Deep Neural Network Augments Performance of Junior Residents in Diagnosing COVID-19 Pneumonia on Chest Radiographs.

Chest X-rays (CXRs) are essential in the preliminary radiographic assessment of patients affected by COVID-19. Junior residents, as the first point-of-contact in the diagnostic process, are expected to interpret these CXRs accurately. We aimed to assess the effectiveness of a deep neural network in distinguishing COVID-19 from other types of pneumonia, and to determine its potential contribution to improving the diagnostic precision of less experienced residents. A total of 5051 CXRs were utilized to develop and assess an artificial intelligence (AI) model capable of performing three-class classification, namely non-pneumonia, non-COVID-19 pneumonia, and COVID-19 pneumonia. Additionally, an external dataset comprising 500 distinct CXRs was examined by three junior residents with differing levels of training. The CXRs were evaluated both with and without AI assistance. The AI model demonstrated impressive performance, with an Area under the ROC Curve (AUC) of 0.9518 on the internal test set and 0.8594 on the external test set, which improves the AUC score of the current state-of-the-art algorithms by 1.25% and 4.26%, respectively. When assisted by the AI model, the performance of the junior residents improved in a manner that was inversely proportional to their level of training. Among the three junior residents, two showed significant improvement with the assistance of AI. This research highlights the novel development of an AI model for three-class CXR classification and its potential to augment junior residents' diagnostic accuracy, with validation on external data to demonstrate real-world applicability. In practical use, the AI model effectively supported junior residents in interpreting CXRs, boosting their confidence in diagnosis. While the AI model improved junior residents' performance, a decline in performance was observed on the external test compared to the internal test set. This suggests a domain shift between the patient dataset and the external dataset, highlighting the need for future research on test-time training domain adaptation to address this issue.

Choline and trimethylamine N-oxide supplementation in normal chow diet and western diet promotes the development of atherosclerosis in Apoe -/- mice through different mechanisms.

Trimethylamine N-oxide (TMAO), a gut microbiota-dependent metabolite, has been shown to aggravate cardiovascular disease. However, the mechanisms of TMAO in the setting of cardiovascular disease progress remain unclear. Here, we aim to investigate the effects of TMAO on atherosclerosis (AS) development and the underlying mechanisms. Apoe -/- mice received choline or TMAO supplementation in a normal diet and a western diet for 12 weeks. Choline or TMAO supplementation in both normal diet and western diet significantly promoted plaque progression in Apoe-/- mice. Besides, serum lipids levels and inflammation response in the aortic root were enhanced by choline or TMAO supplementation. In particular, choline or TMAO supplementation in the western diet changed intestinal microbiota composition and bile acid metabolism. Therefore, choline or TMAO supplementation may promote AS by modulating gut microbiota in mice fed with a western diet and by other mechanisms in mice given a normal diet, even choline or TMAO supplementation in a normal diet can promote AS.

Diagnostic Performance of a Deep Learning Model Deployed at a National COVID-19 Screening Facility for Detection of Pneumonia on Frontal Chest Radiographs.

(1) Background: Chest radiographs are the mainstay of initial radiological investigation in this COVID-19 pandemic. A reliable and readily deployable artificial intelligence (AI) algorithm that detects pneumonia in COVID-19 suspects can be useful for screening or triage in a hospital setting. This study has a few objectives: first, to develop a model that accurately detects pneumonia in COVID-19 suspects; second, to assess its performance in a real-world clinical setting; and third, by integrating the model with the daily clinical workflow, to measure its impact on report turn-around time. (2) Methods: The model was developed from the NIH Chest-14 open-source dataset and fine-tuned using an internal dataset comprising more than 4000 CXRs acquired in our institution. Input from two senior radiologists provided the reference standard. The model was integrated into daily clinical workflow, prioritising abnormal CXRs for expedited reporting. Area under the receiver operating characteristic curve (AUC), F1 score, sensitivity, and specificity were calculated to characterise diagnostic performance. The average time taken by radiologists in reporting the CXRs was compared against the mean baseline time taken prior to implementation of the AI model. (3) Results: 9431 unique CXRs were included in the datasets, of which 1232 were ground truth-labelled positive for pneumonia. On the "live" dataset, the model achieved an AUC of 0.95 (95% confidence interval (CI): 0.92, 0.96) corresponding to a specificity of 97% (95% CI: 0.97, 0.98) and sensitivity of 79% (95% CI: 0.72, 0.84). No statistically significant degradation of diagnostic performance was encountered during clinical deployment, and report turn-around time was reduced by 22%. (4) Conclusion: In real-world clinical deployment, our model expedites reporting of pneumonia in COVID-19 suspects while preserving diagnostic performance without significant model drift.

CAFT: a deep learning-based comprehensive abdominal fat analysis tool for large cohort studies.

BACKGROUND: There is increasing appreciation of the association of obesity beyond co-morbidities, such as cancers, Type 2 diabetes, hypertension, and stroke to also impact upon the muscle to give rise to sarcopenic obesity. Phenotypic knowledge of obesity is crucial for profiling and management of obesity, as different fat-subcutaneous adipose tissue depots (SAT) and visceral adipose tissue depots (VAT) have various degrees of influence on metabolic syndrome and morbidities. Manual segmentation is time consuming and laborious. Study focuses on the development of a deep learning-based, complete data processing pipeline for MRI-based fat analysis, for large cohort studies which include (1) data augmentation and preprocessing (2) model zoo (3) visualization dashboard, and (4) correction tool, for automated quantification of fat compartments SAT and VAT. METHODS: Our sample comprised 190 healthy community-dwelling older adults from the Geri-LABS study with mean age of 67.85 ± 7.90 years, BMI 23.75 ± 3.65 kg/m2, 132 (69.5%) female, and mainly Chinese ethnicity. 3D-modified Dixon T1-weighted gradient-echo MR images were acquired. Residual global aggregation-based 3D U-Net (RGA-U-Net) and standard 3D U-Net were trained to segment SAT, VAT, superficial and deep subcutaneous adipose tissue depots (SSAT and DSAT). Manual segmentation from 26 subjects was used as ground truth during training. Data augmentations, random bias, noise and ghosting were carried out to increase the number of training datasets to 130. Segmentation accuracy was evaluated using Dice and Hausdorff metrics. RESULTS: The accuracy of segmentation was SSAT:0.92, DSAT:0.88 and VAT:0.9. Average Hausdorff distance was less than 5 mm. Automated segmentation significantly correlated R2 > 0.99 (p < 0.001) with ground truth for all 3-fat compartments. Predicted volumes were within ± 1.96SD from Bland-Altman analysis. CONCLUSIONS: DL-based, comprehensive SSAT, DSAT, and VAT analysis tool showed high accuracy and reproducibility and provided a comprehensive fat compartment composition analysis and visualization in less than 10 s.

Caffeic acid prevents non-alcoholic fatty liver disease induced by a high-fat diet through gut microbiota modulation in mice.

Caffeic acid (CA) is derived from many plants and may have the ability to reduce hepatic lipid accumulation. The gut microbiota produces lipopolysaccharides and further influences hepatic lipid metabolism, and thus plays an important role in the development of nonalcoholic fatty liver disease (NAFLD). However, whether the beneficial effects of CA are associated with the gut microbiota remains unclear. The present study aimed to investigate the benefits of experimental treatment with CA on the gut microbiota and metabolic functions in a mouse model of NAFLD. In this study, C57BL/6J mice received a high-fat diet (HFD) for 8 weeks and were then fed a HFD supplemented with or without CA for another 8 weeks. HFD induced obesity and increased accumulation of intrahepatic lipids, serum biochemical parameters and gene expression related to lipid metabolism. Microbiota composition was determined via 16S rRNA sequencing, and analysis revealed that HFD led to dysbiosis, accompanied by endotoxemia and low-grade inflammation. CA reverted the imbalance in the gut microbiota and related lipopolysaccharide-mediated inflammation, thus inhibiting deregulation of lipid metabolism-related gene expression. Our results support the possibility that CA can be used as a therapeutic approach for obesity-associated NAFLD via its anti-inflammatory and prebiotic integrative response.

[Establishment of a novel mouse mode of elastase-induced chronic obstructive pulmonary disease related osteoporosis].

OBJECTIVE: To establish and evaluate the model of chronic obstructive pulmonary disease (COPD) with osteoporosis induced by elastase in mice. METHODS: Twenty four healthy female 8-week-old C57BL / 6 mice (weighing about 18 g) were randomly divided into three groups. The control group was given intratracheal drip of normal saline, the experimental group 1 and the experimental group 2 were given intratracheal drip of elastase, the control group and the experimental group 1 were kept for 8 weeks and then killed, the experimental group 2 was kept for 12 weeks and then killed. HE staining was used to evaluate the histopathological changes of lung and tibia in the control and experimental groups. The levels of serum inflammatory factors and broncho alveolar lavage factors (BALF) were detected by ELISA. Micro CT was used to detect the bone mass related parameters of mouse femur. The expression of osteoclastic and osteogenic genes was detected by real-time fluorescence quantitative PCR. RESULTS: Lung histopathology showed that the structure of alveoli in the experimental group was disordered, the walls of alveoli became thin or broken, and the alveoli cavity expanded. IL-6 and TNF-α in BALF were significantly higher than those in control group (P<0.001), while IL-1ß and TNF-α in serum inflammatory factors were significantly higher than those in control group (P<0.001). BV / TV(bone volume fraction), TB.Th(average bone trabecular thickness) and TB.N(average bone trabecular number) in the experimental group were significantly lower than those in the control group (P<0.05), TB.Sp (average bone trabecular separation) and BS / BV (bone surface area fraction) in the experimental group were significantly higher than those in the control group (P<0.01). Compared with the control group, the expression of osteoclast related marker genes increased in the experimental group (P<0.05), but decreased in the experimental group(P<0.05). The results of experiment 1 and experiment 2 were time-dependent. CONCLUSION: In this study, elastase was used to construct a COPD model with osteoporosis successfully, which provides a suitable animal model for the future study of the pathogenesis of COPD with osteoporosis.

Antioxidants Attenuate Oxidative Stress-Induced Hidden Blood Loss in Rats.

OBJECTIVE: Hidden blood loss (HBL), commonly seen after total knee or hip arthroplasty, causes postoperative anemia even after reinfusion or blood transfusion based on the visible blood loss volume. Recent studies demonstrated that oxidative stress might be involved in HBL. However, whether the antioxidants proanthocyanidin (PA) or hydrogen water (HW) can ameliorate HBL remains poorly understood. The aim of this study was to evaluate the effects of PA and HW on HBL. MATERIALS AND METHODS: A rat HBL model was established through administration of linoleic acid with or without treatment with PA or HW. The levels of hemoglobin (Hb), red blood cell (RBC) count, superoxide dismutase (SOD) activity, glutathione peroxidase (GSH-PX) activity, malondialdehyde (MDA), and ferryl Hb were measured. RESULTS: RBC and Hb values as well as the activity of SOD and GSH-PX were reduced after administration of linoleic acid, which was ameliorated by treatment with PA or HW. In addition, the quantity of MDA was significantly decreased with the administration of PA or HW. CONCLUSION: PA and HW could ameliorate HBL in a rat model by reducing oxidative stress, suggesting that they might be used as a novel therapeutic approach in the prophylaxis or treatment of HBL in clinics.

The effect of thalassemia on erythrocyte reference intervals in a representative Han Chinese adult population.

BACKGROUND: The data from apparently healthy individuals with thalassemia has been demonstrated to have an effect on the reference intervals for the erythrocyte indices in areas with a high incidence of thalassemia. METHODS: Six clinical centers screened apparently healthy individuals using a questionnaire and a physical examination. Then, the qualified reference individuals were selected by hematological indices and a genotypic thalassemia diagnosis. Statistical comparisons were conducted for the erythrocyte reference intervals in the Chinese population with and without thalassemia. The constituent ratios and the mean (SD) of erythrocyte indices according to the thalassemia genotype were calculated. The relationship between the MCV values and the thalassemia genotype was also estimated. RESULTS: 4,636 reference individuals were included using hematological indices and genotypic thalassemia screening. The results of the erythrocyte reference intervals for individuals in Guangzhou with thalassemia demonstrated that the RBC, MCV, and MCH values significantly differed by gender compared with other regions (p < 0.01). In contrast, for individuals without thalassemia, the results tended to be similar and clinically acceptable. In addition, the results of the erythrocyte indices revealed significant differences among α-thalassemia patients, ß-thalassemia patients, and the control group. CONCLUSIONS: Apparently healthy individuals with thalassemia in the high prevalence zone of thalassemia could not be excluded by the questionnaire, physical examination or laboratory indices (Fe < 6 µmol/L, Hb < 90 g/L). The screening of genotypic thalassemia based on the MCV or MCH values to exclude unqualified individuals is the most effective way to obtain accurate and reliable reference intervals for the erythrocyte indices.

Delay in seeking medical evaluations and predictors of self-efficacy among women with newly diagnosed breast cancer: a longitudinal study.

BACKGROUND: Delaying a diagnosis of breast cancer directly and positively impacts survival. Self-efficacy has been shown to be a causal mechanism in a wide range of health behaviors, a measurable trait that predicts behavior across domains, which is strong associated with psychological variables. However, factors predicting self-efficacy of women with suspected breast cancer who delayed or did not delay seeking a breast cancer diagnosis over time have not been identified. OBJECTIVES: To examine the differences between women who delay and women who did not delay seeking a cancer diagnosis, and key factors predicting self-efficacy over time among women with newly-diagnosed breast cancer. DESIGN: Descriptive, longitudinal design over 2 months following breast cancer diagnostic evaluation. SETTING: A medical center is located in southern Taiwan. PARTICIPANTS: Eighty women with suspected breast cancer were approached and 67 subjects with a positive diagnosis of breast cancer were recruited. METHODS: Subjects were categorized into women who delayed their diagnosis and women who did not delay their diagnosis. A battery of 5 standardized questionnaires including self-efficacy, anxiety and depression, personality, spiritual support and hope was completed at the first three clinic visits. RESULTS: Stage of cancer, trait extroversion/neuroticism and spiritual support were significantly different between groups (p<0.05). Subjects who did not delay (ß=-1.613, p<0.05), and time that histology results were provided (ß=-2.4333, p<0.001) had a significantly predicted negative change in self-efficacy compared to the group that delayed. Hope at the first clinic visit contributed to the change in self-efficacy over time (ß=0.391, p<0.001). CONCLUSIONS: Personal factors affecting a woman's delay in obtaining medical assessment of breast cancer confirmation. Hope impacts self-efficacy of women with suspected breast cancer and interventions to enhance hope during the early stages of breast cancer evaluation require further study.

[Relationship between fractional esterification rate of high density lipoprotein cholesterol and coronary artery disease].

OBJECTIVE: To assess the relationship between fractional esterification rate of high density lipoprotein cholesterol (FER(HDL)) and coronary artery disease. METHODS: A total of 131 hospitalized patients underwent coronary angiography due to chest pain were included in the study and patients were divided into CAD group (n = 76) and non CAD group (n = 55) according to coronary angiogram. Clinical and laboratory data including biochemical laboratory, FER(HDL) and lipid subclasses were analyzed. RESULTS: The FER(HDL) value of CAD group was significantly higher than that of the non CAD group (21.70 ± 8.73 vs. 18.65 ± 6.26, P < 0.05). There was an increased trend of FER(HDL) with numbers of diseased coronary arteries, significant difference was evidenced between non CAD group and 3-vessel group (18.65 ± 6.26 vs. 24.00 ± 9.22, P < 0.05). FER(HDL) was positively correlated with TG (r = 0.647, P < 0.001), LDLb-C(r = 0.441, P < 0.001) and negatively correlated with HDL-C (r = -0.708, P < 0.001) and HDL(2)-C (r = -0.748, P < 0.001). CONCLUSION: Our data showed that the values of FER(HDL) were significantly increased in CAD patients and correlated with the severity of the CAD.

[National external quality assessment and comparability of assays for tumor markers measurements].

OBJECTIVE: To evaluate the performance of tumor markers (TM) measurements in clinical laboratories by external quality assessment (EQA) and investigate the comparability of assays for TM. METHODS: Ten quality control sera specimens were distributed to 586 laboratories by global Express Mail Services (EMS) in March 2008 and tested twice with 5 specimens each time. Analytes were total prostate specific antigen (PSA), free PSA, carcinoembryonic antigen (CEA), alpha fetoprotein (AFP), human chorionic gonadotrophin (HCG), beta-HCG, carbohydrate antigen 19-9 (CA19-9), cancer antigen 15-3 (CA 15-3), cancer antigen 125 (CA125), beta-2-microglobulin and ferritin. The collected data were divided into peer groups according to analyzers or methods and the median of peer group was adopted as the target value (TV) separately after outlier removal. Two standard deviations of the median were set as the limit of difference. RESULTS: The first TM EQA results of 2008 showed that the pass percentage of all participating laboratories ranged from 87.3% (CA125) to 95.5% (beta-2-microglobulin). And the second batch ranged from 83.5% (HCG) to 94.0% (beta-2-microglobulin). The coefficient variances (CVs) of intra-group values determined by automatic analyzers were lesser than 15% for each test of every specimen. The CVs of radioimmunoassay (RIA) and enzyme-linked immunosorbent assay (ELISA) were over 20% and 50% respectively. The inter-group medians of 9 tests showed CVs>20% with HCG 13.4% and ferritin 15.7%. The CV of paired medians of some automatic analyzers was small and showed no statistical significance (all Z<1.890, all P>0.05). CONCLUSION: The analytical performance of automatic analyzers is superior to RIA and ELISA. There is an excellent comparability within automatic analyzers for TM measurements and a lack of comparability within RIA and ELISA. Noncomparability is found in over 80% of TM assays.

[Effects of slopes on nitrogen transport along with runoff from sloping plots on a lateritic red soil amended with sewage sludge].

The effects of different slopes on nitrogen transport along with runoff from sloping plots amended with sewage sludge on a lateritic red soil were studied under simulated rainfall conditions. When the sludge was broadcasted and mixed with surface soils (BM), the MTN (total nitrogen of mixing sample), STN (total nitrogen of settled sample), TPN (total particulate nitrogen), TSN (total suspended nitrogen), TDN (total dissolved nitrogen) and NH4(+) -N concentrations and nitrogen loss amounts in runoff of all treatments were highest at 1 day or 18 days after application. The highest concentrations and the loss amounts of MTN and STN in the slope runoff for the BM treatment increased with slope degree, showing increasing pollution risks to the surface waters. The STN concentration and loss amounts from the 25 degrees plots were 126.1 mg x L(-1) and 1788.6 mg x m(-2), respectively, being 4.6 times and 5.8 times of the corresponding values from the 10 degrees plots, respectively. Then the concentrations and the loss amounts of nitrogen (except NO3(-) -N) from the BM plots diminished rapidly first and then tended to be stable with dwindling differences between the slopes. The loss of MTN and STN in early runoff (1 day and 18 days) accounted for 68.6% -73.4% and 62.3% -66.7% of the cumulative loss amounts during the experimental period for all the broadcasted treatments. Runoff loss coefficients of MTN increased in the order of 20 degrees > 25 degrees > 15 degrees > 10 degrees. Nitrogen was largely lost in dissolved species while large portion of NH4(+) -N was lost with particulates.

External quality assessment on detection of hepatitis C virus RNA in clinical laboratories of China.

BACKGROUND: As with many studies carried out in European countries, a quality assurance program has been established by the National Center for Clinical Laboratories in China (NCCL). The results showed that the external quality assessment significantly improves laboratory performance for quantitative evaluation of hepatitis C virus (HCV) RNA. METHODS: Serum panels were delivered twice annually to the clinical laboratories which performed HCV RNA detection in China. Each panel made up of 5 coded samples. All laboratories were requested to carry out the detection within the required time period and report on testing results which contained qualitative and/or quantitative test findings, reagents used and relevant information about apparatus. All the positive samples were calibrated against the first International Standard for HCV RNA in a collaborative study and the range of comparison target value (TG) designated as +/- 0.5 log. RESULTS: The numbers of laboratories reporting on qualitative testing results for the first and second time external quality assessment were 168 and 167 in the year of 2003 and increased to 209 and 233 in 2007; the numbers of laboratories reporting on quantitative testing results were 134 and 147 in 2003 and rose to 340 and 339 in 2007. Deviation between the mean value for quantitative results at home in 2003 and the target value was above 0.5 log, which was comparatively high. By 2007, the target value was close to the national average except for the low concentrated specimens (10(3) IU/ml). The percentage of results within the range of GM +/- 0.5 log(10) varied from 8.2% to 93.5%. Some laboratories had some difficulties in the exact quantification of the lowest (3.00 log IU/ml) as well as of the highest viral levels (6.37 log IU/ml) values, very near to the limits of the dynamic range of the assays. CONCLUSIONS: The comparison of these results with the previous study confirms that a regular participation in external quality assessment (EQA) assures the achievement of a high proficiency level in the diagnosis of HCV infection. During the 5-year external quality assessment, sensitivity and accuracy of detection in most of the clinical laboratories have been evidently improved and the quality of kits has also been substantially improved.

[Establishment of the first national standards for nucleic acid amplification technology assay for HBV DNA].

OBJECTIVES: To establish a Chinese national standard for a nucleic acid test (NAT) for HBV DNA. METHODS: The candidate sample of HBV DNA positive plasma was diluted with HBV-negative human plasma. The sample was lyophilised with a concentration of approximately 300,000 copies/ml. The measurement methods used included Roche Amplicor assay (version 2.0) and real-time PCR. The lyophilised preparation was calibrated by the international standard (NIBSC code: 97/746) from NIBSC. RESULTS: The quantity of this lyophilised preparation was (1.29+/-0.24) x 10(5)IU/ml in comparison with the international standard for HBV DNA 97/746. The stability test indicated that the sample was stable at room temperature (20 to 25 degrees C) for 2 weeks and at 37 degrees C for at least 1 week. Long-term stability was observed at 2 to 8 degrees C for 6 months and at -20 degrees C for more than 2 years with no significant changes. The vial-to-vial imprecision rate was 3.53%. CONCLUSION: Based on the results of this study, our lyophilized sample can be used as a standard in China for the nucleic acid test (NAT) for HBV DNA.

Lyophilized standards for the calibration of real time PCR assay for hepatitis C virus RNA.

BACKGROUND: Since October 1997, an international standard for hepatitis C virus (HCV) nucleic acid amplification technology assay, 96/790, has been available. We compared a series of lyophilized standards with known HCV RNA concentrations against the international standard in fluorescence quantitative PCR detection. METHODS: A series of lyophilized sera were calibrated by ROCHE COBAS AMPLICOR HCV Monitor test against the international standard and sent to various manufacturers to analyse the samples using their own kits. Then calibration curves from the series were compared with that obtained from the external standard calibration curve with the manufacture's series. RESULTS: The standard calibration curve with the series of lyophilized serum showed an excellent correlation (R(2) > 0.98), slope and intercept that were similar to those from the manufacture's series. When the standard calibration curve from the series of lyophilized standards were used to define the values of the given sample, lower coefficients of variation between kits from different manufactures were obtained. CONCLUSION: The results showed that the lyophilized standards could be used to setup the standard calibration curve for clinical HCV RNA quantitative PCR detection.