Multiple RF classifier for the hippocampus segmentation: Method and validation on EADC-ADNI Harmonized Hippocampal Protocol. Issue 8 (December 2015)

• Record Type: Journal Article
• Authors: Inglese, P.; Amoroso, N.; Boccardi, M.; Bocchetta, M.; Bruno, S.; Chincarini, A.; Errico, R.; Frisoni, G.B.; Maglietta, R.; Redolfi, A.; Sensi, F.; Tangaro, S.; Tateo, A.; Bellotti, R.
• Is Part Of: Physica medica. Volume 31:Issue 8(2015)
• Page Start: 1085
• Page End: 1091
• Publication Date: 2015-12
• Subjects: Hippocampus segmentation -- Random forest classifier -- Alzheimer's disease; Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2015.08.003
• Languages: English
• ISSNs: 1120-1797

Abstract

Highlights: We present a voxel-by-voxel based algorithm for automated hippocampal segmentation. A Multiple Random Forests algorithm was used to classify each voxel. Structural MRI from EADC-ADNI Harmonized Protocol labels is used as gold standard. The method is validated on a cohort of 50 T1 MRI scans for diagnosis of dementia. A test on an independent database of MRI scans has been presented. Abstract: The hippocampus has a key role in a number of neurodegenerative diseases, such as Alzheimer's Disease. Here we present a novel method for the automated segmentation of the hippocampus from structural magnetic resonance images (MRI), based on a combination of multiple classifiers. The method is validated on a cohort of 50 T1 MRI scans, comprehending healthy control, mild cognitive impairment, and Alzheimer's Disease subjects. The preliminary release of the EADC-ADNI Harmonized Protocol training labels is used as gold standard. The fully automated pipeline consists of a registration using an affine transformation, the extraction of a local bounding box, and the classification of each voxel in two classes (background and hippocampus). The classification is performed slice-by-slice along each of the three orthogonal directions of the 3D-MRI using a Random Forest (RF) classifier, followed by a fusion of the three full segmentations. Dice coefficients obtained by multiple RF (0.87 ± 0.03) are larger than those obtained by a single monolithic RF applied to the entire bounding box, and are comparable to state-of-the-art. A test on an external cohort of 50 T1 MRI scans shows that the presented method is robust and reliable. Additionally, a comparison of local changes in the morphology of the hippocampi between the three subject groups is performed. Our work showed that a multiple classification approach can be implemented for the segmentation for the measurement of volume and shape changes of the hippocampus with diagnostic purposes.