A comparison of automatic multi-tissue segmentation methods of the human fetal brain using the FeTA Dataset.
Year: 2020
Type: article
Abstract: It is critical to quantitatively analyse the developing human fetal brain in order to fully understand neurodevelopment in both normal fetuses and those with congenital disorders.To facilitate this analysis, automatic multi-tissue fetal brain segmentation algorithms are needed, which in turn requires open databases of segmented fetal brains.Here we introduce a publicly available database of 50 manually segmented pathological and non-pathological fetal magnetic resonance brain volume reconstructions across a range of gestational ages (20 to 33 weeks) into 7 different tissue categories (external cerebrospinal fluid, grey matter, white matter, ventricles, cerebellum, deep grey matter, brainstem/spinal cord).In addition, we quantitatively evaluate the accuracy of several automatic multi-tissue segmentation algorithms of the developing human fetal brain.Four research groups participated, submitting a total of 10 algorithms, demonstrating the benefits the database for the development of automatic algorithms. Background & SummaryCongenital disorders are one of the leading causes of neonatal mortality worldwide (http://www.who.int/gho/child_health/mortality/causes/en/).In order to detect and treat congenital disorders, prenatal maternal and fetal healthcare is of the utmost importance.Fetal imaging with ultrasound has been an essential tool in prenatal care for many years.Recently, fetal magnetic resonance imaging (MRI) has started to emerge as an important supplemental tool for providing information about the developing fetus when the ultrasound image is unclear, or for more accurate diagnosis of congenital disorders [1][2][3] .Fetal MRI, especially of the brain, has been shown to be useful in the diagnosis of many congenital disorders such as spina bifida, intrauterine growth retardation, congenital heart disease, lissencephaly, and corpus callosum anomalies [3][4][5][6][7] .Fetal MRI is a challenging imaging modality requiring both clinical and technical expertise as the fetus is not sedated and is able to move freely.Ultra-fast MRI sequences such as T2-weighted single shot fast spin echo (ssFSE) acquire low resolution images very quickly and can be acquired in all planes.Super-resolution (SR) reconstruction algorithms can then be applied in order to combine several low-resolution images into a single high resolution volume, incorporating outlier rejection and motion correction strategies [8][9][10][11][12][13][14] , which can then be used for further quantitative analysis.Automated quantification of the highly complex and rapidly changing brain morphology in MRI data could improve the diagnostic and decisionmaking process.An initial step for the quantification of fetal brain morphology and volumetry is image segmentation.It is clinically relevant to analyse information such as shape or volume of the developing cortex, cerebellum, brainstem, white matter and cerebrospinal fluid spaces, as many congenital disorders cause subtle changes to these tissue compartments 6,15,16 .Existing growth data is mainly based on normally developing brains [17][18][19] , and we lack growth data for many pathologies and congenital disorders.The automatic segmentation of the developing human brain would be a first step in being able to perform such an analysis, as manual segmentation is both timeconsuming and prone to human error.However, segmenting these SR fetal brain volumes is still challenging due to rapid changing anatomy of the fetal brain, remaining motion or blurry artefacts and partial volume effects.
Authors: Kelly Payette, Priscille de Dumast, Hamza Kebiri, Ivan Ezhov, Johannes C. Paetzold, Suprosanna Shit, Asim Iqbal, Romesa Khan, Raimund Kottke, Patrice Grehten, Hui Ji, Levente Lánczi, Marianna Nagy, Mónika Béresová, Thi Dao Nguyen, Giancarlo Natalucci, Theofanis Karayannis, Bjoern Menze, Meritxell Bach Cuadra, András Jakab
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Keywords: fetal brain segmentation, fetal brain MRI, brain development, congenital disorders, super-resolution reconstruction, T2-weighted turbo spin echo, segmentation algorithm evaluation, cerebrospinal fluid, grey matter, white matter, ventricles, cerebellum, deep gray matter, prenatal diagnosis, gestational age
Open Access status: green
OpenAlex ID: W3097323379
