PURPOSE: Septic shock is associated with high mortality and poor outcomes among sepsis patients with coagulopathy. Although traditional statistical methods or machine learning (ML) algorithms have been proposed to predict septic shock, these potential approaches have never been systematically compared. The present work aimed to develop and compare models to predict septic shock among patients with sepsis. METHODS: It is a retrospective cohort study based on 484 patients with sepsis who were admitted to our intensive care units between May 2018 and November 2022. Patients from the 908th Hospital of Chinese PLA Logistical Support Force and Nanchang Hongdu Hospital of Traditional Chinese Medicine were respectively allocated to training (n=311) and validation (n=173) sets. All clinical and laboratory data of sepsis patients characterized by comprehensive coagulation indexes were collected. We developed 5 models based on ML algorithms and 1 model based on a traditional statistical method to predict septic shock in the training cohort. The performance of all models was assessed using the area under the receiver operating characteristic curve and calibration plots. Decision curve analysis was used to evaluate the net benefit of the models. The validation set was applied to verify the predictive accuracy of the models. This study also used Shapley additive explanations method to assess variable importance and explain the prediction made by a ML algorithm. RESULTS: Among all patients, 37.2% experienced septic shock. The characteristic curves of the 6 models ranged from 0.833 to 0.962 and 0.630 to 0.744 in the training and validation sets, respectively. The model with the best prediction performance was based on the support vector machine (SVM) algorithm, which was constructed by age, tissue plasminogen activator-inhibitor complex, prothrombin time, international normalized ratio, white blood cells, and platelet counts. The SVM model showed good calibration and discrimination and a greater net benefit in decision curve analysis. CONCLUSION The SVM algorithm may be superior to other ML and traditional statistical algorithms for predicting septic shock. Physicians can better understand the reliability of the predictive model by Shapley additive explanations value analysis.
Humans ; Shock, Septic/blood* ; Machine Learning ; Male ; Female ; Retrospective Studies ; Middle Aged ; Aged ; Sepsis/complications* ; ROC Curve ; Cohort Studies ; Adult ; Intensive Care Units ; Algorithms ; Blood Coagulation ; Critical Illness

PURPOSE: To screen laboratory markers with predictive value in early spinal surgical site infections (SSI) that are diagnosed within 30 days postoperatively. METHODS: Patients who underwent surgical treatment for internal spinal fixation between March 2022 and March 2023 in our hospital were retrospectively studied. The inclusion criteria were aged >18 years, undergoing internal fixation surgery, complete medical records with >30 days of postoperative follow-up, diagnosis was made within 30 days postoperatively, and an informed consent form was obtained. The exclusion criteria were abnormal white blood cell count or neutrophil percentage in the preoperative blood routine and combined diseases that may affect the C-reactive protein (CRP) or procalcitonin (PCT) values, including lower respiratory tract infection, renal insufficiency, and liver disease. We collected patients' personal information, surgical information, and blood laboratory data, including CRP, PCT, lymphocyte-neutrophil ratio, platelet-neutrophil ratio, and routine blood tests on preoperative and postoperative days 3, 5, and 7, from these patients. These data were statistically analyzed to determine which laboratory markers were statistically significant. The diagnostic value and optimal diagnostic threshold of these laboratory markers were further determined by receiver operating characteristic curve analysis. RESULTS: A total of 106 patients were enrolled in this study, of whom 8 patients were diagnosed with early SSI. A total of 4 laboratory markers were screened, namely, CRP on postoperative day 7 (optimal diagnostic threshold of ≥64.1 mg/L, sensitivity of 100%, specificity of 76.5%, area under the curve (AUC) of 0.908), PCT on postoperative day 7 (optimal diagnostic threshold of ≥0.2 ng/mL, sensitivity of 87.5%, specificity of 94.1%, AUC of 0.967), lymphocyte count on postoperative day 5 (optimal diagnostic threshold of ≤0.67 × 10⁹/L, sensitivity of 50%, specificity of 95.9%, AUC of 0.760), and lymphocyte count on postoperative day 7 (optimal diagnostic threshold of ≤1.32 × 10⁹/L, sensitivity of 87.5%, specificity of 55.1%, AUC of 0.721). CONCLUSION We concluded that CRP and PCT levels on postoperative day 7 and lymphocyte counts on postoperative days 5 and 7 are useful markers in screening for early spinal SSI.
Humans ; Retrospective Studies ; Male ; Female ; Biomarkers/blood* ; Middle Aged ; C-Reactive Protein/analysis* ; Surgical Wound Infection/blood* ; Procalcitonin/blood* ; Adult ; Aged ; Postoperative Period ; ROC Curve ; Predictive Value of Tests ; Spine/surgery*

PURPOSE: To analyze risk factors for severe trauma and establish a nomogram for early risk prediction, to improve the early identification of severe trauma. METHODS: This study was conducted on the patients treated in 81 trauma treatment institutions in Gansu province from 2020 to 2022. Patients were grouped by year, with 5364 patients from 2020 to 2021 as the training set and 1094 newly admitted patients in 2020 as the external validation set. Based on the injury severity score (ISS), patients in the training set were classified into 2 subgroups of the severe trauma group (n = 478, ISS scores ≥25) and the non-severe trauma group (n = 4886, ISS scores <25). Univariate and binary logistic regression analyses were employed to identify independent risk factors for severe trauma. Subsequently, a predictive model was developed using the R software environment. Furthermore, the model was subjected to internal and external validation via the Hosmer-Lemeshow test and receiver operating characteristic curve analysis. RESULTS: In total, 6458 trauma patients were included in this study. Initially, this study identified several independent risk factors for severe trauma, including multiple traumatic injuries (polytrauma), external hemorrhage, elevated shock index, elevated respiratory rate, decreased peripheral oxygen saturation, and decreased Glasgow coma scale score (all p < 0.05). For internal validation, the area under the receiver operating characteristic curve was 0.914, with the sensitivity and specificity of 88.4% and 87.6%, respectively; while for external validation, the area under the receiver operating characteristic curve was 0.936, with the sensitivity and specificity of 84.6% and 93.7%, respectively. In addition, a good model fitting was observed through the Hosmer-Lemeshow test and calibration curve analysis (p > 0.05). CONCLUSION This study establishes a nomogram for early risk prediction of severe trauma, which is suitable for primary healthcare institutions in underdeveloped western China. It facilitates early triage and quantitative assessment of trauma severity by clinicians prior to clinical interventions.
Humans ; Nomograms ; Male ; Female ; Wounds and Injuries/diagnosis* ; Risk Factors ; Middle Aged ; Adult ; Injury Severity Score ; Risk Assessment ; ROC Curve ; Aged ; Logistic Models ; China ; Glasgow Coma Scale

Parameters of peripheral blood cell have been shown as the potential predictors of erectile dysfunction (ED). To investigate the clinical significance of hematological parameters for predicting the risk of rapid ejaculation, we established a rat copulatory model on the basis of ejaculation distribution theory. Blood samples from different ejaculatory groups were collected for peripheral blood cell counts and serum serotonin (5-HT) tests. Meanwhile, the relationship between hematological parameters and ejaculatory behaviors was assessed. Final analysis included 11 rapid ejaculators, 10 normal ejaculators, and 10 sluggish ejaculators whose complete data were available. The platelet (PLT) count in rapid ejaculators was significantly lower than that in normal and sluggish ejaculators, whereas the platelet distribution width (PDW) and mean platelet volume (MPV) were significantly greater in rapid ejaculators. Multivariate logistic regression analysis and receiver operating characteristic (ROC) curve analysis showed that the PLT was an independent protective factor for rapid ejaculation. Meanwhile, rapid ejaculators were found to have the lowest serum 5-HT compared to normal and sluggish ejaculators ( P < 0.001). Furthermore, there was a positive correlation between the PLT and serum 5-HT ( r = 0.662, P < 0.001), indicating that the PLT could indirectly reflect the serum 5-HT concentration. In addition, we assessed the association between the PLT and ejaculatory parameters. There was a negative correlation between ejaculation frequency (EF) and the PLT ( r = -0.595, P < 0.001), whereas there was a positive correlation between ejaculation latency (EL) and the PLT ( r = 0.740, P < 0.001). This study indicated that the PLT might be a useful and convenient diagnostic marker for predicting the risk of rapid ejaculation.
Male ; Animals ; Ejaculation/physiology* ; Rats ; Platelet Count ; Pilot Projects ; Serotonin/blood* ; Biomarkers/blood* ; Mean Platelet Volume ; Rats, Sprague-Dawley ; ROC Curve ; Erectile Dysfunction/physiopathology*

We propose a strategy to reduce unnecessary prostate biopsies in Chinese patients with total prostate-specific antigen (tPSA) >10 ng ml -1 and Prostate Imaging Reporting and Data System (PI-RADS) scores between 1 and 3. Clinical data derived from 517 patients of The First Affiliated Hospital of USTC (Hefei, China) from January 2020 to December 2023 who met the screening criteria for the study were retrospectively collected. Independent predictors were identified via univariate and multivariate logistic regression analysis. The diagnostic capacity of clinical variables was evaluated using the receiver operating characteristic (ROC) curves and area under the curve (AUC). A prostate biopsy strategy was developed via risk stratification. Of the 517 patients, 17/348 (4.9%) with PI-RADS 1-2 were diagnosed with clinically significant prostate cancer (csPCa), and 27/169 (16.0%) patients with PI-RADS 3 were diagnosed with csPCa. The appropriate prostate-specific antigen density (PSAD) cut-off values were 0.45 ng ml -2 for PI-RADS 1-2 patients and 0.3 ng ml -2 for PI-RADS 3 patients. The appropriate prostate volume (PV) cut-off values were 40 ml for PI-RADS 1-2 patients and 50 ml for PI-RADS 3 patients. The prostate biopsy strategy based on PSAD and PV developed in this study can reduce unnecessary prostate biopsies in patients with tPSA >10 ng ml -1 and PI-RADS 1-3. In the study, 66.5% (344/517) patients did not need to undergo prostate biopsy, at the expense of missing only 1.7% (6/344) patients with csPCa.
Humans ; Male ; Prostatic Neoplasms/diagnostic imaging* ; Prostate-Specific Antigen/blood* ; Aged ; Middle Aged ; Retrospective Studies ; Prostate/diagnostic imaging* ; Unnecessary Procedures/statistics & numerical data* ; Biopsy/statistics & numerical data* ; China ; ROC Curve

OBJECTIVES: To investigate the risk factors for plastic bronchitis (PB) in children with macrolide-unresponsive Mycoplasma pneumoniae pneumonia (MUMPP) and to establish a nomogram prediction model. METHODS: A retrospective analysis was conducted on 178 children with MUMPP who underwent bronchoscopy from January to December 2023. According to the presence or absence of PB, the children were divided into a PB group (49 children) and a non-PB group (129 children). The predictive factors for the development of PB in children with MUMPP were analyzed, and a nomogram prediction model was established. The model was assessed in terms of discriminatory ability, accuracy, and clinical effectiveness. RESULTS: The multivariate logistic regression analysis showed that older age and higher levels of lactate dehydrogenase and fibrinogen were closely associated with the development of PB in children with MUMPP (P<0.05). A nomogram model established based on these factors had an area under the receiver operating characteristic curve of 0.733 (95%CI: 0.651-0.816, P<0.001) and showed a good discriminatory ability. The Hosmer-Lemeshow goodness-of-fit test indicated that the predictive model had a good degree of fit (P>0.05), and the decision curve analysis showed that the model had a good clinical application value. CONCLUSIONS The risk nomogram model established based on age and lactate dehydrogenase and fibrinogen levels has good discriminatory ability, accuracy, and predictive efficacy for predicting the development of PB in children with MUMPP.
Retrospective Studies ; Risk Factors ; Nomograms ; Mycoplasma pneumoniae/isolation & purification* ; Pneumonia, Mycoplasma/microbiology* ; Bronchitis/microbiology* ; Macrolides/therapeutic use* ; Drug Resistance, Bacterial ; Bronchoscopy ; Area Under Curve ; ROC Curve ; Fibrinogen/analysis* ; Age Factors ; Humans ; Male ; Female ; Infant ; Child, Preschool ; Child ; Adolescent ; L-Lactate Dehydrogenase/blood*

OBJECTIVES: To explore the predictive factors for non-response to intravenous immunoglobulin (IVIG) in children with Kawasaki disease (KD) and to establish an IVIG non-response prediction scoring model for the Sichuan region. METHODS: A retrospective study was conducted by collecting clinical data from children with KD admitted to four tertiary hospitals in Sichuan Province between 2019 and 2023. Among them, 940 children responded to IVIG, while 74 children did not respond. Multivariate logistic regression analysis was used to identify the predictive factors for non-response to IVIG and to establish a predictive scoring model. The model's effectiveness was assessed using the receiver operating characteristic curve (ROC) and validated with an independent dataset. RESULTS: Multivariate logistic regression analysis showed that the platelet-to-lymphocyte ratio (PLR), hemoglobin (Hb), serum creatinine, aspartate aminotransferase (AST), and platelet count (PLT) were closely related to non-response to IVIG in children with KD (P<0.05). Based on these indicators, a predictive scoring model was established: PLR > 199, 0.4 points; Hb ≤ 116 g/L, 4 points; AST > 58 U/L, 0.2 points; serum creatinine > 38 µmol/L, 3.9 points; PLT count ≤ 275 × 10⁹/L, 0.3 points. Using this model, children with KD were scored, and a total score greater than 4.3 was considered high risk of non-response to IVIG. The sensitivity of the model in predicting non-response to IVIG was 77.0%, specificity was 65.7%, and the area under the ROC curve was 0.746 (95%CI: 0.688-0.805). CONCLUSIONS The predictive scoring model based on PLR, Hb, serum creatinine, AST, and PLT demonstrates good predictive performance for non-response to IVIG in children with KD in the Sichuan region and can serve as a reference for clinical decision-making.
Humans ; Mucocutaneous Lymph Node Syndrome/blood* ; Immunoglobulins, Intravenous/therapeutic use* ; Male ; Female ; Retrospective Studies ; Child, Preschool ; Infant ; Logistic Models ; Child ; Platelet Count ; ROC Curve

OBJECTIVES: To study the risk factors for hypoxemia in children with severe Mycoplasma pneumoniae pneumonia (SMPP). METHODS: A retrospective collection of clinical data from children diagnosed with SMPP at the Third Affiliated Hospital of Zhengzhou University from June to December 2023 was conducted. The patients were categorized into hypoxemia and non-hypoxemia groups. Logistic regression analysis was used to assess the risk factors for hypoxemia, and receiver operating characteristic (ROC) curve analysis was employed to analyze the diagnostic performance of various indicators. RESULTS: A total of 113 children with SMPP were included. Univariate logistic regression analysis showed that ferritin, aspartate aminotransferase, creatinine, creatine kinase isoenzyme, lactate dehydrogenase, alpha-hydroxybutyrate dehydrogenase, immunoglobulin G, complement C3, complement C4, age, extrapulmonary complications, and a chest computed tomography (CT) scan showing a bronchiolitis pattern were significant factors for hypoxemia in children with SMPP (P<0.05). Multivariate logistic regression analysis revealed that elevated ferritin levels, presence of extrapulmonary complications, and a bronchiolitis pattern on lung CT were independent risk factors for hypoxemia in these patients (P<0.05). The ROC curve analysis indicated that the combination of these three indicators for predicting hypoxemia had a sensitivity of 71.9%, a specificity of 95.1%, and an area under the curve of 0.888 (95%CI: 0.809-0.968). CONCLUSIONS In children with SMPP, when there are elevated ferritin levels, a bronchiolitis pattern on chest CT, and the presence of extrapulmonary complications, there should be a high level of vigilance for the potential development of hypoxemia.
Humans ; Pneumonia, Mycoplasma/complications* ; Male ; Female ; Risk Factors ; Child, Preschool ; Hypoxia/etiology* ; Retrospective Studies ; Child ; Logistic Models ; Infant ; ROC Curve ; Adolescent

INTRODUCTION: Obstructive sleep apnoea (OSA) is a serious but underdiagnosed condition. Demand for the gold standard diagnostic polysomnogram (PSG) far exceeds its availability. More efficient diagnostic methods are needed, even in tertiary settings. Machine learning (ML) models have strengths in disease prediction and early diagnosis. We explored the use of ML with oximetry, demographic and anthropometric data to diagnose OSA. METHODS: A total of 2,996 patients were included for modelling and divided into test and training sets. Seven commonly used supervised learning algorithms were trained with the data. Sensitivity (recall), specificity, positive predictive value (PPV) (precision), negative predictive value, area under the receiver operating characteristic curve (AUC) and F1 measure were reported for each model. RESULTS: In the best performing four-class model (neural network model predicting no, mild, moderate or severe OSA), a prediction of moderate and/or severe disease had a combined PPV of 94%; one out of 335 patients had no OSA and 19 had mild OSA. In the best performing two-class model (logistic regression model predicting no-mild vs. moderate-severe OSA), the PPV for moderate-severe OSA was 92%; two out of 350 patients had no OSA and 26 had mild OSA. CONCLUSION Our study showed that the prediction of moderate-severe OSA in a tertiary setting with an ML approach is a viable option to facilitate early identification of OSA. Prospective studies with home-based oximeters and analysis of other oximetry variables are the next steps towards formal implementation.
Humans ; Oximetry/methods* ; Sleep Apnea, Obstructive/diagnosis* ; Male ; Female ; Middle Aged ; Machine Learning ; Polysomnography ; Adult ; Anthropometry ; ROC Curve ; Aged ; Algorithms ; Predictive Value of Tests ; Sensitivity and Specificity ; Neural Networks, Computer ; Demography

INTRODUCTION: In this study, we aimed to compare the performance of a convolutional neural network (CNN)-based deep learning model that was trained on a dataset of normal and abnormal paediatric elbow radiographs with that of paediatric emergency department (ED) physicians on a binomial classification task. METHODS: A total of 1,314 paediatric elbow lateral radiographs (patient mean age 8.2 years) were retrospectively retrieved and classified based on annotation as normal or abnormal (with pathology). They were then randomly partitioned to a development set (993 images); first and second tuning (validation) sets (109 and 100 images, respectively); and a test set (112 images). An artificial intelligence (AI) model was trained on the development set using the EfficientNet B1 network architecture. Its performance on the test set was compared to that of five physicians (inter-rater agreement: fair). Performance of the AI model and the physician group was tested using McNemar test. RESULTS: The accuracy of the AI model on the test set was 80.4% (95% confidence interval [CI] 71.8%-87.3%), and the area under the receiver operating characteristic curve (AUROC) was 0.872 (95% CI 0.831-0.947). The performance of the AI model vs. the physician group on the test set was: sensitivity 79.0% (95% CI: 68.4%-89.5%) vs. 64.9% (95% CI: 52.5%-77.3%; P = 0.088); and specificity 81.8% (95% CI: 71.6%-92.0%) vs. 87.3% (95% CI: 78.5%-96.1%; P = 0.439). CONCLUSION The AI model showed good AUROC values and higher sensitivity, with the P-value at nominal significance when compared to the clinician group.
Humans ; Deep Learning ; Child ; Retrospective Studies ; Male ; Female ; Radiography/methods* ; ROC Curve ; Elbow/diagnostic imaging* ; Neural Networks, Computer ; Child, Preschool ; Elbow Joint/diagnostic imaging* ; Emergency Service, Hospital ; Adolescent ; Infant ; Artificial Intelligence