Manually segmented template library for 8-year-old pediatric brain MRI data with 16 subcortical structures

Abstract

Abstract. Manual segmentation of anatomy in brain MRI data taken to be the closest to the "gold standard" in quality is often used in automated registration-based segmentation paradigms for transfer of template labels onto the unlabeled MRI images. This study presents a library of template data with 16 subcortical structures in the central brain area which were manually labeled for MRI data from 22 children (8 male, mean age ¼ 8 AE 0.6 years). The lateral ventricle, thalamus, caudate, putamen, hippocampus, cerebellum, third vevntricle, fourth ventricle, brainstem, and corpuscallosum were segmented by two expert raters. Cross-validation experiments with randomized template subset selection were conducted to test for their ability to accurately segment MRI data under an automated segmentation pipeline. A high value of the dice similarity coefficient (0.86 AE 0.06, min ¼ 0.74, max ¼ 0.96) and small Hausdorff distance (3.33 AE 4.24, min ¼ 0.63, max ¼ 25.24) of the automated segmentation against the manual labels was obtained on this template library data. Additionally, comparison with segmentation obtained from adult templates showed significant improvement in accuracy with the use of an age-matched library in this cohort. A manually delineated pediatric template library such as the one described here could provide a useful benchmark for testing segmentation algorithms.

Morphological alterations in the caudate, putamen, pallidum, and thalamus in Parkinson's disease

“…This segmentation method has already been shown to provide high quality segmentation for segmentation of caudate, thalamus, putamen, and hippocampus (average dice coefficient = 0.85). Another validation study in the context of segmentation of pediatric MRI images found the dice coefficients of 0.89 for thalamus, 0.89 for caudate and 0.87 for putamen, thereby emphasizing accurate segmentation results (Garg et al, 2014 ).…”

“…The choice of templates for segmentation is known to impact segmentation quality (Garg et al, 2014 ). In our study, we used templates ( n = 6) from the UBC cohort which were manually segmented by a neuro-anatomy expert.…”