ABSTRACT
Segmentation of the fetus from 2-dimensional (2D) magnetic resonance imaging (MRI) can aid radiologists with clinical decision making for disease diagnosis. Machine learning can facilitate this process of automatic segmentation, making diagnosis more accurate and user independent. We propose a deep learning (DL) framework for 2D fetal MRI segmentation using a Cross Attention Squeeze Excitation Network (CASE-Net) for research and clinical applications. CASE-Net is an end-to-end segmentation architecture with relevant modules that are evidence based. The goal of CASE-Net is to emphasize localization of contextual information that is relevant in biomedical segmentation, by combining attention mechanisms with squeeze-and-excitation (SE) blocks. This is a retrospective study with 34 patients. Our experiments have shown that our proposed CASE-Net achieved the highest segmentation Dice score of 87.36%, outperforming other competitive segmentation architectures.
Subject(s)
Image Processing, Computer-Assisted , Neural Networks, Computer , Fetus , Humans , Magnetic Resonance Imaging , Retrospective StudiesABSTRACT
OBJECTIVE: To provide a review and quantitative analysis of the available fetal MR imaging phantoms. MATERIALS AND METHODS: A literature search was conducted across Pubmed, Google Scholar, and Ryerson University Library databases to identify fetal MR imaging phantoms. Phantoms were graded on a semi-quantitative scale in regards to four evaluation categories: (1) anatomical accuracy in size and shape, (2) dielectric conductivity similar to the simulated tissue, (3) relaxation times similar to simulated tissue, and (4) physiological motion similar to fetal gross body, cardiovascular, and breathing motion. This was followed by statistical analysis to identify significant findings. RESULTS: Seventeen fetal phantoms were identified and had an average overall percentage accuracy of 26%, with anatomical accuracy being satisfied the most (56%) and physiological motion the least (7%). Phantoms constructed using 3D printing were significantly more accurate than conventionally constructed phantoms. DISCUSSION: Currently available fetal phantoms lack accuracy and motion simulation. 3D printing may lead to higher accuracy compared with traditional manufacturing. Future research needs to focus on properly simulating both fetal anatomy and physiological motion to produce a phantom that is appropriate for fetal MRI sequence development and optimization.
Subject(s)
Magnetic Resonance Imaging/methods , Phantoms, Imaging , Prenatal Diagnosis/methods , Computer Simulation , Equipment Design , Female , Heart , Humans , Imaging, Three-Dimensional , Lung , Motion , Printing, Three-Dimensional , Reproducibility of Results , Respiration , UterusABSTRACT
OBJECTIVE: Zika virus (ZIKV) infection is a vector-borne disease that can be transmitted sexually and vertically. The vertical transmission of the virus may lead to congenital Zika syndrome in infants. The aim of this study is to conduct a systematic review and meta-analysis of published reports documenting the prevalence of congenital Zika-related disorders in infants of mothers infected with ZIKV during pregnancy. METHODS: We conducted a comprehensive search in Ovid MEDLINE, Ovid MEDLINE (R) Epub ahead of print, Embase, Embase Classic and Web of Science databases to identify human studies reporting prevalence of congenital disorders in infants of ZIKV-infected mothers. RESULTS: We identified 25 reports selected for inclusion in the current study (n = 4683 subjects). The majority of the studies were from South American high-risk countries. Only one third of the identified studies were conducted in the United States. Clinical maternal symptoms included maculopapular rash (76.9%), arthralgia (46.4%), fever (45.5%) and headache (31.8%) with myalgia and conjunctivitis only presented in 25% of the cases. The most prevalent congenital disorder in the newborns was brain calcifications (42.6; 95% CI, 30.8-54.4), followed by ventriculomegaly (21.8; 95% CI, 15.2-28.4), joint abnormalities (13.2; 95% CI, 9.4-18.2), ocular abnormalities (4.2; 95% CI, 1.0-7.5) and microcephaly (3.9; 95% CI, 2.4-5.4). CONCLUSION: The current study highlights the high prevalence of a range of congenital disorders in newborns of mothers infected with ZIKV. It warrants developing studies to further clarify the mechanisms by which each of these disorders occurs in response to the viral infection during pregnancy and its vertical transmission to the infants.
