Prognostic prediction of sepsis patient using transformer with skip connected token for tabular data

Sepsis is known as a common syndrome in intensive care units (ICU), and severe sepsis and septic shock are among the leading causes of death worldwide. The purpose of this study is to develop a deep learning model that supports clinicians in efficiently managing sepsis patients in the ICU by predicting mortality, ICU length of stay (>14 days), and hospital length of stay (>30 days). The proposed model was developed using 591 retrospective data with 16 tabular data related to a sequential organ failure assessment (SOFA) score. To analyze tabular data, we designed the modified architecture of the transformer that has achieved extraordinary success in the field of languages and computer vision tasks in recent years. The main idea of the proposed model is to use a skip-connected token, which combines both local (feature-wise token) and global (classification token) information as the output of a transformer encoder. The proposed model was compared with four machine learning models (ElasticNet, Extreme Gradient Boosting [XGBoost]), and Random Forest) and three deep learning models (Multi-Layer Perceptron [MLP], transformer, and Feature-Tokenizer transformer [FT-Transformer]) and achieved the best performance (mortality, area under the receiver operating characteristic (AUROC) 0.8047; ICU length of stay, AUROC 0.8314; hospital length of stay, AUROC 0.7342). We anticipate that the proposed model architecture will provide a promising approach to predict the various clinical endpoints using tabular data such as electronic health and medical records.

Conditional Convolution Projecting Latent Vectors on Condition-Specific Space

Despite rapid advancements over the past several years, the conditional generative adversarial networks (cGANs) are still far from being perfect. Although one of the major concerns of the cGANs is how to provide the conditional information to the generator, there are not only no ways considered as the optimal solution but also a lack of related research. This brief presents a novel convolution layer, called the conditional convolution (cConv) layer, which incorporates the conditional information into the generator of the generative adversarial networks (GANs). Unlike the most general framework of the cGANs using the conditional batch normalization (cBN) that transforms the normalized feature maps after convolution, the proposed method directly produces conditional features by adjusting the convolutional kernels depending on the conditions. More specifically, in each cConv layer, the weights are conditioned in a simple but effective way through filter-wise scaling and channel-wise shifting operations. In contrast to the conventional methods, the proposed method with a single generator can effectively handle condition-specific characteristics. The experimental results on CIFAR, LSUN, and ImageNet datasets show that the generator with the proposed cConv layer achieves a higher quality of conditional image generation than that with the standard convolution layer.

Simple Yet Effective Way for Improving the Performance of GAN

In adversarial learning, the discriminator often fails to guide the generator successfully since it distinguishes between real and generated images using silly or nonrobust features. To alleviate this problem, this brief presents a simple but effective way that improves the performance of the generative adversarial network (GAN) without imposing the training overhead or modifying the network architectures of existing methods. The proposed method employs a novel cascading rejection (CR) module for discriminator, which extracts multiple nonoverlapped features in an iterative manner using the vector rejection operation. Since the extracted diverse features prevent the discriminator from concentrating on nonmeaningful features, the discriminator can guide the generator effectively to produce images that are more similar to the real images. In addition, since the proposed CR module requires only a few simple vector operations, it can be readily applied to existing frameworks with marginal training overheads. Quantitative evaluations on various data sets, including CIFAR-10, CelebA, CelebA-HQ, LSUN, and tiny-ImageNet, confirm that the proposed method significantly improves the performance of GAN and conditional GAN in terms of the Frechet inception distance (FID), indicating the diversity and visual appearance of the generated images.

PEPSI++: Fast and Lightweight Network for Image Inpainting

Among the various generative adversarial network (GAN)-based image inpainting methods, a coarse-to-fine network with a contextual attention module (CAM) has shown remarkable performance. However, due to two stacked generative networks, the coarse-to-fine network needs numerous computational resources, such as convolution operations and network parameters, which result in low speed. To address this problem, we propose a novel network architecture called parallel extended-decoder path for semantic inpainting (PEPSI) network, which aims at reducing the hardware costs and improving the inpainting performance. PEPSI consists of a single shared encoding network and parallel decoding networks called coarse and inpainting paths. The coarse path produces a preliminary inpainting result to train the encoding network for the prediction of features for the CAM. Simultaneously, the inpainting path generates higher inpainting quality using the refined features reconstructed via the CAM. In addition, we propose Diet-PEPSI that significantly reduces the network parameters while maintaining the performance. In Diet-PEPSI, to capture the global contextual information with low hardware costs, we propose novel rate-adaptive dilated convolutional layers that employ the common weights but produce dynamic features depending on the given dilation rates. Extensive experiments comparing the performance with state-of-the-art image inpainting methods demonstrate that both PEPSI and Diet-PEPSI improve the qualitative scores, i.e., the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM), as well as significantly reduce hardware costs, such as computational time and the number of network parameters.

Unsupervised Deep Contrast Enhancement with Power Constraint for OLED Displays

Various power-constrained contrast enhance-ment (PCCE) techniques have been applied to an organic light emitting diode (OLED) display for reducing the pow-er demands of the display while preserving the image qual-ity. In this paper, we propose a new deep learning-based PCCE scheme that constrains the power consumption of the OLED displays while enhancing the contrast of the displayed image. In the proposed method, the power con-sumption is constrained by simply reducing the brightness a certain ratio, whereas the perceived visual quality is pre-served as much as possible by enhancing the contrast of the image using a convolutional neural network (CNN). Furthermore, our CNN can learn the PCCE technique without a reference image by unsupervised learning. Ex-perimental results show that the proposed method is supe-rior to conventional ones in terms of image quality assess-ment metrics such as a visual saliency-induced index (VSI) and a measure of enhancement (EME).1.

Relationship between gamma-glutamyltransferase and metabolic syndrome in a Korean population

AIMS

To investigate associations between gamma-glutamyltransferase (GGT) and components of metabolic syndrome (MS), insulin resistance and inflammatory markers in the Korean population.

METHODS

The 3508 subjects enrolled in this survey participated in the Korean Rural Genomic Cohort (KRGC) study. Written consent was obtained from the local ethical committee. Of these participants, 1437 were men (mean age 56.9 +/- 7.9 years) and 2071 were women (mean age 55.8 +/- 8.1 years). We measured GGT levels and various biochemical markers. To examine insulin resistance status, we used the homeostasis assessment method for insulin resistance (HOMA-IR). For inflammatory marker, we used C-reactive protein (CRP) levels.

RESULTS

Serum GGT levels were significantly higher in the MS group compared to the healthy patient group [23 (5-1403) vs. 19 (5-1920) IU/l; P = 0.01]. The prevalence of MS and adjusted relative risk were both significantly increased from the lowest to highest GGT quartiles; these results persisted after adjustments for multiple confounders. Positive correlations were established between GGT and HOMA-IR or CRP.

CONCLUSION

These results suggest that GGT levels may be a surrogate marker of insulin resistance, inflammation and MS.

Analysis and screening of combinatorial libraries using mass spectrometry

Mass spectrometry is a highly selective and high throughput analytical technique that is ideally suited for the identification and purity determination of large numbers of compounds prepared using combinatorial chemistry or for the dereplication of natural products. Compounds may be characterized based on molecular weight, elemental composition and structural features based on fragmentation patterns. When coupled to a separation technique such as high-performance liquid chromatography (HPLC) or capillary electrophoresis, mass spectrometric applications may be expanded to include analysis of complex mixtures. However, the slower speed of the separation step reduces the throughput of the analysis. This review concerns the application of mass spectrometry to the characterization of combinatorial libraries and the screening of library and natural product mixtures. Strategies to enhance the throughput of LC-MS are discussed including fast HPLC and parallel LC-MS. Also, mass spectrometry-based screening methods are described including frontal affinity chromatography-mass spectrometry, gel permeation chromatography LC-MS, direct electrospray mass spectrometry of receptor-ligand complexes, affinity chromatography-mass spectrometry, and pulsed ultrafiltration mass spectrometry.

Five cases of microphthalmia with other ocular malformations

We report five cases of complex microphthalmia with other ocular malformations in infants or children, which were evaluated to investigate the relationship between the corneal diameters and total axial length. The size of the globe was measured by using computerized tomographic scans (CT scan), A-scan ultrasonography, or magnetic resonance imaging (MRI). There is a limited range of well-described malformation, including anterior or posterior segment dysgenesis or combined pathology such as corneal opacity, small cornea, iris hypoplasia, lens dislocation, cataract, chorioretinal coloboma, persistent hyperplastic primary vitreous (PHPV), retinal dysplasia, and intraocular tumor. Corneal diameters were correlated significantly with total axial length (r2 = 0.88) and decreased linearly as the total axial length decreased in these cases. However, there was no relationship seen between the total axial length and posterior segment length (r2 = -0.06). The results of this study may aid the clinical ophthalmologist to accurately understand or assess microphthamia combined with other ocular malformations.

Analysis of 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors using liquid chromatography-electrospray mass spectrometry

Employing high-performance liquid chromatography-electrospray mass spectrometry, we describe a new assay for monitoring 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase activity. Incubations were carried out with HMG-CoA reductase (rat liver), HMG-CoA and NADPH, and terminated by the addition of HCl. The reaction product, mevalonolactone, and internal standard, were extracted with ethyl acetate, dissolved in methanol, and analyzed by LC-MS. Using an isocratic mobile phase of 10% acetonitrile and 0.1% formic acid (flow-rate, 0.2 ml/min), the protonated molecules of mevalonolactone at m/z 131 and internal standard, beta,beta-dimethyl-gamma-(hydroxymethyl)-gamma-butyrolactone, at m/z 145, were detected using selected ion monitoring. The limit of detection was approximately 6.5 pg, and the limit of quantitation was approximately 16.3 pg. Extraction recovery was >90%. The relative standard deviations for intra- and inter-day assays were approximately 4.1+/-2.7 and 9.4+/-3.4%, respectively. Mevalonolactone was examined over a period of 3 days and found to be stable. Using this assay, lovastatin and mevastatin inhibited HMG-CoA reductase activity with IC50 values 0.24+/-0.02 and 2.16+/-0.31 microM, respectively. These methods offer some advantages over those reported previously which employ radiolabeled substrate and products, and should be useful in searching for compounds that could lower serum cholesterol or alter cell growth and differentiation.

Structural and functional consequences of inactivation of human glutathione S-transferase P1-1 mediated by the catechol metabolite of equine estrogens, 4-hydroxyequilenin

The inactivation mechanism(s) of human glutathione S-transferase P1-1 (hGST P1-1) by the catechol metabolite of Premarin estrogens, 4-hydroxyequilenin (4-OHEN), was (were) studied by means of site-directed mutagenesis, electrospray ionization mass spectrometric analysis, titration of free thiol groups, kinetic studies of irreversible inhibition, and analysis of band patterns on nonreducing sodium dodecyl sulfate--polyacrylamide gel electrophoresis (SDS-PAGE). The four cysteines (Cys 14, Cys 47, Cys 101, and Cys 169 in the primary sequence) in hGST P1-1 are susceptible to electrophilic attack and/or oxidative damage leading to loss of enzymatic activity. To investigate the role of cysteine residues in the 4-OHEN-mediated inactivation of this enzyme, one or a combination of cysteine residues was replaced by alanine residues (C47A, C101A, C47A/C101A, C14A/C47A/C101A, and C47A/C101A/C169A mutants). Mutation of Cys 47 decreased the affinity for the substrate GSH but not for the cosubstrate 1-chloro-2,4-dinitrobenzene (CDNB). However, the Cys 47 mutation did not significantly affect the rate of catalysis since V(max) values of the mutants were similar or higher compared to that of wild type. Electrospray ionization mass spectrometric analyses of wild-type and mutant enzymes treated with 4-OHEN showed that a single molecule of 4-OHEN-o-quinone attached to the proteins, with the exception of the C14A/C47A/C101A mutant where no covalent adduct was detected. 4-OHEN also caused oxidative damage as demonstrated by the appearance of disulfide-bonded species on nonreducing SDS--PAGE and protection of 4-OHEN-mediated enzyme inhibition by free radical scavengers. The studies of thiol group titration and irreversible kinetic experiments indicated that the different cysteines have distinct reactivity for 4-OHEN; Cys 47 was the most reactive thiol group whereas Cys 169 was resistant to modification. These results demonstrate that hGST P1-1 is inactivated by 4-OHEN through two possible mechanisms: (1) covalent modification of cysteine residues and (2) oxidative damage leading to proteins inactivated by disulfide bond formation.

Pharmacokinetics of deguelin, a cancer chemopreventive agent in rats

PURPOSE

To study the pharmacokinetics of deguelin, a naturally occurring potential cancer chemopreventive agent, in rats.

METHODS

[3H]Deguelin was administered intravenously (i.v.) under anesthesia, and blood samples were collected over 24 h. [3H]Deguelin and metabolites were extracted from plasma with ethyl acetate, and quantified by HPLC. Data were analyzed with the WinNolin pharmacokinetic software package to determine pharmacokinetic parameters. A three-compartment first-order elimination model was used to fit the plasma concentration-time curve. In addition, deguelin concentrations in tissues after i.v. and intragastric (i.g.) administration were determined by HPLC, and excretion (feces and urine) was evaluated over a 5-day period after i.g. administration.

RESULTS

Deguelin exhibited a mean residence time (MRT) of 6.98 h and terminal half-life (t1/2(gamma)) of 9.26 h. The area under the curve (AUC) and total clearance (Cl) were 57.3 ng.h/ml and 4.37 l/h per kg, respectively, with an apparent volume of distribution (V) and volume of distribution at steady-state (Vss) of 3.421 l/kg and 30.46 l/kg, respectively. Following i.v. administration, the relative levels of tissue distribution were as follows: heart > fat > mammary gland > colon > liver > kidney > brain > lung. Following i.g. administration, the relative levels of tissue distribution were as follows: perirenal fat > heart > mammary gland > colon > kidney > liver > lung > brain > skin. Within 5 days of i.g. administration, about 58.1% of the [3H]deguelin was eliminated via the feces and 14.4% via the urine. Approximately 1.7% of unchanged deguelin was found in the feces, and 0.4% in the urine.

CONCLUSIONS

An initial pharmacokinetic investigation of deguelin showed that this rotenoid has a relatively long MRT and half-life in plasma in the rat. The compound distributed in the tissues and excreted as metabolites, mainly via the feces.

Cytotoxic polyacetylenes from the twigs of Ochanostachys amentacea

Bioassay-guided investigation of the twigs of Ochanostachys amentacea using LNCaP (hormone-dependent human prostate cancer) cells as a monitor led to the isolation of three alkynes, the known (S)-17-hydroxy-9,11,13,15-octadecatetraynoic acid (minquartynoic acid, 1) and two novel analogues, (S)-17,18-dihydroxy-9,11,13,15-octadecatetraynoic acid (2) and (S)-17-hydroxy-15E-octadecen-9,11,13-triynoic acid (3). Compounds 1-3 were tested against a panel of human tumor cell lines and found to be significantly cytotoxic.

A cytotoxic diterpene from Alomia myriadenia

An extract of the aerial parts from Alomia myriadenia Schultz-Bip. ex Baker (Asteraceae) showed significant cytotoxicity against a panel of human cancer cell lines in a screening of extracts from Brazilian Atlantic Forest plant species. Employing a bioassay-linked HPLC-electrospray/MS method, followed by semi-preparative HPLC, the active component was isolated and characterized as a mixture of epimers of the labdane diterpene 12S,16-dihydroxy-ent-labda-7,13-dien-15,16-olide.

Cytotoxic constituents of the roots of Ekmanianthe longiflora

Bioactivity-directed fractionation of the CHCl(3) extract of the roots of Ekmanianthe longiflora resulted in the isolation of three new natural products, (2R,3R,4R)-3,4-dihydro-3, 4-dihydroxy-2-(3-methyl-2-butenyl)-1(2H)-naphthalenone (1), (2S,3R, 4R)-3,4-dihydro-3, 4-dihydroxy-2-(3-methyl-2-butenyl)-1(2H)-naphthalenone (2), and (2R, 3aR,9R,9aR)-9-hydroxy-2-(1-hydroxy-1-methylethyl)-2,3,3a,4,9 , 9a-hexahydro-naphtho[2,3-b]furan-4-one (3), together with the known compounds 2-(1-hydroxyethyl)naphtho[2,3-b]furan-4,9-quinone (4), 2-acetylnaphtho[2,3-b]furan-4,9-quinone (5), dehydro-iso-alpha-lapachone (6), alpha-lapachone (7), catalponol, and epi-catalponol. The structures of 1-3 were determined using a combination of NMR spectroscopic techniques, and the absolute configurations of compounds 1 and 2 were obtained using Mosher ester methodology. Compounds 4-6 showed significant cytotoxicity in a panel of human cancer cells. alpha-Lapachone (7) exhibited only marginal activity, and catalponol and epi-catalponol were inactive in this regard. When tested at 72 mg/kg/injection in an in vivo mouse P-388 leukemia system, compound 4 was inactive (110% T/C).

The visible man: three-dimensional interactive musculoskeletal anatomic atlas of the lower extremity

A personal computer-based interactive musculoskeletal anatomic atlas of the lower extremity has been created by using the Visible Human Male data set. A semiautomatic segmentation program was developed by using an intelligent scissors approach and shape-based interpolation, thus considerably reducing the laborious work of the segmentation and labeling process. Manual contour extractions at 3-mm section intervals and shape-based interpolations of intervening sections of the musculoskeletal structures of the lower extremity were performed. For interactive and realistic three-dimensional display, an efficient binary volume rendering method was developed that introduces the concept of shear-warp factorization and applies a newly developed normal calculation technique. Binary volume rendering reconstructs various structures from a series of two-dimensional sections in a few seconds, thus enabling real-time manipulations of the computerized atlas. All of the muscles, tendons, and bones of the lower extremity have been segmented and labeled. The volume-based three-dimensional interactive atlas supports various interactions including rotation, removal, highlighting with artificial colors, arbitrary cutting operation, transparent view, and descriptive knowledge representation. In addition, browsing through the two-dimensional images of transverse, coronal, and sagittal views with labeling and segmentation information is possible.

Determination of betaine in Lycium chinense fruits by liquid chromatography-electrospray ionization mass spectrometry

A rapid and sensitive high-performance liquid chromatography-electrospray mass spectrometric method has been developed for the determination of betaine in Lycium chinense fruits. Betaine was analyzed on a system consisting of a NH2 stationary phase and a mobile phase of water-acetonitrile (25:75) by isocratic elution for 40 min. Betaine was identified and quantitated by electrospray ionization mass spectrometry with selected ion monitoring of the protonated ion [Betaine+H]+ and clustered ions [nBetaines+H]+. The limit of detection for betaine by this method was ca. 0.2 ng/ml and the relative standard deviations of the assay (intra- and inter-day) were less than 8.1%.

Determination of betulinic acid in mouse blood, tumor and tissue homogenates by liquid chromatography-electrospray mass spectrometry

A rapid and sensitive high-performance liquid chromatography-electrospray MS method has been developed to determine tissue distribution of betulinic acid in mice. The method involved deproteinization of these samples with 2.5 volumes (v/w) of acetonitrile-ethanol (1:1) and then 5 microl aliquots of the supernatant were injected onto a C18 reversed-phase column coupled with an electrospray MS system. The mobile phase employed isocratic elution with 80% acetonitrile for 10 min; the flow-rate was 0.7 ml/min. The column effluent was analyzed by selected ion monitoring for the negative pseudo-molecular ion of betulinic acid [M-H]- at m/z 455. The limit of detection for betulinic acid in biological samples by this method was approximately 1.4 pg and the coefficients of variation of the assay (intra- and inter-day) were generally low (below 9.1%). When athymic mice bearing human melanoma were treated with betulinic acid (500 mg/kg, i.p.), distribution was as follows: tumor, 452.2 +/- 261.2 microg/g; liver, 233.9 +/- 80.3 microg/g; lung, 74.8 +/- 63.7 microg/g; kidney, 95.8 +/- 122.8 microg/g; blood, 1.8 +/- 0.5 microg/ml. No interference was noted due to endogenous substances. These methods of analysis should be of value in future studies related to the development and characterization of betulinic acid.

Collaborative workspace for multimedia medical conferencing

We propose an approach for collaborative workspace management in medical conferencing. A collaborative workspace is a virtual data space shared between medical experts for working out solutions collaboratively while conferencing. Our approach provides medical users with an integrated view of various kinds of multimedia patient data and a unified control over the workspace. For data navigation and conferencing, a tree-like navigation tool, which we named the patient record tree, is provided. And we classify patient data, which is the object of medical collaborative works, into six basic types, and provide a view template for displaying each of these types.

Inhibition of aromatase activity by flavonoids

In searching for potent cancer chemopreventive agents from synthetic or natural products, 28 randomly selected flavonoids were screened for inhibitory effects against partially purified aromatase prepared from human placenta. Over 50% of the flavonoids significantly inhibited aromatase activity, with greatest activity being demonstrated with apigenin (IC50: 0.9 microg/mL), chrysin (IC50: 1.1 microg/mL), and hesperetin (IC50: 1.0 microg/mL).

Determination of aloesin in plasma by high-performance liquid chromatography as fluorescent 9-anthroyl derivative

A sensitive high-performance liquid chromatographic (HPLC) method for the determination of aloesin in plasma was developed. After solid-phase extraction from plasma and derivatization of aloesin and compound AD-1, which was prepared from aloesin as a internal standard, with 9-anthroylnitrile in the presence of quinuclidine, the derivatives were separated on a inertsil ODS-3 column using acetonitrile/methanol/water (3:1:6) as a mobile phase, and detected fluorimetrically at 460 nm with excitation at 360 nm. The detection limit of aloesin was 3.2 ng/ml in plasma (S/N = 3).

The development of a new method to detect the adulteration of commercial aloe gel powders

Simple and accurate methods to detect the adulteration of commercial aloe gel powder were developed. Crude polysaccharide in aloe gel powder was isolated by precipitating with excess ethyl alcohol and total hexose in isolated polysaccharide was determined by Dubois assay. After hydrolysis of non-dialysable polysaccharides, resultant free sugar was determined by gas chromatography for sugar recognition and ash contents was considered simultaneously. In some products, the content of ash was very low while the content of total hexose was very high. And polysaccharides of these products revealed typical dextran pattern, therefore, these products could be identified that adulterated with commercial maltodextrin. The content of maltodextrin in adulterated product was determined by HPLC and TLC analysis which could be adopted as a part of a certification process.

Neoaloesin A: a new C-glucofuranosyl chromone from Aloe barbadensis

The first C-glucofuranosyl compound, named as neoaloesin A, was isolated from the leaves of Aloe barbadensis. Its structure was determined to be 8-alpha-D-glucofuranosyl-7-hydroxy-5-methyl-2-(2-oxopropyl)-4 H-1-benzopyran-4-one on the basis of chemical and spectral evidence.