DBF-Net: a semi-supervised dual-task balanced fusion network for segmenting infected regions from lung CT images.

Accurate segmentation of infected regions in lung computed tomography (CT) images is essential to improve the timeliness and effectiveness of treatment for coronavirus disease 2019 (COVID-19). However, the main difficulties in developing of lung lesion segmentation in COVID-19 are still the fuzzy boundary of the lung-infected region, the low contrast between the infected region and the normal trend region, and the difficulty in obtaining labeled data. To this end, we propose a novel dual-task consistent network framework that uses multiple inputs to continuously learn and extract lung infection region features, which is used to generate reliable label images (pseudo-labels) and expand the dataset. Specifically, we periodically feed multiple sets of raw and data-enhanced images into two trunk branches of the network; the characteristics of the lung infection region are extracted by a lightweight double convolution (LDC) module and fusiform equilibrium fusion pyramid (FEFP) convolution in the backbone. According to the learned features, the infected regions are segmented, and pseudo-labels are made based on the semi-supervised learning strategy, which effectively alleviates the semi-supervised problem of unlabeled data. Our proposed semi-supervised dual-task balanced fusion network (DBF-Net) creates pseudo-labels on the COVID-SemiSeg dataset and the COVID-19 CT segmentation dataset. Furthermore, we perform lung infection segmentation on the DBF-Net model, with a segmentation sensitivity of 70.6% and specificity of 92.8%. The results of the investigation indicate that the proposed network greatly enhances the segmentation ability of COVID-19 infection.

DBF-Net: a semi-supervised dual-task balanced fusion network for segmenting infected regions from lung CT images

Accurate segmentation of infected regions in lung computed tomography (CT) images is essential to improve the timeliness and effectiveness of treatment for coronavirus disease 2019 (COVID-19). However, the main difficulties in developing of lung lesion segmentation in COVID-19 are still the fuzzy boundary of the lung-infected region, the low contrast between the infected region and the normal trend region, and the difficulty in obtaining labeled data. To this end, we propose a novel dual-task consistent network framework that uses multiple inputs to continuously learn and extract lung infection region features, which is used to generate reliable label images (pseudo-labels) and expand the dataset. Specifically, we periodically feed multiple sets of raw and data-enhanced images into two trunk branches of the network;the characteristics of the lung infection region are extracted by a lightweight double convolution (LDC) module and fusiform equilibrium fusion pyramid (FEFP) convolution in the backbone. According to the learned features, the infected regions are segmented, and pseudo-labels are made based on the semi-supervised learning strategy, which effectively alleviates the semi-supervised problem of unlabeled data. Our proposed semi-supervised dual-task balanced fusion network (DBF-Net) creates pseudo-labels on the COVID-SemiSeg dataset and the COVID-19 CT segmentation dataset. Furthermore, we perform lung infection segmentation on the DBF-Net model, with a segmentation sensitivity of 70.6% and specificity of 92.8%. The results of the investigation indicate that the proposed network greatly enhances the segmentation ability of COVID-19 infection.

Short-term developmental outcomes in neonates born to mothers with COVID-19 from Wuhan, China.

BACKGROUND: Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an emerging disease. The consequences of SARS-CoV-2 exposure in infants remain unknown. Therefore, this study aims to investigate whether neonates born to mothers with COVID-19 have adverse brain development. METHODS: This multicenter observational study was conducted at two designated maternal and children's hospitals in Hubei Province, mainland China from February 1, 2020 to May 15, 2020. Neonates born to mothers with COVID-19 were enrolled. Brain magnetic resonance imaging (MRI) findings, and volumes of grey and white matters, and physical growth parameters were observed at 44 weeks corrected gestational age. RESULTS: Of 72 neonates born to mothers with COVID-19, 8 (11%) were diagnosed with COVID-19, 8 (11%) were critically ill, and no deaths were reported. Among the eight neonates that underwent brain MRI at corrected gestational age of 44 weeks, five neonates were diagnosed with COVID-19. Among these five neonates, three presented abnormal MRI findings including abnormal signal in white matter and delayed myelination in newborn 2, delayed myelination and brain dysplasia in newborn 3, and abnormal signal in the bilateral periventricular in newborn 5. The other three neonates without COVID-19 presented no significantly changes of brain MRI findings and the volumes of grey matter and white matter compared to those of healthy newborns at the equivalent age (P > 0.05). Physical growth parameters for weight, length, and head circumference at gestational age of 44 weeks were all above the 3rd percentile for all neonates. CONCLUSIONS: Some of the neonates born to mothers with COVID-19 had abnormal brain MRI findings but these neonates did not appear to have poor physical growth. These findings may provide the information on the follow-up schedule on the neonates exposed to SARS-CoV-2, but further study is required to evaluate the association between the abnormal MRI findings and the exposure to SARS-CoV-2.

Effects of SARS-CoV-2 infection on neuroimaging and neurobehavior in neonates.

BACKGROUND: We collected neonatal neurological, clinical, and imaging data to study the neurological manifestations and imaging characteristics of neonates with coronavirus disease 2019 (COVID-19). METHODS: This case-control study included newborns diagnosed with COVID-19 in Wuhan, China from January 2020 to July 2020. All included newborns had complete neurological evaluations and head magnetic resonance imaging. We normalized the extracted T2-weighted imaging data to a standard neonate template space, and segmented them into gray matter, white matter, and cerebrospinal fluid. The comparison of gray matter volume was conducted between the two groups. RESULTS: A total of five neonates with COVID-19 were included in this study. The median reflex scores were 2 points lower in the infected group than in the control group (P = 0.0094), and the median orientation and behavior scores were 2.5 points lower in the infected group than in the control group (P = 0.0008). There were also significant differences between the two groups in the total scale score (P = 0.0426). The caudate nucleus, parahippocampal gyrus, and thalamus had the strongest correlations with the Hammersmith neonatal neurologic examination (HNNE) score, and the absolute correlation coefficients between the gray matter volumes and each part of the HNNE score were all almost greater than 0.5. CONCLUSIONS: We first compared the neurological performance of neonates with and without COVID-19 by quantitative neuroimaging and neurological examination methods. Considering the limited numbers of patients, more studies focusing on the structural or functional aspects of the virus in the central nervous system in different age groups will be carried out in the future.