1
|
Alzubaidi M, Shah U, Agus M, Househ M. FetSAM: Advanced Segmentation Techniques for Fetal Head Biometrics in Ultrasound Imagery. IEEE OPEN JOURNAL OF ENGINEERING IN MEDICINE AND BIOLOGY 2024; 5:281-295. [PMID: 38766538 PMCID: PMC11100952 DOI: 10.1109/ojemb.2024.3382487] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/30/2024] [Revised: 03/05/2024] [Accepted: 03/24/2024] [Indexed: 05/22/2024] Open
Abstract
Goal: FetSAM represents a cutting-edge deep learning model aimed at revolutionizing fetal head ultrasound segmentation, thereby elevating prenatal diagnostic precision. Methods: Utilizing a comprehensive dataset-the largest to date for fetal head metrics-FetSAM incorporates prompt-based learning. It distinguishes itself with a dual loss mechanism, combining Weighted DiceLoss and Weighted Lovasz Loss, optimized through AdamW and underscored by class weight adjustments for better segmentation balance. Performance benchmarks against prominent models such as U-Net, DeepLabV3, and Segformer highlight its efficacy. Results: FetSAM delivers unparalleled segmentation accuracy, demonstrated by a DSC of 0.90117, HD of 1.86484, and ASD of 0.46645. Conclusion: FetSAM sets a new benchmark in AI-enhanced prenatal ultrasound analysis, providing a robust, precise tool for clinical applications and pushing the envelope of prenatal care with its groundbreaking dataset and segmentation capabilities.
Collapse
Affiliation(s)
- Mahmood Alzubaidi
- College of Science and EngineeringHamad Bin Khalifa UniversityDoha34110Qatar
| | - Uzair Shah
- College of Science and EngineeringHamad Bin Khalifa UniversityDoha34110Qatar
| | - Marco Agus
- College of Science and EngineeringHamad Bin Khalifa UniversityDoha34110Qatar
| | - Mowafa Househ
- College of Science and EngineeringHamad Bin Khalifa UniversityDoha34110Qatar
| |
Collapse
|
2
|
Jafrasteh B, Lubián-López SP, Benavente-Fernández I. A deep sift convolutional neural networks for total brain volume estimation from 3D ultrasound images. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2023; 242:107805. [PMID: 37738840 DOI: 10.1016/j.cmpb.2023.107805] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 12/21/2022] [Revised: 08/04/2023] [Accepted: 09/07/2023] [Indexed: 09/24/2023]
Abstract
Preterm infants are a highly vulnerable population. The total brain volume (TBV) of these infants can be accurately estimated by brain ultrasound (US) imaging which enables a longitudinal study of early brain growth during Neonatal Intensive Care (NICU) admission. Automatic estimation of TBV from 3D images increases the diagnosis speed and evades the necessity for an expert to manually segment 3D images, which is a sophisticated and time consuming task. We develop a deep-learning approach to estimate TBV from 3D ultrasound images. It benefits from deep convolutional neural networks (CNN) with dilated residual connections and an additional layer, inspired by the fuzzy c-Means (FCM), to further separate the features into different regions, i.e. sift layer. Therefore, we call this method deep-sift convolutional neural networks (DSCNN). The proposed method is validated against three state-of-the-art methods including AlexNet-3D, ResNet-3D, and VGG-3D, for TBV estimation using two datasets acquired from two different ultrasound devices. The results highlight a strong correlation between the predictions and the observed TBV values. The regression activation maps are used to interpret DSCNN, allowing TBV estimation by exploring those pixels that are more consistent and plausible from an anatomical standpoint. Therefore, it can be used for direct estimation of TBV from 3D images without needing further image segmentation.
Collapse
Affiliation(s)
- Bahram Jafrasteh
- Biomedical Research and Innovation Institute of Cádiz (INiBICA), Puerta del Mar University, Cádiz, Spain.
| | - Simón Pedro Lubián-López
- Biomedical Research and Innovation Institute of Cádiz (INiBICA), Puerta del Mar University, Cádiz, Spain; Division of Neonatology, Department of Paediatrics, Puerta del Mar University Hospital, Cádiz, Spain.
| | - Isabel Benavente-Fernández
- Biomedical Research and Innovation Institute of Cádiz (INiBICA), Puerta del Mar University, Cádiz, Spain; Division of Neonatology, Department of Paediatrics, Puerta del Mar University Hospital, Cádiz, Spain; Area of Paediatrics, Department of Child and Mother Health and Radiology, Medical School, University of Cádiz, Cádiz, Spain.
| |
Collapse
|
3
|
Benavente-Fernández I, Steggerda SJ, Liem KD, Lubián-López S, de Vries LS. Ultrasonographic Estimation of Ventricular Volume in Infants Born Preterm with Posthemorrhagic Ventricular Dilatation: A Nested Substudy of the Randomized Controlled Early Versus Late Ventricular Intervention Study (ELVIS) Trial. J Pediatr 2023; 261:113578. [PMID: 37353143 DOI: 10.1016/j.jpeds.2023.113578] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 01/13/2023] [Revised: 05/29/2023] [Accepted: 06/16/2023] [Indexed: 06/25/2023]
Abstract
OBJECTIVE To study the potential role of ventricular volume (VV) estimation in the management of posthemorrhagic ventricular dilatation related to the need for ventriculoperitoneal (VP)-shunt insertion and 2-year neurodevelopmental outcome in infants born preterm. STUDY DESIGN We included 59 patients from the Early vs Late Ventricular Intervention Study from 4 participating centers. VV was manually segmented in 209 3-dimensional ultrasound scans and estimated from 2-dimensional ultrasound linear measurements in a total of 1226 ultrasounds. We studied the association of both linear measurements and VV to the need for VP shunt and 2-year neurodevelopmental outcome in the overall cohort and in the 29 infants who needed insertion of a reservoir. We used general estimating equations to account for repeated measures per individual. RESULTS Maximum pre-reservoir VV (β coefficient = 0.185, P = .0001) and gestational age at birth (β = -0.338; P = .0001) were related to the need for VP shunt. The estimated optimal single VV measurement cut point of 17 cm3 correctly classified 79.31% with an area under the curve of 0.76 (CI 95% 0.74-0.79). Maximum VV (β = 0.027; P = .012) together with VP shunt insertion (β = 3.773; P = .007) and gestational age (β = -0.273; P = .0001) were related to cognitive outcome at 2 years. Maximum ventricular index and anterior horn width before reservoir insertion were independently associated with the need of VP shunt and the proposed threshold groups in the Early vs Late Ventricular Intervention Study trial were associated with long-term outcome. CONCLUSIONS Pre-reservoir VV measurements were associated with the need for VP-shunt insertion and 2-year cognitive outcome among infants born preterm with posthemorrhagic ventricular dilatation. TRIAL REGISTRATION ISRCTN43171322.
Collapse
Affiliation(s)
- Isabel Benavente-Fernández
- Area of Paediatrics, Department of Child and Mother Health and Radiology, Medical School, University of Cádiz, Cádiz, Spain; Division of Neonatology, Department of Paediatrics, Puerta del Mar University Hospital, Cádiz, Spain; Biomedical Research and Innovation Institute of Cádiz (INiBICA) Research Unit, Puerta del Mar University, Cádiz, Spain.
| | - Sylke J Steggerda
- Division of Neonatology, Department of pediatrics, Leiden University Medical Center, Willem-Alexander Children's Hospital, Leiden, The Netherlands
| | - Kian D Liem
- Division of Neonatology, Department of Paediatrics, Radboud University Medical Centre Nijmegen, Amalia Children's Hospital, Nijmegen, The Netherlands
| | - Simón Lubián-López
- Division of Neonatology, Department of Paediatrics, Puerta del Mar University Hospital, Cádiz, Spain; Biomedical Research and Innovation Institute of Cádiz (INiBICA) Research Unit, Puerta del Mar University, Cádiz, Spain
| | - Linda S de Vries
- Division of Neonatology, Department of pediatrics, Leiden University Medical Center, Willem-Alexander Children's Hospital, Leiden, The Netherlands; Department of Neonatology, Wilhelmina Children's Hospital, Utrecht, The Netherlands
| |
Collapse
|
4
|
López-Arango G, Deguire F, Agbogba K, Boucher MA, Knoth IS, El-Jalbout R, Côté V, Damphousse A, Kadoury S, Lippé S. Impact of brain overgrowth on sensorial learning processing during the first year of life. Front Hum Neurosci 2022; 16:928543. [PMID: 35927999 PMCID: PMC9344916 DOI: 10.3389/fnhum.2022.928543] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/25/2022] [Accepted: 06/29/2022] [Indexed: 11/13/2022] Open
Abstract
Macrocephaly is present in about 2–5% of the general population. It can be found as an isolated benign trait or as part of a syndromic condition. Brain overgrowth has been associated with neurodevelopmental disorders such as autism during the first year of life, however, evidence remains inconclusive. Furthermore, most of the studies have involved pathological or high-risk populations, but little is known about the effects of brain overgrowth on neurodevelopment in otherwise neurotypical infants. We investigated the impact of brain overgrowth on basic perceptual learning processes (repetition effects and change detection response) during the first year of life. We recorded high density electroencephalograms (EEG) in 116 full-term healthy infants aged between 3 and 11 months, 35 macrocephalic (14 girls) and 81 normocephalic (39 girls) classified according to the WHO head circumference norms. We used an adapted oddball paradigm, time-frequency analyses, and auditory event-related brain potentials (ERPs) to investigate differences between groups. We show that brain overgrowth has a significant impact on repetition effects and change detection response in the 10–20 Hz frequency band, and in N450 latency, suggesting that these correlates of sensorial learning processes are sensitive to brain overgrowth during the first year of life.
Collapse
Affiliation(s)
- Gabriela López-Arango
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Department of Neurosciences, Montreal University, Montreal, QC, Canada
- *Correspondence: Gabriela López-Arango,
| | - Florence Deguire
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Department of Psychology, Montreal University, Montreal, QC, Canada
| | - Kristian Agbogba
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Polytechnique Montreal, Montreal, QC, Canada
| | | | - Inga S. Knoth
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
| | - Ramy El-Jalbout
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Department of Medical Imaging, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
| | - Valérie Côté
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
| | - Amélie Damphousse
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Department of Medical Imaging, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
| | | | - Sarah Lippé
- Research Center, Sainte-Justine Hospital, Montreal University, Montreal, QC, Canada
- Department of Psychology, Montreal University, Montreal, QC, Canada
- Sarah Lippé,
| |
Collapse
|
5
|
Martin M, Sciolla B, Sdika M, Quétin P, Delachartre P. Automatic segmentation and location learning of neonatal cerebral ventricles in 3D ultrasound data combining CNN and CPPN. Comput Biol Med 2021; 131:104268. [PMID: 33639351 DOI: 10.1016/j.compbiomed.2021.104268] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/16/2020] [Revised: 02/09/2021] [Accepted: 02/09/2021] [Indexed: 10/22/2022]
Abstract
Preterm neonates are highly likely to suffer from ventriculomegaly, a dilation of the Cerebral Ventricular System (CVS). This condition can develop into life-threatening hydrocephalus and is correlated with future neuro-developmental impairments. Consequently, it must be detected and monitored by physicians. In clinical routing, manual 2D measurements are performed on 2D ultrasound (US) images to estimate the CVS volume but this practice is imprecise due to the unavailability of 3D information. A way to tackle this problem would be to develop automatic CVS segmentation algorithms for 3D US data. In this paper, we investigate the potential of 2D and 3D Convolutional Neural Networks (CNN) to solve this complex task and propose to use Compositional Pattern Producing Network (CPPN) to enable Fully Convolutional Networks (FCN) to learn CVS location. Our database was composed of 25 3D US volumes collected on 21 preterm nenonates at the age of 35.8±1.6 gestational weeks. We found that the CPPN enables to encode CVS location, which increases the accuracy of the CNNs when they have few layers. Accuracy of the 2D and 3D FCNs reached intraobserver variability (IOV) in the case of dilated ventricles with Dice of 0.893±0.008 and 0.886±0.004 respectively (IOV = 0.898±0.008) and with volume errors of 0.45±0.42 cm3 and 0.36±0.24 cm3 respectively (IOV = 0.41±0.05 cm3). 3D FCNs were more accurate than 2D FCNs in the case of normal ventricles with Dice of 0.797±0.041 against 0.776±0.038 (IOV = 0.816±0.009) and volume errors of 0.35±0.29 cm3 against 0.35±0.24 cm3 (IOV = 0.2±0.11 cm3). The best segmentation time of volumes of size 320×320×320 was obtained by a 2D FCN in 3.5±0.2 s.
Collapse
Affiliation(s)
- Matthieu Martin
- Univ Lyon, INSA-Lyon, Université Claude Bernard Lyon 1, UJM-Saint Etienne, CNRS, Inserm, CREATIS UMR 5220, U1206, F-69621, LYON, France.
| | - Bruno Sciolla
- Univ Lyon, INSA-Lyon, Université Claude Bernard Lyon 1, UJM-Saint Etienne, CNRS, Inserm, CREATIS UMR 5220, U1206, F-69621, LYON, France
| | - Michaël Sdika
- Univ Lyon, INSA-Lyon, Université Claude Bernard Lyon 1, UJM-Saint Etienne, CNRS, Inserm, CREATIS UMR 5220, U1206, F-69621, LYON, France
| | | | - Philippe Delachartre
- Univ Lyon, INSA-Lyon, Université Claude Bernard Lyon 1, UJM-Saint Etienne, CNRS, Inserm, CREATIS UMR 5220, U1206, F-69621, LYON, France
| |
Collapse
|
6
|
Vermeulen MJ, Burkhardt W, Fritze A, Roelants J, Mense L, Willemsen S, Rüdiger M. Reference Charts for Neonatal Cranial Volume Based on 3D Laser Scanning to Monitor Head Growth. Front Pediatr 2021; 9:654112. [PMID: 34123964 PMCID: PMC8192695 DOI: 10.3389/fped.2021.654112] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 01/15/2021] [Accepted: 05/07/2021] [Indexed: 11/23/2022] Open
Abstract
Background: Postnatal brain growth is an important predictor of neurodevelopmental outcome in preterm infants. A new reliable proxy for brain volume is cranial volume, which can be measured routinely by 3-D laser scanning. The aim of this study was to develop reference charts for normal cranial volume in newborn infants at different gestational ages starting from late preterm for both sexes. Methods: Cross-sectional cohort study in a German university hospital, including singleton, clinically stable, neonates born after 34 weeks of gestation. Cranial volume was measured in the first week of life by a validated 3-D laser scanner. Cranial volume data was modeled to calculate percentile values by gestational age and birth weight and to develop cranial volume reference charts for girls and boys separately. Results: Of the 1,703 included infants, 846 (50%) were female. Birth weights ranged from 1,370 to 4,830 grams (median 3,370). Median cranial volume ranged from 320 [interquartile range (IQR) 294-347] ml at 34 weeks to 469 [IQR 442-496] ml at 42 weeks and was higher in boys than in girls. Conclusions: This study presents the first reference charts of cranial volume which can be used in clinical practice to monitor brain growth between 34 and 42 weeks gestation in infants.
Collapse
Affiliation(s)
- Marijn Jorien Vermeulen
- Division of Neonatology, Department of Pediatrics, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Wolfram Burkhardt
- Division of Neonatology and Pediatric Intensive Care Medicine, Department of Pediatrics, Medizinische Fakultät der Technischen Universität, Dresden, Germany
| | - Anne Fritze
- Division of Neonatology and Pediatric Intensive Care Medicine, Department of Pediatrics, Medizinische Fakultät der Technischen Universität, Dresden, Germany
| | - Jorine Roelants
- Division of Neonatology, Department of Pediatrics, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Lars Mense
- Division of Neonatology and Pediatric Intensive Care Medicine, Department of Pediatrics, Medizinische Fakultät der Technischen Universität, Dresden, Germany
| | - Sten Willemsen
- Division of Biostatistics, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Mario Rüdiger
- Division of Neonatology and Pediatric Intensive Care Medicine, Department of Pediatrics, Medizinische Fakultät der Technischen Universität, Dresden, Germany.,Saxonian Center for Feto-Neonatal Health, Technische Universität Dresden, Dresden, Germany
| |
Collapse
|