Machine Learning Models of Survival Prediction in Trauma Patients

Prediction of mortality among severely injured trauma patients A comparison between TRISS and machine learning-based predictive models

BACKGROUND

Given the huge impact of trauma on hospital systems around the world, several attempts have been made to develop predictive models for the outcomes of trauma victims. The most used, and in many studies most accurate predictive model, is the "Trauma Score and Injury Severity Score" (TRISS). Although it has proven to be fairly accurate and is widely used, it has faced criticism for its inability to classify more complex cases. In this study, we aimed to develop machine learning models that better than TRISS could predict mortality among severely injured trauma patients, something that has not been studied using data from a nationwide register before.

METHODS

Patient data was collected from the national trauma register in Sweden, SweTrau. The studied period was from the 1st of January 2015 to 31st of December 2019. After feature selection and multiple imputation of missing data three machine learning (ML) methods (Random Forest, eXtreme Gradient Boosting, and a Generalized Linear Model) were used to create predictive models. The ML models and TRISS were then tested on predictive ability for 30-day mortality.

RESULTS

The ML models were well-calibrated and outperformed TRISS in all the tested measurements. Among the ML models, the eXtreme Gradient Boosting model performed best with an AUC of 0.91 (0.88-0.93).

CONCLUSION

This study showed that all the developed ML-based prediction models were superior to TRISS for the prediction of trauma mortality.

Predicting factors for extremity fracture among border-fall patients using machine learning computing

Background

The factors causing the injuries sustained from falls at US-Mexican border include falls from border wall or fence, fleeing from border patrols, ejecting from vehicle, and others. This study aimed to determine the factors leading to anatomical injuries and to identify the importance of factors leading to limb fracture and internal organ injuries.

Methods

A total of 178 patients who sustained musculoskeletal injuries or internal organ injuries and were admitted to our hospital were included in this retrospective study. Factors indexed for analysis included demographics, comorbidities, and falling mechanic factors. Correlations between anatomical injuries and mechanical injuries were analyzed. Multilayer perceptron neural network (MPNN) was used to identify predictive factors and to stratify the importance of these factors leading to injuries. The SPSS software was used for statistical analysis and predictive factor analysis.

Results

The extremity fracture was associated with border wall/fence fall (p = 0.001) and fleeing (p = 0.002). The spine fracture was correlated with bridge jump/fall (p = 0.007), fence jump/fall (p = 0.026). The vehicle ejecting/MVA was correlated with head injury (P < 0.001), chest injury (P < 0.001), and abdominal injury p < 0.001). MNPP stratify the importance of factor causing injury with multiple factor considered.

Conclusion

The various injury factors caused different anatomical injuries. Multifactorial assessment associated with these injuries can improve the accuracy of diagnosis and develop a predictive model for clinical applications.

Development and Application of a Novel Machine Learning Model Predicting Pancreatic Cancer-Specific Mortality

Precise prognostication is vital for guiding treatment decisions in people diagnosed with pancreatic cancer. Existing models depend on predetermined variables, constraining their effectiveness. Our objective was to explore a novel machine learning approach to enhance a prognostic model for predicting pancreatic cancer-specific mortality and, subsequently, to assess its performance against Cox regression models. Datasets were retrospectively collected and analyzed for 9,752 patients diagnosed with pancreatic cancer and with surgery performed. The primary outcomes were the mortality of patients with pancreatic carcinoma at one year, three years, and five years. Model discrimination was assessed using the concordance index (C-index), and calibration was assessed using Brier scores. The Survival Quilts model was compared with Cox regression models in clinical use, and decision curve analysis was done. The Survival Quilts model demonstrated robust discrimination for one-year (C-index 0.729), three-year (C-index 0.693), and five-year (C-index 0.672) pancreatic cancer-specific mortality. In comparison to Cox models, the Survival Quilts models exhibited a higher C-index up to 32 months but displayed inferior performance after 33 months. A subgroup analysis was conducted, revealing that within the subset of individuals without metastasis, the Survival Quilts models showcased a significant advantage over the Cox models. In the cohort with metastatic pancreatic cancer, Survival Quilts outperformed the Cox model before 24 months but exhibited a weaker performance after 25 months. This study has developed and validated a novel machine learning-based Survival Quilts model to predict pancreatic cancer-specific mortality that outperforms the Cox regression model.

On Scene Injury Severity Prediction (OSISP) model for trauma developed using the Swedish Trauma Registry

BACKGROUND

Providing optimal care for trauma, the leading cause of death for young adults, remains a challenge e.g., due to field triage limitations in assessing a patient's condition and deciding on transport destination. Data-driven On Scene Injury Severity Prediction (OSISP) models for motor vehicle crashes have shown potential for providing real-time decision support. The objective of this study is therefore to evaluate if an Artificial Intelligence (AI) based clinical decision support system can identify severely injured trauma patients in the prehospital setting.

METHODS

The Swedish Trauma Registry was used to train and validate five models - Logistic Regression, Random Forest, XGBoost, Support Vector Machine and Artificial Neural Network - in a stratified 10-fold cross validation setting and hold-out analysis. The models performed binary classification of the New Injury Severity Score and were evaluated using accuracy metrics, area under the receiver operating characteristic curve (AUC) and Precision-Recall curve (AUCPR), and under- and overtriage rates.

RESULTS

There were 75,602 registrations between 2013-2020 and 47,357 (62.6%) remained after eligibility criteria were applied. Models were based on 21 predictors, including injury location. From the clinical outcome, about 40% of patients were undertriaged and 46% were overtriaged. Models demonstrated potential for improved triaging and yielded AUC between 0.80-0.89 and AUCPR between 0.43-0.62.

CONCLUSIONS

AI based OSISP models have potential to provide support during assessment of injury severity. The findings may be used for developing tools to complement field triage protocols, with potential to improve prehospital trauma care and thereby reduce morbidity and mortality for a large patient population.

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

Objective

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.

Methods

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.

Results

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.

Conclusion

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.

Analysis of the role of glucose metabolism-related genes in dilated cardiomyopathy based on bioinformatics

BACKGROUND

Dilated cardiomyopathy (DCM) is a prevalent condition with diverse etiologies, including viral infection, autoimmune response, and genetic factors. Despite the crucial role of energy metabolism in cardiac function, therapeutic targets for key genes in DCM's energy metabolism remain scarce.

METHODS

Our study employed the GSE79962 and GSE42955 datasets from the Gene Expression Omnibus (GEO) database for myocardial tissue sample collection and target gene identification via differential gene expression screening. Using various R packages, GSEA software, and the STRING database, we conducted data analysis, gene set enrichment, and protein-protein interaction predictions. The least absolute shrinkage and selection operator (LASSO) and Support Vector Machine (SVM) algorithms aided in feature gene selection, while the predictive model's efficiency was evaluated via the receiver operating characteristic (ROC) curve analysis. We used the non-negative matrix factorization (NMF) method for molecular typing and the cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) algorithm for predicting immune cell infiltration.

RESULTS

The DLAT and LDHA genes may regulate the immune microenvironment of DCM by influencing activated dendritic cells, activated mast cells, and M0 macrophages, respectively. The BPGM, DLAT, PGM2, ADH1A, ADH1C, LDHA, and PFKM genes may regulate m6A methylation in DCM by affecting the ZC3H13, ALKBH5, RBMX, HNRNPC, METTL3, and YTHDC1 genes. Further regulatory mechanism analysis suggested that PFKM, DLAT, PKLR, PGM2, LDHA, BPGM, ADH1A, and ADH1C could be involved in the development of cardiomyopathy by regulating the Toll-like receptor signaling pathway.

CONCLUSIONS

PFKM, DLAT, PKLR, PGM2, LDHA, BPGM, ADH1A, and ADH1C may serve as potential targets for guiding the diagnosis, treatment, and follow-up of DCM.

Machine learning-based prediction of in-ICU mortality in pneumonia patients

Conventional severity-of-illness scoring systems have shown suboptimal performance for predicting in-intensive care unit (ICU) mortality in patients with severe pneumonia. This study aimed to develop and validate machine learning (ML) models for mortality prediction in patients with severe pneumonia. This retrospective study evaluated patients admitted to the ICU for severe pneumonia between January 2016 and December 2021. The predictive performance was analyzed by comparing the area under the receiver operating characteristic curve (AU-ROC) of ML models to that of conventional severity-of-illness scoring systems. Three ML models were evaluated: (1) logistic regression with L2 regularization, (2) gradient-boosted decision tree (LightGBM), and (3) multilayer perceptron (MLP). Among the 816 pneumonia patients included, 223 (27.3%) patients died. All ML models significantly outperformed the Simplified Acute Physiology Score II (AU-ROC: 0.650 [0.584-0.716] vs 0.820 [0.771-0.869] for logistic regression vs 0.827 [0.777-0.876] for LightGBM 0.838 [0.791-0.884] for MLP; P < 0.001). In the analysis for NRI, the LightGBM and MLP models showed superior reclassification compared with the logistic regression model in predicting in-ICU mortality in all length of stay in the ICU subgroups; all age subgroups; all subgroups with any APACHE II score, PaO₂/FiO₂ ratio < 200; all subgroups with or without history of respiratory disease; with or without history of CVA or dementia; treatment with mechanical ventilation, and use of inotropic agents. In conclusion, the ML models have excellent performance in predicting in-ICU mortality in patients with severe pneumonia. Moreover, this study highlights the potential advantages of selecting individual ML models for predicting in-ICU mortality in different subgroups.

The Addition of the Geriatric Nutritional Risk Index to the Prognostic Scoring Systems Did Not Improve Mortality Prediction in Trauma Patients in the Intensive Care Unit

Background

Malnutrition is prevalent among critically ill patients and has been associated with a poor prognosis. This study sought to determine whether the addition of a nutritional indicator to the various variables of prognostic scoring models can improve the prediction of mortality among trauma patients in the intensive care unit (ICU).

Methods

This study's cohort included 1,126 trauma patients hospitalized in the ICU between January 1, 2018, and December 31, 2021. Two nutritional indicators, the prognostic nutrition index (PNI), a calculation based on the serum albumin concentration and peripheral blood lymphocyte count, and the geriatric nutritional risk index (GNRI), a calculation based on the serum albumin concentration and the ratio of current body weight to ideal body weight, were examined for their association with the mortality outcome. The significant nutritional indicator was served as an additional variable in prognostic scoring models of the Trauma and Injury Severity Score (TRISS), the Acute Physiology and Chronic Health Evaluation (APACHE II), and the mortality prediction models (MPM II) at admission, 24, 48, and 72 h in the mortality outcome prediction. The predictive performance was determined by the area under the receiver operating characteristic curve.

Results

Multivariate logistic regression revealed that GNRI (OR, 0.97; 95% CI, 0.96-0.99; p=0.007), but not PNI (OR, 0.99; 95% CI, 0.97-1.02; p=0.518), was independent risk factor for mortality. However, none of these predictive scoring models showed a significant improvement in prediction when the GNRI variable is incorporated.

Conclusions

The addition of GNRI as a variable to the prognostic scoring models did not significantly enhance the performance of the predictors.

Hospital mortality prediction in traumatic injuries patients: comparing different SMOTE-based machine learning algorithms

BACKGROUND

Trauma is one of the most critical public health issues worldwide, leading to death and disability and influencing all age groups. Therefore, there is great interest in models for predicting mortality in trauma patients admitted to the ICU. The main objective of the present study is to develop and evaluate SMOTE-based machine-learning tools for predicting hospital mortality in trauma patients with imbalanced data.

METHODS

This retrospective cohort study was conducted on 126 trauma patients admitted to an intensive care unit at Besat hospital in Hamadan Province, western Iran, from March 2020 to March 2021. Data were extracted from the medical information records of patients. According to the imbalanced property of the data, SMOTE techniques, namely SMOTE, Borderline-SMOTE1, Borderline-SMOTE2, SMOTE-NC, and SVM-SMOTE, were used for primary preprocessing. Then, the Decision Tree (DT), Random Forest (RF), Naive Bayes (NB), Artificial Neural Network (ANN), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) methods were used to predict patients' hospital mortality with traumatic injuries. The performance of the methods used was evaluated by sensitivity, specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), accuracy, Area Under the Curve (AUC), Geometric Mean (G-means), F1 score, and P-value of McNemar's test.

RESULTS

Of the 126 patients admitted to an ICU, 117 (92.9%) survived and 9 (7.1%) died. The mean follow-up time from the date of trauma to the date of outcome was 3.98 ± 4.65 days. The performance of ML algorithms is not good with imbalanced data, whereas the performance of SMOTE-based ML algorithms is significantly improved. The mean area under the ROC curve (AUC) of all SMOTE-based models was more than 91%. F1-score and G-means before balancing the dataset were below 70% for all ML models except ANN. In contrast, F1-score and G-means for the balanced datasets reached more than 90% for all SMOTE-based models. Among all SMOTE-based ML methods, RF and ANN based on SMOTE and XGBoost based on SMOTE-NC achieved the highest value for all evaluation criteria.

CONCLUSIONS

This study has shown that SMOTE-based ML algorithms better predict outcomes in traumatic injuries than ML algorithms. They have the potential to assist ICU physicians in making clinical decisions.

Science fiction or clinical reality: a review of the applications of artificial intelligence along the continuum of trauma care

Artificial intelligence (AI) and machine learning describe a broad range of algorithm types that can be trained based on datasets to make predictions. The increasing sophistication of AI has created new opportunities to apply these algorithms within within trauma care. Our paper overviews the current uses of AI along the continuum of trauma care, including injury prediction, triage, emergency department volume, assessment, and outcomes. Starting at the point of injury, algorithms are being used to predict severity of motor vehicle crashes, which can help inform emergency responses. Once on the scene, AI can be used to help emergency services triage patients remotely in order to inform transfer location and urgency. For the receiving hospital, these tools can be used to predict trauma volumes in the emergency department to help allocate appropriate staffing. After patient arrival to hospital, these algorithms not only can help to predict injury severity, which can inform decision-making, but also predict patient outcomes to help trauma teams anticipate patient trajectory. Overall, these tools have the capability to transform trauma care. AI is still nascent within the trauma surgery sphere, but this body of the literature shows that this technology has vast potential. AI-based predictive tools in trauma need to be explored further through prospective trials and clinical validation of algorithms.

Artificial intelligence and machine learning for hemorrhagic trauma care

Artificial intelligence (AI), a branch of machine learning (ML) has been increasingly employed in the research of trauma in various aspects. Hemorrhage is the most common cause of trauma-related death. To better elucidate the current role of AI and contribute to future development of ML in trauma care, we conducted a review focused on the use of ML in the diagnosis or treatment strategy of traumatic hemorrhage. A literature search was carried out on PubMed and Google scholar. Titles and abstracts were screened and, if deemed appropriate, the full articles were reviewed. We included 89 studies in the review. These studies could be grouped into five areas: (1) prediction of outcomes; (2) risk assessment and injury severity for triage; (3) prediction of transfusions; (4) detection of hemorrhage; and (5) prediction of coagulopathy. Performance analysis of ML in comparison with current standards for trauma care showed that most studies demonstrated the benefits of ML models. However, most studies were retrospective, focused on prediction of mortality, and development of patient outcome scoring systems. Few studies performed model assessment via test datasets obtained from different sources. Prediction models for transfusions and coagulopathy have been developed, but none is in widespread use. AI-enabled ML-driven technology is becoming integral part of the whole course of trauma care. Comparison and application of ML algorithms using different datasets from initial training, testing and validation in prospective and randomized controlled trials are warranted for provision of decision support for individualized patient care as far forward as possible.

Distinct Subtypes of Hepatorenal Syndrome and Associated Outcomes as Identified by Machine Learning Consensus Clustering

BACKGROUND

The utilization of multi-dimensional patient data to subtype hepatorenal syndrome (HRS) can individualize patient care. Machine learning (ML) consensus clustering may identify HRS subgroups with unique clinical profiles. In this study, we aim to identify clinically meaningful clusters of hospitalized patients for HRS using an unsupervised ML clustering approach.

METHODS

Consensus clustering analysis was performed based on patient characteristics in 5564 patients primarily admitted for HRS in the National Inpatient Sample from 2003-2014 to identify clinically distinct HRS subgroups. We applied standardized mean difference to evaluate key subgroup features, and compared in-hospital mortality between assigned clusters.

RESULTS

The algorithm revealed four best distinct HRS subgroups based on patient characteristics. Cluster 1 patients (n = 1617) were older, and more likely to have non-alcoholic fatty liver disease, cardiovascular comorbidities, hypertension, and diabetes. Cluster 2 patients (n = 1577) were younger and more likely to have hepatitis C, and less likely to have acute liver failure. Cluster 3 patients (n = 642) were younger, and more likely to have non-elective admission, acetaminophen overdose, acute liver failure, to develop in-hospital medical complications and organ system failure, and to require supporting therapies, including renal replacement therapy, and mechanical ventilation. Cluster 4 patients (n = 1728) were younger, and more likely to have alcoholic cirrhosis and to smoke. Thirty-three percent of patients died in hospital. In-hospital mortality was higher in cluster 1 (OR 1.53; 95% CI 1.31-1.79) and cluster 3 (OR 7.03; 95% CI 5.73-8.62), compared to cluster 2, while cluster 4 had comparable in-hospital mortality (OR 1.13; 95% CI 0.97-1.32).

CONCLUSIONS

Consensus clustering analysis provides the pattern of clinical characteristics and clinically distinct HRS phenotypes with different outcomes.

Predicting pneumonia during hospitalization in flail chest patients using machine learning approaches

Objective

Pneumonia is a common pulmonary complication of flail chest, causing high morbidity and mortality rates in affected patients. The existing methods for identifying pneumonia have low accuracy, and their use may delay antimicrobial therapy. However, machine learning can be combined with electronic medical record systems to identify information and assist in quick clinical decision-making. Our study aimed to develop a novel machine-learning model to predict pneumonia risk in flail chest patients.

Methods

From January 2011 to December 2021, the electronic medical records of 169 adult patients with flail chest at a tertiary teaching hospital in an urban level I Trauma Centre in Chongqing were retrospectively analysed. Then, the patients were randomly divided into training and test sets at a ratio of 7:3. Using the Fisher score, the best subset of variables was chosen. The performance of the seven models was evaluated by computing the area under the receiver operating characteristic curve (AUC). The output of the XGBoost model was shown using the Shapley Additive exPlanation (SHAP) method.

Results

Of 802 multiple rib fracture patients, 169 flail chest patients were eventually included, and 86 (50.80%) were diagnosed with pneumonia. The XGBoost model performed the best among all seven machine-learning models. The AUC of the XGBoost model was 0.895 (sensitivity: 84.3%; specificity: 80.0%).Pneumonia in flail chest patients was associated with several features: systolic blood pressure, pH value, blood transfusion, and ISS.

Conclusion

Our study demonstrated that the XGBoost model with 32 variables had high reliability in assessing risk indicators of pneumonia in flail chest patients. The SHAP method can identify vital pneumonia risk factors, making the XGBoost model's output clinically meaningful.

Decision support by machine learning systems for acute management of severely injured patients: A systematic review

Introduction

Treating severely injured patients requires numerous critical decisions within short intervals in a highly complex situation. The coordination of a trauma team in this setting has been shown to be associated with multiple procedural errors, even of experienced care teams. Machine learning (ML) is an approach that estimates outcomes based on past experiences and data patterns using a computer-generated algorithm. This systematic review aimed to summarize the existing literature on the value of ML for the initial management of severely injured patients.

Methods

We conducted a systematic review of the literature with the goal of finding all articles describing the use of ML systems in the context of acute management of severely injured patients. MESH search of Pubmed/Medline and Web of Science was conducted. Studies including fewer than 10 patients were excluded. Studies were divided into the following main prediction groups: (1) injury pattern, (2) hemorrhage/need for transfusion, (3) emergency intervention, (4) ICU/length of hospital stay, and (5) mortality.

Results

Thirty-six articles met the inclusion criteria; among these were two prospective and thirty-four retrospective case series. Publication dates ranged from 2000 to 2020 and included 32 different first authors. A total of 18,586,929 patients were included in the prediction models. Mortality was the most represented main prediction group (n = 19). ML models used were artificial neural network ( n = 15), singular vector machine (n = 3), Bayesian network (n = 7), random forest (n = 6), natural language processing (n = 2), stacked ensemble classifier [SuperLearner (SL), n = 3], k-nearest neighbor (n = 1), belief system (n = 1), and sequential minimal optimization (n = 2) models. Thirty articles assessed results as positive, five showed moderate results, and one article described negative results to their implementation of the respective prediction model.

Conclusions

While the majority of articles show a generally positive result with high accuracy and precision, there are several requirements that need to be met to make the implementation of such models in daily clinical work possible. Furthermore, experience in dealing with on-site implementation and more clinical trials are necessary before the implementation of ML techniques in clinical care can become a reality.

COVID Mortality Prediction with Machine Learning Methods: A Systematic Review and Critical Appraisal

More than a year has passed since the report of the first case of coronavirus disease 2019 (COVID), and increasing deaths continue to occur. Minimizing the time required for resource allocation and clinical decision making, such as triage, choice of ventilation modes and admission to the intensive care unit is important. Machine learning techniques are acquiring an increasingly sought-after role in predicting the outcome of COVID patients. Particularly, the use of baseline machine learning techniques is rapidly developing in COVID mortality prediction, since a mortality prediction model could rapidly and effectively help clinical decision-making for COVID patients at imminent risk of death. Recent studies reviewed predictive models for SARS-CoV-2 diagnosis, severity, length of hospital stay, intensive care unit admission or mechanical ventilation modes outcomes; however, systematic reviews focused on prediction of COVID mortality outcome with machine learning methods are lacking in the literature. The present review looked into the studies that implemented machine learning, including deep learning, methods in COVID mortality prediction thus trying to present the existing published literature and to provide possible explanations of the best results that the studies obtained. The study also discussed challenging aspects of current studies, providing suggestions for future developments.

Association of Platelets and White Blood Cells Subtypes with Trauma Patients' Mortality Outcome in the Intensive Care Unit

Background: White blood cell (WBC) subtypes have been suggested to reflect patients’ immune-inflammatory status. Furthermore, the derived ratio of platelets and WBC subtypes, including monocyte-to-lymphocyte ratio (MLR), neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR), is proposed to be associated with patient outcome. Therefore, this study aimed to identify the association of platelets and white blood cells subtypes with the mortality outcome of trauma patients in the intensive care unit (ICU). Method: The medical information from 2854 adult trauma patients admitted to the ICU between 1 January 2009 and 31 December 2019 were retrospectively retrieved from the Trauma Registry System and classified into two groups: the survivors group (n = 2524) and the death group (n = 330). The levels of monocytes, neutrophils, lymphocytes, platelets, and blood-drawn laboratory data detected upon patient arrival to the emergency room and the derived MLR, NLR, and PLR were calculated. Multivariate logistic regression analysis was used to determine the independent effects of univariate predictive variables on mortality occurrence. Result: The results revealed the patients who died had significantly lower platelet counts (175,842 ± 61,713 vs. 206,890 ± 69,006/μL, p < 0.001) but higher levels of lymphocytes (2458 ± 1940 vs. 1971 ± 1453/μL, p < 0.001) than the surviving patients. However, monocyte and neutrophil levels were not significantly different between the death and survivor groups. Moreover, dead patients had a significantly lower PLR than survivors (124.3 ± 110.3 vs. 150.6 ± 106.5, p < 0.001). However, there was no significant difference in MLR or NLR between the dead patients and the survivors. Multivariate logistic regression revealed that male gender, old age, pre-existing hypertension, coronary artery disease and end-stage renal disease, lower Glasgow Coma Scale (GCS), higher Injury Severity Score (ISS), higher level of lymphocytes and lower level of red blood cells and platelets, longer activated partial thromboplastin time (aPTT), and lower level of PLR were independent risk factors associated with higher odds of trauma patient mortality outcome in the ICU. Conclusion: This study revealed that a higher lymphocyte count, lower platelet count, and a lower PLR were associated with higher risk of death in ICU trauma patients.

Intelligent prediction of RBC demand in trauma patients using decision tree methods

BACKGROUND

The vital signs of trauma patients are complex and changeable, and the prediction of blood transfusion demand mainly depends on doctors' experience and trauma scoring system; therefore, it cannot be accurately predicted. In this study, a machine learning decision tree algorithm [classification and regression tree (CRT) and eXtreme gradient boosting (XGBoost)] was proposed for the demand prediction of traumatic blood transfusion to provide technical support for doctors.

METHODS

A total of 1371 trauma patients who were diverted to the Emergency Department of the First Medical Center of Chinese PLA General Hospital from January 2014 to January 2018 were collected from an emergency trauma database. The vital signs, laboratory examination parameters and blood transfusion volume were used as variables, and the non-invasive parameters and all (non-invasive + invasive) parameters were used to construct an intelligent prediction model for red blood cell (RBC) demand by logistic regression (LR), CRT and XGBoost. The prediction accuracy of the model was compared with the area under the curve (AUC).

RESULTS

For non-invasive parameters, the LR method was the best, with an AUC of 0.72 [95% confidence interval (CI) 0.657-0.775], which was higher than the CRT (AUC 0.69, 95% CI 0.633-0.751) and the XGBoost (AUC 0.71, 95% CI 0.654-0.756, P < 0.05). The trauma location and shock index are important prediction parameters. For all the prediction parameters, XGBoost was the best, with an AUC of 0.94 (95% CI 0.893-0.981), which was higher than the LR (AUC 0.80, 95% CI 0.744-0.850) and the CRT (AUC 0.82, 95% CI 0.779-0.853, P < 0.05). Haematocrit (Hct) is an important prediction parameter.

CONCLUSIONS

The classification performance of the intelligent prediction model of red blood cell transfusion in trauma patients constructed by the decision tree algorithm is not inferior to that of the traditional LR method. It can be used as a technical support to assist doctors to make rapid and accurate blood transfusion decisions in emergency rescue environment, so as to improve the success rate of patient treatment.

Machine Learning and Surgical Outcomes Prediction: A Systematic Review

BACKGROUND

Machine learning (ML) has garnered increasing attention as a means to quantitatively analyze the growing and complex medical data to improve individualized patient care. We herein aim to critically examine the current state of ML in predicting surgical outcomes, evaluate the quality of currently available research, and propose areas of improvement for future uses of ML in surgery.

METHODS

A systematic review was conducted in accordance with the Preferred Reporting Items for a Systematic Review and Meta-Analysis (PRISMA) checklist. PubMed, MEDLINE, and Embase databases were reviewed under search syntax "machine learning" and "surgery" for papers published between 2015 and 2020.

RESULTS

Of the initial 2677 studies, 45 papers met inclusion and exclusion criteria. Fourteen different subspecialties were represented with neurosurgery being most common. The most frequently used ML algorithms were random forest (n = 19), artificial neural network (n = 17), and logistic regression (n = 17). Common outcomes included postoperative mortality, complications, patient reported quality of life and pain improvement. All studies which compared ML algorithms to conventional studies which used area under the curve (AUC) to measure accuracy found improved outcome prediction with ML models.

CONCLUSIONS

While still in its early stages, ML models offer surgeons an opportunity to capitalize on the myriad of clinical data available and improve individualized patient care. Limitations included heterogeneous outcome and imperfect quality of some of the papers. We therefore urge future research to agree upon methods of outcome reporting and require basic quality standards.

A Real-Time Artificial Intelligence-Assisted System to Predict Weaning from Ventilator Immediately after Lung Resection Surgery

Assessment of risk before lung resection surgery can provide anesthesiologists with information about whether a patient can be weaned from the ventilator immediately after surgery. However, it is difficult for anesthesiologists to perform a complete integrated risk assessment in a time-limited pre-anesthetic clinic. We retrospectively collected the electronic medical records of 709 patients who underwent lung resection between 1 January 2017 and 31 July 2019. We used the obtained data to construct an artificial intelligence (AI) prediction model with seven supervised machine learning algorithms to predict whether patients could be weaned immediately after lung resection surgery. The AI model with Naïve Bayes Classifier algorithm had the best testing result and was therefore used to develop an application to evaluate risk based on patients' previous medical data, to assist anesthesiologists, and to predict patient outcomes in pre-anesthetic clinics. The individualization and digitalization characteristics of this AI application could improve the effectiveness of risk explanations and physician-patient communication to achieve better patient comprehension.

Predicting 2-Day Mortality of Thrombocytopenic Patients Based on Clinical Laboratory Data Using Machine Learning

BACKGROUND

Clinical laboratories have traditionally used a single critical value for thrombocytopenic events. This system, however, could lead to inaccuracies and inefficiencies, causing alarm fatigue and compromised patient safety.

OBJECTIVES

This study shows how machine learning (ML) models can provide auxiliary information for more accurate identification of critical thrombocytopenic patients when compared with the traditional notification system.

RESEARCH DESIGN

A total of 50,505 patients' platelet count and other 26 additional laboratory datasets of each thrombocytopenic event were used to build prediction models. Conventional logistic regression and ML methods, including random forest (RF), artificial neural network, stochastic gradient descent (SGD), naive Bayes, support vector machine, and decision tree, were applied to build different models and evaluated.

RESULTS

Models using logistic regression [area under the curve (AUC)=0.842], RF (AUC=0.859), artificial neural network (AUC=0.867), or SGD (AUC=0.826) achieved the desired average AUC>0.80. The highest positive predictive value was obtained by the SGD model in the testing data (72.2%), whereas overall, the RF model showed higher sensitivity and total positive predictions in both the training and testing data and outperformed other models. The positive 2-day mortality predictive rate of RF methods is as high as 46.1%-significantly higher than using the traditional notification system at only 14.8% [χ2(1)=81.66, P<0.001].

CONCLUSIONS

This study demonstrates a data-driven ML approach showing a significantly more accurate 2-day mortality prediction after a critical thrombocytopenic event, which can reinforce the accuracy of the traditional notification system.

Machine learning combining CT findings and clinical parameters improves prediction of length of stay and ICU admission in torso trauma

OBJECTIVE

To develop machine learning (ML) models capable of predicting ICU admission and extended length of stay (LOS) after torso (chest, abdomen, or pelvis) trauma, by using clinical and/or imaging data.

MATERIALS AND METHODS

This was a retrospective study of 840 adult patients admitted to a level 1 trauma center after injury to the torso over the course of 1 year. Clinical parameters included age, sex, vital signs, clinical scores, and laboratory values. Imaging data consisted of any injury present on CT. The two outcomes of interest were ICU admission and extended LOS, defined as more than the median LOS in the dataset. We developed and tested artificial neural network (ANN) and support vector machine (SVM) models, and predictive performance was evaluated by area under the receiver operating characteristic (ROC) curve (AUC).

RESULTS

The AUCs of SVM and ANN models to predict ICU admission were up to 0.87 ± 0.03 and 0.78 ± 0.02, respectively. The AUCs of SVM and ANN models to predict extended LOS were up to 0.80 ± 0.04 and 0.81 ± 0.05, respectively. Predictions based on imaging alone or imaging with clinical parameters were consistently more accurate than those based solely on clinical parameters.

CONCLUSIONS

The best performing models incorporated imaging findings and outperformed those with clinical findings alone. ML models have the potential to help predict outcomes in trauma by integrating clinical and imaging findings, although further research may be needed to optimize their performance.

KEY POINTS

• Artificial neural network and support vector machine-based models were used to predict the intensive care unit admission and extended length of stay after trauma to the torso. • Our input data consisted of clinical parameters and CT imaging findings derived from radiology reports, and we found that combining the two significantly enhanced the prediction of both outcomes with either model. • The highest accuracy (83%) and highest area under the receiver operating characteristic curve (0.87) were obtained for artificial neural networks and support vector machines, respectively, by combining clinical and imaging features in the prediction of intensive care unit admission.

Machine learning-based prediction models for accidental hypothermia patients

Background

Accidental hypothermia is a critical condition with high risks of fatal arrhythmia, multiple organ failure, and mortality; however, there is no established model to predict the mortality. The present study aimed to develop and validate machine learning-based models for predicting in-hospital mortality using easily available data at hospital admission among the patients with accidental hypothermia.

Method

This study was secondary analysis of multi-center retrospective cohort study (J-point registry) including patients with accidental hypothermia. Adult patients with body temperature 35.0 °C or less at emergency department were included. Prediction models for in-hospital mortality using machine learning (lasso, random forest, and gradient boosting tree) were made in development cohort from six hospitals, and the predictive performance were assessed in validation cohort from other six hospitals. As a reference, we compared the SOFA score and 5A score.

Results

We included total 532 patients in the development cohort [N = 288, six hospitals, in-hospital mortality: 22.0% (64/288)], and the validation cohort [N = 244, six hospitals, in-hospital mortality 27.0% (66/244)]. The C-statistics [95% CI] of the models in validation cohorts were as follows: lasso 0.784 [0.717–0.851] , random forest 0.794[0.735–0.853], gradient boosting tree 0.780 [0.714–0.847], SOFA 0.787 [0.722–0.851], and 5A score 0.750[0.681–0.820]. The calibration plot showed that these models were well calibrated to observed in-hospital mortality. Decision curve analysis indicated that these models obtained clinical net-benefit.

Conclusion

This multi-center retrospective cohort study indicated that machine learning-based prediction models could accurately predict in-hospital mortality in validation cohort among the accidental hypothermia patients. These models might be able to support physicians and patient’s decision-making. However, the applicability to clinical settings, and the actual clinical utility is still unclear; thus, further prospective study is warranted to evaluate the clinical usefulness.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40560-021-00525-z.

Performance of Prognostic Scoring Systems in Trauma Patients in the Intensive Care Unit of a Trauma Center

BACKGROUND

Prediction of mortality outcomes in trauma patients in the intensive care unit (ICU) is important for patient care and quality improvement. We aimed to measure the performance of 11 prognostic scoring systems for predicting mortality outcomes in trauma patients in the ICU.

METHODS

Prospectively registered data in the Trauma Registry System from 1 January 2016 to 31 December 2018 were used to extract scores from prognostic scoring systems for 1554 trauma patients in the ICU. The following systems were used: the Trauma and Injury Severity Score (TRISS); the Acute Physiology and Chronic Health Evaluation (APACHE II); the Simplified Acute Physiology Score (SAPS II); mortality prediction models (MPM II) at admission, 24, 48, and 72 h; the Multiple Organ Dysfunction Score (MODS); the Sequential Organ Failure Assessment (SOFA); the Logistic Organ Dysfunction Score (LODS); and the Three Days Recalibrated ICU Outcome Score (TRIOS). Predictive performance was determined according to the area under the receiver operator characteristic curve (AUC).

RESULTS

MPM II at 24 h had the highest AUC (0.9213), followed by MPM II at 48 h (AUC: 0.9105). MPM II at 24, 48, and 72 h (0.8956) had a significantly higher AUC than the TRISS (AUC: 0.8814), APACHE II (AUC: 0.8923), SAPS II (AUC: 0.9044), MPM II at admission (AUC: 0.9063), MODS (AUC: 0.8179), SOFA (AUC: 0.7073), LODS (AUC: 0.9013), and TRIOS (AUC: 0.8701). There was no significant difference in the predictive performance of MPM II at 24 and 48 h (p = 0.37) or at 72 h (p = 0.10).

CONCLUSIONS

We compared 11 prognostic scoring systems and demonstrated that MPM II at 24 h had the best predictive performance for 1554 trauma patients in the ICU.

Predicting Acute Kidney Injury after Cardiac Surgery by Machine Learning Approaches

Cardiac surgery-associated AKI (CSA-AKI) is common after cardiac surgery and has an adverse impact on short- and long-term mortality. Early identification of patients at high risk of CSA-AKI by applying risk prediction models allows clinicians to closely monitor these patients and initiate effective preventive and therapeutic approaches to lessen the incidence of AKI. Several risk prediction models and risk assessment scores have been developed for CSA-AKI. However, the definition of AKI and the variables utilized in these risk scores differ, making general utility complex. Recently, the utility of artificial intelligence coupled with machine learning, has generated much interest and many studies in clinical medicine, including CSA-AKI. In this article, we discussed the evolution of models established by machine learning approaches to predict CSA-AKI.

Artificial Intelligence in the Intensive Care Unit

This article is one of ten reviews selected from the Annual Update in Intensive Care and Emergency Medicine 2020. Other selected articles can be found online at https://www.biomedcentral.com/collections/annualupdate2020. Further information about the Annual Update in Intensive Care and Emergency Medicine is available from http://www.springer.com/series/8901.

Current applications of machine learning in the screening and diagnosis of glaucoma: a systematic review and Meta-analysis

AIM

To compare the effectiveness of two well described machine learning modalities, ocular coherence tomography (OCT) and fundal photography, in terms of diagnostic accuracy in the screening and diagnosis of glaucoma.

METHODS

A systematic search of Embase and PubMed databases was undertaken up to 1^st of February 2019. Articles were identified alongside their reference lists and relevant studies were aggregated. A Meta-analysis of diagnostic accuracy in terms of area under the receiver operating curve (AUROC) was performed. For the studies which did not report an AUROC, reported sensitivity and specificity values were combined to create a summary ROC curve which was included in the Meta-analysis.

RESULTS

A total of 23 studies were deemed suitable for inclusion in the Meta-analysis. This included 10 papers from the OCT cohort and 13 from the fundal photos cohort. Random effects Meta-analysis gave a pooled AUROC of 0.957 (95%CI=0.917 to 0.997) for fundal photos and 0.923 (95%CI=0.889 to 0.957) for the OCT cohort. The slightly higher accuracy of fundal photos methods is likely attributable to the much larger database of images used to train the models (59 788 vs 1743).

CONCLUSION

No demonstrable difference is shown between the diagnostic accuracy of the two modalities. The ease of access and lower cost associated with fundal photo acquisition make that the more appealing option in terms of screening on a global scale, however further studies need to be undertaken, owing largely to the poor study quality associated with the fundal photography cohort.

Impact of Adapting the Abbreviated Injury Scale (AIS)-2005 from AIS-1998 on Injury Severity Scores and Clinical Outcome

BACKGROUND

In recent years, several versions of the Abbreviated Injury Scale (AIS) were updated and published. It was reported that the codeset in the dictionary of AIS-2005 had significant change from that of AIS-1998. This study was designed to evaluate the potential impact of adapting the AIS-2005 codeset from the AIS-1998 in an established trauma system of a single level I trauma center. The patients' outcome was measured in different Injury Severity Score (ISS) strata according to the double-coded injuries in a three-year period.

METHODS

The double-coded injuries sustained by 7520 trauma patients between 1 January, 2016, and 31 December, 2018, in a level I trauma center were used to compare the patient injury characteristics and outcomes between AIS-1998 and AIS-2005 and under different ISS strata, defined as <16 (mild to moderate injury), 16-24 (severe injury), and >24 (critical injury).

RESULTS

The mean ISS was significantly lower using AIS-2005 than using AIS-1998 (7.5 ± 6.3 vs. 8.3 ± 7.1, respectively, p < 0.001). AIS-2005 scores in the body regions of the head/neck (2.94 ± 1.08 vs. 3.40 ± 1.15, respectively, p < 0.001) and extremity (2.19 ± 0.56 vs. 2.24 ± 0.58, respectively, p < 0.001), but not in other body regions, were significantly lower than AIS-1998 scores. The critically injured patients (ISS >24), but not severely injured patients or patients with mild-to-moderate injury, coded by AIS-2005 had a significantly higher mortality rate (34.2% vs. 26.2%, respectively, p = 0.031) than did patients coded by AIS-1998. The rate of intensive care unit admission was significantly higher for patients in all ISS strata after adapting AIS-2005 as the scoring system than after adapting AIS-1998. Regarding patients with major trauma, which was defined as ISS > 15, the number of patients with major trauma in this study was 17.0% (n = 1276) for AIS-1998 and 9.7% (n = 733) for AIS-2005. As a consequence, the mortality rate of patients with major trauma was significantly higher in AIS-2005 than in AIS-1998 (15.4% vs. 9.1%, respectively, p < 000.1).

CONCLUSIONS

In this study, we revealed that the adaptation of AIS-2005 from AIS-1998 had resulted in a significant decrease of severity scores in the measurement of the same injuries. The number of head/neck injuries classified as 16-24 was the key difference between AIS-1998 and AIS-2005. Furthermore, critically injured patients who had ISS > 24 coded by AIS-2005 had significantly higher mortality rates than did the patients coded by AIS-1998. This study also indicated that a direct comparison of the measurements that are generated from these two AIS versions can produce misleading results.

Insights into Amyotrophic Lateral Sclerosis from a Machine Learning Perspective

Objective: Amyotrophic lateral sclerosis (ALS) disease state prediction usually assumes linear progression and uses a classifier evaluated by its accuracy. Since disease progression is not linear, and the accuracy measurement cannot tell large from small prediction errors, we dispense with the linearity assumption and apply ordinal classification that accounts for error severity. In addition, we identify the most influential variables in predicting and explaining the disease. Furthermore, in contrast to conventional modeling of the patient’s total functionality, we also model separate patient functionalities (e.g., in walking or speaking). Methods: Using data from 3772 patients from the Pooled Resource Open-Access ALS Clinical Trials (PRO-ACT) database, we introduce and train ordinal classifiers to predict patients’ disease state in their last clinic visit, while accounting differently for different error severities. We use feature-selection methods and the classifiers themselves to determine the most influential variables in predicting the disease from demographic, clinical, and laboratory data collected in either the first, last, or both clinic visits, and the Bayesian network classifier to identify interrelations among these variables and their relations with the disease state. We apply these methods to model each of the patient functionalities. Results: We show the error distribution in ALS state prediction and demonstrate that ordinal classifiers outperform classifiers that do not account for error severity. We identify clinical and lab test variables influential to prediction of different ALS functionalities and their interrelations, and specific value combinations of these variables that occur more frequently in patients with severe deterioration than in patients with mild deterioration and vice versa. Conclusions: Ordinal classification of ALS state is superior to conventional classification. Identification of influential ALS variables and their interrelations help explain disease mechanism. Modeling of patient functionalities separately allows relation of variables and their connections to different aspects of the disease as may be expressed in different body segments.