Caution: shortcomings of traditional segmentation methods from magnetic resonance imaging brain scans intended for 3-dimensional surface modelling in children with pathology

Abstract

This technical innovation assesses the adaptability of some common automated segmentation tools on abnormal pediatric magnetic resonance (MR) brain scans. We categorized 35 MR scans by pathologic features: (1) "normal"; (2) "atrophy"; (3) "cavity"; (4) "other." The following three tools, (1) Computational Anatomy Toolbox version 12 (CAT12); (2) Statistical Parametic Mapping version 12 (SPM12); and (3) MRTool, were tested on each scan-with default and adjusted settings. Success was determined by radiologist consensus on the surface accuracy. Automated segmentation failed in scans demonstrating severe surface brain pathology. Segmentation of the "cavity" group was ineffective, with success rates of 23.1% (CAT12), 69.2% (SPM12) and 46.2% (MRTool), even with refined settings and manual edits. Further investigation is required to improve this workflow and automated segmentation methodology for complex surface pathology.

Keywords: Brain; Children; Hypoxic ischemic injury; Magnetic resonance imaging; Printing; Segmentation; Three-dimensional.

Fig. 1

A 5-year-and-8-month-old girl with a normal magnetic resonance imaging brain scan. a, b Two successful segmentations produced using the Computational Anatomy Toolbox version 12 segmentation method in the 3-dimensional Standard Tessellation Language. The volume model (a) is visually higher in quality and demonstrates more cortical gyral and sulcal detail than the volume model (b) which shows flattened and less defined gyri. Gyral and sulcal detail is an important component of brain surface demonstration

Fig. 2

A 10-year-and-11-month-old girl who suffered a perinatal hypoxic ischemic injury with a watershed pattern of injury. a–c Coronal (a), sagittal (b) and axial (c) T1W magnetic resonance images show bilateral and symmetric cortical and subcortical atrophy in the posterior and perisylvian watershed regions (arrows). d–f Oblique (d), lateral (e) and vertex (f) views of the 3-dimensional (D) mesh virtual model with the corresponding areas of posterior and perisylvian watershed atrophy seen as prominence of the surface sulci with decreased size of the gyri (arrows). g–i Oblique (g), posterior vertex (h) and center vertex (i) photographic views of the corresponding 3-D print models show the affected areas (arrows)

Fig. 3

An 11-year-and-6-month-old boy with combined acute-profound and partial-prolonged hypoxic ischemic injury sustained perinatally. a Axial T1W magnetic resonance image shows focal asymmetric atrophy with a wide right peri-Sylvian fissure (white arrow). There is an expanded posterior body of the right lateral ventricle approximating the surface of the right parietal lobe (due to regional atrophy) and a very thin cortical ribbon (black arrow) accurately depicted in the 3-dimensional (D) images. b, c Vertex (b) and right-sided lateral (c) views of the 3-D mesh show bilateral asymmetric atrophy of the posterior peri-Sylvian and peri-Rolandic regions (brackets). The right is more affected than the left, although with an intact cortical rim, as evidenced by the expanded and smooth cortex overlying the occipital horn of the right lateral ventricle (black arrow in b). The smooth cortex is also seen in  (c)  (black arrow). Interhemispheric widening (asterisk in b) is a sign of atrophy. There is also atrophy of the superior and middle regions of the right peri-Sylvian fissure, which represents the watershed between all three major vessels supplying the brain (small white arrows) with associated widening of the peri-Sylvian fissure (large white arrow in c).  d, e Right (d) and left (e) lateral views of the 3-D print model accurately show the corresponding areas of involvement. There is widening of the peri-Rolandic (small white arrow) and peri-Sylvian (large white arrow) regions. The 3-D model accurately depicts the thin cortical mantle overlying the occipital horn of the right lateral ventricle (black arrow in d). The area of atrophy in (e) at the junction of the peri-Sylvian fissure, medial peri-Rolandic region and posterior intervascular watershed (arrow) is in keeping with a combined partial prolonged and acute profound hypoxic ischemic injury

Fig. 4

Segmentation into separate tissue classes (gray matter, white matter and cerebrospinal fluid) then combining relevant classes (gray and white matter only) into a volume structure MRI magnetic resonance imaging

Fig. 5

A 2-year-and-2-month-old boy who suffered an acute-profound perinatal hypoxic ischemic injury. a–c Multiplanar coronal (a), sagittal (b) and axial (c) reconstructions of T1-weighted magnetic resonance images demonstrate bilateral, relatively symmetric peri-Rolandic volume loss (arrows in a and b). d High-fidelity three-dimensional surface mesh achieved by generating a standard tesselation language model using the Statistical Parametric Mapping version 12 technique accurately demonstrates the fine cortical detail at the atrophic peri-Rolandic regions (arrows)

Fig. 6

Standard tessellation language surface model examples depicting unsuccessful segmentation of an MRI of the brain in a 5-year-and-2-month-old boy with default settings for (a) Computational Anatomy Toolbox version 12, (b) Statistical Parametric Mapping version 12, and (c) MRTool segmentations. a Shows a large open hole in the volume (arrow). b Compared to (a), this image shows a  smaller hole in the volume (arrow). c There is diffuse exclusion of outer gray matter (incorrect segmentation of the gray matter) in the entire volume, as demonstrated by the extreme irregular sulcal pattern when compared to the models in (a) and (b)

Fig. 7

A 10-year-and-5-month-old girl with a previous global partial-prolonged hypoxic ischemic injury affecting the parasagittal watershed regions. a–c Multiplanar coronal (a), sagittal (b) and axial (c) reconstructions of T1 weighted magnetic resonance images demonstrate bilateral parasagittal cystic encephalomalacia (arrows) and very thin residual cortex of the superior frontal lobes. d Vertex view of an initial standard tessellation language (STL) surface model which was unsuccessful in demonstrating the thin cortical mantle of the right frontal lobe, leaving a large hole with a view into the cystic cavity involving the right frontal lobe white matter (asterisk). e Vertex view of a subsequent usable STL surface model created by patching the hole using manual editing (asterisk). Significant parasagittal watershed (between brackets) atrophy is accurately demonstrated (arrows). f Vertex view of the final 3-D print model shows the interhemispheric widening (arrow) and the area that was patched within the right frontal lobe (asterisk)