OBJECTIVES: To construct an intelligent analysis model for classifying fetal orientation during intrapartum ultrasound videos based on multi-feature fusion. METHODS: The proposed model consists of the Input, Backbone Network and Classification Head modules. The Input module carries out data augmentation to improve the sample quality and generalization ability of the model. The Backbone Network was responsible for feature extraction based on Yolov8 combined with CBAM, ECA, PSA attention mechanism and AIFI feature interaction module. The Classification Head consists of a convolutional layer and a softmax function to output the final probability value of each class. The images of the key structures (the eyes, face, head, thalamus, and spine) were annotated with frames by physicians for model training to improve the classification accuracy of the anterior occipital, posterior occipital, and transverse occipital orientations. RESULTS: The experimental results showed that the proposed model had excellent performance in the tire orientation classification task with the classification accuracy reaching 0.984, an area under the PR curve (average accuracy) of 0.993, and area under the ROC curve of 0.984, and a kappa consistency test score of 0.974. The prediction results by the deep learning model were highly consistent with the actual classification results. CONCLUSIONS The multi-feature fusion model proposed in this study can efficiently and accurately classify fetal orientation in intrapartum ultrasound videos.
Humans ; Female ; Ultrasonography, Prenatal/methods* ; Pregnancy ; Fetus/diagnostic imaging* ; Neural Networks, Computer ; Video Recording

Objective To explore the role of an artificial intelligence(AI)model based on real-time object detection network in fetal facial ultrasound examination.Methods With the normal fetal facial ultrasound standard plane(FFUSP)at 20-24 weeks of gestation as the research object,a FFUSP recognition model based on real-time object detection network was constructed.The recognition accuracy of the model for FFUSP and the anatomical structures were analyzed,and the clinical value was evaluated by analyzing its performance in identifying FFUSP in 119 cases of fetal ultrasound images.Results The overall precision,recall rate,mAP@.5 and mAP@.5:.95 of the AI model were 97.8%,98.5%,98.1%and 61.0%,respectively.The clinical validation showed that the AI model had a sensitivity,specificity,positive predictive value,negative predictive value and accuracy of 100.0%,98.5%,87.4%,100.0%and 98.7%for facial anatomy recognition,and the results were highly consistent with the classification of fetal ultrasound experts(k=0.925,P<0.001).The recognition accuracy of the model for 3 types of standard planes reached 100%;and the average speed of dynamic video detection was 33.93 frames per second.Conclusion The FFUSP recognition model based on real-time object detection network exhibits excellent performance,and it can be applied to real-time ultrasound diagnosis,teaching and intelligent quality evaluation.

Objective:To investigate the risk factors associated with mortality in patients with severe traumatic liver injury (TLI) and to establish and validate an early prediction model for mortality.Methods:A retrospective cohort study was conducted to analyze the clinical data of 273 patients with severe TLI admitted to the ICU from the medical information mart for the intensive care-IV (MIMIC-IV) database. The cohort consisted of 176 males and 97 females, with age ranging from 18 to 83 years [35.6 years(25.7,57.5)years]. The patients were divided into two groups based on in-hospital mortality: the survival group (253 patients, 92.7%) and the death group (20 patients, 7.3%). The two groups were compared with regards to gender, age, cause and type of injury, treatment method, massive blood transfusion, comorbidities as well as vital signs and laboratory tests measured within 24 hours of ICU admission. Univariate analysis was used to screen for risk factors associated with mortality in severe TLI patients. Independent risk factors for mortality were determined using multivariate Logistic regression analysis. Lasso regression was used to screen for predictors of mortality, and a nomogram prognostic model was then established through a multivariate Logistic regression analysis. The area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the discrimination of the model, while the Hosmer-Lemeshow goodness-of-fit test and calibration curve were used to evaluate the calibration of the model. The model′s clinical applicability was evaluated through decision curve analysis (DCA). Internal validation was performed by the 200 Bootstrap samples, and external validation was performed by using 163 patients with severe TLI from the emergency ICU collaborative research database (eICU-CRD). Finally, the predictive efficacy of the nomogram model was compared to other trauma or severity scores.Results:Univariate analysis showed that the age, cause of injury, massive blood transfusion, chronic liver disease and laboratory tests measured within 24 hours of ICU admission, including temperature, systolic blood pressure, diastolic blood pressure, mean arterial pressure, shock index, platelets, red blood cell distribution width (RDW), mean red blood cell hemoglobin concentration (MCHC), blood glucose, blood urea nitrogen, creatinine, anion gap, bicarbonate, prothrombin time (PT), activated partial thromboplastin time (APTT) and international normalized ratio (INR) were associated with the mortality of severe TLI patients ( P<0.05 or 0.01). Multivariate Logistic regression analysis revealed that age ( OR=1.08, 95% CI 1.03, 1.12, P<0.01), body temperature <36 ℃ ( OR=8.00, 95% CI 2.17, 29.53, P<0.01), shock index ( OR=9.59, 95% CI 1.76, 52.18, P<0.01) and anion gap ( OR=1.32, 95% CI 1.15, 1.53, P<0.01) were significantly associated with mortality in severe TLI patients. Lasso regression analysis selected 7 predictors, including age, body temperature<36 ℃, shock index, anion gap, chronic liver disease, creatinine and APTT. Based on these 7 predictors, a nomogram prediction model was developed. The AUC of the nomogram for predicting mortality was 0.96 (95% CI 0.94, 0.99), and the Hosmer-Lemeshow goodness-of-fit test indicated a good fit ( P>0.05). The calibration curve demonstrated excellent consistency between the predicted and actual probabilities, and DCA demonstrated that the model had good clinical net benefit at all risk threshold probability ranges. Internal validation confirmed the stability of the model ( AUC=0.96, 95% CI 0.92, 0.98), and external validation demonstrated good generalization ability ( AUC=0.95, 95% CI 0.91, 0.98). Moreover, the nomogram exhibited superior predictive efficacy compared with injury severity score (ISS), revised trauma score (RTS), trauma injury severity score (TRISS), sequential organ failure score (SOFA), acute physiological score III (APS III), Logistic organ dysfunction score (LODS), Oxford acute severity of illness score (OASIS) and simplified acute physiological score II (SAPS II). Conclusions:Age, body temperature <36 ℃, shock index and anion gap are independent risk factors for mortality in severe TLI patients. A nomogram prognosis model based on 7 predictors, namely age, body temperature <36 ℃, shock index, anion gap, chronic liver disease, creatinine and APTT exhibits good predictive efficacy and robustness, and is contributive to accurately assess the risk of mortality in severe TLI patients at an early stage.

Objective:To compare the prediction efficiency of traditional linear regression model and four machine learning models on the learning behavior of clinical medical postgraduates, and to explore the pros and cons and applicability of different prediction models.Methods:A total of 6,922 clinical medical postgraduates were surveyed, their comprehensive learning behavior scores were obtained through the learning behavior scale. In the training set, Lasso linear regression and artificial neural network, decision tree, Bootstrap random forest, and lifting tree were used to build prediction models respectively. The above models were used to predict the validation set data and compare the prediction efficiency.Results:The comprehensive learning behavior score of clinical medical postgraduates was (3.31±0.54) points, and the overall compliance rate was 74.02%. In the linear regression model, the influence of age, school level, degree type, learning interest, pressure and satisfaction on learning behavior were statistically significant. In the prediction of validation set, the sensitivity, specificity, and accuracy of the linear regression model were 0.484, 0.914, and 0.801, respectively. The indexes of the four machine learning models were higher than those of the traditional linear regression model, and the Bootstrap random forest had the highest elevation.Conclusion:The linear regression model has a good prediction effect on learning behavior, and machine learning is superior to linear regression model in terms of accuracy of prediction. However, traditional linear regression models are superior to machine learning models in computational efficiency and interpretability.

Objective:To review the development and progress of geriatric research in China from the perspective of funding from the National Natural Science Foundation of China, and to provide supporting data for further development of the field.Methods:Based on project funding data of the National Natural Science Foundation of China for geriatric medicine from 2008 to 2017, the number and composition of projects, funding amount and funded institutions were statistically analyzed, and the overall trend of change was examined.The research hotspots in recent years were summarized by using word frequency analysis.Results:From 2008 to 2017, a total of 446 projects were approved in the field of geriatric medicine.Both the number of projects and the amount of funding showed an increasing trend.Project types were gradually expanded and the composition was constantly optimized.Since 2014, the fund types each year included 6 or more and breakthroughs in major projects were achieved.East China remained the leading region in the number of projects and amount of funding, with Beijing, Shanghai and Zhejiang receiving the highest numbers of project approvals and accounting for more than 10% of the national total in the number of project approvals and the amount of funding.During this period, 114 institutions received funding, with 74 winning 2 or more projects.In word frequency analysis, senescence(229 times), cell(161 times)and protein(120 times)were the three most frequently used keywords.Conclusions:In the field of geriatric medicine, the project approval level from the National Natural Science Foundation of China has significantly improved in quantity and quality, but it has always been under-represented in the branch of medical sciences.The regional and institutional distribution of projects is unbalanced and tends to skew in favor of institutions with prominent academic disciplines.Basic research is the main type of funded projects.Geriatric research is centered on the mechanisms of aging.

Objective:To investigate the clinicopathological features, differential diagnosis and prognosis of Merkel cell carcinoma (MCC).Methods:The clinical and pathological data of 10 patients with MCC were collected at the 940th Hospital of PLA. The histological characteristics were examined. Immunohistochemical EnVision method was used to detect thyroid transcription factor-1 (TTF1), broad-spectrum cytokeratin (CKpan), CK20, S-100, Ki-67, CD56, chromogranin A, synaptophysin and other markers in the 10 cases.Results:Intradermal MCC of the skin showed a nested, cord-like, cribriform distribution, polygonal cells, uniform size, and lack of cytoplasm. Tumor cell nuclei were large and round, with clear nuclear membranes, fine and scattered chromatin, absence of nucleoli, and mitotic figures of 10 per 50 high power fields. Among them, one patient had sarcoma and squamous cell carcinoma in situ, one patient had squamous cell carcinoma in situ, and one patient had unique cell morphology. Immunohistochemical staining showed that all cancer cells expressed CKpan, synaptophysin and CD56. There were seven cases with perinuclear dot-like positivity of CK20. Six MCCs expressed chromogranin A to varying degrees, while 2 MCCs were weakly positive for p63. The nuclear positive index in the Ki-67 hotspot area was 60%.Conclusion:The histology of MCC varies. Rendering a correct diagnosis of MCC requires adequate sampling, close correlation with clinical history and rational use of immunohistochemical staining. The treatment requires standardized surgery, postoperative radiotherapy and multimodal chemotherapy. Immunotherapy may replace the traditional treatment in the future.

Objective:To study the risk factors for ipsilateral severe hearing impairment in patients with hemifacial spasm (HFS) after microvascular decompression (MVD).Methods:MVD was performed in 3700 patients with HFS, admitted to our hospital from October 2007 to August 2020; according to the existence of ipsilateral severe hearing impairment, these patients were divided into severe hearing impairment group and non-severe hearing impairment group. The clinical data of these patients were compared. Multivariate linear regression analysis was used to determine the independent influencing factors for ipsilateral severe hearing impairment.Results:Forty-five patients (1.2%) had ipsilateral severe hearing impairment after MVD; no one got recovery of hearing impairment during the follow-up period (0.6-11.8 years, 6.3 years in average). As compared with those in the non-severe hearing impairment group, patients in the severe hearing impairment group had significantly older age, significantly higher percentages of male patients, and patients with left HFS, hypertension, and diabetes mellitus, statistically higher percentage of patients having small posterior fossa volume, arachnoid thickening and adhesion, and vertebral artery compression, significantly lower percentage of patients with anterior inferior cerebellar artery compression, significantly higher percentage of patients with arteriosclerosis of offending arteries and difficult decompression ( P<0.05). Multivariate linear regression analysis revealed that hypertension, vertebral artery compression, arteriosclerosis of offending artery and difficult decompression were independent risk factors for severe hearing impairment in patients with HFS after MVD. Conclusion:It's difficult to get recovery for severe hearing impairment in patients with HFS after MVD; this complication is much common in patients with hypertension, vertebral artery compression, arteriosclerosis of offending artery or difficult decompression.

Objective:To investigate the correlation between neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) and central lymph node metastasis (CLNM) in patients with cN0 papillary thyroid microcarcinoma (PTMC).Methods:The clinicopathological data of PTMC patients confirmed by surgery and pathology in the 81 st Military Hospital of People′s Liberation Army from 2016 to 2019 were collected, and the relationship between preoperative NLR, PLR levels and postoperative PTMC CLNM were analyzed. Logistic regression analysis was used for multivariate analysis. Receiver operating characteristic (ROC) curve was used to determine the cutoff value of NLR and PLR. The interaction relative excess risk was used to analyze the relationship between NLR, PLR and CLNM. Results:Among 220 patients with cN0 stage PTMC, 92 were CLNM. The ROC curve showed that when the cutoff value of NLR was 2.5 and the cutoff value of PLR was 175, the highest Youden index was 0.318 and 0.264, respectively. NLR and PLR were both related to CLNM ( P<0.05). The tumor long diameter, multifocality, NLR≥2.5 and PLR≥175 were independent impact factors of CLNM ( P<0.05). The results of the interaction showed that the relative excess risk of the interaction was 5.531 (95% CI: 0.160, 10.901, P=0.016), the attribution ratio was 0.512 (95% CI: 0.230, 0.794, P=0.009), and the synergy index was 2.294 (95% CI: 1.492, 4.579, P=0.022), suggested that NLR and PLR had an interactive effect, and these two synergistically promoted CLNM. Conclusions:NLR and PLR are independent risk factors for cN0 stage PTMC CLNM. When NLR≥2.5 and PLR≥175, preventive central lymph node dissection should be routinely performed.

Objective:To investigate the correlation between neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) and central lymph node metastasis (CLNM) in patients with cN0 papillary thyroid microcarcinoma (PTMC).Methods:The clinicopathological data of PTMC patients confirmed by surgery and pathology in the 81 st Military Hospital of People′s Liberation Army from 2016 to 2019 were collected, and the relationship between preoperative NLR, PLR levels and postoperative PTMC CLNM were analyzed. Logistic regression analysis was used for multivariate analysis. Receiver operating characteristic (ROC) curve was used to determine the cutoff value of NLR and PLR. The interaction relative excess risk was used to analyze the relationship between NLR, PLR and CLNM. Results:Among 220 patients with cN0 stage PTMC, 92 were CLNM. The ROC curve showed that when the cutoff value of NLR was 2.5 and the cutoff value of PLR was 175, the highest Youden index was 0.318 and 0.264, respectively. NLR and PLR were both related to CLNM ( P<0.05). The tumor long diameter, multifocality, NLR≥2.5 and PLR≥175 were independent impact factors of CLNM ( P<0.05). The results of the interaction showed that the relative excess risk of the interaction was 5.531 (95% CI: 0.160, 10.901, P=0.016), the attribution ratio was 0.512 (95% CI: 0.230, 0.794, P=0.009), and the synergy index was 2.294 (95% CI: 1.492, 4.579, P=0.022), suggested that NLR and PLR had an interactive effect, and these two synergistically promoted CLNM. Conclusions:NLR and PLR are independent risk factors for cN0 stage PTMC CLNM. When NLR≥2.5 and PLR≥175, preventive central lymph node dissection should be routinely performed.

The present clinical doctorate system has some irrationalities in the connection with professional qualification . According to analyses of main problems in recruitment , training objective , curriculum design and clinical examination, we proposed some suggestions for carrying forward the comprehensive pilot system: clarify the training objective and improve the overall understanding of reform;improve the examination condition and standardize the recruitment process; improve the curriculum system and reform the teaching method; strengthen the way of quality examination and supervision; attach more attention on constructing tutor teams and explore the way of joint guidance.