests ► Framework evaluated by application to neuroimaging and biological data from ADNI ► Random forests provide consistent pairwise similarity measures for various modalities ► Classification based on MR volumes, FDG-PET intensities, CSF biomarkers, genetic data ► Multi-modality classification out-performs that based on any individual modality