Validation Set Research Articles

Deep Learning Model for Grading and Localization of Lumbar Disc Herniation on Magnetic Resonance Imaging.

Methods for grading and localization of lumbar disc herniation (LDH) on MRI are complex, time-consuming, and subjective. Utilizing deep learning (DL) models as assistance would mitigate such complexities. To develop an interpretable DL model capable of grading and localizing LDH. Retrospective. 1496 patients (M/F: 783/713) were evaluated, and randomly divided into training (70%), validation (10%), and test (20%) sets. 1.5T MRI for axial T2-weighted sequences (spin echo). The training set was annotated by three spinal surgeons using the Michigan State University classification to train the DL model. The test set was annotated by a spinal surgery expert (as ground truth labels), and two spinal surgeons (comparison with the trained model). An external test set was employed to evaluate the generalizability of the DL model. Calculated intersection over union (IoU) for detection consistency, utilized Gwet's AC1 to assess interobserver agreement, and evaluated model performance based on sensitivity and specificity, with statistical significance set at P < 0.05. The DL model achieved high detection consistency in both the internal test dataset (grading: mean IoU 0.84, recall 99.6%; localization: IoU 0.82, recall 99.5%) and external test dataset (grading: 0.72, 98.0%; localization: 0.71, 97.6%). For internal testing, the DL model (grading: 0.81; localization: 0.76), Rater 1 (0.88; 0.82), and Rater 2 (0.86; 0.83) demonstrated results highly consistent with the ground truth labels. The overall sensitivity of the DL model was 87.0% for grading and 84.0% for localization, while the specificity was 95.5% and 94.4%. For external testing, the DL model showed an appreciable decrease in consistency (grading: 0.69; localization: 0.66), sensitivity (77.2%; 76.7%), and specificity (92.3%; 91.8%). The classification capabilities of the DL model closely resemble those of spinal surgeons. For future improvement, enriching the diversity of cases could enhance the model's generalization. Stage 2.

Prediction of Responsiveness to PD-L1/PD-1 Inhibitors Using miRNA Profiles Associated With PD-L1 Expression in Lung Adenocarcinoma and Squamous Cell Carcinoma.

MicroRNAs (miRNAs) regulate programmed cell death ligand 1 (PD-L1) and play a crucial role in tumor immune response. However, the relationship between miRNA expression patterns and PD-L1 remains unclear in both lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). We investigated PD-L1-related miRNAs that can predict treatment response in patients treated with PD-L1/PD-1 inhibitors. We selected miRNAs that were correlated with PD-L1 expression within the LUAD and LUSC datasets obtained from The Cancer Genome Atlas (TCGA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC). We validated whether the miRNA profile could be used to predict the prognosis of patients treated with PD-L1/PD-1 inhibitors. Based on four public datasets, we selected 66 and 23 miRNAs associated with PD-L1 expression in LUAD and LUSC, respectively. From the above miRNAs, we identified 5 miRNAs in LUSC and 1 miRNA in LUAD that could predict the response to PD-L1/PD-1 inhibitors in a validation set of patients treated with PD-L1/PD-1 inhibitors. In LUSC, the miRNA profile exhibited a high predictive capability for the response to PD-L1/PD-1 treatment [area under the curve (AUC)=0.963] and accurately predicted prognosis (p=0.031). In LUAD, the miRNA profile was relatively less predictive than in LUSC (AUC=0.691 and p=0.213). Additionally, we observed variations in the PD-L1-associated miRNA profiles, as well as in the associated pathways, between LUAD and LUSC. The PD-L1-associated miRNA profile may predict treatment response in LUSC patients treated with PD-L1/PD-1 inhibitors and help select the PD-L1/PD-1 inhibitor treatment group.

Crack Recognition in Masonry Structures: CNN Models with Limited Data Sets

This research focuses on the task of Masonry Wall Crack Identification using limited data, employing state-of-the-art Convolutional Neural Network (CNN) models. The models investigated include VGG16, MobileNetV2, Xception, and DenseNet121. The dataset, consisting of 946 masonry wall images containing cracks, is used to evaluate the effectiveness of each model in this specific domain. The training set comprises 642 images, the validation set consists of 90 images, and the test set includes 214 images. The models are pretrained on large-scale datasets to extract robust features and are then fine-tuned on the masonry wall crack dataset. Among the models, DenseNet121 stands out, achieving a commendable accuracy of 85.98% in accurately identifying masonry wall cracks. This result underscores the efficacy of DenseNet121 for the challenging task of crack identification in masonry structures using limited data. This study not only contributes to the field of structural health monitoring but also emphasizes the practicality of employing CNN models for real-world applications, particularly in the critical domain of masonry crack identification.

Triglyceride-inflammation score established on account of random survival forest for predicting survival in patients with nasopharyngeal carcinoma: a retrospective study

ObjectiveThis study aimed to establish an effective prognostic model based on triglyceride and inflammatory markers, including neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and platelet-to-lymphocyte ratio (PLR), to predict overall survival (OS) in patients with nasopharyngeal carcinoma (NPC). Additionally, we aimed to explore the interaction and mediation between these biomarkers in their association with OS.MethodsA retrospective review was conducted on 259 NPC patients who had blood lipid markers, including triglyceride and total cholesterol, as well as parameters of peripheral blood cells measured before treatment. These patients were followed up for over 5 years, and randomly divided into a training set (n=155) and a validation set (n=104). The triglyceride-inflammation (TI) score was developed using the random survival forest (RSF) algorithm. Subsequently, a nomogram was created. The performance of the prognostic model was measured by the concordance index (C-index), time-dependent receiver operating characteristic (ROC) curve, and decision curve analysis (DCA). The interaction and mediation between the biomarkers were further analyzed. Bioinformatics analysis based on the GEO dataset was used to investigate the association between triglyceride metabolism and immune cell infiltration.ResultsThe C-index of the TI score was 0.806 in the training set, 0.759 in the validation set, and 0.808 in the entire set. The area under the curve of time-dependent ROC of TI score in predicting survival at 1, 3, and 5 years were 0.741, 0.847, and 0.871 respectively in the training set, and 0.811, 0.837, and 0.758 in the validation set, then 0.771, 0.848, and 0.862 in the entire set, suggesting that TI score had excellent performance in predicting OS in NPC patients. Patients with stage T1-T2 or M0 had significantly lower TI scores, NLR, and PLR, and higher LMR compared to those with stage T3-T3 or M1, respectively. The nomogram, which integrated age, sex, clinical stage, and TI score, demonstrated good clinical usefulness and predictive ability, as evaluated by the DCA. Significant interactions were found between triglyceride and NLR and platelet, but triglyceride did not exhibit any medicating effects in the inflammatory markers. Additionally, NPC tissues with active triglyceride synthesis exhibited high immune cell infiltration.ConclusionThe TI score based on RSF represents a potential prognostic factor for NPC patients, offering convenience and economic advantages. The interaction between triglyceride and NLR may be attributed to the effect of triglyceride metabolism on immune response.

Comprehensive analysis and prediction model of mitophagy and ferroptosis in primary immune thrombocytopenia.

Primary immune thrombocytopenia (ITP) is linked to specific pathogenic mechanisms, yet its relationship with mitophagy and ferroptosis is poorly understood. This study aimed to identify new biomarkers and explore the role of mitophagy and ferroptosis in ITP pathogenesis. Techniques such as differential analysis, Mfuzz expression pattern clustering, machine learning, gene set enrichment analysis, single-cell RNA sequencing (scRNA-seq) and immune infiltration analysis were employed to investigate the molecular pathways of pivotal genes. Two-sample Mendelian randomization (TSMR) assessed the causal effects in ITP. Key genes identified in the training set included GABARAPL1, S100A8, LIN28A, and GDF9, which demonstrated diagnostic potential in validation sets. Functional analysis indicated these genes' involvement in ubiquitin phosphorylation, PPAR signalling pathway and T-cell differentiation. Immune infiltration analysis revealed increased macrophage presence in ITP, related to the critical genes. scRNA-seq indicated reduced GABARAPL1 expression in ITP bone marrow macrophages. TSMR linked S100A8 with ITP diagnosis, presenting an OR of 0.856 (95% CI = 0.736-0.997, p = 0.045). The study pinpointed four central genes, GABARAPL1, S100A8, LIN28A, and GDF9, tied to mitophagy and ferroptosis in ITP. It posits that diminished GABARAPL1 expression may disrupts ubiquitin phosphorylation and PPAR signalling, impairing mitophagy and inhibiting ferroptosis, leading to immune imbalance.

A Data-Driven Computational Framework for Assessing the Risk of Placental Exposure to Environmental Chemicals.

A computational framework based on placental gene networks was proposed in this work to improve the accuracy of the placental exposure risk assessment of environmental compounds. The framework quantitatively characterizes the ability of compounds to cross the placental barrier by systematically considering the interaction and pathway-level information on multiple placental transporters. As a result, probability scores were generated for 307 compounds crossing the placental barrier based on this framework. These scores were then used to categorize the compounds into different levels of transplacental transport range, creating a gradient partition. These probability scores not only facilitated a more intuitive understanding of a compound's ability to cross the placental barrier but also provided valuable information for predicting potential placental disruptors. Compounds with probability scores greater than 90% were considered to have significant transplacental transport potential, whereas those with probability scores less than 80% were classified as unlikely to cross the placental barrier. Furthermore, external validation set results showed that the probability score could accurately predict the compounds known to cross the placental barrier. In conclusion, the computational framework proposed in this study enhances the intuitive understanding of the ability of compounds to cross the placental barrier and opens up new avenues for assessing the placental exposure risk of compounds.

Prediction of soil organic carbon using machine learning techniques and geospatial data for sustainable agriculture

Sustainable agriculture revolves around soil organic carbon (SOC), which is essential for numerous soil functions and ecological attributes. Farmers are interested in conserving and adding additional soil organic carbon to certain fields in order to improve soil health and productivity. The relationship between soil and environment that has been discovered and standardized throughout time has enhanced the progress of digital soil-mapping techniques; therefore, a variety of machine learning techniques are used to predict soil properties. Studies are thriving at how effectively each machine learning method maps and predicts SOC, especially at high spatial resolutions. To predict SOC of soil at a 30 m resolution, four machine learning models—Random Forest, Support Vector Machine, Adaptive Boosting, and k-Nearest Neighbour were used. For model evaluation, two error metrics, namely R2 and RMSE have been used. The findings demonstrated that the calibration and validation sets’ descriptive statistics sufficiently resembled the entire set of data. The range of the calculated SOC content was 0.06 to 1.76 %. According to the findings of the study, Random Forest showed good results for both cases, i.e. evaluation using cross validation and without cross validation. Using cross validation, RF confirmed highest R2 as 0.5278 and lowest RMSE as 0.1683 for calibration dataset while without cross validation it showed R2 as 0.8612 and lowest RMSE as 0.0912 for calibration dataset. The generated soil maps will help farmers adopt precise knowledge for decisions that will increase farm productivity and provide food security through the sustainable use of nutrients and the agricultural environment.

Pan-cancer analysis implicates novel insights of lactate metabolism into immunotherapy response prediction and survival prognostication

BackgroundImmunotherapy has emerged as a potent clinical approach for cancer treatment, but only subsets of cancer patients can benefit from it. Targeting lactate metabolism (LM) in tumor cells as a method to potentiate anti-tumor immune responses represents a promising therapeutic strategy.MethodsPublic single-cell RNA-Seq (scRNA-seq) cohorts collected from patients who received immunotherapy were systematically gathered and scrutinized to delineate the association between LM and the immunotherapy response. A novel LM-related signature (LM.SIG) was formulated through an extensive examination of 40 pan-cancer scRNA-seq cohorts. Then, multiple machine learning (ML) algorithms were employed to validate the capacity of LM.SIG for immunotherapy response prediction and survival prognostication based on 8 immunotherapy transcriptomic cohorts and 30 The Cancer Genome Atlas (TCGA) pan-cancer datasets. Moreover, potential targets for immunotherapy were identified based on 17 CRISPR datasets and validated via in vivo and in vitro experiments.ResultsThe assessment of LM was confirmed to possess a substantial relationship with immunotherapy resistance in 2 immunotherapy scRNA-seq cohorts. Based on large-scale pan-cancer data, there exists a notably adverse correlation between LM.SIG and anti-tumor immunity as well as imbalance infiltration of immune cells, whereas a positive association was observed between LM.SIG and pro-tumorigenic signaling. Utilizing this signature, the ML model predicted immunotherapy response and prognosis with an AUC of 0.73/0.80 in validation sets and 0.70/0.87 in testing sets respectively. Notably, LM.SIG exhibited superior predictive performance across various cancers compared to published signatures. Subsequently, CRISPR screening identified LDHA as a pan-cancer biomarker for estimating immunotherapy response and survival probability which was further validated using immunohistochemistry (IHC) and spatial transcriptomics (ST) datasets. Furthermore, experiments demonstrated that LDHA deficiency in pancreatic cancer elevated the CD8+ T cell antitumor immunity and improved macrophage antitumoral polarization, which in turn enhanced the efficacy of immunotherapy.ConclusionsWe unveiled the tight correlation between LM and resistance to immunotherapy and further established the pan-cancer LM.SIG, holds the potential to emerge as a competitive instrument for the selection of patients suitable for immunotherapy.

Application of machine learning algorithms to identify people with low bone density

BackgroundOsteoporosis is becoming more common worldwide, imposing a substantial burden on individuals and society. The onset of osteoporosis is subtle, early detection is challenging, and population-wide screening is infeasible. Thus, there is a need to develop a method to identify those at high risk for osteoporosis.ObjectiveThis study aimed to develop a machine learning algorithm to effectively identify people with low bone density, using readily available demographic and blood biochemical data.MethodsUsing NHANES 2017–2020 data, participants over 50 years old with complete femoral neck BMD data were selected. This cohort was randomly divided into training (70%) and test (30%) sets. Lasso regression selected variables for inclusion in six machine learning models built on the training data: logistic regression (LR), support vector machine (SVM), gradient boosting machine (GBM), naive Bayes (NB), artificial neural network (ANN) and random forest (RF). NHANES data from the 2013–2014 cycle was used as an external validation set input into the models to verify their generalizability. Model discrimination was assessed via AUC, accuracy, sensitivity, specificity, precision and F1 score. Calibration curves evaluated goodness-of-fit. Decision curves determined clinical utility. The SHAP framework analyzed variable importance.ResultsA total of 3,545 participants were included in the internal validation set of this study, of whom 1870 had normal bone density and 1,675 had low bone density Lasso regression selected 19 variables. In the test set, AUC was 0.785 (LR), 0.780 (SVM), 0.775 (GBM), 0.729 (NB), 0.771 (ANN), and 0.768 (RF). The LR model has the best discrimination and a better calibration curve fit, the best clinical net benefit for the decision curve, and it also reflects good predictive power in the external validation dataset The top variables in the LR model were: age, BMI, gender, creatine phosphokinase, total cholesterol and alkaline phosphatase.ConclusionThe machine learning model demonstrated effective classification of low BMD using blood biomarkers. This could aid clinical decision making for osteoporosis prevention and management.

Analysis of factors affecting intraoperative hemorrhage during percutaneous nephrolithotomy and establishment of nomogram model.

Intraoperative hemorrhage is an important factor affecting intraoperative safety and postoperative patient recovery in percutaneous nephrolithotomy (PCNL). This study aimed to identify the factors that influence intraoperative hemorrhage during PCNL and develop a predictive nomogram model based on these factors.A total of 118 patients who underwent PCNL at the Department of Urology, The Affiliated Huai'an No.1 People's Hospital of Nanjing Medical University from January 2021 to September 2023 was included in this study. The patients were divided into a hemorrhage group (58 cases) and a control group (60 cases) based on the decrease in hemoglobin levels after surgery. The clinical data of all patients were collected, and both univariate analysis and multivariate logistic regression analysis were conducted to identify the independent risk factors for intraoperative hemorrhage during PCNL. The independent risk factors were used to construct a nomogram model using R software. Additionally, receiver operating characteristic (ROC) curves, calibration curves and decision curve analysis (DCA) were utilized to evaluate the model.Multivariate logistic regression analysis revealed that diabetes, long operation time and low psoas muscle mass index (PMI) were independent risk factors for intraoperative hemorrhage during PCNL (P < 0.05). A nomogram model was developed incorporating these factors, and the areas under the ROC curve (AUCs) in the training set and validation set were 0.740 (95% CI: 0.637-0.843) and 0.742 (95% CI: 0.554-0.931), respectively. The calibration curve and Hosmer-Lemeshow test (P = 0.719) of the model proved that the model was well fitted and calibrated. The results of the DCA showed that the model had high value for clinical application.Diabetes, long operation time and low PMI were found to be independent risk factors for intraoperative hemorrhage during PCNL. The nomogram model based on these factors can be used to predict the risk of intraoperative hemorrhage, which is beneficial for perioperative intervention in high-risk groups to improve the safety of surgery and reduce the incidence of postoperative complications.

APPLICATION VALUE OF CT INDEXES COMBINED WITH TUMOR MARKERS AND 7-TAAB IN THE DIAGNOSIS OF INVASIVE LUNG CANCER BASED ON IMPROVED ARTIFICIAL INTELLIGENCE TECHNOLOGY

Objective: To investigate the efficacy of artificially intelligent CT metrics combined with tumor markers and 7-TAABs in predicting the infiltrative nature of mixed ground-glass nodules in the lungs to provide a basis for rational clinical decision-making. Materials and Methods: The clinical data of patients admitted to Changshu Hospital affiliated with Soochow University (Hospital 1) and the Eighth People’s Hospital of Tongzhou District, Nantong City (Hospital 2) from January 2022 to June 2023 were retrospectively analyzed, and a total of 427 cases were included in the study based on the nadir criteria. Three hundred twenty cases from Hospital 1 were randomly grouped into a training set ([Formula: see text]) and an internal validation set ([Formula: see text]), and 107 cases from Hospital 2 were used for the external validation set. The correlations between patients’ clinicoradiological characteristics, tumor markers, and seven tumor autoantibodies (7-TAABs) were analyzed by the Mantel test and further by univariate and multivariate logistic regression analyses to screen the risk factors for the development of infiltrative tumors in the ground-glass nodes and to establish a joint model, which was combined with the clinicoradiological model and the model of the serological indexes in the internal and external validation sets were performed to compare the subject work characteristics (ROC) curves, clinical calibration curves, and clinical decision curves to evaluate their predictive efficacy and clinical utility. Results: Of the 427 cases included, 174 were in the noninfiltrating group, and 253 were in the infiltrating group. There was a significant relationship between tumor markers, seven tumor autoantibodies, and lung cancer infiltrability, and after multifactorial logistic regression analysis, a bronchial air sign, probability of malignancy, nodule diameter, percentage of solidity, VEGF, proGRP, GAGE7, and 7-TAABs were the risk factors for determining the development of infiltrative lesions in early-stage lung nodules. The clinicoradiological model was superior to the serological index model on both internal and external validation sets (AUC values were: 0.926 versus 0.839, [Formula: see text]; 0.895 versus 0.751, [Formula: see text]), and the combined model was slightly superior to the clinicoradiological model on both internal and external validation sets (AUC values were: 0.928 versus 0.926, [Formula: see text]; 0.914 versus 0.895, [Formula: see text]). The combined model performed optimally in terms of calibration accuracy and net clinical benefit, suggesting that a model constructed based on AI CT metrics combined with tumor markers and 7-TAABs would be beneficial to clinicians in predicting the infiltrative nature of early-stage lung cancer. Conclusion: The column-line diagram model built based on CT indexes improved by artificial intelligence technology combined with tumor markers and 7-TAAB can better predict the infiltration of lung tumors, which has specific clinical application value.

To develop a prognostic model for neoadjuvant immunochemotherapy efficacy in esophageal squamous cell carcinoma by analyzing the immune microenvironment

ObjectiveThe choice of neoadjuvant therapy for esophageal squamous cell carcinoma (ESCC) is controversial. This study aims to provide a basis for clinical treatment selection by establishing a predictive model for the efficacy of neoadjuvant immunochemotherapy (NICT).MethodsA retrospective analysis of 30 patients was conducted, divided into Response and Non-response groups based on whether they achieved major pathological remission (MPR). Differences in genes and immune microenvironment between the two groups were analyzed through next-generation sequencing (NGS) and multiplex immunofluorescence (mIF). Variables most closely related to therapeutic efficacy were selected through LASSO regression and ROC curves to establish a predictive model. An additional 48 patients were prospectively collected as a validation set to verify the model’s effectiveness.ResultsNGS suggested seven differential genes (ATM, ATR, BIVM-ERCC5, MAP3K1, PRG, RBM10, and TSHR) between the two groups (P &amp;lt; 0.05). mIF indicated significant differences in the quantity and location of CD3+, PD-L1+, CD3+PD-L1+, CD4+PD-1+, CD4+LAG-3+, CD8+LAG-3+, LAG-3+ between the two groups before treatment (P &amp;lt; 0.05). Dynamic mIF analysis also indicated that CD3+, CD8+, and CD20+ all increased after treatment in both groups, with a more significant increase in CD8+ and CD20+ in the Response group (P &amp;lt; 0.05), and a more significant decrease in PD-L1+ (P &amp;lt; 0.05). The three variables most closely related to therapeutic efficacy were selected through LASSO regression and ROC curves: Tumor area PD-L1+ (AUC= 0.881), CD3+PD-L1+ (AUC= 0.833), and CD3+ (AUC= 0.826), and a predictive model was established. The model showed high performance in both the training set (AUC= 0.938) and the validation set (AUC= 0.832). Compared to the traditional CPS scoring criteria, the model showed significant improvements in accuracy (83.3% vs 70.8%), sensitivity (0.625 vs 0.312), and specificity (0.937 vs 0.906).ConclusionNICT treatment may exert anti-tumor effects by enriching immune cells and activating exhausted T cells. Tumor area CD3+, PD-L1+, and CD3+PD-L1+ are closely related to therapeutic efficacy. The model containing these three variables can accurately predict treatment outcomes, providing a reliable basis for the selection of neoadjuvant treatment plans.

Risk stratification for infection during immunosuppressive therapy in patients with lupus nephritis: A nested case-control study.

The current prediction models for the risk of infection during immunosuppressive treatment for lupus nephritis (LN) lack a prediction time window and have poor pertinence. This study aimed to develop a risk stratification to predict infection during immunosuppressive therapy in patients with LN. This retrospective nested case-control study collected patients with LN treated with immunosuppressive therapy between 2014 and 2022 in the Nephrology ward in Huashan Hospital affiliated to Fudan University and Huashan Hospital Baoshan Branch. Cases were defined as patients who experienced infection during the follow-up period; patients were eligible as controls if they did not have infection during the follow-up period. There were 53 patients with infection by CTCAE V5.0 grade ≥2. According to the 1:3 nested matching, the 53 patients with infection were matched with 159 controls. In the multivariable logistic regression model, the change rate of fibrinogen (OR = 0.97, 95% CI: 0.94-0.99, p = 0.008), leukopenia (OR = 8.68, 95% CI: 1.16-301.72, p = 0.039), and reduced albumin (OR = 6.25, 95% CI: 1.38-28.24, p = 0.017) were independently associated with infection. The AUC of the ROC curve in the validation set of the multivariable logistic regression model in the internal random sampling was 0.864. The scores range from -2 to 10. The infection risk stratification ranges from 2.8% at score -2 to 97.5% at score 10. A risk stratification was built to predict the risk of infection in patients with LN undergoing immunosuppressive therapy.

Advancing Alzheimer's research: Radiomics visualization of the default mode network in cerebral perfusion imaging.

Alzheimer's disease, an irreversible neurological condition, demands timely diagnosis for effective clinical intervention. This study employs radiomics analysis to assess image features in default mode network cerebral perfusion imaging among individuals with cognitive impairment. A radiomics analysis of cerebral perfusion imaging was conducted on 117 patients with cognitive impairment. They were divided into training and validation sets in a 7:3 ratio. Least Absolute Shrinkage and Selection Operator (LASSO) and Random Forest were employed to select and model image features, followed by logistic regression analysis of LASSO and Random Forest results. Diagnostic performance was assessed by calculating the area under the curve (AUC). In the training set, LASSO achieved AUC of 0.978, Random Forest had an AUC of 0.933. In the validation set, LASSO had AUC of 0.859, Random Forest had AUC of 0.986. By conducting Logistic Regression analysis in combination with LASSO and Random Forest, we identified a total of five radiomics features, with four related to morphology and one to textural features, originating from the medial prefrontal cortex and middle temporal gyrus. In the training set, Logistic Regression achieved AUC of 0.911, while in the validation set, it attained AUC of 0.925. The medial prefrontal cortex and middle temporal gyrus are the two brain regions within the default mode network that hold the highest significance for Alzheimer's disease diagnosis. Radiomics analysis contributes to the clinical assessment of Alzheimer's disease by delving into image data to extract deeper layers of information.

Construction and validation of machine learning algorithm for predicting depression among home-quarantined individuals during the large-scale COVID-19 outbreak: based on Adaboost model

ObjectivesCOVID-19 epidemics often lead to elevated levels of depression. To accurately identify and predict depression levels in home-quarantined individuals during a COVID-19 epidemic, this study constructed a depression prediction model based on multiple machine learning algorithms and validated its effectiveness.MethodsA cross-sectional method was used to examine the depression status of individuals quarantined at home during the epidemic via the network. Characteristics included variables on sociodemographics, COVID-19 and its prevention and control measures, impact on life, work, health and economy after the city was sealed off, and PHQ-9 scale scores. The home-quarantined subjects were randomly divided into training set and validation set according to the ratio of 7:3, and the performance of different machine learning models were compared by 10-fold cross-validation, and the model algorithm with the best performance was selected from 15 models to construct and validate the depression prediction model for home-quarantined subjects. The validity of different models was compared based on accuracy, precision, receiver operating characteristic (ROC) curve, and area under the ROC curve (AUC), and the best model suitable for the data framework of this study was identified.ResultsThe prevalence of depression among home-quarantined individuals during the epidemic was 31.66% (202/638), and the constructed Adaboost depression prediction model had an ACC of 0.7917, an accuracy of 0.7180, and an AUC of 0.7803, which was better than the other 15 models on the combination of various performance measures. In the validation sets, the AUC was greater than 0.83.ConclusionsThe Adaboost machine learning algorithm developed in this study can be used to construct a depression prediction model for home-quarantined individuals that has better machine learning performance, as well as high effectiveness, robustness, and generalizability.

Development and validation of a model for prediction of septic shock in neonates with sepsis.

This study aimed to develop and validate a model for prediction of septic shock in neonates with sepsis. This retrospective study included early-onset septic neonates in the Renmin Hospital of Wuhan University between January 2017 and June 2021. The neonates were divided into the training set and the validation set in a ratio of 7:3, and further categorized into septic shock group and none-shock group according to presence or absence of shock symptoms. A total of 406 septic neonates were enrolled, including 217 in septic shock group. Sex (odds ratio [OR] = 0.092, 95% confidence interval [CI]: 0.012 to 0.683, P = 0.020), C-reactive protein at 6 h (OR = 8.475, 95% CI: 3.154 to 22.774, P < 0.001), serum amyloid A at 6 h (OR = 1.179, 95% CI: 1.094 to 1.269, P < 0.01), white blood cells at 6 h (OR = 0.173, 95% CI: 0.092 to 0.326, P < 0.001), platelets at 6 h (OR = 0.985, 95% CI: 0.975 to 0.995, P < 0.001), and Ca2+ at 6 h (OR = 1.44x1011, 95% CI: 2.70 x106 to 7.70 x1015, P < 0.001) were identified as independent risk factors for septic shock and were further included in the nomogram. The area under the receiver operator characteristic curve were 0.873 and 0.920 in training and validation sets, respectively. A predictive model for early diagnosis of septic shock in neonates was developed and initially validated in this study, allowing for timely intervention.

Identification of common mechanisms and biomarkers of atrial fibrillation and heart failure based on machine learning.

Atrial fibrillation (AF) is the most common arrhythmia. Heart failure (HF) is a disease caused by heart dysfunction. The prevalence of AF and HF were progressively increasing over time. The co-existence of AF and HF presents a significant therapeutic challenge. In order to provide new ideas for the diagnosis of AF and HF, it is necessary to carry out biomarker related studies. The training set and validation set data of AF and HF patient samples were downloaded from the GEO database, 'limma' was used to compare the differences in gene expression levels between the disease group and the normal group to screen for differentially expressed genes (DEGs). Weighted correlation network analysis (WGCNA) identified the modules with the highest positive correlation with AF and HF. Functional enrichment and PPI network construction of key genes were carried out. Biomarkers were screened by machine learning. The infiltration of immune cells in AF and HF groups was evaluated by R-packet 'CIBERSORT'. The miRNA network was constructed and potential therapeutic agents for biomarker genes were predicted through the drugbank database. Through WGCNA analysis, it was found that the modules most positively correlated with AF and HF were MEturquoise (r=0.21, P value=0.09) and MEbrown (r=0.62, P value=8e-12), respectively. We screened 25 genes that were highly correlated with both AF and HF. Lasso regression analysis results showed 7 and 20 core genes in AF and HF groups, respectively. The top 20 important genes in AF and HF groups were obtained as core genes by RF model analysis. Four biomarkers were obtained after the intersection of core genes in four groups, namely, GLUL, NCF2, S100A12, and SRGN. The diagnostic efficacy of four genes in AF validation sets was good (AUC: GLUL 0.76, NCF2 0.64, S100A12 0.68, and SRGN 0.76), as well as in the HF validation set (AUC: GLUL 0.76, NCF2 0.84, S100A12 0.92, and SRGN 0.68). The highest correlation with neutrophils was observed for GLUL, NCF2, and S100A12, while SRGN exhibited the strongest correlation with T cells CD4 memory resting in the AF group. GLUL, NCF2, S100A12, and SRGN were most associated with neutrophils in the HF group. A total of 101 miRNAs were predicted by four genes, and GLUL, NCF2, and S100A12 predicted a total of 10 potential therapeutic agents. We identified four biological markers that are highly correlated with AF and HF, namely, GLUL, NCF2, S100A12, and SRGN. Our findings provide theoretical basis for the clinical diagnosis and treatment of AF and HF.

Risk factors and predictive nomograms for bedside emergency endoscopic treatment following endotracheal intubation in cirrhotic patients with esophagogastric variceal bleeding

Data on emergency endoscopic treatment following endotracheal intubation in patients with esophagogastric variceal bleeding (EGVB) remain limited. This retrospective study aimed to explore the efficacy and risk factors of bedside emergency endoscopic treatment following endotracheal intubation in severe EGVB patients admitted in Intensive Care Unit. A total of 165 EGVB patients were enrolled and allocated to training and validation sets in a randomly stratified manner. Univariate and multivariate logistic regression analyses were used to identify independent risk factors to construct nomograms for predicting the prognosis related to endoscopic hemostasis failure rate and 6-week mortality. In result, white blood cell counts (p = 0.03), Child-Turcotte-Pugh (CTP) score (p = 0.001) and comorbid shock (p = 0.005) were selected as independent clinical predictors of endoscopic hemostasis failure. High CTP score (p = 0.003) and the presence of gastric varices (p = 0.009) were related to early rebleeding after emergency endoscopic treatment. Furthermore, the 6-week mortality was significantly associated with MELD scores (p = 0.002), the presence of hepatic encephalopathy (p = 0.045) and postoperative rebleeding (p < 0.001). Finally, we developed practical nomograms to discern the risk of the emergency endoscopic hemostasis failure and 6-week mortality for EGVB patients. In conclusion, our study may help identify severe EGVB patients with higher hemostasis failure rate or 6-week mortality for earlier implementation of salvage treatments.

Development and validation of a risk nomogram to estimate risk of hyponatremia after spinal cord injury: A retrospective single-center study

Objective This study aimed to establish a nomogram-based assessment for predicting the risk of hyponatremia after spinal cord injury (SCI). Design The study is a retrospective single-center study. Participants SCI patients hospitalized in the First Affiliated Hospital of Guangxi Medical University. Setting The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi, China. Methods We performed a retrospective clinical study to collect SCI patients hospitalized in the First Affiliated Hospital of Guangxi Medical University from 2016 to 2020. Based on their clinical scores, the SCI patients were grouped as either hyponatremic or non-hyponatremic, SCI patients in 2016–2019 were identified as the training set, and patients in 2020 were identified as the test set. A nomogram was generated, the calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) were used to validate the model. Results A total of 895 SCI patients were retrieved. After excluding patients with incomplete data, 883 patients were finally included in this study and used to construct the nomograms. The indicators used in the nomogram included sex, completeness of SCI, pneumonia, urinary tract infection, fever, constipation, white blood cell (WBC), albumin and serum Ca2+. These indices were determined by the least absolute shrinkage and selection operator (LASSO) regression analysis. The C-index of the model was 0.81, the area under the curve (AUC) of the training set was 0.82(Cl:0.79–0.85), and the validation set was 0.79(Cl:0.73–0.85). Conclusions Nomogram has good predictive ability, sex, completeness of SCI, pneumonia, urinary tract infection, fever, constipation, WBC, albumin and serum Ca2+ were predictors of hyponatremia after SCI.

A Hybrid Model for Fetal Growth Restriction Assessment by Automatic Placental Radiomics on T2-Weighted MRI and Multifeature Fusion.

MRI-based placental analyses have been used to improve fetal growth restriction (FGR) assessment by complementing ultrasound-based measurements. However, these are still limited by time-consuming manual annotation in MRI data and the lack of mother-based information. To develop and validate a hybrid model for accurate FGR assessment by automatic placental radiomics on T2-weighted imaging (T2WI) and multifeature fusion. Retrospective. 274 pregnant women (29.5   4.0 years) from two centers were included and randomly divided into training (N = 119), internal test (N = 40), time-independent validation (N = 43), and external validation (N = 72) sets. 1.5-T, T2WI half-Fourier acquisition single-shot turbo spin-echo pulse sequence. First, the placentas on T2WI were manually annotated, and a deep learning model was developed to automatically segment the placentas. Then, the radiomic features were extracted from the placentas and selected by three-step feature selection. In addition, fetus-based measurement features and mother-based clinical features were obtained from ultrasound examinations and medical records, respectively. Finally, a hybrid model based on random forest was constructed by fusing these features, and further compared with models based on other machine learning methods and different feature combinations. The performances of placenta segmentation and FGR assessment were evaluated by Dice similarity coefficient (DSC) and the area under the receiver operating characteristic curve (AUROC), respectively. A P-value <0.05 was considered statistically significant. The placentas were automatically segmented with an average DSC of 90.0%. The hybrid model achieved an AUROC of 0.923, 0.931, and 0.880 on the internal test, time-independent validation, and external validation sets, respectively. The mother-based clinical features resulted in significant performance improvements for FGR assessment. The proposed hybrid model may be able to assess FGR with high accuracy. Furthermore, information complementation based on placental, fetal, and maternal features could also lead to better FGR assessment performance. Stage 2.