Development and Implementation of a Corriedale Ovine Brain Atlas for Use in Atlas-Based Segmentation

Abstract

Segmentation is the process of partitioning an image into subdivisions and can be applied to medical images to isolate anatomical or pathological areas for further analysis. This process can be done manually or automated by the use of image processing computer packages. Atlas-based segmentation automates this process by the use of a pre-labelled template and a registration algorithm. We developed an ovine brain atlas that can be used as a model for neurological conditions such as Parkinson's disease and focal epilepsy. 17 female Corriedale ovine brains were imaged in-vivo in a 1.5T (low-resolution) MRI scanner. 13 of the low-resolution images were combined using a template construction algorithm to form a low-resolution template. The template was labelled to form an atlas and tested by comparing manual with atlas-based segmentations against the remaining four low-resolution images. The comparisons were in the form of similarity metrics used in previous segmentation research. Dice Similarity Coefficients were utilised to determine the degree of overlap between eight independent, manual and atlas-based segmentations, with values ranging from 0 (no overlap) to 1 (complete overlap). For 7 of these 8 segmented areas, we achieved a Dice Similarity Coefficient of 0.5-0.8. The amygdala was difficult to segment due to its variable location and similar intensity to surrounding tissues resulting in Dice Coefficients of 0.0-0.2. We developed a low resolution ovine brain atlas with eight clinically relevant areas labelled. This brain atlas performed comparably to prior human atlases described in the literature and to intra-observer error providing an atlas that can be used to guide further research using ovine brains as a model and is hosted online for public access.

Fig 1. Methods Flowchart.

To create a template, an input image (template subject) undergoes a sequence of processing steps (black boxes) such as pre-processing, template construction, labelling and registration to finally output a labelled image. Red boxes represent input and output images.

Fig 2. Atlas—Template and Associated Labels.

Segmented Atlas showing (1) Ventricles, (2) Motor Cortices, (3) Hippocampi, (4) Thalami, (5) Caudate (6) Amygdala and (7) Cerebellum illustrating the (a) transverse, (b) parasagittal, (d) dorsal planes and (c) 3d rendering of labels on a ventrolateral view with rostral to the left of the image.

Fig 3. The (a) skulled template and (b) deskulling mask.

Variability displayed in extracortical areas reinforces the reason why deskulling is an important step in pre-processing.

Fig 4. Dice Similarity Coefficients of all structures in the low resolution template.

Values range from 0.12 for the amygdala to 0.75 for the cerebellum. Red = Ventricles, Green = Right Motor Cortex, Dark Blue = Left Motor Cortex, Yellow = Hippocampus, Light Blue = Thalamus, Purple = Caudate Nucleus, Peach = Amygdala, Grey = Cerebellum, Brown = Intraobserver Error. Error bars show the standard deviation of the mean for all compared structures. Intraobserver Error was calculated by comparing manual segmentations of all seven labelled structures in the four test subjects. Number of test subjects (n = 4).

Fig 5. Box plot displaying the average of all comparisons performed across all segmented structures.

Shows the median Dice Similarity Coefficient to be 0.61 with the False Positive and False Negative values having a very large variation. This indicates a wide spread in the accuracy of segmentation with regards to various structures. The Jaccard Coefficient is as expected lower than the Dice proportionally due to single use of the intersect in its calculation leading to a more precise comparison.