Automatic segmentation of ventricular volume by 3D ultrasonography in post haemorrhagic ventricular dilatation among preterm infants

Gontard, Lionel C.; Pizarro, J.; Sanz-Peña, Borja; López, Simón Pedro Lubián; Benavente‐Fernández, Isabel

doi:10.1038/s41598-020-80783-3

Cited by 12 publications

(20 citation statements)

References 23 publications

(21 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As 3D US is not widely available we have focused on its utility to make 2D US measurements reliable compared to a 3D US manual segmentation and to both linear and volumetric MRI measures. However, this study reinforces the potential role 3D US has in the NICU which has been previously recognized by other authors ( 26 , 27 ) and by our group ( 14 , 16 , 28 , 29 ). 3D allows navigation through the three planes once a whole brain acquisition has been obtained, is faster than 2D US and allows review offline of any possible section of interest instead of having a static 2D image saved ( 30 – 32 ).…”

Section: Discussionsupporting

confidence: 92%

Ultrasonographic Estimation of Total Brain Volume: 3D Reliability and 2D Estimation. Enabling Routine Estimation During NICU Admission in the Preterm Infant

et al. 2021

Self Cite

View full text Add to dashboard Cite

Objectives: The aim of this study is to explore if manually segmented total brain volume (TBV) from 3D ultrasonography (US) is comparable to TBV estimated by magnetic resonance imaging (MRI). We then wanted to test 2D based TBV estimation obtained through three linear axes which would enable monitoring brain growth in the preterm infant during admission.Methods: We included very low birth weight preterm infants admitted to our neonatal intensive care unit (NICU) with normal neuroimaging findings. We measured biparietal diameter, anteroposterior axis, vertical axis from US and MRI and TBV from both MRI and 3D US. We calculated intra- and interobserver agreement within and between techniques using the intraclass correlation coefficient and Bland-Altman methodology. We then developed a multilevel prediction model of TBV based on linear measurements from both US and MRI, compared them and explored how they changed with increasing age. The multilevel prediction model for TBV from linear measures was tested for internal and external validity and we developed a reference table for ease of prediction of TBV.Results: We used measurements obtained from 426 US and 93 MRI scans from 118 patients. We found good intra- and interobserver agreement for all the measurements. US measurements were reliable when compared to MRI, including TBV which achieved excellent agreement with that of MRI [ICC of 0.98 (95% CI 0.96–0.99)]. TBV estimated through 2D measurements of biparietal diameter, anteroposterior axis, and vertical axis was comparable among both techniques. We estimated the population 95% confidence interval for the mean values of biparietal diameter, anteroposterior axis, vertical axis, and total brain volume by post-menstrual age. A TBV prediction table based on the three axes is proposed to enable easy implementation of TBV estimation in routine 2D US during admission in the NICU.Conclusions: US measurements of biparietal diameter, vertical axis, and anteroposterior axis are reliable. TBV segmented through 3D US is comparable to MRI estimated TBV. 2D US accurate estimation of TBV is possible through biparietal diameter, vertical, and anteroposterior axes.

show abstract

Section: Discussionsupporting

confidence: 92%

Ultrasonographic Estimation of Total Brain Volume: 3D Reliability and 2D Estimation. Enabling Routine Estimation During NICU Admission in the Preterm Infant

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…15,36 A few deep learning-based methods have been reported for segmentation of ventricles from 3D US images. Gontard et al 16 developed a 2D SegNet CNN with 152 3D US images of 10 patients and 230 ventricles, providing a mean intersection over union (IoU) of 0.54 on the training data. However, their validation results used only 83 3D US images from six patients (122 ventricles), and they did not report on distancebased metrics.…”

Section: Discussionmentioning

confidence: 99%

“…In addition, models of the 3D Attention U‐Net and 3D U‐Net++ with shape priors were tested on the two‐ventricle images only to determine if adding shape regularization can further improve those models alone. A 2D SegNet with the VGG16 architecture backbone, same as the one implemented by Gontard et al., 16 was tested and compared for one experiment 17 . The 2D SegNet was implemented using image slices along the sagittal plane, which was the same plane used when creating the manual annotations.…”

Section: Methodsmentioning

confidence: 99%

“…A threshold was applied and the largest objects in the mask were retained to remove potential outliers. The final proposed 3D ensemble was compared to each standalone 3D model performance, the 2D SegNet used by Gontard et al 16 and the 2D multiplane U-Net developed by Szentimrey et al 19 All models were trained under the same conditions and data split.…”

Section: Overview Of Segmentation Pipelinementioning

confidence: 99%

“…[12][13][14][15] Two such studies 13,14 required a user to initialize the background and foreground voxels before solving a convex optimized surface evolution function using a continuous max-flow algorithm. A 2D SegNet model was developed by Gontard et al 16 using the VGG16 encoder-decoder architecture for 3D US segmentation; however, the model interpreted the data in 2D slices only, disregarding valuable 3D information. 17 Additionally, a 2D convolutional neural network (CNN) approach by Martin et al 18 segmented cerebral ventricles on 15 3D US images only using the U-Net architecture.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automated 3D U‐net based segmentation of neonatal cerebral ventricles from 3D ultrasound images

et al. 2022

View full text Add to dashboard Cite

Background: Intraventricular hemorrhaging (IVH) within cerebral lateral ventricles affects 20-30% of very low birth weight infants (<1500 g). As the ventricles increase in size, the intracranial pressure increases, leading to posthemorrhagic ventricle dilatation (PHVD), an abnormal enlargement of the head. The most widely used imaging tool for measuring IVH and PHVD is cranial two-dimensional (2D) ultrasound (US). Estimating volumetric changes over time with 2D US is unreliable due to high user variability when locating the same anatomical location at different scanning sessions. Compared to 2D US, threedimensional (3D) US is more sensitive to volumetric changes in the ventricles and does not suffer from variability in slice acquisition. However, 3D US images require segmentation of the ventricular surface, which is tedious and time-consuming when done manually. Purpose: A fast, automated ventricle segmentation method for 3D US would provide quantitative information in a timely manner when monitoring IVH and PHVD in pre-term neonates. To this end, we developed a fast and fully automated segmentation method to segment neonatal cerebral lateral ventricles from 3D US images using deep learning. Methods: Our method consists of a 3D U-Net ensemble model composed of three U-Net variants,each highlighting various aspects of the segmentation task such as the shape and boundary of the ventricles. The ensemble is made of a U-Net++, attention U-Net, and U-Net with a deep learning-based shape prior combined using a mean voting strategy. We used a dataset consisting of 190 3D US images, which was separated into two subsets, one set of 87 images contained both ventricles, and one set of 103 images contained only one ventricle (caused by limited field-of -view during acquisition).We conducted fivefold crossvalidation to evaluate the performance of the models on a larger amount of test data; 165 test images of which 75 have two ventricles (two-ventricle images) and 90 have one ventricle (one-ventricle images). We compared these results to each stand-alone model and to previous works including,2D multiplane U-Net and 2D SegNet models. Results: Using fivefold cross-validation, the ensemble method reported a Dice similarity coefficient (DSC) of 0.720 ± 0.074, absolute volumetric difference (VD) of 3.7 ± 4.1 cm 3 , and a mean absolute surface distance (MAD) of 1.14 ± 0.41 mm on 75 two-ventricle test images. Using 90 test images with a single ventricle, the model after cross-validation reported DSC, VD, and MAD values of 0.806 ± 0.111, 3.5 ± 2.9 cm 3 , and 1.37 ± 1.70 mm, respectively. Compared to alternatives, the proposed ensemble yielded a higher accuracy in segmentation on both test data sets. Our method required approximately 5 s to segment 1034

show abstract

Automatic brain segmentation in preterm infants with post‐hemorrhagic hydrocephalus using 3D Bayesian U‐Net

Largent

Asis-Cruz

Kapse

et al. 2022

Human Brain Mapping

View full text Add to dashboard Cite

Post‐hemorrhagic hydrocephalus (PHH) is a severe complication of intraventricular hemorrhage (IVH) in very preterm infants. PHH monitoring and treatment decisions rely heavily on manual and subjective two‐dimensional measurements of the ventricles. Automatic and reliable three‐dimensional (3D) measurements of the ventricles may provide a more accurate assessment of PHH, and lead to improved monitoring and treatment decisions. To accurately and efficiently obtain these 3D measurements, automatic segmentation of the ventricles can be explored. However, this segmentation is challenging due to the large ventricular anatomical shape variability in preterm infants diagnosed with PHH. This study aims to (a) propose a Bayesian U‐Net method using 3D spatial concrete dropout for automatic brain segmentation (with uncertainty assessment) of preterm infants with PHH; and (b) compare the Bayesian method to three reference methods: DenseNet, U‐Net, and ensemble learning using DenseNets and U‐Nets. A total of 41 T2‐weighted MRIs from 27 preterm infants were manually segmented into lateral ventricles, external CSF, white and cortical gray matter, brainstem, and cerebellum. These segmentations were used as ground truth for model evaluation. All methods were trained and evaluated using 4‐fold cross‐validation and segmentation endpoints, with additional uncertainty endpoints for the Bayesian method. In the lateral ventricles, segmentation endpoint values for the DenseNet, U‐Net, ensemble learning, and Bayesian U‐Net methods were mean Dice score = 0.814 ± 0.213, 0.944 ± 0.041, 0.942 ± 0.042, and 0.948 ± 0.034 respectively. Uncertainty endpoint values for the Bayesian U‐Net were mean recall = 0.953 ± 0.037, mean negative predictive value = 0.998 ± 0.005, mean accuracy = 0.906 ± 0.032, and mean AUC = 0.949 ± 0.031. To conclude, the Bayesian U‐Net showed the best segmentation results across all methods and provided accurate uncertainty maps. This method may be used in clinical practice for automatic brain segmentation of preterm infants with PHH, and lead to better PHH monitoring and more informed treatment decisions.

show abstract

Automatic segmentation of ventricular volume by 3D ultrasonography in post haemorrhagic ventricular dilatation among preterm infants

Cited by 12 publications

References 23 publications

Ultrasonographic Estimation of Total Brain Volume: 3D Reliability and 2D Estimation. Enabling Routine Estimation During NICU Admission in the Preterm Infant

Ultrasonographic Estimation of Total Brain Volume: 3D Reliability and 2D Estimation. Enabling Routine Estimation During NICU Admission in the Preterm Infant

Automated 3D U‐net based segmentation of neonatal cerebral ventricles from 3D ultrasound images

Automatic brain segmentation in preterm infants with post‐hemorrhagic hydrocephalus using 3D Bayesian U‐Net

Contact Info

Product

Resources

About