No abstract

The present study aimed to evaluate the value of serum amyloid A (SAA) in coronavirus disease 2019 (COVID-19) and compared the efficacy of SAA and C-reactive protein (CRP) in predicting the severity and recovery of COVID-19. A retrospective study was conducted on COVID-19 patients hospitalized in Wuhan No. 1 Hospital (Hubei, China) from January 21, 2020 to March 4, 2020. A two-way ANOVA analysis was used to compare the serum CRP and SAA levels between mild group and severe group during hospitalization days. Linear regression was used to analyze the relationship between the serum CRP, SAA levels and treatment days in recovered patients. The Logistic regression analysis and the area under curve (AUC) were calculated to determine the probability for predicting the severity and recovery of COVID-19. The severe group displayed higher CRP and SAA levels compared with the mild group during hospitalization (P<0.001). Logistic regression indicated that SAA and CRP were independent risk factors for the severity of COVID-19. The corresponding AUC of CRP and SAA values for severity of COVID-19 were 0.804 and 0.818, respectively. Linear regression analysis revealed that CRP and SAA levels were negatively correlated with treatment days in recovered patients (r=-0.761,-0.795, respectively). Logistic regression demonstrated that SAA was an independent factor for predicting the recovery of COVID-19. However, CRP could not predict the recovery of COVID-19. The corresponding AUC of SAA for the recovery of COVID-19 was 0.923. The results of the present study indicated that SAA can be considered to be a biomarker for predicting the severity and recovery of COVID-19.

Objectives: This study was designed to investigate the feasibility of apparent diffusion coefficient (ADC) values in evaluating normal fetal brain development from gestational week 24 up to term age. Methods: Diffusion-weighted imaging (DWI) was performed on 40 normal fetuses (with normal results on sonography and normal fetal MRI results), with two b-values of 0 and 600 s/mm² in the three (x, y, z) orthogonal axes. Ten regions of interest (ROIs) were manually placed symmetrically in the bilateral frontal white matter (FWM), occipital white matter (OWM), thalamus (THAL), basal ganglia (BG), and cerebellar hemispheres (CH). ADC values of the ten ROIs in all subjects were measured by two radiologists independently. One-way ANOVA was used to calculate the differences among the five regions in the fetal brain and linear regression analysis was used to evaluate the correlation between ADC values and gestational age (GA). p < 0.05 was considered significantly different. Results: Mean GA was 31.3 ± 3.9 (range 24-41) weeks. The overall mean ADC values (×10^-6 mm²/s) of the fetuses were 1,800 ± 214 (FWM), 1,400 ± 100 (BG), 1,300 ± 126 (THAL), 1,700 ± 133 (OWM) and 1,400 ± 155 (CH), respectively. The ADC value of BG was not significantly different from those of THAL and CH, while the other four ROIs had significant differences with each other. The ADC values of BG, THAL, OWM and CH had strong negative correlations with increasing GA (R were -0.568, -0.716, -0.830 and -0.700, respectively, all p < 0.01), OWM declined fastest with GA, followed by CH and THAL, the slowest being BG. The ADC value of FWM had no significant change with GA (p = 0.366). Conclusions: The measurement of ADC values is feasible to evaluate fetal brain development with high reliability and reproducibility.

A comprehensive interpretation of remote sensing images involves not only remote sensing object recognition but also the recognition of spatial relations between objects. Especially in the case of different objects with the same spectrum, the spatial relationship can help interpret remote sensing objects more accurately. Compared with traditional remote sensing object recognition methods, deep learning has the advantages of high accuracy and strong generalizability regarding scene classification and semantic segmentation. However, it is difficult to simultaneously recognize remote sensing objects and their spatial relationship from end-to-end only relying on present deep learning networks. To address this problem, we propose a multi-scale remote sensing image interpretation network, called the MSRIN. The architecture of the MSRIN is a parallel deep neural network based on a fully convolutional network (FCN), a U-Net, and a long short-term memory network (LSTM). The MSRIN recognizes remote sensing objects and their spatial relationship through three processes. First, the MSRIN defines a multi-scale remote sensing image caption strategy and simultaneously segments the same image using the FCN and U-Net on different spatial scales so that a two-scale hierarchy is formed. The output of the FCN and U-Net are masked to obtain the location and boundaries of remote sensing objects. Second, using an attention-based LSTM, the remote sensing image captions include the remote sensing objects (nouns) and their spatial relationships described with natural language. Finally, we designed a remote sensing object recognition and correction mechanism to build the relationship between nouns in captions and object mask graphs using an attention weight matrix to transfer the spatial relationship from captions to objects mask graphs. In other words, the MSRIN simultaneously realizes the semantic segmentation of the remote sensing objects and their spatial relationship identification end-to-end. Experimental results demonstrated that the matching rate between samples and the mask graph increased by 67.37 percentage points, and the matching rate between nouns and the mask graph increased by 41.78 percentage points compared to before correction. The proposed MSRIN has achieved remarkable results. the existing studies do not address the interpretation of spatial relationships between remote sensing objects, which limits the understanding of remote sensing objects, especially when the phenomenon of different objects with the same spectrum in remote sensing appears.The phenomenon of different objects with the same spectrum in remote sensing is quite common. It is difficult to identify objects only by their own textures, spectra, and shape information. Object identification requires multi-scale semantic information and spatially adjacent objects to assist in decision-making. The spatial relationship between remote sensing objects is of great significance to the recognition of remote sensing objects when different objects have the same spectru...