Supervised machine learning for the early prediction of acute respiratory distress syndrome (ARDS)

Machine learning-based forecast of Helmet-CPAP therapy failure in Acute Respiratory Distress Syndrome patients

BACKGROUND AND OBJECTIVE

Helmet-Continuous Positive Airway Pressure (H-CPAP) is a non-invasive respiratory support that is used for the treatment of Acute Respiratory Distress Syndrome (ARDS), a severe medical condition diagnosed when symptoms like profound hypoxemia, pulmonary opacities on radiography, or unexplained respiratory failure are present. It can be classified as mild, moderate or severe. H-CPAP therapy is recommended as the initial treatment approach for mild ARDS. Even though the efficacy of H-CPAP in managing patients with moderate-to-severe hypoxemia remains unclear, its use has increased for these cases in response to the emergence of the COVID-19 Pandemic. Using the electronic medical records (EMR) from the Pulmonology Department of Vimercate Hospital, in this study we develop and evaluate a Machine Learning (ML) system able to predict the failure of H-CPAP therapy on ARDS patients.

METHODS

The Vimercate Hospital EMR provides demographic information, blood tests, and vital parameters of all hospitalizations of patients who are treated with H-CPAP and diagnosed with ARDS. This data is used to create a dataset of 622 records and 38 features, with 70%-30% split between training and test sets. Different ML models such as SVM, XGBoost, Neural Network, Random Forest, and Logistic Regression are iteratively trained in a cross-validation fashion. We also apply a feature selection algorithm to improve predictions quality and reduce the number of features.

RESULTS AND CONCLUSIONS

The SVM and Neural Network models proved to be the most effective, achieving final accuracies of 95.19% and 94.65%, respectively. In terms of F1-score, the models scored 88.61% and 87.18%, respectively. Additionally, the SVM and XGBoost models performed well with a reduced number of features (23 and 13, respectively). The PaO2/FiO2 Ratio, C-Reactive Protein, and O2 Saturation resulted as the most important features, followed by Heartbeats, White Blood Cells, and D-Dimer, in accordance with the clinical scientific literature.

Accuracy of artificial intelligence algorithms in predicting acute respiratory distress syndrome: a systematic review and meta-analysis

BACKGROUND

Acute respiratory distress syndrome (ARDS) is a serious threat to human life. Hence, early and accurate diagnosis and treatment are crucial for patient survival. This meta-analysis evaluates the accuracy of artificial intelligence in the early diagnosis of ARDS and provides guidance for future research and applications.

METHODS

A search on PubMed, Embase, Cochrane, Web of Science, CNKI, Wanfang, Chinese Biomedical Literature (CBM), and VIP databases was systematically conducted, from their establishment to November 2023, to obtain eligible studies for the analysis and evaluation of the predictive effect of AI on ARDS. The retrieved literature was screened according to inclusion and exclusion criteria, the quality of the included studies was assessed using QUADAS-2, and the included studies were statistically analyzed.

RESULTS

Among the 2, 996 studies, 33 were included in this meta-analysis, which showed that the pooled sensitivity of AI in predicting ARDS was 0.81 (0.76-0.85), the pooled specificity was 0.88 (0.84-0.91), and the area under the receiver operating characteristic curve (AUC) was 0.91 (0.88-0.93). The analyzed studies included 28 models, with a pooled sensitivity of 0.79 (0.76-0.82), a pooled specificity of 0.85 (0.83-0.88), and an AUC of 0.89 (0.86-0.91). In the subgroup analysis, the pooled AUC of the AI models ANN, CNN, LR, RF, SVM, and XGB were 0.86 (0.83-0.89), 0.91 (0.88-0.93), 0.86 (0.83-0.89), and 0.89 (0.86-0.91), 0.90 (0.87-0.92), 0.93 (0.90-0.95), respectively. In an additional subgroup analysis, we evaluated the predictive performance of the AI models trained using different predictors. This meta-analysis was registered in PROSPERO (CRD42023491546).

CONCLUSION

AI has good sensitivity and specificity for predicting ARDS, indicating a high clinical application value. Algorithmic models such as CNN, SVM, and XGB have improved prediction performance. The subgroup analysis revealed that the model trained using images combined with other predictors had the best predictive performance.

The development of a C5.0 machine learning model in a limited data set to predict early mortality in patients with ARDS undergoing an initial session of prone positioning

BACKGROUND

Acute Respiratory Distress Syndrome (ARDS) has a high morbidity and mortality. One therapy that can decrease mortality is ventilation in the prone position (PP). Patients undergoing PP are amongst the sickest, and there is a need for early identification of patients at particularly high risk of death. These patients may benefit from an in-depth review of treatment or consideration of rescue therapies. We report the development of a machine learning model trained to predict early mortality in patients undergoing prone positioning as part of the management of their ARDS.

METHODS

Prospectively collected clinical data were analysed retrospectively from a single tertiary ICU. The records of patients who underwent an initial session of prone positioning whilst receiving invasive mechanical ventilation were identified (n = 131). The decision to perform prone positioning was based on the criteria in the PROSEVA study. A C5.0 classifier algorithm with adaptive boosting was trained on data gathered before, during, and after initial proning. Data was split between training (85% of data) and testing (15% of data). Hyperparameter tuning was achieved through a grid-search using a maximal entropy configuration. Predictions for 7-day mortality after initial proning session were made on the training and testing data.

RESULTS

The model demonstrated good performance in predicting 7-day mortality (AUROC: 0.89 training, 0.78 testing). Seven variables were used for prediction. Sensitivity was 0.80 and specificity was 0.67 on the testing data set. Patients predicted to survive had 13.3% mortality, while those predicted to die had 66.67% mortality. Among patients in whom the model predicted patient would survive to day 7 based on their response, mortality at day 7 was 13.3%. Conversely, if the model predicted the patient would not survive to day 7, mortality was 66.67%.

CONCLUSIONS

This proof-of-concept study shows that with a limited data set, a C5.0 classifier can predict 7-day mortality from a number of variables, including the response to initial proning, and identify a cohort at significantly higher risk of death. This can help identify patients failing conventional therapies who may benefit from a thorough review of their management, including consideration of rescue treatments, such as extracorporeal membrane oxygenation. This study shows the potential of a machine learning model to identify ARDS patients at high risk of early mortality following PP. This information can guide clinicians in tailoring treatment strategies and considering rescue therapies. Further validation in larger cohorts is needed.

Machine Learning Tools for Acute Respiratory Distress Syndrome Detection and Prediction

Machine learning (ML) tools for acute respiratory distress syndrome (ARDS) detection and prediction are increasingly used. Therefore, understanding risks and benefits of such algorithms is relevant at the bedside. ARDS is a complex and severe lung condition that can be challenging to define precisely due to its multifactorial nature. It often arises as a response to various underlying medical conditions, such as pneumonia, sepsis, or trauma, leading to widespread inflammation in the lungs. ML has shown promising potential in supporting the recognition of ARDS in ICU patients. By analyzing a variety of clinical data, including vital signs, laboratory results, and imaging findings, ML models can identify patterns and risk factors associated with the development of ARDS. This detection and prediction could be crucial for timely interventions, diagnosis and treatment. In summary, leveraging ML for the early prediction and detection of ARDS in ICU patients holds great potential to enhance patient care, improve outcomes, and contribute to the evolving landscape of precision medicine in critical care settings. This article is a concise definitive review on artificial intelligence and ML tools for the prediction and detection of ARDS in critically ill patients.

A Clinician's Guide to Understanding Bias in Critical Clinical Prediction Models

This narrative review focuses on the role of clinical prediction models in supporting informed decision-making in critical care, emphasizing their 2 forms: traditional scores and artificial intelligence (AI)-based models. Acknowledging the potential for both types to embed biases, the authors underscore the importance of critical appraisal to increase our trust in models. The authors outline recommendations and critical care examples to manage risk of bias in AI models. The authors advocate for enhanced interdisciplinary training for clinicians, who are encouraged to explore various resources (books, journals, news Web sites, and social media) and events (Datathons) to deepen their understanding of risk of bias.

Tree-based ensemble machine learning models in the prediction of acute respiratory distress syndrome following cardiac surgery: a multicenter cohort study

BACKGROUND

Acute respiratory distress syndrome (ARDS) after cardiac surgery is a severe respiratory complication with high mortality and morbidity. Traditional clinical approaches may lead to under recognition of this heterogeneous syndrome, potentially resulting in diagnosis delay. This study aims to develop and external validate seven machine learning (ML) models, trained on electronic health records data, for predicting ARDS after cardiac surgery.

METHODS

This multicenter, observational cohort study included patients who underwent cardiac surgery in the training and testing cohorts (data from Nanjing First Hospital), as well as those patients who had cardiac surgery in a validation cohort (data from Shanghai General Hospital). The number of important features was determined using the sliding windows sequential forward feature selection method (SWSFS). We developed a set of tree-based ML models, including Decision Tree, GBDT, AdaBoost, XGBoost, LightGBM, Random Forest, and Deep Forest. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and Brier score. The SHapley Additive exPlanation (SHAP) techinque was employed to interpret the ML model. Furthermore, a comparison was made between the ML models and traditional scoring systems. ARDS is defined according to the Berlin definition.

RESULTS

A total of 1996 patients who had cardiac surgery were included in the study. The top five important features identified by the SWSFS were chronic obstructive pulmonary disease, preoperative albumin, central venous pressure_T4, cardiopulmonary bypass time, and left ventricular ejection fraction. Among the seven ML models, Deep Forest demonstrated the best performance, with an AUC of 0.882 and a Brier score of 0.809 in the validation cohort. Notably, the SHAP values effectively illustrated the contribution of the 13 features attributed to the model output and the individual feature's effect on model prediction. In addition, the ensemble ML models demonstrated better performance than the other six traditional scoring systems.

CONCLUSIONS

Our study identified 13 important features and provided multiple ML models to enhance the risk stratification for ARDS after cardiac surgery. Using these predictors and ML models might provide a basis for early diagnostic and preventive strategies in the perioperative management of ARDS patients.

Early predictive values of clinical assessments for ARDS mortality: a machine-learning approach

Acute respiratory distress syndrome (ARDS) is a devastating critical care syndrome with significant morbidity and mortality. The objective of this study was to evaluate the predictive values of dynamic clinical indices by developing machine-learning (ML) models for early and accurate clinical assessment of the disease prognosis of ARDS. We conducted a retrospective observational study by applying dynamic clinical data collected in the ARDSNet FACTT Trial (n = 1000) to ML-based algorithms for predicting mortality. In order to compare the significance of clinical features dynamically, we further applied the random forest (RF) model to nine selected clinical parameters acquired at baseline and day 3 independently. An RF model trained using clinical data collected at day 3 showed improved performance and prognostication efficacy (area under the curve [AUC]: 0.84, 95% CI: 0.78-0.89) compared to baseline with an AUC value of 0.72 (95% CI: 0.65-0.78). Mean airway pressure (MAP), bicarbonate, age, platelet count, albumin, heart rate, and glucose were the most significant clinical indicators associated with mortality at day 3. Thus, clinical features collected early (day 3) improved performance of integrative ML models with better prognostication for mortality. Among these, MAP represented the most important feature for ARDS patients' early risk stratification.

ARDSFlag: an NLP/machine learning algorithm to visualize and detect high-probability ARDS admissions independent of provider recognition and billing codes

BACKGROUND

Despite the significance and prevalence of acute respiratory distress syndrome (ARDS), its detection remains highly variable and inconsistent. In this work, we aim to develop an algorithm (ARDSFlag) to automate the diagnosis of ARDS based on the Berlin definition. We also aim to develop a visualization tool that helps clinicians efficiently assess ARDS criteria.

METHODS

ARDSFlag applies machine learning (ML) and natural language processing (NLP) techniques to evaluate Berlin criteria by incorporating structured and unstructured data in an electronic health record (EHR) system. The study cohort includes 19,534 ICU admissions in the Medical Information Mart for Intensive Care III (MIMIC-III) database. The output is the ARDS diagnosis, onset time, and severity.

RESULTS

ARDSFlag includes separate text classifiers trained using large training sets to find evidence of bilateral infiltrates in radiology reports (accuracy of 91.9%±0.5%) and heart failure/fluid overload in radiology reports (accuracy 86.1%±0.5%) and echocardiogram notes (accuracy 98.4%±0.3%). A test set of 300 cases, which was blindly and independently labeled for ARDS by two groups of clinicians, shows that ARDSFlag generates an overall accuracy of 89.0% (specificity = 91.7%, recall = 80.3%, and precision = 75.0%) in detecting ARDS cases.

CONCLUSION

To our best knowledge, this is the first study to focus on developing a method to automate the detection of ARDS. Some studies have developed and used other methods to answer other research questions. Expectedly, ARDSFlag generates a significantly higher performance in all accuracy measures compared to those methods.

Digital health delivery in respiratory medicine: adjunct, replacement or cause for division?

Digital medicine is already well established in respiratory medicine through remote monitoring digital devices which are used in the day-to-day care of patients with asthma, COPD and sleep disorders. Image recognition software, deployed in thoracic radiology for many applications including lung cancer screening, is another application of digital medicine. Used as clinical decision support, this software will soon become part of day-to-day practice once concerns regarding generalisability have been addressed. Embodied in the electronic health record, digital medicine also plays a substantial role in the day-to-day clinical practice of respiratory medicine. Given the considerable work the electronic health record demands from clinicians, the next tangible impact of digital medicine may be artificial intelligence that aids administration, makes record keeping easier and facilitates better digital communication with patients. Future promises of digital medicine are based on their potential to analyse and characterise the large amounts of digital clinical data that are collected in routine care. Offering the potential to predict outcomes and personalise therapy, there is much to be excited by in this new epoch of innovation. However, these digital tools are by no means a silver bullet. It remains uncertain whether, let alone when, the promises of better models of personalisation and prediction will translate into clinically meaningful and cost-effective products for clinicians.

A systematic review of machine learning models for management, prediction and classification of ARDS

AIM

Acute respiratory distress syndrome or ARDS is an acute, severe form of respiratory failure characterised by poor oxygenation and bilateral pulmonary infiltrates. Advancements in signal processing and machine learning have led to promising solutions for classification, event detection and predictive models in the management of ARDS.

METHOD

In this review, we provide systematic description of different studies in the application of Machine Learning (ML) and artificial intelligence for management, prediction, and classification of ARDS. We searched the following databases: Google Scholar, PubMed, and EBSCO from 2009 to 2023. A total of 243 studies was screened, in which, 52 studies were included for review and analysis. We integrated knowledge of previous work providing the state of art and overview of explainable decision models in machine learning and have identified areas for future research.

RESULTS

Gradient boosting is the most common and successful method utilised in 12 (23.1%) of the studies. Due to limitation of data size available, neural network and its variation is used by only 8 (15.4%) studies. Whilst all studies used cross validating technique or separated database for validation, only 1 study validated the model with clinician input. Explainability methods were presented in 15 (28.8%) of studies with the most common method is feature importance which used 14 times.

CONCLUSION

For databases of 5000 or fewer samples, extreme gradient boosting has the highest probability of success. A large, multi-region, multi centre database is required to reduce bias and take advantage of neural network method. A framework for validating with and explaining ML model to clinicians involved in the management of ARDS would be very helpful for development and deployment of the ML model.

A pretrain-finetune approach for improving model generalizability in outcome prediction of acute respiratory distress syndrome patients

BACKGROUND

Early prediction of acute respiratory distress syndrome (ARDS) of critically ill patients in intensive care units (ICUs) has been intensively studied in the past years. Yet a prediction model trained on data from one hospital might not be well generalized to other hospitals. It is therefore essential to develop an accurate and generalizable ARDS prediction model adaptive to different hospital or medical centers.

METHODS

We analyzed electronic medical records of 200,859 and 50,920 hospitalized patients within 24 h after being diagnosed with ARDS from the Philips eICU Institute (eICU-CRD) and the Medical Information Mart for Intensive Care (MIMIC-IV) dataset, respectively. Patients were sorted into three groups, including rapid death, long stay, and recovery, based on their condition or outcome between 24 and 72 h after ARDS diagnosis. To improve prediction performance and generalizability, a "pretrain-finetune" approach was applied, where we pretrained models on the eICU-CRD dataset and performed model finetuning using only a part (35%) of the MIMIC-IV dataset, and then tested the finetuned models on the remaining data from the MIMIC-IV dataset. Well-known machine-learning algorithms, including logistic regression, random forest, extreme gradient boosting, and multilayer perceptron neural networks, were employed to predict ARDS outcomes. Prediction performance was evaluated using the area under the receiver-operating characteristic curve (AUC).

RESULTS

Results show that, in general, multilayer perceptron neural networks outperformed the other models. The use of pretrain-finetune yielded improved performance in predicting ARDS outcomes achieving a micro-AUC of 0.870 for the MIMIC-IV dataset, an improvement of 0.046 over the pretrain model.

CONCLUSIONS

The proposed pretrain-finetune approach can effectively improve model generalizability from one to another dataset in ARDS prediction.

Comparison between XGboost model and logistic regression model for predicting sepsis after extremely severe burns

OBJECTIVE

To compare an Extreme Gradient Boosting (XGboost) model with a multivariable logistic regression (LR) model for their ability to predict sepsis after extremely severe burns.

METHODS

For this observational study, patient demographic and clinical information were collected from medical records. The two models were evaluated using area under curve (AUC) of the receiver operating characteristic (ROC) curve.

RESULTS

Of the 103 eligible patients with extremely severe burns, 20 (19%) were in the sepsis group, and 83 (81%) in the non-sepsis group. The LR model showed that age, admission time, body index (BI), fibrinogen, and neutrophil to lymphocyte ratio (NLR) were risk factors for sepsis. Comparing AUC of the ROC curves, the XGboost model had a higher predictive performance (0.91) than the LR model (0.88). The SHAP visualization tool indicated fibrinogen, NLR, BI, and age were important features of sepsis in patients with extremely severe burns.

CONCLUSIONS

The XGboost model was superior to the LR model in predictive efficacy. Results suggest that, fibrinogen, NLR, BI, and age were correlated with sepsis after extremely severe burns.

Acute Respiratory Distress Syndrome: Definition, Diagnosis, and Routine Management

Acute respiratory distress syndrome (ARDS) is an acute inflammatory lung injury characterized by severe hypoxemic respiratory failure, bilateral opacities on chest imaging, and low lung compliance. ARDS is a heterogeneous syndrome that is the common end point of a wide variety of predisposing conditions, with complex pathophysiology and underlying mechanisms. Routine management of ARDS is centered on lung-protective ventilation strategies such as low tidal volume ventilation and targeting low airway pressures to avoid exacerbation of lung injury, as well as a conservative fluid management strategy.

ARDS Mortality Prediction Model Using Evolving Clinical Data and Chest Radiograph Analysis

INTRODUCTION

Within primary ARDS, SARS-CoV-2-associated ARDS (C-ARDS) emerged in late 2019, reaching its peak during the subsequent two years. Recent efforts in ARDS research have concentrated on phenotyping this heterogeneous syndrome to enhance comprehension of its pathophysiology.

METHODS AND RESULTS

A retrospective study was conducted on C-ARDS patients from April 2020 to February 2021, encompassing 110 participants with a mean age of 63.2 ± 11.92 (26-83 years). Of these, 61.2% (68) were male, and 25% (17) experienced severe ARDS, resulting in a mortality rate of 47.3% (52). Ventilation settings, arterial blood gases, and chest X-ray (CXR) were evaluated on the first day of invasive mechanical ventilation and between days two and three. CXR images were scrutinized using a convolutional neural network (CNN). A binary logistic regression model for predicting C-ARDS mortality was developed based on the most influential variables: age, PaO2/FiO2 ratio (P/F) on days one and three, CNN-extracted CXR features, and age. Initial performance assessment on test data (23 patients out of the 110) revealed an area under the receiver operating characteristic (ROC) curve of 0.862 with a 95% confidence interval (0.654-0.969).

CONCLUSION

Integrating data available in all intensive care units enables the prediction of C-ARDS mortality by utilizing evolving P/F ratios and CXR. This approach can assist in tailoring treatment plans and initiating early discussions to escalate care and extracorporeal life support. Machine learning algorithms for imaging classification can uncover otherwise inaccessible patterns, potentially evolving into another form of ARDS phenotyping. The combined features of these algorithms and clinical variables demonstrate superior performance compared to either element alone.

Chinese experts' consensus on the application of intensive care big data

The development of intensive care medicine is inseparable from the diversified monitoring data. Intensive care medicine has been closely integrated with data since its birth. Critical care research requires an integrative approach that embraces the complexity of critical illness and the computational technology and algorithms that can make it possible. Considering the need of standardization of application of big data in intensive care, Intensive Care Medicine Branch of China Health Information and Health Care Big Data Society, Standard Committee has convened expert group, secretary group and the external audit expert group to formulate Chinese Experts' Consensus on the Application of Intensive Care Big Data (2022). This consensus makes 29 recommendations on the following five parts: Concept of intensive care big data, Important scientific issues, Standards and principles of database, Methodology in solving big data problems, Clinical application and safety consideration of intensive care big data. The consensus group believes this consensus is the starting step of application big data in the field of intensive care. More explorations and big data based retrospective research should be carried out in order to enhance safety and reliability of big data based models of critical care field.

The implication of targeting PD-1:PD-L1 pathway in treating sepsis through immunostimulatory and anti-inflammatory pathways

Sepsis currently remains a major contributor to mortality in the intensive care unit (ICU), with 48.9 million cases reported globally and a mortality rate of 22.5% in 2017, accounting for almost 20% of all-cause mortality worldwide. This highlights the urgent need to improve the understanding and treatment of this condition. Sepsis is now recognized as a dysregulation of the host immune response to infection, characterized by an excessive inflammatory response and immune paralysis. This dysregulation leads to secondary infections, multiple organ dysfunction syndrome (MODS), and ultimately death. PD-L1, a co-inhibitory molecule expressed in immune cells, has emerged as a critical factor in sepsis. Numerous studies have found a significant association between the expression of PD-1/PD-L1 and sepsis, with a particular focus on PD-L1 expressed on neutrophils recently. This review explores the role of PD-1/PD-L1 in immunostimulatory and anti-inflammatory pathways, illustrates the intricate link between PD-1/PD-L1 and sepsis, and summarizes current therapeutic approaches against PD-1/PD-L1 in the treatment and prognosis of sepsis in preclinical and clinical studies.

Critical evaluation of established risk prediction models for acute respiratory distress syndrome in adult patients: A systematic review and meta-analysis

AIM

To assess the performance of validated prediction models for acute respiratory distress syndrome (ARDS) by systematic review and meta-analysis.

METHODS

Eight databases (Medline, CINAHL, Embase, The Cochrane Library, CNKI, WanFang Data, Sinomed, and VIP) were searched up to March 26, 2023. Studies developed and validated a prediction model for ARDS in adult patients were included. Items on study design, incidence, derivation methods, predictors, discrimination, and calibration were collected. The risk of bias was assessed by the Prediction model Risk of Bias Assessment Tool. Models with a reported area under the curve of the receiver operating characteristic (AUC) metric were analyzed.

RESULTS

A total of 25 studies were retrieved, including 48 unique prediction models. Discrimination was reported in all studies, with AUC ranging from 0.701 to 0.95. Emerged AUC value of the logistic regression model was 0.837 (95% CI: 0.814 to 0.859). Besides, the value in the ICU group was 0.856 (95% CI: 0.812 to 0.899), the acute pancreatitis group was 0.863 (95% CI: 0.844 to 0.882), and the postoperation group was 0.835 (95% CI: 0.808 to 0.861). In total, 24 of the included studies had a high risk of bias, which was mostly due to the improper methods in predictor screening (13/24), model calibration assessment (9/24), and dichotomization of continuous predictors (6/24).

CONCLUSIONS

This study shows that most prediction models for ARDS are at high risk of bias, and the discrimination ability of the model is excellent. Adherence to standardized guidelines for model development is necessary to derive a prediction model of value to clinicians.

Leveling Up: A Review of Machine Learning Models in the Cardiac ICU

Machine learning has emerged as a significant tool to augment the medical decision-making process. Studies have steadily accrued detailing algorithms and models designed using machine learning to predict and anticipate pathologic states. The cardiac intensive care unit is an area where anticipation is crucial in the division between life and death. In this paper, we aim to review important studies describing the utility of machine learning algorithms to describe the future of artificial intelligence in the cardiac intensive care unit, especially in regards to the prediction of successful ventilatory weaning, acute respiratory distress syndrome, arrhythmia, and acute kidney injury.

Clinician Trust in Artificial Intelligence: What is Known and How Trust Can Be Facilitated

Predictive analytics based on artificial intelligence (AI) offer clinicians the opportunity to leverage big data available in electronic health records (EHR) to improve clinical decision-making, and thus patient outcomes. Despite this, many barriers exist to facilitating trust between clinicians and AI-based tools, limiting its current impact. Potential solutions are available at both the local and national level. It will take a broad and diverse coalition of stakeholders, from health-care systems, EHR vendors, and clinical educators to regulators, researchers and the patient community, to help facilitate this trust so that the promise of AI in health care can be realized.

Artificial Intelligence-Driven Respiratory Distress Syndrome Prediction for Very Low Birth Weight Infants: Korean Multicenter Prospective Cohort Study

BACKGROUND

Respiratory distress syndrome (RDS) is a disease that commonly affects premature infants whose lungs are not fully developed. RDS results from a lack of surfactant in the lungs. The more premature the infant is, the greater is the likelihood of having RDS. However, even though not all premature infants have RDS, preemptive treatment with artificial pulmonary surfactant is administered in most cases.

OBJECTIVE

We aimed to develop an artificial intelligence model to predict RDS in premature infants to avoid unnecessary treatment.

METHODS

In this study, 13,087 very low birth weight infants who were newborns weighing less than 1500 grams were assessed in 76 hospitals of the Korean Neonatal Network. To predict RDS in very low birth weight infants, we used basic infant information, maternity history, pregnancy/birth process, family history, resuscitation procedure, and test results at birth such as blood gas analysis and Apgar score. The prediction performances of 7 different machine learning models were compared, and a 5-layer deep neural network was proposed in order to enhance the prediction performance from the selected features. An ensemble approach combining multiple models from the 5-fold cross-validation was subsequently developed.

RESULTS

Our proposed ensemble 5-layer deep neural network consisting of the top 20 features provided high sensitivity (83.03%), specificity (87.50%), accuracy (84.07%), balanced accuracy (85.26%), and area under the curve (0.9187). Based on the model that we developed, a public web application that enables easy access for the prediction of RDS in premature infants was deployed.

CONCLUSIONS

Our artificial intelligence model may be useful for preparations for neonatal resuscitation, particularly in cases involving the delivery of very low birth weight infants, as it can aid in predicting the likelihood of RDS and inform decisions regarding the administration of surfactant.

Developing an Artificial Intelligence-Based Representation of a Virtual Patient Model for Real-Time Diagnosis of Acute Respiratory Distress Syndrome

Acute Respiratory Distress Syndrome (ARDS) is a condition that endangers the lives of many Intensive Care Unit patients through gradual reduction of lung function. Due to its heterogeneity, this condition has been difficult to diagnose and treat, although it has been the subject of continuous research, leading to the development of several tools for modeling disease progression on the one hand, and guidelines for diagnosis on the other, mainly the "Berlin Definition". This paper describes the development of a deep learning-based surrogate model of one such tool for modeling ARDS onset in a virtual patient: the Nottingham Physiology Simulator. The model-development process takes advantage of current machine learning and data-analysis techniques, as well as efficient hyperparameter-tuning methods, within a high-performance computing-enabled data science platform. The lightweight models developed through this process present comparable accuracy to the original simulator (per-parameter R2 > 0.90). The experimental process described herein serves as a proof of concept for the rapid development and dissemination of specialised diagnosis support systems based on pre-existing generalised mechanistic models, making use of supercomputing infrastructure for the development and testing processes and supported by open-source software for streamlined implementation in clinical routines.

To Establish an Early Prediction Model for Acute Respiratory Distress Syndrome in Severe Acute Pancreatitis Using Machine Learning Algorithm

OBJECTIVE

To develop binary and quaternary classification prediction models in patients with severe acute pancreatitis (SAP) using machine learning methods, so that doctors can evaluate the risk of patients with acute respiratory distress syndrome (ARDS) and severe ARDS at an early stage.

METHODS

A retrospective study was conducted on SAP patients hospitalized in our hospital from August 2017 to August 2022. Logical Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), and eXtreme Gradient Boosting (XGB) were used to build the binary classification prediction model of ARDS. Shapley Additive explanations (SHAP) values were used to interpret the machine learning model, and the model was optimized according to the interpretability results of SHAP values. Combined with the optimized characteristic variables, four-class classification models, including RF, SVM, DT, XGB, and Artificial Neural Network (ANN), were constructed to predict mild, moderate, and severe ARDS, and the prediction effects of each model were compared.

RESULTS

The XGB model showed the best effect (AUC = 0.84) in the prediction of binary classification (ARDS or non-ARDS). According to SHAP values, the prediction model of ARDS severity was constructed with four characteristic variables (PaO₂/FiO₂, APACHE II, SOFA, AMY). Among them, the overall prediction accuracy of ANN is 86%, which is the best.

CONCLUSIONS

Machine learning has a good effect in predicting the occurrence and severity of ARDS in SAP patients. It can also provide a valuable tool for doctors to make clinical decisions.

Computational Simulation of Virtual Patients Reduces Dataset Bias and Improves Machine Learning-Based Detection of ARDS from Noisy Heterogeneous ICU Datasets

Goal: Machine learning (ML) technologies that leverage large-scale patient data are promising tools predicting disease evolution in individual patients. However, the limited generalizability of ML models developed on single-center datasets, and their unproven performance in real-world settings, remain significant constraints to their widespread adoption in clinical practice. One approach to tackle this issue is to base learning on large multi-center datasets. However, such heterogeneous datasets can introduce further biases driven by data origin, as data structures and patient cohorts may differ between hospitals. Methods: In this paper, we demonstrate how mechanistic virtual patient (VP) modeling can be used to capture specific features of patients' states and dynamics, while reducing biases introduced by heterogeneous datasets. We show how VP modeling can be used for data augmentation through identification of individualized model parameters approximating disease states of patients with suspected acute respiratory distress syndrome (ARDS) from observational data of mixed origin. We compare the results of an unsupervised learning method (clustering) in two cases: where the learning is based on original patient data and on data derived in the matching procedure of the VP model to real patient data. Results: More robust cluster configurations were observed in clustering using the model-derived data. VP model-based clustering also reduced biases introduced by the inclusion of data from different hospitals and was able to discover an additional cluster with significant ARDS enrichment. Conclusions: Our results indicate that mechanistic VP modeling can be used to significantly reduce biases introduced by learning from heterogeneous datasets and to allow improved discovery of patient cohorts driven exclusively by medical conditions.

Discriminating Acute Respiratory Distress Syndrome from other forms of respiratory failure via iterative machine learning

Acute Respiratory Distress Syndrome (ARDS) is associated with high morbidity and mortality. Identification of ARDS enables lung protective strategies, quality improvement interventions, and clinical trial enrolment, but remains challenging particularly in the first 24 hours of mechanical ventilation. To address this we built an algorithm capable of discriminating ARDS from other similarly presenting disorders immediately following mechanical ventilation. Specifically, a clinical team examined medical records from 1263 ICU-admitted, mechanically ventilated patients, retrospectively assigning each patient a diagnosis of "ARDS" or "non-ARDS" (e.g., pulmonary edema). Exploiting data readily available in the clinical setting, including patient demographics, laboratory test results from before the initiation of mechanical ventilation, and features extracted by natural language processing of radiology reports, we applied an iterative pre-processing and machine learning framework. The resulting model successfully discriminated ARDS from non-ARDS causes of respiratory failure (AUC = 0.85) among patients meeting Berlin criteria for severe hypoxia. This analysis also highlighted novel patient variables that were informative for identifying ARDS in ICU settings.

Early Prediction of All-Cause Clinical Deterioration in General Wards Patients: Development and Validation of a Biomarker-Based Machine Learning Model Derived From Rapid Response Team Activations

OBJECTIVE

The aim of the study is to evaluate the performance of a biomarker-based machine learning (ML) model (not including vital signs) derived from reviewed rapid response team (RRT) activations in predicting all-cause deterioration in general wards patients.

DESIGN

This is a retrospective single-institution study. All consecutive adult patients' cases on noncritical wards identified by RRT calls occurring at least 24 hours after patient admission, between April 2018 and June 2020, were included. The cases were reviewed and labeled for clinical deterioration by a multidisciplinary expert consensus panel. A supervised learning approach was adopted based on a set of biomarkers and demographic data available in the patient's electronic medical record (EMR).

SETTING

The setting is a 250-bed tertiary university hospital with a basic EMR, with adult (>18 y) patients on general wards.

PATIENTS

The study analyzed the cases of 514 patients for which the RRT was activated. Rapid response teams were extracted from the hospital telephone log data. Two hundred eighteen clinical deterioration cases were identified in these patients after expert chart review and complemented by 146 "nonevent" cases to build the training and validation data set.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

The best performance was achieved with the random forests algorithm, with a maximal area under the receiver operating curve of 0.90 and F1 score of 0.85 obtained at prediction time T0-6h, slightly decreasing but still acceptable (area under the receiver operating curve, >0.8; F1 score, >0.75) at T0-42h. The system outperformed most classical track-and-trigger systems both in terms of prediction performance and prediction horizon.

CONCLUSIONS

In hospitals with a basic EMR, a biomarker-based ML model could be used to predict clinical deterioration in general wards patients earlier than classical track-and-trigger systems, thus enabling appropriate clinical interventions for patient safety and improved outcomes.

Multi-Modal Data Analysis for Pneumonia Status Prediction Using Deep Learning (MDA-PSP)

Evaluating several vital signs and chest X-ray (CXR) reports regularly to determine the recovery of the pneumonia patients at general wards is a challenge for doctors. A recent study shows the identification of pneumonia by the history of symptoms and signs including vital signs, CXR, and other clinical parameters, but they lack predicting the recovery status after starting treatment. The goal of this paper is to provide a pneumonia status prediction system for the early affected patient’s discharge from the hospital within 7 days or late discharge more than 7 days. This paper aims to design a multimodal data analysis for pneumonia status prediction using deep learning classification (MDA-PSP). We have developed a system that takes an input of vital signs and CXR images of the affected patient with pneumonia from admission day 1 to day 3. The deep learning then classifies the health status improvement or deterioration for predicting the possible discharge state. Therefore, the scope is to provide a highly accurate prediction of the pneumonia recovery on the 7th day after 3-day treatment by the SHAP (SHapley Additive exPlanation), imputation, adaptive imputation-based preprocessing of the vital signs, and CXR image feature extraction using deep learning based on dense layers-batch normalization (BN) with class weights for the first 7 days’ general ward patient in MDA-PSP. A total of 3972 patients with pneumonia were enrolled by de-identification with an adult age of 71 mean ± 17 sd and 64% of them were male. After analyzing the data behavior, appropriate improvement measures are taken by data preprocessing and feature vectorization algorithm. The deep learning method of Dense-BN with SHAP features has an accuracy of 0.77 for vital signs, 0.92 for CXR, and 0.75 for the combined model with class weights. The MDA-PSP hybrid method-based experiments are proven to demonstrate higher prediction accuracy of 0.75 for pneumonia patient status. Henceforth, the hybrid methods of machine and deep learning for pneumonia patient discharge are concluded to be a better approach.

Early prediction of moderate-to-severe condition of inhalation-induced acute respiratory distress syndrome via interpretable machine learning

Background

Several studies have investigated the correlation between physiological parameters and the risk of acute respiratory distress syndrome (ARDS), in addition, etiology-associated heterogeneity in ARDS has become an emerging topic quite recently; however, the intersection between the two, which is early prediction of target conditions in etiology-specific ARDS, has not been well-studied. We aimed to develop and validate a machine-learning model for the early prediction of moderate-to-severe condition of inhalation-induced ARDS.

Methods

Clinical expertise was applied with data-driven analysis. Using data from electronic intensive care units (retrospective derivation cohort) and the three most accessible vital signs (i.e. heart rate, temperature, and respiratory rate) together with feature engineering, we applied a random forest approach during the time window of 90 h that ended 6 h prior to the onset of moderate-to-severe respiratory failure (the ratio of partial pressure of arterial oxygen to fraction of inspired oxygen ≤ 200 mmHg).

Results

The trained random forest classifier was validated using two independent validation cohorts, with an area under the curve of 0.9127 (95% confidence interval 0.8713–0.9542) and 0.9026 (95% confidence interval 0.8075–1), respectively. A Stable and Interpretable RUle Set (SIRUS) was used to extract rules from the RF to provide guidelines for clinicians. We identified several predictive factors, including resp_96h_6h_min < 9, resp_96h_6h_mean ≥ 16.1, HR_96h_6h_mean ≥ 102, and temp_96h_6h_max > 100, that could be used for predicting inhalation-induced ARDS (moderate-to-severe condition) 6 h prior to onset in critical care units. (‘xxx_96h_6h_min/mean/max’: the minimum/mean/maximum values of the xxx vital sign collected during a 90 h time window beginning 96 h prior to the onset of ARDS and ending 6 h prior to the onset from every recorded blood gas test).

Conclusions

This newly established random forest‑based interpretable model shows good predictive ability for moderate-to-severe inhalation-induced ARDS and may assist clinicians in decision-making, as well as facilitate the enrolment of patients in prevention programmes to improve their outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-022-01963-7.

Considerations for the implementation of machine learning into acute care settings

INTRODUCTION

Management of patients in the acute care setting requires accurate diagnosis and rapid initiation of validated treatments; therefore, this setting is likely to be an environment in which cognitive augmentation of the clinician's provision of care with technology rooted in artificial intelligence, such as machine learning (ML), is likely to eventuate.

SOURCES OF DATA

PubMed and Google Scholar with search terms that included ML, intensive/critical care unit, electronic health records (EHR), anesthesia information management systems and clinical decision support were the primary sources for this report.

AREAS OF AGREEMENT

Different categories of learning of large clinical datasets, often contained in EHRs, are used for training in ML. Supervised learning uses algorithm-based models, including support vector machines, to pair patients' attributes with an expected outcome. Unsupervised learning uses clustering algorithms to define to which disease grouping a patient's attributes most closely approximates. Reinforcement learning algorithms use ongoing environmental feedback to deterministically pursue likely patient outcome.

AREAS OF CONTROVERSY

Application of ML can result in undesirable outcomes over concerns related to fairness, transparency, privacy and accountability. Whether these ML technologies irrevocably change the healthcare workforce remains unresolved.

GROWING POINTS

Well-resourced Learning Health Systems are likely to exploit ML technology to gain the fullest benefits for their patients. How these clinical advantages can be extended to patients in health systems that are neither well-endowed, nor have the necessary data gathering technologies, needs to be urgently addressed to avoid further disparities in healthcare.

State of the Art of Machine Learning-Enabled Clinical Decision Support in Intensive Care Units: Literature Review

Background

Modern clinical care in intensive care units is full of rich data, and machine learning has great potential to support clinical decision-making. The development of intelligent machine learning–based clinical decision support systems is facing great opportunities and challenges. Clinical decision support systems may directly help clinicians accurately diagnose, predict outcomes, identify risk events, or decide treatments at the point of care.

Objective

We aimed to review the research and application of machine learning–enabled clinical decision support studies in intensive care units to help clinicians, researchers, developers, and policy makers better understand the advantages and limitations of machine learning–supported diagnosis, outcome prediction, risk event identification, and intensive care unit point-of-care recommendations.

Methods

We searched papers published in the PubMed database between January 1980 and October 2020. We defined selection criteria to identify papers that focused on machine learning–enabled clinical decision support studies in intensive care units and reviewed the following aspects: research topics, study cohorts, machine learning models, analysis variables, and evaluation metrics.

Results

A total of 643 papers were collected, and using our selection criteria, 97 studies were found. Studies were categorized into 4 topics—monitoring, detection, and diagnosis (13/97, 13.4%), early identification of clinical events (32/97, 33.0%), outcome prediction and prognosis assessment (46/97, 47.6%), and treatment decision (6/97, 6.2%). Of the 97 papers, 82 (84.5%) studies used data from adult patients, 9 (9.3%) studies used data from pediatric patients, and 6 (6.2%) studies used data from neonates. We found that 65 (67.0%) studies used data from a single center, and 32 (33.0%) studies used a multicenter data set; 88 (90.7%) studies used supervised learning, 3 (3.1%) studies used unsupervised learning, and 6 (6.2%) studies used reinforcement learning. Clinical variable categories, starting with the most frequently used, were demographic (n=74), laboratory values (n=59), vital signs (n=55), scores (n=48), ventilation parameters (n=43), comorbidities (n=27), medications (n=18), outcome (n=14), fluid balance (n=13), nonmedicine therapy (n=10), symptoms (n=7), and medical history (n=4). The most frequently adopted evaluation metrics for clinical data modeling studies included area under the receiver operating characteristic curve (n=61), sensitivity (n=51), specificity (n=41), accuracy (n=29), and positive predictive value (n=23).

Conclusions

Early identification of clinical and outcome prediction and prognosis assessment contributed to approximately 80% of studies included in this review. Using new algorithms to solve intensive care unit clinical problems by developing reinforcement learning, active learning, and time-series analysis methods for clinical decision support will be greater development prospects in the future.

Machine Learning Model for Predicting Acute Respiratory Failure in Individuals With Moderate-to-Severe Traumatic Brain Injury

Background: There is a high incidence of acute respiratory failure (ARF) in moderate or severe traumatic brain injury (M-STBI), worsening outcomes. This study aimed to design a predictive model for ARF.

Methods: Adult patients with M-STBI [3 ≤ Glasgow Coma Scale (GCS) ≤ 12] with a definite history of brain trauma and abnormal head on CT images, obtained from September 2015 to May 2017, were included. Patients with age >80 years or <18 years, multiple injuries with TBI upon admission, or pregnancy (in women) were excluded. Two models based on machine learning extreme gradient boosting (XGBoost) or logistic regression, respectively, were developed for predicting ARF within 48 h upon admission. These models were evaluated by out-of-sample validation. The samples were assigned to the training and test sets at a ratio of 3:1.

Results: In total, 312 patients were analyzed including 132 (42.3%) patients who had ARF. The GCS and the Marshall CT score, procalcitonin (PCT), and C-reactive protein (CRP) on admission significantly predicted ARF. The novel machine learning XGBoost model was superior to logistic regression model in predicting ARF [area under the receiver operating characteristic (AUROC) = 0.903, 95% CI, 0.834–0.966 vs. AUROC = 0.798, 95% CI, 0.697–0.899; p < 0.05].

Conclusion: The XGBoost model could better predict ARF in comparison with logistic regression-based model. Therefore, machine learning methods could help to develop and validate novel predictive models.

Multi-Task Learning with Recurrent Neural Networks for ARDS Prediction using only EHR Data: Model Development and Validation Study (Preprint)

Background

Acute respiratory distress syndrome (ARDS) is a condition that is often considered to have broad and subjective diagnostic criteria and is associated with significant mortality and morbidity. Early and accurate prediction of ARDS and related conditions such as hypoxemia and sepsis could allow timely administration of therapies, leading to improved patient outcomes.

Objective

The aim of this study is to perform an exploration of how multilabel classification in the clinical setting can take advantage of the underlying dependencies between ARDS and related conditions to improve early prediction of ARDS in patients.

Methods

The electronic health record data set included 40,703 patient encounters from 7 hospitals from April 20, 2018, to March 17, 2021. A recurrent neural network (RNN) was trained using data from 5 hospitals, and external validation was conducted on data from 2 hospitals. In addition to ARDS, 12 target labels for related conditions such as sepsis, hypoxemia, and COVID-19 were used to train the model to classify a total of 13 outputs. As a comparator, XGBoost models were developed for each of the 13 target labels. Model performance was assessed using the area under the receiver operating characteristic curve. Heat maps to visualize attention scores were generated to provide interpretability to the neural networks. Finally, cluster analysis was performed to identify potential phenotypic subgroups of patients with ARDS.

Results

The single RNN model trained to classify 13 outputs outperformed the individual XGBoost models for ARDS prediction, achieving an area under the receiver operating characteristic curve of 0.842 on the external test sets. Models trained on an increasing number of tasks resulted in improved performance. Earlier prediction of ARDS nearly doubled the rate of in-hospital survival. Cluster analysis revealed distinct ARDS subgroups, some of which had similar mortality rates but different clinical presentations.

Conclusions

The RNN model presented in this paper can be used as an early warning system to stratify patients who are at risk of developing one of the multiple risk outcomes, hence providing practitioners with the means to take early action.

Artificial intelligence–aided diagnosis model for acute respiratory distress syndrome combining clinical data and chest radiographs

Objective

The aim of this study was to develop an artificial intelligence–based model to detect the presence of acute respiratory distress syndrome (ARDS) using clinical data and chest X-ray (CXR) data.

Method

The transfer learning method was used to train a convolutional neural network (CNN) model with an external image dataset to extract the image features. Then, the last layer of the model was fine-tuned to determine the probability of ARDS. The clinical data were trained using three machine learning algorithms—eXtreme Gradient Boosting (XGB), random forest (RF), and logistic regression (LR)—to estimate the probability of ARDS. Finally, ensemble-weighted methods were proposed that combined the image model and the clinical data model to estimate the probability of ARDS. An analysis of the importance of clinical features was performed to explore the most important features in detecting ARDS. A gradient-weighted class activation mapping (Grad-CAM) model was used to explain what our CNN sees and understands when making a decision.

Results

The proposed ensemble-weighted methods improved the performances of the ARDS classifiers (XGB + CNN, area under the curve [AUC] = 0.916; RF + CNN, AUC = 0.920; LR + CNN, AUC = 0.920; XGB + RF + LR + CNN, AUC = 0.925). In addition, the ML model using clinical data to present the top 15 important features to identify the risk factors of ARDS.

Conclusion

This study developed combined machine learning models with clinical data and CXR images to detect ARDS. According to the results of the Shapley Additive exPlanations values and the Grad-CAM techniques, an explicable ARDS diagnosis model is suitable for a real-life scenario.

The Use of Artificial Intelligence and Machine Learning in Surgery: A Comprehensive Literature Review

Interest in the use of artificial intelligence (AI) and machine learning (ML) in medicine has grown exponentially over the last few years. With its ability to enhance speed, precision, and efficiency, AI has immense potential, especially in the field of surgery. This article aims to provide a comprehensive literature review of artificial intelligence as it applies to surgery and discuss practical examples, current applications, and challenges to the adoption of this technology. Furthermore, we elaborate on the utility of natural language processing and computer vision in improving surgical outcomes, research, and patient care.

eARDS: A multi-center validation of an interpretable machine learning algorithm of early onset Acute Respiratory Distress Syndrome (ARDS) among critically ill adults with COVID-19

We present an interpretable machine learning algorithm called ‘eARDS’ for predicting ARDS in an ICU population comprising COVID-19 patients, up to 12-hours before satisfying the Berlin clinical criteria. The analysis was conducted on data collected from the Intensive care units (ICU) at Emory Healthcare, Atlanta, GA and University of Tennessee Health Science Center, Memphis, TN and the Cerner^® Health Facts Deidentified Database, a multi-site COVID-19 EMR database. The participants in the analysis consisted of adults over 18 years of age. Clinical data from 35,804 patients who developed ARDS and controls were used to generate predictive models that identify risk for ARDS onset up to 12-hours before satisfying the Berlin criteria. We identified salient features from the electronic medical record that predicted respiratory failure among this population. The machine learning algorithm which provided the best performance exhibited AUROC of 0.89 (95% CI = 0.88–0.90), sensitivity of 0.77 (95% CI = 0.75–0.78), specificity 0.85 (95% CI = 085–0.86). Validation performance across two separate health systems (comprising 899 COVID-19 patients) exhibited AUROC of 0.82 (0.81–0.83) and 0.89 (0.87, 0.90). Important features for prediction of ARDS included minimum oxygen saturation (SpO₂), standard deviation of the systolic blood pressure (SBP), O₂ flow, and maximum respiratory rate over an observational window of 16-hours. Analyzing the performance of the model across various cohorts indicates that the model performed best among a younger age group (18–40) (AUROC = 0.93 [0.92–0.94]), compared to an older age group (80+) (AUROC = 0.81 [0.81–0.82]). The model performance was comparable on both male and female groups, but performed significantly better on the severe ARDS group compared to the mild and moderate groups. The eARDS system demonstrated robust performance for predicting COVID19 patients who developed ARDS at least 12-hours before the Berlin clinical criteria, across two independent health systems.

Predicting Duration of Mechanical Ventilation in Acute Respiratory Distress Syndrome Using Supervised Machine Learning

Background: Acute respiratory distress syndrome (ARDS) is an intense inflammatory process of the lungs. Most ARDS patients require mechanical ventilation (MV). Few studies have investigated the prediction of MV duration over time. We aimed at characterizing the best early scenario during the first two days in the intensive care unit (ICU) to predict MV duration after ARDS onset using supervised machine learning (ML) approaches. Methods: For model description, we extracted data from the first 3 ICU days after ARDS diagnosis from patients included in the publicly available MIMIC-III database. Disease progression was tracked along those 3 ICU days to assess lung severity according to Berlin criteria. Three robust supervised ML techniques were implemented using Python 3.7 (Light Gradient Boosting Machine (LightGBM); Random Forest (RF); and eXtreme Gradient Boosting (XGBoost)) for predicting MV duration. For external validation, we used the publicly available multicenter database eICU. Results: A total of 2466 and 5153 patients in MIMIC-III and eICU databases, respectively, received MV for >48 h. Median MV duration of extracted patients was 6.5 days (IQR 4.4–9.8 days) in MIMIC-III and 5.0 days (IQR 3.0–9.0 days) in eICU. LightGBM was the best model in predicting MV duration after ARDS onset in MIMIC-III with a root mean square error (RMSE) of 6.10–6.41 days, and it was externally validated in eICU with RMSE of 5.87–6.08 days. The best early prediction model was obtained with data captured in the 2nd day. Conclusions: Supervised ML can make early and accurate predictions of MV duration in ARDS after onset over time across ICUs. Supervised ML models might have important implications for optimizing ICU resource utilization and high acute cost reduction of MV.

Semi-supervised deep learning from time series clinical data for acute respiratory distress syndrome prediction: model development and validation study

Background

A high number of patients who are hospitalized with COVID-19 develop acute respiratory distress syndrome (ARDS).

Objective

In response to the need for clinical decision support tools to help manage the next pandemic during the early stages (ie, when limited labeled data are present), we developed machine learning algorithms that use semisupervised learning (SSL) techniques to predict ARDS development in general and COVID-19 populations based on limited labeled data.

Methods

SSL techniques were applied to 29,127 encounters with patients who were admitted to 7 US hospitals from May 1, 2019, to May 1, 2021. A recurrent neural network that used a time series of electronic health record data was applied to data that were collected when a patient’s peripheral oxygen saturation level fell below the normal range (<97%) to predict the subsequent development of ARDS during the remaining duration of patients’ hospital stay. Model performance was assessed with the area under the receiver operating characteristic curve and area under the precision recall curve of an external hold-out test set.

Results

For the whole data set, the median time between the first peripheral oxygen saturation measurement of <97% and subsequent respiratory failure was 21 hours. The area under the receiver operating characteristic curve for predicting subsequent ARDS development was 0.73 when the model was trained on a labeled data set of 6930 patients, 0.78 when the model was trained on the labeled data set that had been augmented with the unlabeled data set of 16,173 patients by using SSL techniques, and 0.84 when the model was trained on the entire training set of 23,103 labeled patients.

Conclusions

In the context of using time-series inpatient data and a careful model training design, unlabeled data can be used to improve the performance of machine learning models when labeled data for predicting ARDS development are scarce or expensive.

Generalization in Clinical Prediction Models: The Blessing and Curse of Measurement Indicator Variables

Supplemental Digital Content is available in the text.

OBJECTIVE:

Specific factors affecting generalizability of clinical prediction models are poorly understood. Our main objective was to investigate how measurement indicator variables affect external validity in clinical prediction models for predicting onset of vasopressor therapy.

DESIGN:

We fit logistic regressions on retrospective cohorts to predict vasopressor onset using two classes of variables: seemingly objective clinical variables (vital signs and laboratory measurements) and more subjective variables denoting recency of measurements.

SETTING:

Three cohorts from two tertiary-care academic hospitals in geographically distinct regions, spanning general inpatient and critical care settings.

PATIENTS:

Each cohort consisted of adult patients (age greater than or equal to 18 yr at time of hospitalization), with lengths of stay between 6 and 600 hours, and who did not receive vasopressors in the first 6 hours of hospitalization or ICU admission. Models were developed on each of the three derivation cohorts and validated internally on the derivation cohort and externally on the other two cohorts.

INTERVENTIONS:

None.

MEASUREMENTS AND MAIN RESULTS:

The prevalence of vasopressors was 0.9% in the general inpatient cohort and 12.4% and 11.5% in the two critical care cohorts. Models utilizing both classes of variables performed the best in-sample, with C-statistics for predicting vasopressor onset in 4 hours of 0.862 (95% CI, 0.844–0.879), 0.822 (95% CI, 0.793–0.852), and 0.889 (95% CI, 0.880–0.898). Models solely using the subjective variables denoting measurement recency had poor external validity. However, these practice-driven variables helped adjust for differences between the two hospitals and led to more generalizable models using clinical variables.

CONCLUSIONS:

We developed and externally validated models for predicting the onset of vasopressors. We found that practice-specific features denoting measurement recency improved local performance and also led to more generalizable models if they are adjusted for during model development but discarded at validation. The role of practice-specific features such as measurement indicators in clinical prediction modeling should be carefully considered if the goal is to develop generalizable models.

Utilizing Artificial Intelligence in Critical Care: Adding A Handy Tool to Our Armamentarium

We have witnessed rapid advancement in technology over the last few decades. With the advent of artificial intelligence (AI), newer avenues have opened for researchers. AI has added an entirely new dimension to this technological boom. Researchers in medical science have been excited about the tantalizing prospect of utilizing AI for the benefit of patient care. Lately, we have come across studies trying to test and validate various models based on AI to improve patient care strategies in critical care medicine as well. Thus, in this review, we will attempt to succinctly review current literature discussing AI in critical care medicine and analyze its future utility based on prevailing evidence.

Novel criteria to classify ARDS severity using a machine learning approach

Background

Usually, arterial oxygenation in patients with the acute respiratory distress syndrome (ARDS) improves substantially by increasing the level of positive end-expiratory pressure (PEEP). Herein, we are proposing a novel variable [PaO₂/(FiO₂xPEEP) or P/FP_E] for PEEP ≥ 5 to address Berlin’s definition gap for ARDS severity by using machine learning (ML) approaches.

Methods

We examined P/FP_E values delimiting the boundaries of mild, moderate, and severe ARDS. We applied ML to predict ARDS severity after onset over time by comparing current Berlin PaO₂/FiO₂ criteria with P/FP_E under three different scenarios. We extracted clinical data from the first 3 ICU days after ARDS onset (N = 2738, 1519, and 1341 patients, respectively) from MIMIC-III database according to Berlin criteria for severity. Then, we used the multicenter database eICU (2014–2015) and extracted data from the first 3 ICU days after ARDS onset (N = 5153, 2981, and 2326 patients, respectively). Disease progression in each database was tracked along those 3 ICU days to assess ARDS severity. Three robust ML classification techniques were implemented using Python 3.7 (LightGBM, RF, and XGBoost) for predicting ARDS severity over time.

Results

P/FP_E ratio outperformed PaO₂/FiO₂ ratio in all ML models for predicting ARDS severity after onset over time (MIMIC-III: AUC 0.711–0.788 and CORR 0.376–0.566; eICU: AUC 0.734–0.873 and CORR 0.511–0.745).

Conclusions

The novel P/FP_E ratio to assess ARDS severity after onset over time is markedly better than current PaO₂/FiO₂ criteria. The use of P/FP_E could help to manage ARDS patients with a more precise therapeutic regimen for each ARDS category of severity.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13054-021-03566-w.

Immune Deregulation in Sepsis and Septic Shock: Reversing Immune Paralysis by Targeting PD-1/PD-L1 Pathway

Sepsis remains a major problem for human health worldwide, thereby manifesting high rates of morbidity and mortality. Sepsis, once understood as a monophasic sustained hyperinflammation, is currently recognized as a dysregulated host response to infection, with both hyperinflammation and immunoparalysis occurring simultaneously from the earliest stages of sepsis, involving multiple organ dysfunctions. Despite the recent progress in the understanding of the pathophysiology underlying sepsis, no specific treatment to restore immune dysregulation in sepsis has been validated in clinical trials. In recent years, treatment for immune checkpoints such as the programmed cell death protein 1/programmed death ligand (PD-1/PD-L) pathway in tumor-infiltrating T-lymphocytes has been successful in the field of cancer immune therapy. As immune-paralysis in sepsis involves exhausted T-lymphocytes, future clinical applications of checkpoint inhibitors for sepsis are expected. In addition, the functions of PD-1/PD-L on innate lymphoid cells and the role of exosomal forms of PD-L1 warrant further research. Looking back on the history of repeatedly failed clinical trials of immune modulatory therapies for sepsis, sepsis must be recognized as a difficult disease entity for performing clinical trials. A major obstacle that could prevent effective clinical trials of drug candidates is the disease complexity and heterogeneities; clinically diagnosed sepsis could contain multiple sepsis subgroups that suffer different levels of hyper-inflammation and immune-suppression in distinct organs. Thus, the selection of appropriate more homogenous sepsis subgroup is the key for testing the clinical efficacy of experimental therapies targeting specific pathways in either hyperinflammation and/or immunoparalysis. An emerging technology such as artificial intelligence (AI) may help to identify an immune paralysis subgroup who would best be treated by PD-1/PD-L1 pathway inhibitors.

Transfer learning with chest X-rays for ER patient classification

One of the challenges with urgent evaluation of patients with acute respiratory distress syndrome (ARDS) in the emergency room (ER) is distinguishing between cardiac vs infectious etiologies for their pulmonary findings. We conducted a retrospective study with the collected data of 171 ER patients. ER patient classification for cardiac and infection causes was evaluated with clinical data and chest X-ray image data. We show that a deep-learning model trained with an external image data set can be used to extract image features and improve the classification accuracy of a data set that does not contain enough image data to train a deep-learning model. An analysis of clinical feature importance was performed to identify the most important clinical features for ER patient classification. The current model is publicly available with an interface at the web link: http://nbttranslationalresearch.org/.

Analytics with artificial intelligence to advance the treatment of acute respiratory distress syndrome

Artificial intelligence (AI) has found its way into clinical studies in the era of big data. Acute respiratory distress syndrome (ARDS) or acute lung injury (ALI) is a clinical syndrome that encompasses a heterogeneous population. Management of such heterogeneous patient population is a big challenge for clinicians. With accumulating ALI datasets being publicly available, more knowledge could be discovered with sophisticated analytics. We reviewed literatures with big data analytics to understand the role of AI for improving the caring of patients with ALI/ARDS. Many studies have utilized the electronic medical records (EMR) data for the identification and prognostication of ARDS patients. As increasing number of ARDS clinical trials data is open to public, secondary analysis on these combined datasets provide a powerful way of finding solution to clinical questions with a new perspective. AI techniques such as Classification and Regression Tree (CART) and artificial neural networks (ANN) have also been successfully used in the investigation of ARDS problems. Individualized treatment of ARDS could be implemented with a support from AI as we are now able to classify ARDS into many subphenotypes by unsupervised machine learning algorithms. Interestingly, these subphenotypes show different responses to a certain intervention. However, current analytics involving ARDS have not fully incorporated information from omics such as transcriptome, proteomics, daily activities and environmental conditions. AI technology is assisting us to interpret complex data of ARDS patients and enable us to further improve the management of ARDS patients in future with individual treatment plans.