Objective:To develop CNN-Dinov2, a deep learning-based automatic classification model for Gram-stained images, enabling rapid diagnosis of three prevalent Gram-negative bacilli in bloodstream infections: Escherichia coli ( E.coli), Klebsiella pneumoniae ( K.pneumoniae), and Pseudomonas aeruginosa ( P.aeruginosa). Methods:This evaluation study analyzed 1 425 Gram-stained microscopic images from patients with bloodstream infections at Houjie Hospital, in Dongguan City, collected between January 2023 and January 2024. The images, all positive for blood culture and identified as target strains, were categorized into Escherichia coli (419 images), Klebsiella pneumoniae (411 images), Pseudomonas aeruginosa (413 images), and other Gram-negative bacilli (182 images). They were randomly split into a training set (1 141 images), a validation set (141 images), and a test set (143 images) in an 8∶1∶1 ratio. A hybrid CNN-Dinov2 model was developed by integrating ResNet′s local feature extraction with Dinov2′s global pre-trained features, followed by a fully connected layer. The model was optimized by inputting the preprocessed images and adjusting parameters through loss calculation and backpropagation. AlexNet, Dinov2, and ResNet18 served as control models. The models′ classification performance was assessed using accuracy, precision, weighted F1 score, and recall rate, derived from the confusion matrix. The PR curve and AP value further evaluated each model′s classification capability across the four image categories. Results:The CNN-Dinov2 model achieved a training accuracy of 99.74%, a validation accuracy of 98.12%, and a validation loss of 0.070 6, demonstrating robust generalization without overfitting. Validation metrics revealed superior performance with an accuracy of 98.60%, precision of 98.65%, a weighted F1 score of 98.60%, and a recall rate of 98.60%, outperforming other models. The confusion matrix confirmed its strong classification capability, with the highest sum of diagonal values for identifying four types of bacteria. The macro average precision (AP) values under the precision-recall (PR) curves were all 1, indicating excellent discrimination across all categories. Overall, the CNN-Dinov2 model exhibited the best performance among the four models evaluated.Conclusion:This study successfully developed CNN-Dinov2, an automated classification model for Gram staining images. It offers valuable support for the rapid diagnosis of bloodstream infections caused by Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa, demonstrating practical utility.

Objective:To develop CNN-Dinov2, a deep learning-based automatic classification model for Gram-stained images, enabling rapid diagnosis of three prevalent Gram-negative bacilli in bloodstream infections: Escherichia coli ( E.coli), Klebsiella pneumoniae ( K.pneumoniae), and Pseudomonas aeruginosa ( P.aeruginosa). Methods:This evaluation study analyzed 1 425 Gram-stained microscopic images from patients with bloodstream infections at Houjie Hospital, in Dongguan City, collected between January 2023 and January 2024. The images, all positive for blood culture and identified as target strains, were categorized into Escherichia coli (419 images), Klebsiella pneumoniae (411 images), Pseudomonas aeruginosa (413 images), and other Gram-negative bacilli (182 images). They were randomly split into a training set (1 141 images), a validation set (141 images), and a test set (143 images) in an 8∶1∶1 ratio. A hybrid CNN-Dinov2 model was developed by integrating ResNet′s local feature extraction with Dinov2′s global pre-trained features, followed by a fully connected layer. The model was optimized by inputting the preprocessed images and adjusting parameters through loss calculation and backpropagation. AlexNet, Dinov2, and ResNet18 served as control models. The models′ classification performance was assessed using accuracy, precision, weighted F1 score, and recall rate, derived from the confusion matrix. The PR curve and AP value further evaluated each model′s classification capability across the four image categories. Results:The CNN-Dinov2 model achieved a training accuracy of 99.74%, a validation accuracy of 98.12%, and a validation loss of 0.070 6, demonstrating robust generalization without overfitting. Validation metrics revealed superior performance with an accuracy of 98.60%, precision of 98.65%, a weighted F1 score of 98.60%, and a recall rate of 98.60%, outperforming other models. The confusion matrix confirmed its strong classification capability, with the highest sum of diagonal values for identifying four types of bacteria. The macro average precision (AP) values under the precision-recall (PR) curves were all 1, indicating excellent discrimination across all categories. Overall, the CNN-Dinov2 model exhibited the best performance among the four models evaluated.Conclusion:This study successfully developed CNN-Dinov2, an automated classification model for Gram staining images. It offers valuable support for the rapid diagnosis of bloodstream infections caused by Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa, demonstrating practical utility.

Objective:To investigate the clinical features and prognosis of male with anti-melanoma differentiation-associated gene 5 (MDA5) autoantibody.Methods:The clinical data of 246 patients with DM and anti-MDA5 autoantibodies hospitalized by Jiangsu Myositis Cooperation Group from 2017 to 2020 were collected and retrospectively analyzed. Chi-square test was performed to compared between counting data groups; Quantitative data were expressed by M ( Q1, Q3), and rank sum test was used for comparison between groups; Single factor survival analysis was performed by Kaplan-Meier method and Log rank test; Cox regression analysis were used for multivariate survival analysis. Results:①The male group had a higher proportion of rash at the sun exposure area [67.1%(47/70) vs 52.8%(93/176), χ2=4.18, P=0.041] and V-sign [50.0%(35/70) vs 30.7%(54/176), χ2=8.09, P=0.004] than the female group. The male group had higher levels of creatine kinase [112(18, 981)U/L vs 57 (13.6, 1 433)U/L, Z=-3.50, P<0.001] and ferritin [1 500 (166, 32 716)ng/ml vs 569 (18, 14 839)ng/ml, Z=-5.85, P<0.001] than the female group. The proportion of ILD [40.0%(28/70) vs 59.7%(105/176), χ2=7.82, P=0.020] patients and the red blood cell sedimentation rate[31.0(4.0, 101.5)mm/1 h vs 43.4(5.0, 126.5)mm/1 h, Z=-2.22, P=0.026] in the male group was lower than that of the female group, but the proportion of rapidly progressive interstitial lung disease (PR-ILD) [47.1%(33/70) vs 31.3%(55/176), χ2=5.51, P=0.019] was higher than that of the female group. ②In male patients with positive anti-MDA5 antibodies,the death group had a shorter course of disease[1.0(1.0, 3.0) month vs 2.5(0.5,84) month, Z=-3.07, P=0.002], the incidence of arthritis [16.7%(4/24) vs 42.2%(19/45), χ2=4.60, P=0.032] were low than those in survival group,while aspartate aminotransferase (AST)[64(22.1, 565)U/L vs 51(14,601)U/L, Z=-2.42, P=0.016], lactate dehydrogenase (LDH) [485(24,1 464)U/L vs 352(170, 1 213)U/L, Z=-3.38, P=0.001], C-reactive protein (CRP) [11.6(2.9, 61.7) mg/L vs 4.95(0.6, 86.4) mg/L, Z=-1.96, P=0.050], and ferritin levels [2 000(681, 7 676) vs 1 125 (166, 32 716)ng/ml, Z=-3.18, P=0.001] were higher than those in the survival group, and RP-ILD [95.8%(23/24) vs 22.2%(10/45), χ2=33.99, P<0.001] occurred at a significantly higher rate. ③Cox regression analysis indicated that the course of disease LDH level, and RP-ILD were related factors for the prognosis of male anti-MDA5 antibodies [ HR (95% CI)=0.203(0.077, 0.534), P=0.001; HR (95% CI)=1.002(1.001, 1.004), P=0.003; HR (95% CI)=95.674 (10.872, 841.904), P<0.001]. Conclusion:The clinical manifestations of male anti-MDA5 antibody-positive patients are different from those of female. The incidence of ILD is low, but the proportion of PR-ILD is high. The course of disease, serum LDH level, and RP-ILD are prognostic factors of male anti-MDA5 antibody-positive patients.

Objective Based on HPLC fingerprinting and chemometrics,to evaluate the quality of Schefflera kwangsiensis Merr.from Guangxi.Methods HPLC was used to establish fingerprints of Schefflera kwangsiensis Merr.from ten different origins,and gradient elution was carried out with methanol-0.1％phosphoric acid aqueous solution as mobile phase.Cluster analysis(CA),principal component analysis(PCA)and orthogonal partial least squares-discriminant analysis(OPLS-DA)were applied to evaluate quality.Results The fingerprints of Schefflera kwangsiensis Merr.from ten different origins were established by HPLC,a total of 22 common peaks were calibrated,with a similarity range of 0.922-0.999.Four chromatographic peaks were identified as rhodopsin,4,5-bis-O-caffeoylquinic acid,caffeic acid,and naringin.The samples were classified into four types according to the CA and OPLS-DA.PCA identified four principal components with a cumulative contribution rare of 95.39％.Conclusion The quality of Schefflera kwangsiensis Merr.can be comprehensively evaluated by fingerprinting combined with CA,PCA and OPLS-DA analysis.The Study can provide a reference for improving the quality control and assessment of Schefflera kwangsiensis Merr.

OBJECTIVES: Programmed death 1 (PD-1) associated fulminant type 1 diabetes (PFD) is a rare acute and critical in internal medicine, and its clinical characteristics are still unclear. This study aims to analyze the clinical characteristics of PFD patients to improve clinical diagnosis and treatment. METHODS: We retrospectively analyzed the clinical data of 10 patients with PFD admitted to the Second Xiangya Hospital of Central South University, combined with the data of 66 patients reported in the relevant literature, analyzed and summarized their clinical and immunological characteristics, and compared the patients with PFD with different islet autoantibody status. RESULTS: Combined with our hospital and literature data, a total of 76 patients with PFD were reported, with the age of (60.9±12.1) years old, 60.0% male and body mass index of (22.1±5.2) kg/m2. In 76 patients, the most common tumors were lung cancer (43.4%) and melanoma (22.4%). Among PD-1 inhibitors, the most common drugs are nivolumab (37.5%) and pembrolizumab (38.9%). 82.2% of PFD patients developed diabetes ketoacidosis. The median onset time from PD-1 related inhibitor treatment to hyperglycemia was 95 (36.0, 164.5) d, and the median treatment cycle before the onset of diabetes was 6 (2.3, 8.0) cycles. 26% (19/73) of PFD patients had positive islet autoantibodies, and the proportion of ketoacidosis in the positive group was significantly higher than that in the negative group (100.0% vs 75.0%, P<0.05). The onset time and infusion times of diabetes after PD-1 inhibitor treatment in the autoantibody positive group were significantly lower than those in the autoantibody negative group (28.5 d vs 120.0 d; 2 cycles vs 7 cycles, both P<0.001). CONCLUSIONS After initiation of tumor immunotherapy, it is necessary to be alert to the occurrence of adverse reactions of PFD, and the onset of PFD with islet autoantibody positive is faster and more serious than that of patients with autoantibodies negative. Detection of islet autoantibodies and blood glucose before and after treatment with PD-1 inhibitors is of great value for early warning and prediction of PFD.
Humans ; Male ; Middle Aged ; Aged ; Female ; Diabetes Mellitus, Type 1 ; Programmed Cell Death 1 Receptor ; Immune Checkpoint Inhibitors/therapeutic use* ; Retrospective Studies ; Ketosis ; Autoantibodies

Objective:To analyze the application of Chronic Obstructive Pulmonary Disease (COPD) Screening Questionnaire and pulmonary function test in dust-exposed migrant workers.Methods:In May 2019, 149 cases of dust exposed migrant workers were selected as the research subjects through the free clinic in the countryside. COPD Screening Questionnaire and lung function test were carried out to analyze the high-risk groups and the influencing factors of positive pulmonary function test results.Results:Among 149 cases of dust-exposed migrant workers, 107 (71.8%) were positive for questionnaire screening, 73 (49.0%) were positive for pulmonary function test, 75 (50.3%) were diagnosed with coal worker's pneumoconiosis, and 101 (67.8%) were diagnosed with lung function injury. The positive rate of pulmonary function of migrant workers with positive questionnaire screening results was significantly higher than that of those with negative results ( P<0.05) . The results of multivariate analysis showed that compared with non-pneumoconiosis, the risk of positive pulmonary function test results was higher in dust-exposed migrant workers with stage Ⅲ pneumoconiosis ( OR=16.462, 95% CI: 3.390-79.946; P<0.01) . Compared with non-smoking, the risks of positive pulmonary function test results of dust-exposed migrant workers with smoking index of 11-20 package years and >20 package years were higher ( OR=19.814, 95% CI: 3.854-101.883; OR=9.733, 95% CI: 2.310-41.008; P<0.01) . Conclusion:The risk of COPD in dust-exposed migrant workers is high, so we should strengthen the early examination of the high pneumoconiosis stage and smoking population. The screening questionnaire can better screen out the high-risk groups of COPD, and it can be used as a basic screening tool.

Objective:To analyze the application of Chronic Obstructive Pulmonary Disease (COPD) Screening Questionnaire and pulmonary function test in dust-exposed migrant workers.Methods:In May 2019, 149 cases of dust exposed migrant workers were selected as the research subjects through the free clinic in the countryside. COPD Screening Questionnaire and lung function test were carried out to analyze the high-risk groups and the influencing factors of positive pulmonary function test results.Results:Among 149 cases of dust-exposed migrant workers, 107 (71.8%) were positive for questionnaire screening, 73 (49.0%) were positive for pulmonary function test, 75 (50.3%) were diagnosed with coal worker's pneumoconiosis, and 101 (67.8%) were diagnosed with lung function injury. The positive rate of pulmonary function of migrant workers with positive questionnaire screening results was significantly higher than that of those with negative results ( P<0.05) . The results of multivariate analysis showed that compared with non-pneumoconiosis, the risk of positive pulmonary function test results was higher in dust-exposed migrant workers with stage Ⅲ pneumoconiosis ( OR=16.462, 95% CI: 3.390-79.946; P<0.01) . Compared with non-smoking, the risks of positive pulmonary function test results of dust-exposed migrant workers with smoking index of 11-20 package years and >20 package years were higher ( OR=19.814, 95% CI: 3.854-101.883; OR=9.733, 95% CI: 2.310-41.008; P<0.01) . Conclusion:The risk of COPD in dust-exposed migrant workers is high, so we should strengthen the early examination of the high pneumoconiosis stage and smoking population. The screening questionnaire can better screen out the high-risk groups of COPD, and it can be used as a basic screening tool.

Objective:To investigate the serum levels, diagnosis and prognosis value of endothelin-1(ET-1) in children suffering from lung metastasis of osteosarcoma.Methods:A total of 84 children with osteosarcoma, 67 children with lung metastasis of osteosarcoma and 35 healthy people from January 1, 2013 to January 1, 2018 in Second People′s Hospital of Nanyang were retrospectively included.The serum level of ET-1 was measured by performing enzyme linked immunosorbent assay(ELISA) methods and the influencing factors of serum ET-1 levels in children with lung metastasis of osteosarcoma were conducted by Logistic regression analysis.The clinical value of ET-1 in the prediction of the incidence of lung metastasis in children with osteosarcoma was analyzed by receiver operating characteristic (ROC) curve.Forty-five children with lung metastasis of osteosarcoma were followed up for 18 months and the prognosis value of serum ET-1 levels in children with lung metastasis of osteosarcoma was evaluated by Kaplan-Meier survival analysis. Results:The serum ET-1 level in children with lung metastasis of osteosarcoma was 97.23 (65.13, 134.98) ng/L and significantly higher than osteosarcoma group 60.21 (43.12, 74.63) ng/L and healthy control group 34.45 (12.01, 63.03) ng/L, respectively ( Z=-5.671, -4.92, all P<0.05), with significant differences. Logistic regression analysis proved that lung bilateral involvement ( OR=3.449), numbers of lung metastases (more than 3)( OR=3.449), average diameter of lung metastases (more than 5 cm) ( OR=6.501) and extrapulmonary metastasis ( OR=4.369) were independent risk factors for elevated serum ET-1 levels in children developing lung metastasis of osteosarcoma.The predictive value of ET-1 in the incidence of lung metastasis in children with osteosarcoma was significant (area under ROC curve: 0.841). When the cut-off value was 94.27 ng/L, the sensitivity and specificity were 88.5% and 92.6%, respectively.Survival analysis revealed that higher levels of ET-1 was correlated with poor prognosis ( OR=3.287, 95% CI: 1.119-9.547). Conclusions:The serum levels of ET-1 in children with lung metastasis of osteosarcoma are significantly elevated.ET-1 is a serological marker for the differential diagnosis of lung metastasis of osteosarcoma.Moreover, the higher levels of ET-1 are correlated with poor prognosis in children with lung metastasis of osteosarcoma.

Objective To analyse the changes of cerebral cortical thickness and explore the connectivity of cortical thickness and the clinical syndrome of concomitant strabismus using high-resolution magnetic resonance imaging.Methods Participants were 26 children with primary strabismus and 28 healthy children strictly matched for sex, age, education and socioeconomic level.By using Freesurfer software, the whole-brain-based analysis was perfomrmed to compare the cortical thickness between the two groups.Results Compared with the healthy control group the children with strabismus were found that the cerebral cortex became thinner, including these brain regions related with stereovision,advanced visual-attention,attention and executive control function,as well as brain areas in insula and cingulate circuit.Conclusion Concomitant strabismus in children are not only caused by pathological changes of extraocular muscles,but anatomical changes of the central nervous system, especially the brain areas involved in the advanced visual processing related with visual-motor and visual-attention, and the advanced cognitive brain regions associated with attention and executive control, as well as the anatomical changes in insula and cingulate cortex circuit, which can explain the clinical syndrome of children with concomitant strabismus.

Objective The aim of this study is to understand the impairment and compensation mechanism of brain function in untreated primary insomnia (PI) patients.The approach of fractional amplitude of low frequency fluctuation (fALFF) is used to analyze raw data between the PI patients and the normal control group in resting state using functional magnetic resonance imaging (fMRI).Methods Fifty-nine PI patients,admitted to our hospital from November 2015 to November 2016,and 47 age-,education-,and gender-matched normal healthy subjects were chosen in our study.Pittsburg sleep quality index (PSQI),insomnia severity index (ISI) were employed to evaluate the sleep quality.Self-rating anxiety scale (SAS) and self-rating depression scale (SDS) were employed to evaluate the emotion.Resting state fMRI and fALFF analyses were used to compare the functional regional activities.The correlations of fALFF data with PSQI,SAS and SDS scores were analyzed.Results In PI patients,2 had mild to moderate insomnia,41 had moderate insomnia,and 16 had serious insomnia.ISI scores in the normal healthy subjects were less than 7.The PSQI,SAS,SDS and ISI scores in the PI patients were significantly higher than those in the normal healthy subjects (P＜0.05).As compared with the control group,the PI group had significantly increased fALFF value in the right hippocampus (HIP),right parahippocampa gyms,right amygdala,and bilateral thalamus.The fALFF value was positively correlated PSQI,SASandSDSscores (r=0.582,P=0.000;r=0.617,P=0.000;r=0.653,P=0.000).Conclusion Some brain regions in the PI patients are abnormal in the resting state,which can reflex functional regional activities of PI patients.