Objectives
To propose a transfer learning (TL) radiomics model that efficiently combines the information from gray scale and elastogram ultrasound images for accurate liver fibrosis grading.
Methods
...Totally 466 patients undergoing partial hepatectomy were enrolled, including 401 with chronic hepatitis B and 65 without fibrosis pathologically. All patients received elastography and got liver stiffness measurement (LSM) 2–3 days before surgery. We proposed a deep convolutional neural network by TL to analyze images of gray scale modality (GM) and elastogram modality (EM). The TL process was used for liver fibrosis classification by Inception-V3 network which pretrained on ImageNet. The diagnostic performance of TL and non-TL was compared. The value of single modalities, including GM and EM alone, and multimodalities, including GM + LSM and GM + EM, was evaluated and compared with that of LSM and serological indexes. Receiver operating characteristic curve analysis was performed to calculate the optimal area under the curve (AUC) for classifying fibrosis of S4, ≥ S3, and ≥ S2.
Results
TL in GM and EM demonstrated higher diagnostic accuracy than non-TL, with significantly higher AUCs (all
p
< .01). Single-modal GM and EM both performed better than LSM and serum indexes (all
p
< .001). Multimodal GM + EM was the most accurate prediction model (AUCs are 0.950, 0.932, and 0.930 for classifying S4, ≥ S3, and ≥ S2, respectively) compared with GM + LSM, GM and EM alone, LSM, and biomarkers (all
p
< .05).
Conclusions
Liver fibrosis can be staged by a transfer learning modal based on the combination of gray scale and elastogram ultrasound images, with excellent performance.
Key Points
• Transfer learning consists in applying to a specific deep learning algorithm that pretrained on another relevant problem, expected to reduce the risk of overfitting due to insufficient medical images.
• Liver fibrosis can be staged by transfer learning radiomics with excellent performance.
• The most accurate prediction model of transfer learning by Inception-V3 network is the combination of gray scale and elastogram ultrasound images.
We present 870 m ALMA dust polarization observations of 10 young Class 0/I protostars in the Perseus Molecular Cloud. At ∼0 35 (80 au) resolution, all of our sources show some degree of polarization, ...with most (9/10) showing significantly extended emission in the polarized continuum. Each source has incredibly intricate polarization signatures. In particular, all three disk-candidates have polarization vectors roughly along the minor axis, which is indicative of polarization produced by dust scattering. On ∼100 au scales, the polarization is at a relatively low level ( 1%) and is quite ordered. In sources with significant envelope emission, the envelope is typically polarized at a much higher ( 5%) level and has a far more disordered morphology. We compute the cumulative probability distributions for both the small (disk-scale) and large (envelope-scale) polarization percentage. We find that the two are intrinsically different, even after accounting for the different detection thresholds in the high/low surface brightness regions. We perform Kolmogorov-Smirnov and Anderson-Darling tests on the distributions of angle offsets of the polarization from the outflow axis. We find disk-candidate sources are different from the non-disk-candidate sources. We conclude that the polarization on the 100 au scale is consistent with the signature of dust scattering for disk-candidates and that the polarization on the envelope-scale in all sources may come from another mechanism, most likely magnetically aligned grains.
ABSTRACT Diffuse striations in molecular clouds are preferentially aligned with local magnetic fields, whereas dense filaments tend to be perpendicular to them. When and why this transition occurs ...remain uncertain. To explore the physics behind this transition, we compute the histogram of relative orientation (HRO) between the density gradient and the magnetic field in three-dimensional magnetohydrodynamic (MHD) simulations of prestellar core formation in shock-compressed regions within giant molecular clouds. We find that, in the magnetically dominated (sub-Alfvénic) post-shock region, the gas structure is preferentially aligned with the local magnetic field. For overdense sub-regions with super-Alfvénic gas, their elongation becomes preferentially perpendicular to the local magnetic field. The transition occurs when self-gravitating gas gains enough kinetic energy from the gravitational acceleration to overcome the magnetic support against the cross-field contraction, which results in a power-law increase of the field strength with density. Similar results can be drawn from HROs in projected two-dimensional maps with integrated column densities and synthetic polarized dust emission. We quantitatively analyze our simulated polarization properties, and interpret the reduced polarization fraction at high column densities as the result of increased distortion of magnetic field directions in trans- or super-Alfvénic gas. Furthermore, we introduce measures of the inclination and tangledness of the magnetic field along the line of sight as the controlling factors of the polarization fraction. Observations of the polarization fraction and angle dispersion can therefore be utilized in studying local magnetic field morphology in star-forming regions.
Genotype imputation Li, Yun; Willer, Cristen; Sanna, Serena ...
Annual review of genomics and human genetics,
01/2009, Volume:
10, Issue:
1
Journal Article
Peer reviewed
Open access
Genotype imputation is now an essential tool in the analysis of genome-wide association scans. This technique allows geneticists to accurately evaluate the evidence for association at genetic markers ...that are not directly genotyped. Genotype imputation is particularly useful for combining results across studies that rely on different genotyping platforms but also increases the power of individual scans. Here, we review the history and theoretical underpinnings of the technique. To illustrate performance of the approach, we summarize results from several gene mapping studies. Finally, we preview the role of genotype imputation in an era when whole genome resequencing is becoming increasingly common.
The distant metastasis of cancer cells is a risk factor for tumor lethality and poor prognosis in non-small-cell lung carcinoma (NSCLC). Increased SOX9 expression has been associated with clinical ...stage and poor prognosis in NSCLC, but the molecular mechanisms by which SOX9 promotes metastasis in NSCLC are still unknown.
The relationship between SOX9 expression and T, N, M classification was assessed using the χ
test and Spearman's analysis in 142 immunohistochemically diagnosed specimens of NSCLC. We also generated SOX9-overexpression and SOX9-knockdown cells lines and their corresponding control cell lines by transfection with lentiviral constructs. In vivo assay, SOX9-overexpressing and SOX9-knockdown NSCLC cells were injected in zebrafish to examine distance metastasis. Gene set enrichment analysis (GSEA) was applied to analysis the correlation between SOX9 overexpression and Wnt/β-catenin pathway. Luciferase assay was used to check transcriptional activity of TCF/LEF and western blot and immunofluorescence was employed to detect β-catenin translocation in SOX9-overexpression, SOX9-knockdown and their corresponding control cell lines.
We found that SOX9 overexpression correlates with the T, N and M stage significantly (p = 0.03, 0.000, and 0.032 respectively) in 142 immunohistochemically diagnosed specimens of NSCLC. SOX9 overexpression was found to decrease the expression of the epithelial cell markers E-cadherin and γ-catenin and increase the expression of the mesenchymal cell markers N-cadherin and vimentin. An in vivo assay showed distant metastasis of the SOX9-overexpressing cells, which was not observed in the SOX9-knockdown cells. These findings indicate that SOX9 promotes distant metastasis by promoting EMT in NSCLC cells. GSEA showed that SOX9 overexpression was significantly correlated with the Wnt/β-catenin pathway which was corroborated by the expression of EMT-associated proteins in this pathway and its downstream target genes. SOX9 overexpression was also found to enhance the transcriptional activity of TCF/LEF, promote the nuclear translocation of β-catenin and increase the phosphorylation of GSK3β at Ser9. Further, inhibition of β-catenin suppressed the metastasis-promoting effects of SOX9 overexpression.
This study is the first to report that SOX9 is associated with clinical TNM stage and indicates that SOX9 promotes migration, invasion and the EMT process through the Wnt/β-catenin pathway.
Emission from protostars at centimeter radio wavelengths has been shown to trace the free-free emission arising from ionizing shocks as a result of jets and outflows driven by protostars. Therefore, ...measuring properties of protostars at radio frequencies can provide valuable insights into the nature of their outflows and jets. We present a C-band (4.1 and 6.4 cm) survey of all known protostars (Class 0 and Class I) in Perseus as part of the VLA Nascent Disk and Multiplicity (VANDAM) Survey. We examine the known correlations between radio flux density and protostellar parameters, such as bolometric luminosity and outflow force, for our sample. We also investigate the relationship between radio flux density and far-infrared line luminosities from Herschel. We show that free-free emission most likely originates from J-type shocks; however, the large scatter indicates that those two types of emission probe different time and spatial scales. Using C-band fluxes, we removed an estimation of free-free contamination from the corresponding Ka-band (9 mm) flux densities that primarily probe dust emission from embedded disks. We find that the compact (<1″) dust emission is lower for Class I sources (median dust mass 96 M⊕) relative to Class 0 (248 M⊕), but several times higher than in Class II (5-15 M⊕). If this compact dust emission is tracing primarily the embedded disk, as is likely for many sources, this result provides evidence of decreasing disk masses with protostellar evolution, with sufficient mass for forming giant planet cores primarily at early times.
Taking a relational perspective on the employment relationship, we examined processes (mediation and moderation) linking high-performance human resource practices and productivity and turnover, two ...indicators of organizational performance. Multilevel analysis of data from hotels in the People's Republic of China revealed that service-oriented organizational citizenship behavior (OCB) partially mediated the relationships between high-performance human resource practices and both performance indicators. Unemployment rate moderated the service-oriented OCB--turnover relationship, and business strategy (service quality) moderated the service-oriented OCB--productivity relationship.
Smart wearables are receiving increasing attention. Different forms of wearables have a wide range of power requirements, and lithium-ion batteries are now the most popular energy storage option. ...This paper discusses the trends and challenges of smart wearables. Future wearable technology will be more compact and integrated. For the specificity of future wearables, new energy harvesting technologies that are more practical, durable, and ecologically beneficial are needed. This paper summarizes the research on the human body and environment based energy harvesting techniques, including solar, thermal, Radio frequency (RF) energy, kinetic energy, and biomass energy. We deeply analyze the application potential and limitations of each new energy harvesting technology and summarize the relationship between the energy required by wearables and the new energy harvesting systems. Solar and kinetic energy generation technologies have higher energy density and are suitable for application in outdoor sports wearables. RF energy, thermal energy, and biomass energy have less energy dense and can be used as auxiliary power sources for small wearables. The combination of the energy harvesting system and the micro energy storage unit enables the continuous power supply of wearables in different circumstances of daytime, nighttime, indoor and outdoor. The significance of this work stems from providing guidance for future energy supply methods of wearables.
Display omitted
•A summary of the specifications and energy consumption of popular smart wearables.•Introduces the classification of smart wearables and predicts future trends.•Reviews the external and human energy harvesting methods applicable to wearables.•Combination schemes and applicability of energy harvesting methods are discussed.•Micro-supercapacitors are considered for energy storage of smart wearables.
•An anisotropic elastic–plastic model was developed for predicting rafting behavior.•Variation in mismatch degree has a strong effect on rafting in the elastic stage.•Variation in mismatch degree has ...very slight impact on rafting in the plastic stage.•Three elastic-constant differences have different influences on rafting behavior.
An anisotropic elastic–plastic model was developed to investigate the rafting behavior of Ni-based single crystal superalloys. This model was developed by introducing the plastic constitutive equation of face-centered cubic (FCC) single crystal in the framework of Eshelby's equivalent inclusion theory. The Hill's equivalent stress was calculated when applying a tensile or compressive loading along the 001 direction. The calculated results successfully predict the rafting direction for alloys with a positive or negative mismatch, consistent with pervious experimental and theoretical studies. Moreover, based on this model, the mismatch degree and the elastic-constant differences of the matrix and precipitate phases and their effects on the speed of rafting are carefully discussed. Regardless of a positive or negative mismatch alloy, the larger absolute value of mismatch degree can more effectively accelerate the process of rafting in Ni-based single crystal superalloys. However, when the alloys enter the plastic deformation stage, the variation in mismatch degree slightly affects the speed of rafting. For the elastic-constant differences (C11r′−C11r,C12r′−C12r,and C44r′−C44r), the smaller the value of C11r′−C11r, or the greater the value of C12r′−C12r, the more effective the acceleration of γ′ rafting; whereas the value of C44r′−C44r has no effect on the rafting of alloys. The research results provide a new theory and method for studying the rafting behavior and its influencing factors for Ni-based single crystal superalloys.
Fig. Material model of Ni-based single crystal superalloys and representative volume element (RVE) for micromechanical analysis. (a) Original material microstructure: Cubical γ′ precipitates are uniformly embedded in the γ matrix, and the volume fraction of γ′ phase is 70%. (b) Simplified material model. (c) RVE for micromechanical analysis. (d) Variations in Hill's equivalent stress with the external tensile stress. Display omitted
PDF
