C15 Standard brain template and multi-atlas based segmentation of tghd minipig brain. (13th September 2016)

• Record Type: Journal Article
• Authors: Schubert, Robin; Freisfeld, Frauke; Johnson, Hans; Kim, Eunyoung Regina; Nagelmann, Nina; Motlik, Jan; Schramke, Sarah; Schuldenzucker, Verena; Rieke, Lorena; Matheis, Tamara; Wirsig, Maike; Faber, Cornelius; Reilmann, Ralf
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 87(2016)Supplement 1
• Page Start: A32
• Page End: A32
• Publication Date: 2016-09-13
• Subjects: Animal models -- Brain atlas -- MRI -- Minipig -- Preclinical research; Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2016-314597.95
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : Background: In the TRACK tgHD minipig study, we aim to track longitudinal progression of the Libechov minipig, transgenic for the huntingtin gene. Annual MRI scans are conducted, including anatomical, spectroscopic and diffusion weighted imaging (DWI) sequences. Quantitative analyses require proper segmentation of the brain images. Standard Brain Templates are used as registration target, for intra subject comparison of various MR endpoints. Aims: Primary objective is to provide a multi-atlas for the Libechov minipig for automated segmentation of MR brain images. Secondary objectives are to achieve standard brain templates for various functions, including T1-weighted template and diffusivity templates. Methods: Nine T1 weighted, sagittal acquired MRI volumes of Libechov minipigs were transformed into a common coordinate system and manually skull-stripped. A domestic pig reference atlas was registered on the extracted brains with a landmark and intensity based algorithm. The BRAINSTools library was used for tissue segmentation and structural segmentation. The segmentation was cleaned up manually and merged with Multi-atlas joint label fusion. The T1 weighted images were acquired with a 3 T Philips Achieva scanner. Results: A T1-weighted standard brain volume has been created, with 10 segmented regions of interest, covering the basal ganglia and surroundings. Conclusions: It was shown in previous data, that comprehensive MR assessments are feasible with the Libechov minipig, yielding high quality images. However, more work is needed to manually segment minipig brain images, to expand the multi-atlas to a feasible size. This will open the way to automated segmentation and thus to statistical analyses. This study was supported by a grant of the CHDI Foundation.