IC‐P‐088: Collaboration for a systematic comparison of different techniques to measure subfield volumes: Announcement and first results
Susanne G. Mueller, Paul A. Yushkevich, Lei Wang, Koen Van Leemput, Adam F. Mezher, Juan Eugenio Iglesias, Sandhitsu R. Das, Michael Walter Weiner · Alzheimer s & Dementia · 2013
The hippocampus consists of several histologically and functionally distinct subfields that are differently affected by neurodegenerative processes. Particularly in the early stages subfield-specific atrophy measurements often provide superior diagnostic information compared to global hippocampal atrophy measurements. Consequently several approaches to investigate subfield specific atrophy using in vivo MR imaging have been developed and have been successfully used in small populations. However a systematic performance comparison in a common, large, and well defined data set with different degrees of hippocampal atrophy has yet to be done. The overall goal of this study is therefore to compare the performance of five representative manual and automated T1 and T2 based subfield labeling techniques in a sub-set of the ADNI2 population with mild hippocampal atrophy (MCI, eMCI, amyloid-pos and amyloid-neg controls). Methods: The project has two phases: 1. Preparation: High resolution hippocampal T2 sequence harmonization and implementation at 20 3T Siemens ADNI2 sites. 2. Comparison of the performance of the following approaches: a. T1-based Freesurfer+Large-Diffeomorphic Metric Mapping in combination with shape analysis (1). b. T1-based deformation-based morphometry using a high resolution, high field ex vivo reference atlas (2). c. T1/T2-based Model based subfield segmentation using Bayesian Inference (3). d. T2-based automated multi-atlas segmentation combined with similarity-weighted voting (4). e. Manual subfield parcellation (5). Results: The first phase of the project has been completed and a sequence with the following parameters has been implemented: TR/TE 8020/50 ms 0.4x0.4x2.0 mm resolution. Project-specific data acquisition started late in December 2012 (n = 41) and the first data sets are now being processed. All high resolution images acquired for this project will be made available to the research community on the LONI website. The preliminary findings of this comparison will be presented. Conclusion: This project will highlight the particular strengths and limitations of the different approaches and so provide guidance for the selection of the appropriate approach for different research questions. 1. Hippocampus 2009:19:541–548 2. Neuroimage 2009;44:385–398 3. Hippocampus 2009;19:1658–1661 4. Neuroimage 2010;53:1208–1224 5. Neurobiol Aging 2007;28:719–726.