Machine Learning-based Prediction Model Research Articles

Establishment and Validation of a Machine Learning-Based Prediction Model for Termination of Pregnancy via Cesarean Section.

This study aimed to investigate the risk factors of cesarean section and establish a prediction model for cesarean section based on the characteristics of pregnant women. The clinical characteristics of 2552 singleton pregnant women who delivered a live baby between January 2020 and December 2021 were retrospectively reviewed. They were divided into vaginal delivery group (n = 1850) and cesarean section group (n = 702). These subjects were divided into training set (2020.1-2021.6) and validation set (2021.7-2021.12). In the training set, univariate analysis, Lasso regression, and Boruta were used to screen independent risk factors for cesarean section. Four models, including Logistic Regression (LR), K-Nearest Neighbor (KNN), Classification and Regression Tree (CART), and Random forest (RF), were established in the training set using K-fold cross validation, hyperparameter optimization, and random oversampling techniques. The best model was screened, and Sort graph of feature variables, univariate partial dependency profile, and Break Down profile were delineated. In the validation set, the confusion matrix parameters were calculated, and receiver operating characteristic curve (ROC), precision recall curve (PRC), calibration curve, and clinical decision curve analysis (DCA) were delineated. The risk factors of cesarean section included age and height of women, weight at delivery, weight gain, para, assisted reproduction, abnormal blood glucose during pregnancy, pregnancy hypertension, scarred uterus, premature rupture of membrane (PROM), placenta previa, fetal malposition, thrombocytopenia, floating fetal head, and labor analgesia. RF had the best performance among the four models, and the accuracy of confusion matrix parameters was 0.8956357. The Matthews correlation coefficient (MCC) was 0.753012. The area under ROC (AUC-ROC) was 0.9790787, and the area under PRC (AUC-PRC) was 0.957888. RF prediction model for caesarean section has high discrimination performance, accuracy and consistency, and outstanding generalization ability.

Interpretable Machine Learning-Based Predictive Modeling of Patient Outcomes Following Cardiac Surgery

ObjectiveThe clinical applicability of machine learning predictions of patient outcomes following cardiac surgery remains unclear. We applied machine learning to predict patient outcomes associated with high morbidity and mortality after cardiac surgery and identified the importance of variables to the derived model’s performance. MethodsWe applied machine learning to the Society of Thoracic Surgeons Adult Cardiac Surgery Database to predict post-operative hemorrhage requiring re-operation, venous thromboembolism and stroke. We used permutation feature importance to identify variables important to model performance and a misclassification analysis to study the limitations of the model. ResultsThe study dataset included 662,772 subjects who had cardiac surgery between 2015 and 2017 and 240 variables. Hemorrhage requiring re-operation, venous thromboembolism and stroke occurred in 2.9%, 1.2% and 2.0% of subjects respectively. The model performed remarkably well at predicting all three complications (AUC 0.92-0.97). Pre- and intra-operative variables were not important to model performance. Instead, performance for the prediction of all three outcomes was driven primarily by several post-operative variables, including known risk factors for the complications such as mechanical ventilation and new-onset of post-operative arrhythmias. Many of the post-operative variables important to model performance also increased the risk of subject misclassification, indicating internal validity. ConclusionsA machine learning model accurately and reliably predicts patient outcomes following cardiac surgery. Post-operative, as opposed to pre- or intra-operative variables, are important to model performance. Interventions targeting this period including minimizing the duration of mechanical ventilation and early treatment of new-onset post-operative arrhythmias may help lower the risk of these complications.

Electrochemical promoted dry methane reforming for power and syngas co-generation in solid oxide fuel cells: Experiments, modelling and optimizations

The solid oxide fuel cell (SOFC) combining dry methane reforming (DMR) is an efficient electrochemical power generation device that simultaneously converts greenhouse gases (methane and CO2) into syngas and produces electricity power. The electrochemical promotion of catalysis effect (EPOC) in SOFC is known to be promising for enhancing the syngas conversion e.g. dry methane reforming reaction upon application of electrical currents or potentials. However, traditional DMR catalytic kinetic models were developed from heterogeneous catalysis experimental data, neglecting the EPOC effect and thus fail to accurately predict the DMR catalytic kinetics in SOFC. This study experimentally investigated the EPOC effect on the DMR reaction during SOFC operation, and proposes a machine learning-based predictive model using multiswarm particle swarm optimization algorithm (MSPSO) and back propagating (BP) neural network for the accurate prediction of catalysis performance in DMR-SOFCs under the EPOC. Key parameters including molar flow rate, reaction temperature, and electrical potentials are used as input parameters and CH4/CO2 conversion as output in the predictive model. The MSPSO-BP model exhibits high prediction accuracy with the average error of predicted CH4/CO2 conversion less than 5 %, and the coefficient of determination (R2) values are 0.971 and 0.968. respectively. Sensitivity analysis through the response surface method (RSM) reveals that temperature and electrical potentials are the most important parameters affecting dry methane reforming performance under EPOC. The developed model in this work is the first machine learning-based predictive model for DMR-SOFCs with a focus on EPOC effect and co-generation performance, providing a valuable tool for the optimization and design of future efficient DMR-SOFCs systems.

Machine learning-based prediction models for home discharge in patients with COVID-19: Development and evaluation using electronic health records.

This study aimed to develop and validate predictive models using electronic health records (EHR) data to determine whether hospitalized COVID-19-positive patients would be admitted to alternative medical care or discharged home. We conducted a retrospective cohort study using deidentified data from the University of Florida Health Integrated Data Repository. The study included 1,578 adult patients (≥18 years) who tested positive for COVID-19 while hospitalized, comprising 960 (60.8%) female patients with a mean (SD) age of 51.86 (18.49) years and 618 (39.2%) male patients with a mean (SD) age of 54.35 (18.48) years. Machine learning (ML) model training involved cross-validation to assess their performance in predicting patient disposition. We developed and validated six supervised ML-based prediction models (logistic regression, Gaussian Naïve Bayes, k-nearest neighbors, decision trees, random forest, and support vector machine classifier) to predict patient discharge status. The models were evaluated based on the area under the receiver operating characteristic curve (ROC-AUC), precision, accuracy, F1 score, and Brier score. The random forest classifier exhibited the highest performance, achieving an accuracy of 0.84 and an AUC of 0.72. Logistic regression (accuracy: 0.85, AUC: 0.71), k-nearest neighbor (accuracy: 0.84, AUC: 0.63), decision tree (accuracy: 0.84, AUC: 0.61), Gaussian Naïve Bayes (accuracy: 0.84, AUC: 0.66), and support vector machine classifier (accuracy: 0.84, AUC: 0.67) also demonstrated valuable predictive capabilities. This study's findings are crucial for efficiently allocating healthcare resources during pandemics like COVID-19. By harnessing ML techniques and EHR data, we can create predictive tools to identify patients at greater risk of severe symptoms based on their medical histories. The models developed here serve as a foundation for expanding the toolkit available to healthcare professionals and organizations. Additionally, explainable ML methods, such as Shapley Additive Explanations, aid in uncovering underlying data features that inform healthcare decision-making processes.

Machine learning-based predictive model for the differential diagnosis of ≤ 5 cm gastric stromal tumor and gastric schwannoma based on CT images

The clinical symptoms of ≤ 5 cm gastric stromal tumor (GST) and gastric schwannoma (GS) are similar, but the treatment regimens are different. This study explored the value of computed tomography (CT) combined with machine learning (ML) algorithms to find the best model to discriminate them. A total of 126 patients with GST ≤ 5 cm and 35 patients with GS ≤ 5 during 2013–2022 were included. CT imaging features included qualitative data (tumor location, growth pattern, lobulation, surface ulcer status, necrosis, calcification, and surrounding lymph nodes) and quantitative data [long diameter (LD); short diameter (SD); LD/SD ratio; degree of enhancement (DE); heterogeneous degree (HD)]. Patients were randomly divided into a training set (n = 112) and test set (n = 49) using 7:3 stratified sampling. The univariate and multivariate logistic regression analysis were used to identify independent risk factors. Five ML algorithms were used to build prediction models: Support Vector Machine, k-Nearest Neighbor, Random Forest, Extra Trees, and Extreme Gradient Boosting Machine. The analysis identified that HDv, lobulation, and tumor growth site were independent risk factors (P < 0.05). We should focus on these three imaging features of tumors, which are relatively easy to obtain. The area under the curve for the SVM, KNN, RF, ET, and XGBoost prediction models were, respectively, 0.790, 0.895, 0.978, 0.988, and 0.946 for the training set, and were, respectively, 0.848, 0.892, 0.887, 0.912, and 0.867 for the test set. The CT combined with ML algorithms generated predictive models to improve the differential diagnosis of ≤ 5 cm GST and GS which has important clinical practical value. The Extra Trees algorithm resulted in the optimal model.

Silica sources for arsenic mitigation in rice: machine learning-based predictive modeling and risk assessment.

Arsenic (As) is a well-known human carcinogen, and the consumption of rice is the main pathway for the South Asian people. The study evaluated the impact of the amendments involving CaSiO3, SiO2 nanoparticles, silica solubilizing bacteria (SSB), and rice straw compost (RSC) on mitigation of As toxicity in rice. The translocation of As from soil to cooked rice was tracked, and the results showed that RSC and its combination with SSB were the most effective in reducing As loading in rice grain by 53.2%. To determine the risk of dietary exposure to As, the average daily intake (ADI), hazard quotient (HQ), and incremental lifetime cancer risk (ILCR) were computed. The study observed that the ADI was reduced to one-third (0.24 μg kg-1bw) under RSC+SSB treatments compared to the control. An effective prediction model was established using random forest model and described the accumulation of As by rice grains depend on bioavailable As, P, and Fe which explained 48.5, 5.07%, and 2.6% of the variation in the grain As, respectively. The model anticipates that to produce As benign rice grain, soil should have P and Fe concentration more than 30 mg kg-1 and 12 mg kg-1, respectively if soil As surpasses 2.5 mg kg-1.

Smart manufacturing with transfer learning under limited data: Towards Data-Driven Intelligences

Advanced cyber-physical production methods, often known as smart manufacturing or Industry 4.0, keep benefiting from the combination of cutting-edge artificial intelligence (AI) techniques with sophisticated manufacturing techniques. Yet, the degree to which the deployed predictive models can manage uncertainties and production data changes in the factory over time strongly influences such systems' “smartness” level. Traditional machine learning (ML) approaches often tend to perform badly in situations of configuration modification in a production process without enough fresh data. Conventional machine learning-based predictive modeling methods require large volumes of training data; however, accumulating large amounts of training data in real-world applications such as manufacturing is labor-intensive and costly, especially for small and medium enterprises. Consequently, limited data is a prevalent obstacle for ML techniques. The present study employs a transfer learning (TL) framework to address this issue. Due to its numerous benefits, electrical Discharge Machining (EDM) has been selected as an illustrative case. This study throws light on developing a new prediction model using TL to predict the material removal rate (MRR) of the electrical discharge machining (EDM) process. The input parameters are pulse on time, current, and voltage. The developed model is based on Deep Neural Network (DNN). The proposed method overcomes the limitations of traditional machine learning (ML) models, which require many experimental datasets for accurately predicting the responses. The results show that TL can be used to overcome the issue of limited data in the manufacturing process.

Radiation pneumonia predictive model for radiotherapy in esophageal carcinoma patients

BackgroundThe machine learning models with dose factors and the deep learning models with dose distribution matrix have been used to building lung toxics models for radiotherapy and achieve promising results. However, few studies have integrated clinical features into deep learning models. This study aimed to explore the role of three-dimension dose distribution and clinical features in predicting radiation pneumonitis (RP) in esophageal cancer patients after radiotherapy and designed a new hybrid deep learning network to predict the incidence of RP.MethodsA total of 105 esophageal cancer patients previously treated with radiotherapy were enrolled in this study. The three-dimension (3D) dose distributions within the lung were extracted from the treatment planning system, converted into 3D matrixes and used as inputs to predict RP with ResNet. In total, 15 clinical factors were normalized and converted into one-dimension (1D) matrixes. A new prediction model (HybridNet) was then built based on a hybrid deep learning network, which combined 3D ResNet18 and 1D convolution layers. Machine learning-based prediction models, which use the traditional dosiomic factors with and without the clinical factors as inputs, were also constructed and their predictive performance compared with that of HybridNet using tenfold cross validation. Accuracy and area under the receiver operator characteristic curve (AUC) were used to evaluate the model effect. DeLong test was used to compare the prediction results of the models.ResultsThe deep learning-based model achieved superior prediction results compared with machine learning-based models. ResNet performed best in the group that only considered dose factors (accuracy, 0.78 ± 0.05; AUC, 0.82 ± 0.25), whereas HybridNet performed best in the group that considered both dose factors and clinical factors (accuracy, 0.85 ± 0.13; AUC, 0.91 ± 0.09). HybridNet had higher accuracy than that of Resnet (p = 0.009).ConclusionBased on prediction results, the proposed HybridNet model could predict RP in esophageal cancer patients after radiotherapy with significantly higher accuracy, suggesting its potential as a useful tool for clinical decision-making. This study demonstrated that the information in dose distribution is worth further exploration, and combining multiple types of features contributes to predict radiotherapy response.

Machine learning-based prediction model and visual interpretation for prostate cancer

BackgroundMost prostate cancers(PCa) rely on serum prostate-specific antigen (PSA) testing for biopsy confirmation, but the accuracy needs to be further improved. We need to continue to develop PCa prediction model with high clinical application value.MethodsBenign prostatic hyperplasia (BPH) and prostate cancer data were obtained from the Chinese National Clinical Medical Science Data Center for retrospective analysis. The model was constructed using the XGBoost algorithm, and patients’ age, body mass index (BMI), PSA-related parameters and serum biochemical parameters were used as model variables. Using decision analysis curve (DCA) to evaluate the clinical utility of the models. The shapley additive explanation (SHAP) framework was used to analyze the importance ranking and risk threshold of the variables.ResultsA total of 1915 patients were included in this study, including 823 (43.0%) were BPH patients and 1092 (57.0%) were PCa patients. The XGBoost model provided better performance (AUC 0.82) compared with f/tPSA (AUC 0.75),tPSA (AUC 0.68) and fPSA (AUC 0.61), respectively. Based on SHAP values, f/tPSA was the most important variable, and the top five most important biochemical parameter variables were inorganic phosphorus (P), potassium (K), creatine kinase MB isoenzyme (CKMB), low-density lipoprotein cholesterol (LDL-C), and creatinine (Cre). PCa risk thresholds for these risk markers were f/tPSA (0.13), P (1.29 mmol/L), K (4.29 mmol/L), CKMB ( 11.6U/L), LDL-C (3.05mmol/L) and Cre (74.5-99.1umol/L).ConclusionThe present model has advantages of wide-spread availability and high net benefit, especially for underdeveloped countries and regions. Furthermore, these risk thresholds can assist in the diagnosis and screening of prostate cancer in clinical practice.

Power Prediction Method for Ships Using Data Regression Models

This study proposes machine learning-based prediction models to estimate hull form performance. The developed models can predict the residuary resistance coefficient (CR), wake fraction (wTM), and thrust deduction fraction (t). The multi-layer perceptron and convolutional neural network models, wherein the hull shape was considered as images, were evaluated. A prediction model for the open-water characteristics of the propeller was also generated. The experimental data used in the learning process were obtained from model test results conducted in the Korea Research Institute of Ships and Ocean Engineering towing tank. The prediction results of the proposed models showed good agreement with the model test values. According to the ITTC procedures, the service speed and shaft revolution speed of a ship can be extrapolated from the values obtained from the predictive models. The proposed models demonstrated sufficient accuracy when applied to the sample hull forms based on data not used for training. Thus, they can be implemented in the preliminary design phase of hull forms.

Machine learning-based prediction model of acute kidney injury in patients with acute respiratory distress syndrome

BackgroundAcute kidney injury (AKI) can make cases of acute respiratory distress syndrome (ARDS) more complex, and the combination of the two can significantly worsen the prognosis. Our objective is to utilize machine learning (ML) techniques to construct models that can promptly identify the risk of AKI in ARDS patients.MethodWe obtained data regarding ARDS patients from the Medical Information Mart for Intensive Care III (MIMIC-III) and MIMIC-IV databases. Within the MIMIC-III dataset, we developed 11 ML prediction models. By evaluating various metrics, we visualized the importance of its features using Shapley additive explanations (SHAP). We then created a more concise model using fewer variables, and optimized it using hyperparameter optimization (HPO). The model was validated using the MIMIC-IV dataset.ResultA total of 928 ARDS patients without AKI were included in the analysis from the MIMIC-III dataset, and among them, 179 (19.3%) developed AKI after admission to the intensive care unit (ICU). In the MIMIC-IV dataset, there were 653 ARDS patients included in the analysis, and among them, 237 (36.3%) developed AKI. A total of 43 features were used to build the model. Among all models, eXtreme gradient boosting (XGBoost) performed the best. We used the top 10 features to build a compact model with an area under the curve (AUC) of 0.850, which improved to an AUC of 0.865 after the HPO. In extra validation set, XGBoost_HPO achieved an AUC of 0.854. The accuracy, sensitivity, specificity, positive prediction value (PPV), negative prediction value (NPV), and F1 score of the XGBoost_HPO model on the test set are 0.865, 0.813, 0.877, 0.578, 0.957 and 0.675, respectively. On extra validation set, they are 0.724, 0.789, 0.688, 0.590, 0.851, and 0.675, respectively.ConclusionML algorithms, especially XGBoost, are reliable for predicting AKI in ARDS patients. The compact model maintains excellent predictive ability, and the web-based calculator improves clinical convenience. This provides valuable guidance in identifying AKI in ARDS, leading to improved patient outcomes.

Machine Learning Aided Prediction of Glass-Forming Ability of Metallic Glass

The prediction of the glass-forming ability (GFA) of metallic glasses (MGs) can accelerate the efficiency of their development. In this paper, a dataset was constructed using experimental data collected from the literature and books, and a machine learning-based predictive model was established to predict the GFA. Firstly, a classification model based on the size of the critical diameter (Dmax) was established to determine whether an alloy system could form a glass state, with an accuracy rating of 0.98. Then, regression models were established to predict the crystallization temperature (Tx), glass transition temperature (Tg), and liquidus temperature (Tl) of MGs. The R2 of the prediction model obtained in the test set was greater than 0.89, which showed that the model had good prediction accuracy. The key features used by the regression models were analyzed using variance, correlation, embedding, recursive, and exhaustive methods to select the most important features. Furthermore, to improve the interpretability of the prediction model, feature importance, partial dependence plot (PDP), and individual conditional expectation (ICE) methods were used for visualization analysis, demonstrating how features affect the target variables. Finally, taking Zr-Cu-Ni-Al system MGs as an example, a prediction model was established using a genetic algorithm to optimize the alloy composition for high GFA in the compositional space, achieving the optimal design of alloy composition.

Implications of Classification Models for Patients with Chronic Obstructive Pulmonary Disease

Machine learning-based prediction models have the potential to revamp various industries, and one such promising area is healthcare. This study demonstrates the potential impact of machine learning in healthcare, particularly in managing patients with Chronic Obstructive Pulmonary Disease (COPD). The experimental results showcase the remarkable performance of machine learning models, surpassing doctors' predictions for COPD patients. Among the evaluated models, the Gradient Boosted Decision Tree classifier emerges as the top performer, displaying exceptional classification accuracy, precision, recall, and F1-Score compared to doctors' experience. Notably, the comparison between the best machine learning model and doctors' predictions reveals an interesting pattern: machine learning models tend to be more conservative, resulting in an increased probability of patient recovery.

Outcome assessment for out-of-hospital cardiac arrest patients in Singapore and Japan with initial shockable rhythm

BackgroundSingapore and Osaka in Japan have comparable population sizes and prehospital management; however, the frequency of ECPR differs greatly for out-of-hospital cardiac arrest (OHCA) patients with initial shockable rhythm. Given this disparity, we hypothesized that the outcomes among the OHCA patients with initial shockable rhythm in Singapore were different from those in Osaka. The aim of this study was to evaluate the outcomes of OHCA patients with initial shockable rhythm in Singapore compared to the expected outcomes derived from Osaka data using machine learning-based prediction models.MethodsThis was a secondary analysis of two OHCA databases: the Singapore PAROS database (SG-PAROS) and the Osaka-CRITICAL database from Osaka, Japan. This study included adult (18–74 years) OHCA patients with initial shockable rhythm. A machine learning-based prediction model was derived and validated using data from the Osaka-CRITICAL database (derivation data 2012–2017, validation data 2018–2019), and applied to the SG-PAROS database (2010–2016 data), to predict the risk-adjusted probability of favorable neurological outcomes. The observed and expected outcomes were compared using the observed–expected ratio (OE ratio) with 95% confidence intervals (CI).ResultsFrom the SG-PAROS database, 1,789 patients were included in the analysis. For OHCA patients who achieved return of spontaneous circulation (ROSC) on hospital arrival, the observed favorable neurological outcome was at the same level as expected (OE ratio: 0.905 [95%CI: 0.784–1.036]). On the other hand, for those who had continued cardiac arrest on hospital arrival, the outcomes were lower than expected (shockable rhythm on hospital arrival, OE ratio: 0.369 [95%CI: 0.258–0.499], and nonshockable rhythm, OE ratio: 0.137 [95%CI: 0.065–0.235]).ConclusionThis observational study found that the outcomes for patients with initial shockable rhythm but who did not obtain ROSC on hospital arrival in Singapore were lower than expected from Osaka. We hypothesize this is mainly due to differences in the use of ECPR.

Development of personalized machine learning-based prediction models for short-term postoperative outcomes in patients undergoing cervical laminoplasty.

By predicting short-term postoperative outcomes before surgery, patients undergoing cervical laminoplasty (CLP) surgery could benefit from more accurate patient care strategies that could reduce the likelihood of adverse outcomes. With this study, we developed a series of machine learning (ML) models for predicting short-term postoperative outcomes and integrated them into an open-source online application. National surgical quality improvement program database was utilized to identify individuals who have undergone CLP surgery. The investigated outcomes were prolonged length of stay (LOS), non-home discharges, 30-day readmissions, unplanned reoperations, and major complications. ML models were developed and implemented on a website to predict these three outcomes. A total of 1740 patients that underwent CLP were included in the analysis. Performance evaluation indicated that the top-performing models for each outcome were the models built with TabPFN and LightGBM algorithms. The TabPFN models yielded AUROCs of 0.830, 0.847, and 0.858 in predicting non-home discharges, unplanned reoperations, and major complications, respectively. The LightGBM models yielded AUROCs of 0.812 and 0.817 in predicting prolonged LOS, and 30-day readmissions, respectively. The potential of ML approaches to predict postoperative outcomes following spine surgery is significant. As the volume of data in spine surgery continues to increase, the development of predictive models as clinically relevant decision-making tools could significantly improve risk assessment and prognosis. Here, we present an accessible predictive model for predicting short-term postoperative outcomes following CLP intended to achieve the stated objectives.

Modeling vibrant areas at nighttime: A machine learning-based analytical framework for urban regeneration

Enhancing vibrant areas at nighttime (VAN) is important for promoting urban regeneration. However, the process simulation of the potential impact of urban renewal initiatives on VANs has yet to achieve complete coupling with population mobility, economy, and land utilization. In this study, we developed a simulation framework to simulate the changes in VANs and reveal their potential interconnection with urban regeneration strategies. The research methods involved the use of overlaying multiple data sources, including satellite imagery, thermal and land use maps, and point of interest data, to obtain a collection of nightlife activities in Mianyang City. We were able to identify and grade VANs based on multisource big data, providing support for urban renewal planning, as well as determine the key driving factors and configuration patterns of environmental elements that impact VAN. We also developed a machine learning-based predictive model for urban regeneration and VAN redevelopment. These results show that a vibrant nightlife can be used to regenerate dilapidated urban areas, thus reducing urbanization. Moreover, the simulation method developed in this study has wide applicability in other regions for identifying potential improvements and guiding investment and revitalization efforts in a targeted and effective manner.

Letter to the Editor. Machine learning-based prediction models in neurosurgery.

Letter to the Editor. Machine learning-based prediction models in neurosurgery.

Early plasma proteomic biomarkers and prediction model of acute respiratory distress syndrome after cardiopulmonary bypass: a prospective nested cohort study.

Early recognition of the risk of acute respiratory distress syndrome (ARDS) after cardiopulmonary bypass (CPB) may improve clinical outcomes. The main objective of this study was to identify proteomic biomarkers and develop an early prediction model for CPB-ARDS. We conducted three prospective nested cohort studies of all consecutive patients undergoing cardiac surgery with CPB at XXX Hospital. Plasma proteomic profiling was performed in ARDS patients and matched controls (Cohort 1, April, 2021 to July, 2021) at multiple timepoints: before CPB (T1), at end of CPB (T2), and 24 h after CPB (T3). Then, for Cohort 2 (August, 2021 to July, 2022), biomarker expression was measured and verified in the plasma. Furthermore, lung ischemia/reperfusion injury (LIRI) models and sham-operation were established in 50 rats to explore the tissue-level expression of biomarkers identified in the aforementioned clinical cohort. Subsequently, a machine learning-based prediction model incorporating protein and clinical predictors from Cohort 2 for CPB-ARDS was developed and internally validated. Model performance was externally validated on Cohort 3 (January, 2023 to March, 2023). A total of 709 proteins were identified, with 9, 29, and 35 altered proteins between ARDS cases and controls at T1, T2, and T3, respectively, in Cohort 1. Following quantitative verification of several predictive proteins in Cohort 2, higher levels of thioredoxin domain containing 5 (TXNDC5), cathepsin L (CTSL), and NPC intracellular cholesterol transporter 2 (NPC2) at T2 were observed in CPB-ARDS patients. A dynamic online predictive nomogram was developed based on three proteins (TXNDC5, CTSL, and NPC2) and two clinical risk factors (CPB time, massive blood transfusion), with excellent performance (precision: 83.33%, sensitivity: 93.33%, specificity: 61.16%, F1 score: 85.05%). Mean area under the receiver operating characteristics curve (AUC) of the model after 10-fold cross-validation was 0.839 (95% confidence interval (CI): 0.824 to 0.855). Model discrimination and calibration were maintained during external validation dataset testing, with AUC of 0.820 (95% CI: 0.685 to 0.955) and Brier Score of 0.177 (95% CI: 0.147 to 0.206). Moreover, the considerably overexpressed TXNDC5 and CTSL proteins identified in the plasma of patients with CPB-ARDS, exhibited a significant upregulation in the lung tissue of LIRI rats. This study identified several novel predictive biomarkers, developed and validated a practical prediction tool using biomarker and clinical factor combinations for individual prediction of CPB-ARDS risk. Assessing the plasma TXNDC5, CTSL, and NPC2 levels might identify patients who warrant closer follow-up and intensified therapy for ARDS prevention following major surgery.

A Retrospective Cohort Study: Predicting 90-Day Mortality for ICU Trauma Patients with a Machine Learning Algorithm Using XGBoost Using MIMIC-III Database.

The aim of this study was to develop and validate a machine learning-based predictive model that predicts 90-day mortality in ICU trauma patients. Data of patients with severe trauma were extracted from the Medical Information Mart for Intensive Care III (MIMIC-III) database. The performances of mortality prediction models generated using nine machine learning extreme gradient boosting (XGBoost), logistic regression, random forest, AdaBoost, multilayer perceptron (MLP) neural networks, support vector machine (SVM), light gradient boosting machine (GBM), k nearest neighbors (KNN) and gaussian naive bayes (GNB). The performance of the model was evaluated in terms of discrimination, calibration and clinical application. We found that the accuracy, sensitivity, specificity, PPV, NPV and F1 score of our proposed XGBoost model were 82.8%, 79.7%, 77.6%, 51.2%, 91.5% and 0.624, respectively. Among the nine models, the XGBoost model performed best. Compared with traditional logistic regression, the calibration curves of the XGBoost model and decision curve analysis (DCA) performed well. Our study shows that the XGBoost model outperforms other machine learning models in predicting 90-day mortality in trauma patients. It can be used to assist clinicians in the early identification of mortality risk factors and early intervention to reduce mortality.

Development of Prediction Models for Needs and Resources Service Using Big Data Related to Elderly Care

Objectives: In this study, we aimed to verify the possibility of utilizing the integrated case management big data in Korea to identify the complex care needs of the elderly in the community and link them to appropriate resources through a machine learning-based predictive model. Methods: Three merged datasets were made by combining different combinations of the data from integrated case management, elderly care service, and Community Integrated Care Leading Project conducted in Korea. Due to the issue of multiple selection of the response variables, the analysis considering multi-label classification method was conducted. The performance of the predictive model was evaluated based on the F1 score. Results: (1) ‘Maintaining daily life’ and ‘Health’ labels in the needs, (2) ‘Maintaining physical health’ and ‘food, clothing and shelter’ labels in the problem domain, (3) ‘Physical health’, ‘daily life’, and ‘housing’ labels in the major classification of resources, and (4) ‘meal food support’, ‘disease prevention and health care’, and ‘living environment improvement’ labels in the major classification of resources showed F1 ≥ 0.8. Conclusions: In this study, we verified that the elderly need-resource linkage prediction based on big data for integrated case management was at an appropriate level, and the possibility of supporting integrated case management using this was confirmed.