1. Automatic segmentation of ventricular volume by 3D ultrasonography in post haemorrhagic ventricular dilatation among preterm infants
- Author
-
Isabel Benavente-Fernández, J. Pizarro, Simon P. Lubián López, Lionel C. Gontard, Borja Sanz-Peña, [Gontard,LC] Department of Condensed Matter Physics, University of Cádiz, Puerto Real, Spain. [Gontard,LC, Pizarro,J] Department of Computer Engineering, University of Cádiz, Puerto Real, Spain. [Sanz-Peña,B] Department of Neurosurgery, Puerta del Mar Hospital, Cádiz, Spain. [Lubián López,SP, Benavente-Fernández,I] Department of Pediatrics (Neonatology), Biomedical Research and Innovation Institute of Cádiz (INiBICA) Research Unit, Puerta del Mar Hospital, Cádiz, Spain. [Lubián López,SP, Benavente-Fernández,I] Biomedical Research and Innovation Institute of Cádiz (INiBICA) Research Unit, Puerta del Mar University, Cádiz, Spain. [Lubián López,SP, Benavente-Fernández,I] Foundation for the Development of Neonatal Neurology (Nene), Madrid, Spain. [Benavente-Fernández,I] Department of Maternal and Child Health and Radiology, School of Medicine, University of Cádiz, Cádiz, Spain., This work was supported by the 2017 (PI0052/2017) and 2019 (ITI-0019-2019) ITI-Cadiz integrated territorial initiative for biomedical research European Regional Development Fund (ERDF) 2014–2020. Andalusian Ministry of Health and Families, Spain., Física de la Materia Condensada, Ingeniería Informática, and Materno-Infantil y Radiología
- Subjects
Edad gestacional ,Convolutional neural network ,Anatomy::Body Regions::Breast [Medical Subject Headings] ,Ultrasonografía ,Organisms::Eukaryota::Animals::Chordata::Vertebrates::Mammals::Primates::Haplorhini::Catarrhini::Hominidae::Humans [Medical Subject Headings] ,Recien nacido prematuro ,Engineering ,0302 clinical medicine ,Software Design ,Image Processing, Computer-Assisted ,Medicine ,Breast ,Peso al nacer ,Ultrasonography ,Multidisciplinary ,Analytical, Diagnostic and Therapeutic Techniques and Equipment::Diagnosis::Diagnostic Techniques and Procedures::Diagnostic Imaging::Imaging, Three-Dimensional [Medical Subject Headings] ,Premature infants ,Analytical, Diagnostic and Therapeutic Techniques and Equipment::Diagnosis::Diagnostic Techniques and Procedures::Physical Examination::Body Constitution::Body Weights and Measures::Organ Size [Medical Subject Headings] ,Aprendizaje profundo ,Ventricular dilatation ,Analytical, Diagnostic and Therapeutic Techniques and Equipment::Investigative Techniques::Epidemiologic Methods::Statistics as Topic::Sensitivity and Specificity [Medical Subject Headings] ,Ultrasound ,Information Science::Information Science::Computing Methodologies::Software::Software Design [Medical Subject Headings] ,Gestational age ,Organ Size ,3d ultrasonography ,Anatomy::Cardiovascular System::Heart::Heart Ventricles [Medical Subject Headings] ,Persons::Persons::Age Groups::Infant::Infant, Newborn::Infant, Premature [Medical Subject Headings] ,Neurology ,Diseases::Pathological Conditions, Signs and Symptoms::Pathological Conditions, Anatomical::Dilatation, Pathologic [Medical Subject Headings] ,Cardiology ,Convolutional neural networks ,Infant, Premature ,Dilatation, Pathologic ,medicine.medical_specialty ,Heart Ventricles ,Science ,Red nerviosa ,Sensitivity and Specificity ,Article ,03 medical and health sciences ,Deep Learning ,Imaging, Three-Dimensional ,Analytical, Diagnostic and Therapeutic Techniques and Equipment::Diagnosis::Diagnostic Techniques and Procedures::Diagnostic Imaging::Ultrasonography [Medical Subject Headings] ,Birth weight ,030225 pediatrics ,Internal medicine ,Humans ,Information Science::Information Science::Computing Methodologies::Image Processing, Computer-Assisted [Medical Subject Headings] ,business.industry ,Deep learning ,Infant, Newborn ,Stroke Volume ,Hypertrophy ,Persons::Persons::Age Groups::Infant::Infant, Newborn [Medical Subject Headings] ,Diseases::Pathological Conditions, Signs and Symptoms::Pathological Conditions, Anatomical::Hypertrophy [Medical Subject Headings] ,Phenomena and Processes::Mathematical Concepts::Neural Networks (Computer) [Medical Subject Headings] ,Diseases of the nervous system ,Automatic segmentation ,Ventricular volume ,Neural Networks, Computer ,Artificial intelligence ,Volumen sistólico ,business ,Phenomena and Processes::Circulatory and Respiratory Physiological Phenomena::Cardiovascular Physiological Phenomena::Hemodynamics::Cardiac Output::Stroke Volume [Medical Subject Headings] ,Phenomena and Processes::Physiological Phenomena::Physiological Processes::Growth and Development::Morphogenesis::Embryonic and Fetal Development::Fetal Development::Gestational Age [Medical Subject Headings] ,030217 neurology & neurosurgery ,Analytical, Diagnostic and Therapeutic Techniques and Equipment::Diagnosis::Diagnostic Techniques and Procedures::Physical Examination::Body Constitution::Body Weights and Measures::Body Size::Body Weight::Birth Weight [Medical Subject Headings] - Abstract
To train, evaluate, and validate the application of a deep learning framework in three-dimensional ultrasound (3D US) for the automatic segmentation of ventricular volume in preterm infants with post haemorrhagic ventricular dilatation (PHVD). We trained a 2D convolutional neural network (CNN) for automatic segmentation ventricular volume from 3D US of preterm infants with PHVD. The method was validated with the Dice similarity coefficient (DSC) and the intra-class coefficient (ICC) compared to manual segmentation. The mean birth weight of the included patients was 1233.1 g (SD 309.4) and mean gestational age was 28.1 weeks (SD 1.6). A total of 152 serial 3D US from 10 preterm infants with PHVD were analysed. 230 ventricles were manually segmented. Of these, 108 were used for training a 2D CNN and 122 for validating the methodology for automatic segmentation. The global agreement for manual versus automated measures in the validation data (n = 122) was excellent with an ICC of 0.944 (0.874–0.971). The Dice similarity coefficient was 0.8 (± 0.01). 3D US based ventricular volume estimation through an automatic segmentation software developed through deep learning improves the accuracy and reduces the processing time needed for manual segmentation using VOCAL. 3D US should be considered a promising tool to help deepen our current understanding of the complex evolution of PHVD.
- Published
- 2021