Automated Ischemic Lesion Segmentation in MRI Mouse Brain Data after Transient Middle Cerebral Artery Occlusion

Abstract

Magnetic resonance imaging (MRI) has become increasingly important in ischemic stroke experiments in mice, especially because it enables longitudinal studies. Still, quantitative analysis of MRI data remains challenging mainly because segmentation of mouse brain lesions in MRI data heavily relies on time-consuming manual tracing and thresholding techniques. Therefore, in the present study, a fully automated approach was developed to analyze longitudinal MRI data for quantification of ischemic lesion volume progression in the mouse brain. We present a level-set-based lesion segmentation algorithm that is built using a minimal set of assumptions and requires only one MRI sequence (T2) as input. To validate our algorithm we used a heterogeneous data set consisting of 121 mouse brain scans of various age groups and time points after infarct induction and obtained using different MRI hardware and acquisition parameters. We evaluated the volumetric accuracy and regional overlap of ischemic lesions segmented by our automated method against the ground truth obtained in a semi-automated fashion that includes a highly time-consuming manual correction step. Our method shows good agreement with human observations and is accurate on heterogeneous data, whilst requiring much shorter average execution time. The algorithm developed here was compiled into a toolbox and made publically available, as well as all the data sets.

MRI Mouse Brain Data of Ischemic Lesion after Transient Middle Cerebral Artery Occlusion

“…The novel method developed to segment the ischemic lesion in mouse brains in an automated fashion is described in detail in Mulder et al ( 2017 ) and is available for download from the Software Downloads section of our webpage 2 . The same manuscript describes how the semi-automated segmentations (by two observers) used for validation of the automated algorithm were obtained.…”

“…All files associated with a single scan of the mouse, 24 h after stroke onset (belonging to the 3- to 5-month-old mice age group) that was used as a template/reference image in the registration step of the automated ischemic lesion approach proposed by Mulder et al ( 2017 ), are provided in MetaImage format. The manually drawn labels were propagated to each data set and used to initialize the level-set-based segmentation.…”

“…In this data report we make available to the community a highly variable longitudinal MRI mouse brain data set of ischemic lesion after transient middle cerebral artery occlusion ( t MCAo). Together with the provided semi-automated and automated segmentations, these data can be used to further improve the method proposed by Mulder et al ( 2017 ) and also to serve as a benchmark for comparison between different approaches to segment ischemic lesions in MRI mouse brain data. It can also be used to develop and validate algorithms that further classify the stroke area into core and penumbra.…”

“…Segmentations of the ischemic lesions are provided as well. These were obtained by: (i) two observers using a semi-automated method and (ii) using the novel automated segmentation approach described by Mulder et al ( 2017 ).…”

Automatic Rodent Brain MRI Lesion Segmentation with Fully Convolutional Networks

“…Recently, few automatic approaches were developed to tackle rodent brain lesion segmentation, including statistical models [2], thresholding [4] and level-sets [13]. However, these methods rely on strict assumptions such as the distribution of the data, or they are limited to use a single image modality.…”