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

Largent, A.; Asis-Cruz, Josepheen De; Kapse, Kushal; Barnett, Scott D.; Murnick, Jonathan; Basu, Sudeepta K; Andersen, Nicole; Norman, Stephanie; Andescavage, Nickie; Limperopoulos, Catherine

doi:10.1002/hbm.25762

Cited by 8 publications

(2 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For 3D cUS to play a role in diagnosing and monitoring PHVD in other clinical settings, an efficient and fully automated segmentation algorithm is crucial. Several other groups have adapted this technique, given its reliability in VV estimation 67 , 68 . Nevertheless, we still achieved reliable VV estimates using our current system.…”

Section: Discussionmentioning

confidence: 99%

Three-dimensional cranial ultrasound and functional near-infrared spectroscopy for bedside monitoring of intraventricular hemorrhage in preterm neonates

Kebaya

Stubbs²,

Lo³

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Germinal Matrix-Intraventricular Hemorrhage (GMH-IVH) remains a significant cause of adverse neurodevelopment in preterm infants. Current management relies on 2-dimensional cranial ultrasound (2D cUS) ventricular measurements. Reliable biomarkers are needed to aid in the early detection of posthemorrhagic ventricular dilatation (PHVD) and subsequent neurodevelopment. In a prospective cohort study, we incorporated 3-dimensional (3D) cUS and functional near-infrared spectroscopy (fNIRS) to monitor neonates with GMH-IVH. Preterm neonates (≤ 32 weeks' gestation) were enrolled following a GMH-IVH diagnosis. Neonates underwent sequential measurements: 3D cUS images were manually segmented using in-house software, and the ventricle volumes (VV) were extracted. Multichannel fNIRS data were acquired using a high-density system, and spontaneous functional connectivity (sFC) was calculated. Of the 30 neonates enrolled in the study, 19 (63.3%) had grade I–II and 11 (36.7%) grade III–IV GMH-IVH; of these, 7 neonates (23%) underwent surgical interventions to divert cerebrospinal fluid (CSF). In infants with severe GMH-IVH, larger VV were significantly associated with decreased |sFC|. Our findings of increased VV and reduced sFC suggest that regional disruptions of ventricular size may impact the development of the underlying white matter. Hence, 3D cUS and fNIRS are promising bedside tools for monitoring the progression of GMH-IVH in preterm neonates.

show abstract

Section: Discussionmentioning

confidence: 99%

Three-dimensional cranial ultrasound and functional near-infrared spectroscopy for bedside monitoring of intraventricular hemorrhage in preterm neonates

Kebaya

Stubbs²,

Lo³

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In the medical image segmentation tasks, Largent et al used MC-Dropout and U-Net architecture as baseline models for automatic brain segmentation in preterm infants. The proposed method shows the best segmentation results across all tested methods and produces accurate uncertainty maps [25].…”

Section: ) Recent Bdl Applicationsmentioning

confidence: 94%

On Performance and Calibration of Natural Gradient Langevin Dynamics

et al. 2023

View full text Add to dashboard Cite

Producing deep neural network (DNN) models with calibrated confidence is essential for applications in many fields, such as medical image analysis, natural language processing, and robotics. Modern neural networks have been reported to be poorly calibrated compared with those from a decade ago. The stochastic gradient Langevin dynamics (SGLD) algorithm offers a tractable approximate Bayesian inference applicable to DNN, providing a principled method for learning the uncertainty. A recent benchmark study showed that SGLD could produce a more robust model to covariate shifts than other competing methods. However, vanilla SGLD is also known to be slow, and preconditioning can improve SGLD efficacy. This paper proposes eigenvalue-corrected Kronecker factorization (EKFAC) preconditioned SGLD (EKS-GLD), in which a novel second-order gradient approximation is employed as a preconditioner for the SGLD algorithm. This approach is expected to bring together the advantages of both second-order optimization and the approximate Bayesian method. Experiments were conducted to compare the performance of EKSGLD with existing preconditioning methods and showed that it could achieve higher predictive accuracy and better calibration on the validation set. EKSGLD improved the best accuracy by 3.06% on CIFAR-10 and 4.15% on MNIST, improved the best negative log-likelihood by 16.2% on CIFAR-10 and 11.4% on MNIST, and improved the best thresholded adaptive calibration error by 4.05% on CIFAR-10.

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

Self Cite

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

Cited by 8 publications

References 38 publications

Three-dimensional cranial ultrasound and functional near-infrared spectroscopy for bedside monitoring of intraventricular hemorrhage in preterm neonates

Three-dimensional cranial ultrasound and functional near-infrared spectroscopy for bedside monitoring of intraventricular hemorrhage in preterm neonates

On Performance and Calibration of Natural Gradient Langevin Dynamics

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

Contact Info

Product

Resources

About