OBJECTIVES: We propose a YOLOv11-TDSP model for improving the accuracy of dental abnormality detection on panoramic oral X-ray images. METHODS: The SHSA single-head attention mechanism was integrated with C2PSA in the backbone layer to construct a new C2PSA_SHSA attention mechanism. The computational redundancy was reduced by applying single-head attention to some input channels to enhance the efficiency and detection accuracy of the model. A small object detection layer was then introduced into the head layer to correct the easily missed and false detections of small objects. Two rounds of structured pruning were implemented to reduce the number of model parameters, avoid overfitting, and improve the average precision. Before training, data augmentation techniques such as brightness enhancement and gamma contrast adjustment were employed to enhance the generalization ability of the model. RESULTS: The experiment results showed that the optimized YOLOv11-TDSP model achieved an accuracy of 94.5%, a recall rate of 92.3%, and an average precision of 95.8% for detecting dental abnormalities. Compared with the baseline model YOLOv11n, these metrics were improved by 6.9%, 7.4%, and 5.6%, respectively. The number of parameters and computational cost of the YOLOv11-TDSP model were only 12% and 13% of those of the high-precision YOLOv11x model, respectively. CONCLUSIONS The lightweight YOLOv11-TDSP model is capable of highly accurate identification of various dental diseases on panoramic oral X-ray images.
Radiography, Panoramic/methods* ; Humans ; Algorithms ; Tooth Abnormalities/diagnostic imaging*

Objective:To deeply explore the thematic needs and characteristics of learners regarding course elements based on the review texts of nursing massive open online courses (MOOC), providing a reference for achieving effective alignment between digital nursing education content and learner needs.Methods:Data were collected from the review texts of 112 nursing courses on the Chinese University MOOC platform using a web crawler script written with Python′s Requests library. The collection period spanned from the course launch dates to December 31, 2023. Sentiment analysis and high-frequency words analysis were conducted using Chinese text Nature language processing library, and core themes of learners′ positive and negative reviews were extracted using the latent dirichlet allocation.Results:A corpus of 18 184 nursing MOOC review texts was constructed, with positive sentiment reviews accounting for 89.30% (16 238/18 184) and negative sentiment reviews making up 10.70% (1 946/18 184). Word frequency analysis revealed that most nursing MOOC serve as carriers for blended online and offline teaching models, with students being the primary target audience, though social participants were also involved. The reviews effectively mirrored real-world clinical nursing scenarios. The need of learners was categorized into three major themes: content design and assessment, course resources and teaching strategies, and software applications and platform functionality.Conclusions:This study, leveraging text mining technology, thoroughly investigated the three thematic characteristics of nursing MOOC needs of online learners and proposed targeted optimization recommendations. Future research could incorporate other online teaching platforms and comprehensively construct a sentiment lexicon for nursing online course reviews using big data modeling and machine learning algorithms. These would enable a holistic analysis of the digital nursing education landscape, allowing for precise improvements to address existing shortcomings.

Objective:To deeply explore the thematic needs and characteristics of learners regarding course elements based on the review texts of nursing massive open online courses (MOOC), providing a reference for achieving effective alignment between digital nursing education content and learner needs.Methods:Data were collected from the review texts of 112 nursing courses on the Chinese University MOOC platform using a web crawler script written with Python′s Requests library. The collection period spanned from the course launch dates to December 31, 2023. Sentiment analysis and high-frequency words analysis were conducted using Chinese text Nature language processing library, and core themes of learners′ positive and negative reviews were extracted using the latent dirichlet allocation.Results:A corpus of 18 184 nursing MOOC review texts was constructed, with positive sentiment reviews accounting for 89.30% (16 238/18 184) and negative sentiment reviews making up 10.70% (1 946/18 184). Word frequency analysis revealed that most nursing MOOC serve as carriers for blended online and offline teaching models, with students being the primary target audience, though social participants were also involved. The reviews effectively mirrored real-world clinical nursing scenarios. The need of learners was categorized into three major themes: content design and assessment, course resources and teaching strategies, and software applications and platform functionality.Conclusions:This study, leveraging text mining technology, thoroughly investigated the three thematic characteristics of nursing MOOC needs of online learners and proposed targeted optimization recommendations. Future research could incorporate other online teaching platforms and comprehensively construct a sentiment lexicon for nursing online course reviews using big data modeling and machine learning algorithms. These would enable a holistic analysis of the digital nursing education landscape, allowing for precise improvements to address existing shortcomings.

Objective:To explore the predictive value of the prognostic nutritional index (PNI) in concurrently infected patients with acute-on-chronic liver failure (ACLF).Methods:220 cases with ACLF diagnosed and treated at the First Affiliated Hospital of Xi'an Jiaotong University from January 2011 to December 2016 were selected. Patients were divided into an infection and non-infection group according to whether they had co-infections during the course of the disease. Clinical data differences were compared between the two groups of patients. Binary logistic regression analysis was used to screen out influencing factors related to co-infection. The receiver operating characteristic curve was used to evaluate the predictive value of PNI for ACLF co-infection. The measurement data between groups were compared using the independent sample t-test and the Mann-Whitney U rank sum test. The enumeration data were analyzed using the Fisher exact probability test or the Pearson χ2 test. The Pearson method was performed for correlation analysis. The independent risk factors for liver failure associated with co-infection were analyzed by multivariate logistic analysis. Results:There were statistically significant differences in ascites, hepatorenal syndrome, PNI score, and albumin between the infection and the non-infection group ( P ?

【Objective】 To analyze the independent influencing factors of repeated extracorporeal shock wave lithotripsy (ESWL) in the treatment of upper urinary calculi (UUC), based on which a nomogram model was established to predict the efficacy. 【Methods】 Clinical and imaging data of 203 patients receiving repeated ESWL during Jan.2020 and Dec.2022 were collected, including 117 cases in the successful group and 86 cases in the unsuccessful group.The patients’ age and sex, stone volume (SV), surface area (SA), skin-to-site distance (SSD), maximum CT value, mean stone density (MSD), and stone heterogeneity index (SHI) were compared between the two groups.The independent predictors were analyzed with logistic regression and the meaningful variables (P<0.05) were used to establish a nomogram.The efficacy of the model was evaluated using receiver operating characteristic (ROC) curve and decreasing curve analysis (DCA).Internal validation was also performed. 【Results】 Stepwise regression showed that SV, SSD, MSD and SHI were independent influencing factors (P<0.05).The area under the ROC curve (AUC), optimal threshold, sensitivity and specificity were 0.793 (95%CI: 0.674-0.911), 0.619, 77.1% and 74.0%, respectively.The DCA curve was above two extreme curves.Hosmer-Lemeshow test and calibration curve showed that the nomogram had a good fitting degree (χ²=5.526, P=0.489), and the correction C-index was 0.746. 【Conclusion】 SV, SSD, MSD and SHI are independent predictors of the efficacy of repeated ESWL in the treatment of UUC.The nomogram established based on the above indicators has good predictive efficiency and clinical applicability.

Objective:To evaluate the diagnostic accuracy of a noninvasive physiological parameter-based early sepsis screening model for elderly patients in comparison to the systemic inflammatory response syndrome(SIRS)and quick sequential organ failure assessment(qSOFA)scores.Methods:A retrospective study was conducted using data from the Medical Information Mart for Intensive Care Ⅳ(MIMIC-Ⅳ)database.The study focused on patients who were admitted to the intensive care unit(ICU)within 24 hours and were categorized into septic and non-septic groups based on the presence or absence of sepsis.Baseline data and patient outcomes were recorded.Additionally, the SIRS score and qSOFA scores within 24 hours of ICU admission were calculated.Physiological parameters that showed statistical significance in the univariate analysis included respiratory rate, heart rate, level of consciousness, body temperature, systolic blood pressure, and urine output.These parameters were then included in Logistic regression models.The specificity and sensitivity of the regression model for sepsis screening were calculated, and receiver operating characteristic(ROC)curves were plotted.The areas under the ROC curves(AUCs)of the screening model, SIRS, and qSOFA scoring systems were compared.Results:A total of 53 150 ICU hospitalization records were screened, and 23 681 patients with infection or suspected infection within 24 hours were included.Among them, 18 277 patients had sepsis.The 28-day mortality rate for septic patients was higher compared to non-septic patients(13.5% vs.5.1%, χ2=285.131, P<0.001).The baseline data within 24 hours showed significant differences between the two groups in terms of heart rate, respiratory rate, body temperature, state of consciousness, 24-hour urine output, and systolic blood pressure(all P<0.001).These variables were included in the regression equation: ∑β iX i=2.055+ 0.285(temperature: 0/1)+ 0.172(respiratory rate: 0/1)+ 0.073(heart rate: 0/1)+ 1.204(mental status: 0/1)-0.022(systolic blood pressure)+ 0.227(classification of urine output: 0/1/2), P=1/[1+ EXP(-∑β iX i)].The regression model diagnosed sepsis ROC area in young and middle-aged patients as 0.726(95% CI: 0.718 to 0.735), which was significantly higher than the SIRS score(0.585, 95% CI: 0.576 to 0.595)and the qSOFA score(0.676, 95% CI: 0.667 to 0.685)(both P<0.001).In elderly patients, the regression model diagnosed sepsis ROC area as 0.671(95% CI: 0.663 to 0.679), which was also significantly higher than the SIRS score(0.572, 95% CI: 0.563 to 0.580)and the qSOFA score(0.631, 95% CI: 0.623 to 0.639)(both P<0.001). Conclusions:The early sepsis diagnosis model, which utilizes noninvasive physiological parameters, has shown higher accuracy when compared to the SIRS and qSOFA scores.However, it is important to note that its accuracy is lower in elderly patients as compared to young and middle-aged patients.This indicates the necessity for further optimization of the model in order to improve its performance in diagnosing sepsis in the elderly.

ObjectiveTo investigate the influencing factors for the prognosis of patients with acute－on－chronic liver failure （ACLF）， and to establish a short－term prognostic model. MethodsA retrospective analysis was performed for the baseline clinical data of 247 patients with ACLF who were hospitalized in Department of Infectious Diseases， The First Affiliated Hospital of Xi’an Jiaotong University， from January 2011 to December 2016， and the patients were divided into survival group and death group. The two groups were compared to identify the influencing factors for prognosis； a prognostic model was established， and the receiver operating characteristic （ROC） curve was used to assess its predictive efficacy and determine the optimal cut－off value. The independent－samples t test was used for comparison of normally distributed continuous data between groups， and the Mann－Whitney U rank sum test was used for comparison of non－normally distributed continuous data between groups； the Fisher’s exact test or the Pearson’s chi－square test was used for comparison of categorical data between groups. The univariate and multivariate logistic regression analyses were used to investigate the independent risk factors for 28－ and 90－day prognosis， and the Kaplan－Meier method was used to plot the 28－day survival curves. ResultsA total of 220 patients with ACLF were included based on the inclusion and exclusion criteria； there were 148 patients in the 28－day survival group and 72 patients in the 28－day death group， with a 28－day transplantation－free survival rate of 67.27%； there were 115 patients in the 90－day survival group and 105 patients in the 90－day death group， with a 90－day transplantation－free survival rate of 52.27%. The logistic regression analysis showed that female sex （odds ratio ［OR］=2.149， P=0.030）， high Model for End－Stage Liver Disease （MELD） score （OR=1.120， P<0.001）， and low lymphocyte count （OR=0.411， P=0.002） were independent risk factors for 28－day prognosis， and an LS－MELD model for 28－day prognosis was established as Logit （28－day prognosis）=－3.432+0.765×sex－0.890×lymphocyte count×10^－9+0.113×MELD（1 for male sex and 2 for female sex）. The ROC curve analysis showed that this model had an optimal cut－off value of 0.35， and then the patients were divided into low LS－MELD group （≤0.35） and high LS－MELD group （>0.35）； the low LS－MELD group had a significantly higher 28－day survival rate than the high LS－MELD group （P<0.001）. ConclusionPeripheral blood lymphocyte count combined with sex and MELD score has a certain value in predicting the short－term prognosis of ALCF patients.

Objective:To explore the efficacy and safety of adalimumab (ADA) in Chinese Takayasu arteritis (TAK) patients.Methods:This study prospectively included ten refractory active TAK patients. The clinical, laboratory and imaging data of these patients before and after ADA treatment were collected and compared. Statistical analysis was carried out with the statistical product and service solutions (SPSS) program V.24.0. We performed univariate analysis by Wilcoxon-Mann-Whitney test for numerical data according to normality. The P value <0.05 was interpreted as statistically significant. Results:Ten TAK patients recruited in this study were all female, with a median age of 27.5 [25.0, 36.0] years. Fever or fatigue (6/10, 60%), neck pain (6/10, 60%), and limb claudication (4/10, 40%) were the most common clinical manifestations. The most common pattern of vascular involvement was Numano type V (7/10, 70%). The serum erythrocyte sedimentation rate (ESR) [22.0(4.5, 37.5) mm/1 h vs 7.0(4.5, 31.0) mm/1 h, Z=-2.04, P=0.042] and IL-6 level [12.3(3.6, 57.7) ng/L vs 2.8(2.0, 24.0) ng/L, Z=-2.19, P=0.028] reduced significantly after ADA treatment. The thickness of the common carotid artery intima decreased from 2.7(2.2, 4.0) mm to 2.3(1.6, 3.4) mm after ADA treatment ( Z=-1.99, P=0.046). The lumen diameter of the common carotid artery increased from 0.31(0.20, 0.42) cm to 0.37(0.29, 0.43) cm ( Z=-2.02, P=0.043). The glucocorticoid dosage was reduced from [11.2(6.9, 15.0) mg to 10.0(5.0, 13.1) mg, Z=-1.89, P=0.059]. In the 10 patients, 6 patients (60%) reached complete response, 2 patients (20%) reached the partial response, and the total effective rate was 80%(8/10). Pneumonia in one TAK patient was recorded. Conclusion:The results of this study suggest, that ADA can be used for refractory active TAK patients, and is safe.

Objective: To evaluate the effectiveness of health belief model-based health education intervention in improving blood pressure control of patients with hypertension in community settings. Methods: From September 2016 to September 2017, 400 newly diagnosed patients with hypertension were recruited from 6 community healthcare centers with comparable population size and health services in the Shunyi District of Beijing. All community healthcare centers were randomly assigned to the intervention group (206 patients) and the control group (194 patients). Patients in the intervention group received 3 lectures (20-30 min for each) of health belief model-based health education. Patients in the control group received usual care. The basic characteristics, health beliefs, and health literacy were collected, and blood pressure was measured before and after the intervention, respectively. The difference-in-difference model was used to analyze the change of blood pressure and the influencing factors between two groups before and after the intervention. Results: A total of 134 patients in the intervention group and 129 patients in the control group completed the study. After adjusting for the age, gender, family income, medical insurance, chronic diseases and family history, the score of perceived barriers was increased by 1.65 (P=0.016), and perceived seriousness was decreased by 0.73 (P=0.018). The systolic blood pressure of patients was decreased by 7.37 mmHg (1 mmHg=0.133 kPa, P=0.001) and diastolic blood pressure was decreased by 4.07 mmHg (P=0.014), respectively. The β (95%CI) values were -7.37 (-11.88,-2.86) and -4.07 (-7.30, -0.84). The perceived susceptibility and self-efficacy had a significant influence on the blood pressure of patients (P<0.05). Conclusion Health belief model-based health education intervention could significantly improve the blood pressure control of patients with hypertension in the community settings.

The original version of this article unfortunately contained some mistakes.