Construction and validation of prognostic models in critically Ill patients with sepsis-associated acute kidney injury: interpretable machine learning approach

Web server-based deep learning-driven predictive models for respiratory toxicity of environmental chemicals: Mechanistic insights and interpretability

Respiratory toxicity of chemicals is a common clinical and environmental health concern. Currently, most in silico prediction models for chemical respiratory toxicity are often based on a single or vague toxicity endpoint, and machine learning models always lack interpretability. In this study, we developed eight interpretable deep learning models to predict respiratory toxicity of chemicals, focusing on specific respiratory diseases such as pneumonia, pulmonary edema, respiratory infections, pulmonary embolism and pulmonary arterial hypertension, asthma, bronchospasm, bronchitis, and pulmonary fibrosis. In addition, we integrated data from eight respiratory toxicity endpoints into a comprehensive dataset and developed an overall respiratory system model. Model performance was evaluated using 5-fold cross-validation and external validation, with area under the curve (AUC) and accuracy (ACC) values exceeding 0.85 for all eight toxicity endpoints. To enhance model interpretability, we employed the frequency ratio method to identify key structural fragments in Klekota-Roth fingerprints (KRFP) and utilized SHAP (SHapley Additive exPlanations) game theory analysis to visualize critical features driving model predictions. This study demonstrates the role of interpretable deep learning models in predicting the respiratory toxicity of drugs and their environmental metabolites, offering valuable tools and information for early detection and risk assessment of pharmaceutical compounds and environmental pollutants with respiratory toxicity potential.

The application of machine learning in clinical microbiology and infectious diseases

With the development of artificial intelligence(AI) in computer science and statistics, it has been further applied to the medical field. These applications include the management of infectious diseases, in which machine learning has created inroads in clinical microbiology, radiology, genomics, and the analysis of electronic health record data. Especially, the role of machine learning in microbiology has gradually become prominent, and it is used in etiological diagnosis, prediction of antibiotic resistance, association between human microbiome characteristics and complex host diseases, prognosis judgment, and prevention and control of infectious diseases. Machine learning in the field of microbiology mainly adopts supervised learning and unsupervised learning, involving algorithms from classification and regression to clustering and dimensionality reduction. This Review explains crucial concepts in machine learning for unfamiliar readers, describes machine learning's current applications in clinical microbiology and infectious diseases, and summarizes important approaches clinicians must be aware of when evaluating research using machine learning.

The predictive value of renal vascular resistance index and serum biomarkers for sepsis-associated acute kidney injury: a retrospective study

BACKGROUND

Sepsis-associated acute kidney injury (AKI) presents a significant clinical challenge, necessitating the identification of predictive indicators for early detection and intervention. This retrospective case-control study aimed to investigate the predictive potential of renal vascular resistance index and serum biomarkers in sepsis-associated AKI.

METHODS

A cohort of 108 patients diagnosed with sepsis was separated into two groups-those with acute kidney injury (AKI) and those without-using the diagnostic criteria established by the kidney disease: Improving Global Outcomes (KDIGO) guidelines. Various demographic, clinical, and laboratory parameters were collected, including renal resistive index, serum biomarkers, disease severity scores, and clinical outcomes. Statistical analyses, including t-tests, correlation analysis, receiver operating characteristic (ROC) analysis, and joint model construction, were conducted to evaluate the predictive value of these parameters.

RESULTS

The AKI group exhibited higher APACHE II and SOFA scores compared to the non-AKI group, indicating the association between disease severity scores and the presence of AKI in septic patients. Renal resistive index and several serum biomarkers, including C-reactive protein and procalcitonin, were notably elevated in the AKI group. Correlation analysis demonstrated significant associations between renal vascular resistance index, serological biomarkers, and clinical severity scores. ROC analysis revealed that several parameters, including Renal Resistive Index (AUC = 0.667), C-reactive Protein (CRP, AUC = 0.665), Platelet Count (AUC = 0.666), and Prothrombin Time (AUC = 0.669), demonstrated moderate diagnostic performance for predicting sepsis-associated AKI. These parameters were subsequently incorporated into a joint predictive model, which exhibited robust diagnostic accuracy with an AUC of 0.780, highlighting its potential utility as a reliable predictive tool in clinical practice.

CONCLUSIONS

The study findings underscore the potential for integrating renal vascular parameters and serum biomarkers in clinical risk stratification and early intervention strategies for sepsis-associated AKI.

CLINICAL REGISTRATION

Not applicable.

Interpretable multimodal deep learning model for predicting post-surgical international society of urological pathology grade in primary prostate cancer

PURPOSE

To address heterogeneity in prostate cancer (PCa) pathological grading, we developed an interpretable multimodal fusion model integrating 18F prostate-specific membrane antigen (18F-PSMA)-targeted positron emission tomography/computed tomography (18F-PSMA-PET/CT) imaging features with clinical variables for predicting post-surgical ISUP grade (psISUP ≥ 4 vs. < 4).

METHODS

This retrospective study analyzed 222 patients with PCa (2020-2024) undergoing 18F-PSMA PET/CT. We constructed a deep transfer learning framework incorporating radiomic features from PET/CT and clinical parameters. Model performance was validated against three established methods and preoperative biopsy Gleason scores. Additionally, SHapley Additive exPlanations (SHAP) values elucidated feature contributions, and a radiomic nomogram was developed for clinical translation.

RESULTS

The fusion model achieved superior discrimination in psISUP grading (test set area under the curve (AUC) = 0.850, 95% confidence interval [CI] 0.769-0.932; validation set AUC = 0.833, 95% CI 0.657-1.000), significantly outperforming preoperative Gleason scores. SHAP analysis identified PSMA uptake heterogeneity and PSA density as key predictive features. The nomogram demonstrated clinical interpretability through visualised risk stratification.

CONCLUSION

Our deep learning-based multimodal fusion model enables accurate preoperative prediction of aggressive PCa pathology (ISUP ≥ 4), potentially optimising surgical planning and personalised therapeutic strategies. The interpretable framework enhances clinical trustworthiness in artificial intelligence-assisted decision-making.

Harness machine learning for multiple prognoses prediction in sepsis patients: evidence from the MIMIC-IV database

BACKGROUND

Sepsis, a severe systemic response to infection, frequently results in adverse outcomes, underscoring the urgency for prompt and accurate prognostic tools. Machine learning methods such as logistic regression, random forests, and CatBoost, have shown potential in early sepsis prediction. The study aimed to create and verify a machine learning model capable of early prognostic identification of patients with sepsis in intensive care units (ICUs).

METHODS

Patients adhering to inclusion and exclusion criteria from the MIMIC-IV v2.2 database were divided into a training set and a validation set in a 7:3 ratio. Initially, we employed difference analysis to assess the significance of each variable and subsequently screened relevant features with multinomial logistic regression analysis. Logistic regression, random forest, and CatBoost algorithms were used to construct machine learning models to predict rapid recovery, chronic critical illness, and mortality in sepsis. The models were compared through several evaluation indexes including precision, accuracy, recall, F1 score, and the area under the receiver-operating-characteristic curve(AUC) in the validation set to select the optimal model. The best model was visualized and interpreted utilizing the Shapley Additive explanations method.

RESULTS

13174 sepsis patients were included. Post the screening process,26 clinical features were obtained to develop three distinct machine learning models. CatBoost exhibited superior performance among the three models with a weighted AUC of 0.771. The prognosis with the highest predictive performance was mortality (AUC = 0.804), followed by the prognoses of rapid recovery (AUC = 0.773) and chronic critical illness(AUC = 0.737). Urine output, respiratory rate, and temperature were the top three important features for the whole model prediction.

CONCLUSION

The machine learning model developed leveraging the CatBoost algorithm demonstrates the latent capacity to identify sepsis prognosis early. It also suggests that interventions targeting factors such as urine output, respiratory status, and temperature in the early stage may potentially alter the adverse prognosis of sepsis patients. However, the model will still require further external validation in the future.

Machine learning for risk prediction of acute kidney injury in patients with diabetes mellitus combined with heart failure during hospitalization

This study aimed to develop a machine learning (ML) model for predicting the risk of acute kidney injury (AKI) in diabetic patients with heart failure (HF) during hospitalization. Using data from 1,457 patients in the MIMIC-IV database, the study identified twenty independent risk factors for AKI through LASSO regression and logistic regression. Six ML algorithms were evaluated, including LightGBM, random forest, and neural networks. The LightGBM model demonstrated superior performance with the highest prediction accuracy, with AUC values of 0.973 and 0.804 in the training and validation sets, respectively. The Shapley additive explanations algorithm was used to visualize the model and identify the most relevant features for AKI risk. Clinical impact curves further confirmed the strong discriminatory ability and generalizability of the LightGBM model. This study highlights the potential of ML models, particularly LightGBM, to effectively predict AKI risk in diabetic patients with HF, enabling early identification of high-risk patients and timely interventions to improve prognosis.

Identification of key feature variables and prediction of harmful algal blooms in a water diversion lake based on interpretable machine learning

Harmful algal blooms (HABs) as an increasing environmental problem in lakes, and water diversion has become a common and effective strategy for mitigating HABs. Early and accurate identification of the occurrence of HABs in lakes is essential for scientific guidance of water diversion. Furthermore, the inevitable changes of hydrodynamic and water environment in the receiving area during water diversion make it more challenging to identify the important environmental features of HABs. Therefore, we constructed a machine learning modelling framework suitable for predicting HABs with favorable performance in both non-water diversion and water diversion states. In this study, we collected data from three monitoring sites for the years 2008-2020 (non-water diversion period from 2008 to 2013 and water diversion period from 2014 to 2020) as external validations and six sampling sites for the years 2021-2022 (2021 non-water diversion period and 2022 water diversion period) as internal validation. The CatBoost (AUC = 0.948) model fared best performance was obtained by comparing 10 machine learning models for comprehensive HABs prediction analyses in the external cohorts of Yilong Lake, and the 24 features were reduced to obtain the 8 (Including TP, TN and CODCr, etc.) most important environmental features. In addition, the SHapley Additive explanation (SHAP) method was used to interpret this CatBoost model through a global interpretation that describes the whole features of the model and a local interpretation that details how a certain forecast of HABs is made for a single sample via inputting the individual data. The CatBoost interpretable model also performed well in internal validation and the model has been converted into a convenient application for use by the Bureau of Yilong Lake Administration personnel and researchers. Finally, the results of the PLS-PM explains that water diversion indirectly mitigates HABs mainly through diluting nutrient concentrations. Overall, the final model of this study has a good performance and application benefits in predicting HABs during the non-water diversion period and water diversion period of Yilong Lake, which provides a guideline for water diversion. Furthermore, this study also provides a reference for other similar eutrophic lake water diversion strategies.

Systematic Review and Meta-Analysis of Machine Learning Models for Acute Kidney Injury Risk Classification

Key Points

Background

Artificial intelligence through machine learning models seems to provide accurate and precise AKI risk classification in some clinical settings, but their performance and implementation in real-world settings has not been established.

Methods

PubMed, Excerpta Medica (EMBASE) database, Web of Science, and Scopus were searched until August 2023. Articles reporting on externally validated models for prediction of AKI onset, AKI severity, and post-AKI complications in hospitalized adult and pediatric patients were searched using text words related to AKI, artificial intelligence, and machine learning. Two independent reviewers screened article titles, abstracts, and full texts. Areas under the receiver operating characteristic curves (AUCs) were used to compare model discrimination and pooled using a random-effects model.

Results

Of the 4816 articles initially identified and screened, 95 were included, representing 3.8 million admissions. The Kidney Disease Improving Global Outcomes (KDIGO)-AKI criteria were most frequently used to define AKI (72%). We identified 302 models, with the most common being logistic regression (37%), neural networks (10%), random forest (9%), and eXtreme gradient boosting (9%). The most frequently reported predictors of hospitalized incident AKI were age, sex, diabetes, serum creatinine, and hemoglobin. The pooled AUCs for AKI onset were 0.82 (95% confidence interval, 0.80 to 0.84) and 0.78 (95% confidence interval, 0.76 to 0.80) for internal and external validation, respectively. Pooled AUCs across multiple clinical settings, AKI severities, and post-AKI complications ranged from 0.78 to 0.87 for internal validation and 0.73 to 0.84 for external validation. Although data were limited, results in the pediatric population aligned with those observed in adults. Between-study heterogeneity was high for all outcomes (I2>90%), and most studies presented high risk of bias (86%) according to the Prediction Model Risk of Bias Assessment Tool.

Conclusions

Most externally validated models performed well in predicting AKI onset, AKI severity, and post-AKI complications in hospitalized adult and pediatric populations. However, heterogeneity in clinical settings, study populations, and predictors limits their generalizability and implementation at the bedside.

An interpreting machine learning models to predict amputation risk in patients with diabetic foot ulcers: a multi-center study

Background

Diabetic foot ulcers (DFUs) constitute a significant complication among individuals with diabetes and serve as a primary cause of nontraumatic lower-extremity amputation (LEA) within this population. We aimed to develop machine learning (ML) models to predict the risk of LEA in DFU patients and used SHapley additive explanations (SHAPs) to interpret the model.

Methods

In this retrospective study, data from 1,035 patients with DFUs at Sun Yat-sen Memorial Hospital were utilized as the training cohort to develop the ML models. Data from 297 patients across multiple tertiary centers were used for external validation. We then used least absolute shrinkage and selection operator analysis to identify predictors of amputation. We developed five ML models [logistic regression (LR), support vector machine (SVM), random forest (RF), k-nearest neighbors (KNN) and extreme gradient boosting (XGBoost)] to predict LEA in DFU patients. The performance of these models was evaluated using several metrics, including the area under the receiver operating characteristic curve (AUC), decision curve analysis (DCA), precision, recall, accuracy, and F1 score. Finally, the SHAP method was used to ascertain the significance of the features and to interpret the model.

Results

In the final cohort comprising 1332 individuals, 600 patients underwent amputation. Following hyperparameter optimization, the XGBoost model achieved the best amputation prediction performance with an accuracy of 0.94, a precision of 0.96, an F1 score of 0.94 and an AUC of 0.93 for the internal validation set on the basis of the 17 features. For the external validation set, the model attained an accuracy of 0.78, a precision of 0.93, an F1 score of 0.78, and an AUC of 0.83. Through SHAP analysis, we identified white blood cell counts, lymphocyte counts, and blood urea nitrogen levels as the model's main predictors.

Conclusion

The XGBoost algorithm-based prediction model can be used to dynamically estimate the risk of LEA in DFU patients, making it a valuable tool for preventing the progression of DFUs to amputation.

Construction of a machine learning-based interpretable prediction model for acute kidney injury in hospitalized patients

In this observational study, we used data from 59,936 hospitalized adults to construct a model. For the models constructed with all 53 variables, all five models achieved acceptable performance with the validation cohort, with the extreme gradient boosting (XGBoost) model showing the best predictive efficacy and stability (area under the curve (AUC), 0.9301). For the simpler models constructed with 39 significant variables screened by the random forest recursive feature elimination method, the XGBoost model also had the best performance (AUC, 0.9357). All the models showed significant net returns according to decision analysis curves, and the XGBoost model achieved the optimal results. In addition, the Shapley additive explanation (SHAP) importance matrices revealed that uric acid, colloidal solution, first creatinine value on admission, pulse and albumin represented the top five most important variables for both modeling strategies. With the external validation cohort based on 4022 hospitalized patients, the performance of all models declined, among which the Support vector machine (SVM) model showed the best predictive efficacy (AUC, 0.8230 and 0.8329), followed by the XGBoost model (0.8124 and 0.8316). Thus, our model can predict the occurrence and risk of acute kidney injury (AKI) up to 48 h in advance, enabling clinicians to assess the risk of AKI in hospitalized patients more accurately and intuitively and to develop necessary AKI management strategies.

Research hotspots and frontiers of machine learning in renal medicine: a bibliometric and visual analysis from 2013 to 2024

BACKGROUND

The kidney, an essential organ of the human body, can suffer pathological damage that can potentially have serious adverse consequences on the human body and even affect life. Furthermore, the majority of kidney-induced illnesses are frequently not readily identifiable in their early stages. Once they have progressed to a more advanced stage, they impact the individual's quality of life and burden the family and broader society. In recent years, to solve this challenge well, the application of machine learning techniques in renal medicine has received much attention from researchers, and many results have been achieved in disease diagnosis and prediction. Nevertheless, studies that have conducted a comprehensive bibliometric analysis of the field have yet to be identified.

OBJECTIVES

This study employs bibliometric and visualization analyses to assess the progress of the application of machine learning in the renal field and to explore research trends and hotspots in the field.

METHODS

A search was conducted using the Web of Science Core Collection database, which yielded articles and review articles published from the database's inception to May 12, 2024. The data extracted from these articles and review articles were then analyzed. A bibliometric and visualization analysis was conducted using the VOSviewer, CiteSpace, and Bibliometric (R-Tool of R-Studio) software.

RESULTS

2,358 papers were retrieved and analyzed for this topic. From 2013 to 2024, the number of publications and the frequency of citations in the relevant research areas have exhibited a consistent and notable increase annually. The data set comprises 3734 institutions in 91 countries and territories, with 799 journals publishing the results. The total number of authors contributing to the data set is 14,396. China and the United States have the highest number of published papers, with 721 and 525 papers, respectively. Harvard University and the University of California System exert the most significant influence at the institutional level. Regarding authors, Cheungpasitporn, Wisit, and Thongprayoon Charat of the Mayo Clinic organization were the most prolific researchers, with 23 publications each. It is noteworthy that researcher Breiman I had the highest co-citation frequency. The journal with the most published papers was "Scientific Reports," while "PLoS One" had the highest co-citation frequency. In this field of machine learning applied to renal medicine, the article "A Clinically Applicable Approach to Continuous Prediction of Future Acute Kidney Injury" by Tomasev N et al., published in NATURE in 2019, emerged as the most influential article with the highest co-citation frequency. A keyword and reference co-occurrence analysis reveals that current research trends and frontiers in nephrology are the management of patients with renal disease, prediction and diagnosis of renal disease, imaging of renal disease, and development of personalized treatment plans for patients with renal disease. "Acute kidney injury," "chronic kidney disease," and "kidney tumors" are the most discussed diseases in medical research.

CONCLUSIONS

The field of renal medicine is witnessing a surge in the application of machine learning. On one hand, this study offers a novel perspective on applying machine learning techniques to kidney-related diseases based on bibliometric analysis. This analysis provides a comprehensive overview of the current status and emerging research areas in the field, as well as future trends and frontiers. Conversely, this study furnishes data on collaboration and exchange between countries, regions, institutions, journals, authors, keywords, and reference co-citations. This information can facilitate the advancement of future research endeavors, which aim to enhance interdisciplinary collaboration, optimize data sharing and quality, and further advance the application of machine learning in the renal field.

Breaking new ground: machine learning enhances survival forecasts in hypercapnic respiratory failure

Background

The prognostic prediction of patients with hypercapnic respiratory failure holds significant clinical value. The objective of this study was to develop and validate a predictive model for predicting survival in patients with hypercapnic respiratory failure.

Methods

The study enrolled a total of 697 patients with hypercapnic respiratory failure, including 565 patients from the First People's Hospital of Yancheng in the modeling group and 132 patients from the People's Hospital of Jiangsu Province in the external validation group. The three selected models were random survival forest (RSF), DeepSurv, a deep learning-based survival prediction algorithm, and Cox Proportional Risk (CoxPH). The model's predictive performance was evaluated using the C-index and Brier score. Receiver operating characteristic curve (ROC), area under ROC curve (AUC), and decision curve analysis (DCA) were employed to assess the accuracy of predicting the prognosis for survival at 6, 12, 18, and 24 months.

Results

The RSF model (c-index: 0.792) demonstrated superior predictive ability for the prognosis of patients with hypercapnic respiratory failure compared to both the traditional CoxPH model (c-index: 0.699) and DeepSurv model (c-index: 0.618), which was further validated on external datasets. The Brier Score of the RSF model demonstrated superior performance, consistently measuring below 0.25 at the 6-month, 12-month, 18-month, and 24-month intervals. The ROC curve confirmed the superior discrimination of the RSF model, while DCA demonstrated its optimal clinical net benefit in both the modeling group and the external validation group.

Conclusion

The RSF model offered distinct advantages over the CoxPH and DeepSurv models in terms of clinical evaluation and monitoring of patients with hypercapnic respiratory failure.

A liver CT based nomogram to preoperatively predict lung metastasis secondary to hepatic alveolar echinococcosis

PURPOSE

To develop a nomogram based on liver CT and clinical features to preoperatively predict lung metastasis (LM) secondary to hepatic alveolar echinococcosis (HAE).

METHODS

A total of 291 consecutive HAE patients from Institution A undergoing preoperative abdominal contrast-enhanced CT and chest unenhanced CT were retrospectively reviewed, and were randomly divided into the training and internal validation sets at the 7:3 ratio. 82 consecutive patients from Institution B were enrolled as an external validation set. A nomogram was constructed based on the significant CT and clinical features of HAE from the training set selected by univariable and multivariable analyses to predict LM, and its predictive accuracy was assessed by area under the receiver operating characteristic curve (AUC) and Brier score. Decision-curve analysis was applied to evaluate the clinical effectiveness. This nomogram was verified in two independent validation sets.

RESULTS

Eosinophil (odds ratio [OR] = 9.60; 95 % confidence interval [CI]: 1.80-51.11), lesion size (OR = 1.02; 95 %CI: 1.01-1.04), and moderate-severe invasion of inferior vena cava (IVC) (OR = 5.57; 95 %CI: 1.82-17.10) were independently associated with LM (all P-values < 0.05). The nomogram based on the three independent predictors displayed AUCs of 0.875 (95 %CI, 0.824-0.927), 0.872 (95 %CI, 0.787-0.957) and 0.836 (95 %CI, 0.729-0.943), and Brier score of 0.105, 0.1 and 0.118 in the training, internal validation and external validation sets, respectively. Decision-curve analysis showed good clinical utility.

CONCLUSION

A nomogram based on eosinophil, lesion size and moderate-severe invasion of IVC showed good ability and accuracy for preoperative prediction of LM due to HAE.

Construction of a prognostic prediction model for colorectal cancer based on 5-year clinical follow-up data

Colorectal cancer (CRC) is a prevalent malignant tumor that presents significant challenges to both public health and healthcare systems. The aim of this study was to develop a machine learning model based on five years of clinical follow-up data from CRC patients to accurately identify individuals at risk of poor prognosis. This study included 411 CRC patients who underwent surgery at Yixing Hospital and completed the follow-up process. A modeling dataset containing 73 characteristic variables was established by collecting demographic information, clinical blood test indicators, histopathological results, and additional treatment-related information. Decision tree, random forest, support vector machine, and extreme gradient boosting (XGBoost) models were selected for modeling based on the features identified through recursive feature elimination (RFE). The Cox proportional hazards model was used as the baseline for model comparison. During the model training process, hyperparameters were optimized using a grid search method. The model performance was comprehensively assessed using multiple metrics, including accuracy, F1 score, Brier score, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic curve, calibration curve, and decision curve analysis curve. For the selected optimal model, the decision-making process was interpreted using the SHapley Additive exPlanations (SHAP) method. The results show that the optimal RFE-XGBoost model achieved an accuracy of 0.83 (95% CI 0.76-0.90), an F1 score of 0.81 (95% CI 0.72-0.88), and an area under the receiver operating characteristic curve of 0.89 (95% CI 0.82-0.94). Furthermore, the model exhibited superior calibration and clinical utility. SHAP analysis revealed that increased perioperative transfusion quantity, higher tumor AJCC stage, elevated carcinoembryonic antigen level, elevated carbohydrate antigen 19-9 (CA19-9) level, advanced age, and elevated carbohydrate antigen 125 (CA125) level were correlated with increased individual mortality risk. The RFE-XGBoost model demonstrated excellent performance in predicting CRC patient prognosis, and the application of the SHAP method bolstered the model's credibility and utility.

[Predicting Intensive Care Unit Mortality in Patients With Heart Failure Combined With Acute Kidney Injury Using an Interpretable Machine Learning Model: A Retrospective Cohort Study]

Objective

Heart failure (HF) complicated by acute kidney injury (AKI) significantly impacts patient outcomes, and it is crucial to make early predictions of short-term mortality. This study is focused on developing an interpretable machine learning model to enhance early prediction accuracy in such clinical scenarios.

Methods

This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care Ⅳ (MIMIC-Ⅳ, version 2.0) database. Data from the first 24 hours after admission to the ICU were extracted and divided into a training set (70%) and a validation set (30%). We utilized the SHapley Additive exPlanation (SHAP) method to interpret the workings of an extreme gradient boosting (XGBoost) model and identify key prognostic factors. The XGBoost model's predictive ability was evaluated against three other machine learning models using the area under the curve (AUC) metric, and its interpretation was enhanced using the SHAP method.

Results

The study included 8028 patients with HF complicated by AKI. The XGBoost model outperformed the other models, achieving an AUC of 0.93 (95% confidence interval [CI]: 0.78-0.94; accuracy = 0.89), while neural network model showed the worst performance (AUC = 0.79, 95% CI: 0.77-0.82; accuracy = 0.82). Decision curve analysis showed the superior net benefit of the XGBoost model within the 9% to 60% threshold probabilities. SHAP analysis was performed to identify the top 20 predictors, with age (mean SHAP value 1.29) and Glasgow Coma Scale score (mean SHAP value 1.24) emerging as significant factors.

Conclusions

Our interpretable model offers an enhanced ability to predict mortality risk in HF patients with AKI in ICUs. This model can be used to assist in formulating effective treatment plans and optimizing resource allocation.

Harnessing artificial intelligence in sepsis care: advances in early detection, personalized treatment, and real-time monitoring

Sepsis remains a leading cause of morbidity and mortality worldwide due to its rapid progression and heterogeneous nature. This review explores the potential of Artificial Intelligence (AI) to transform sepsis management, from early detection to personalized treatment and real-time monitoring. AI, particularly through machine learning (ML) techniques such as random forest models and deep learning algorithms, has shown promise in analyzing electronic health record (EHR) data to identify patterns that enable early sepsis detection. For instance, random forest models have demonstrated high accuracy in predicting sepsis onset in intensive care unit (ICU) patients, while deep learning approaches have been applied to recognize complications such as sepsis-associated acute respiratory distress syndrome (ARDS). Personalized treatment plans developed through AI algorithms predict patient-specific responses to therapies, optimizing therapeutic efficacy and minimizing adverse effects. AI-driven continuous monitoring systems, including wearable devices, provide real-time predictions of sepsis-related complications, enabling timely interventions. Beyond these advancements, AI enhances diagnostic accuracy, predicts long-term outcomes, and supports dynamic risk assessment in clinical settings. However, ethical challenges, including data privacy concerns and algorithmic biases, must be addressed to ensure fair and effective implementation. The significance of this review lies in addressing the current limitations in sepsis management and highlighting how AI can overcome these hurdles. By leveraging AI, healthcare providers can significantly enhance diagnostic accuracy, optimize treatment protocols, and improve overall patient outcomes. Future research should focus on refining AI algorithms with diverse datasets, integrating emerging technologies, and fostering interdisciplinary collaboration to address these challenges and realize AI's transformative potential in sepsis care.

Development and validation of a nomogram to predict survival in septic patients with heart failure in the intensive care unit

Heart failure is a common complication in patients with sepsis, and individuals who experience both sepsis and heart failure are at a heightened risk for adverse outcomes. This study aims to develop an effective nomogram model to predict the 7-day, 15-day, and 30-day survival probabilities of septic patients with heart failure in the intensive care unit (ICU). This study extracted the pertinent clinical data of septic patients with heart failure from the Critical Medical Information Mart for Intensive Care (MIMIC-IV) database. Patients were then randomly allocated into a training set and a test set at a ratio of 7:3. Cox proportional hazards regression analysis was used to determine independent risk factors influencing patient prognosis and to develop a nomogram model. The model's efficacy and clinical significance were assessed through metrics such as the concordance index (C-index), time-dependent receiver operating characteristic (ROC), calibration curve, and decision curve analysis (DCA). A total of 5,490 septic patients with heart failure were included in the study. A nomogram model was developed to predict short-term survival probabilities, using 13 variables: age, pneumonia, endotracheal intubation, mechanical ventilation, potassium (K), anion gap (AG), lactate (Lac), activated partial thromboplastin time (APTT), white blood cell count (WBC), red cell distribution width (RDW), hemoglobin-to-red cell distribution width ratio (HRR), Sequential Organ Failure Assessment (SOFA) score, and Charlson Comorbidity Index (CCI). The C-index was 0.730 (95% CI 0.719-0.742) for the training set and 0.761 (95% CI 0.745-0.776) for the test set, indicating strong model accuracy, indicating good model accuracy. Evaluations via the ROC curve, calibration curve, and decision curve analyses further confirmed the model's reliability and utility. This study effectively developed a straightforward and efficient nomogram model to predict the 7-day, 15-day, and 30-day survival probabilities of septic patients with heart failure in the ICU. The implementation of treatment strategies that address the risk factors identified in the model can enhance patient outcomes and increase survival rates.

Explainable PET-Based Habitat and Peritumoral Machine Learning Model for Predicting Progression-free Survival in Clinical Stage IA Pure-Solid Non-small Cell Lung Cancer: A Two-center Study

RATIONALE AND OBJECTIVES

This study aimed to develop and validate machine learning (ML) models utilizing positron emission tomography (PET)-habitat of the tumor and its peritumoral microenvironment to predict progression-free survival (PFS) in patients with clinical stage IA pure-solid non-small cell lung cancer (NSCLC).

MATERIALS AND METHODS

234 Patients who underwent lung resection for NSCLC from two hospitals were reviewed. Radiomic features were extracted from both intratumoral, peritumoral and habitat regions on PET. Univariate and multivariate logistic regression analyses were employed to determine significant clinical variables. Subsequently, a radiomics nomogram was developed by combining the radiomics signature with these identified clinical variables. Kaplan-Meier (KM) analysis was performed to investigate the prognostic value of the nomogram. Shapley Additive Explanations (SHAP) were used to interpret the ML models.

RESULTS

The combination model which contained peritumoral 5 mm and habitat regions radiomics features, clinical variables obtained a strong well-performance, achieving area under the curve (AUC) of 0.905 (95% confidence interval (CI) 0.854-0.957) in the train set and 0.875 (95% CI 0.789-0.962) in the internal validation set. The radiomics signature was significantly associated with PFS, the model significantly discerned high and low-risk patients, and exhibited a significant benefit in the clinical use showed low-risk score given have far longer RFS than those with high-risk score (log-rank P＜0.001).

CONCLUSION

The habitat and peritumoral radiomics signatures serve as an independent biomarker for predicting PFS in patients with early-stage NSCLC, effectively stratified survival risk among patients with clinical stage IA pure-solid non-small cell lung cancer.

Predicting frailty in older patients with chronic pain using explainable machine learning: A cross-sectional study

Frailty is common among older adults with chronic pain, and early identification is crucial in preventing adverse outcomes like falls, disability, and dementia. However, effective tools for identifying frailty in this population remain limited. This study aimed to explore frailty risk factors in older adults with chronic pain and to develop 9 machine learning models for frailty identification. The Shapley Additive Explanations (SHAP) method was used to explain the models. The Random Forest (RF) model performed best with 0.822 accuracy, 0.797 precision, and an AUC of 0.881. The variables in the RF model included: age, BMI, education level, pain duration, number of pain sites, pain level, depression, and Activity of Daily Living (ADL). Pain level, depression, and ADL were the 3 most important variables in the RF model. This model helps healthcare providers to identify frailty early, enabling timely interventions to improve patient outcomes and promote healthy aging.

Acute Kidney Injury Prognosis Prediction Using Machine Learning Methods: A Systematic Review

Rationale & Objective

Accurate estimation of in-hospital outcomes for patients with acute kidney injury (AKI) is crucial for aiding physicians in making optimal clinical decisions. We aimed to review prediction models constructed by machine learning methods for predicting AKI prognosis using administrative databases.

Study Design

A systematic review following PRISMA guidelines.

Setting & Study Populations

Adult patients diagnosed with AKI who are admitted to either hospitals or intensive care units.

Search Strategy & Sources

We searched PubMed, Embase, Web of Science, Scopus, and Cumulative Index to Nursing and Allied Health for studies published between January 1, 2014 and February 29, 2024. Eligible studies employed machine learning models to predict in-hospital outcomes of AKI based on administrative databases.

Data Extraction

Extracted data included prediction outcomes and population, prediction models with performance, feature selection methods, and predictive features.

Analytical Approach

The included studies were qualitatively synthesized with assessments of quality and bias. We calculated the pooled model discrimination of different AKI prognoses using random-effects models.

Results

Of 3,029 studies, 27 studies were eligible for qualitative review. In-hospital outcomes for patients with AKI included acute kidney disease, chronic kidney disease, renal function recovery or kidney failure, and mortality. Compared with models predicting the mortality of patients with AKI during hospitalization, the prediction performance of models on kidney function recovery was less accurate. Meta-analysis showed that machine learning methods outperformed traditional approaches in mortality prediction (area under the receiver operating characteristic curve, 0.831; 95% CI, 0.799-0.859 vs 0.772; 95% CI, 0.744-0.797). The overlapping predictive features for in-hospital mortality identified from ≥6 studies were age, serum creatinine level, serum urea nitrogen level, anion gap, and white blood cell count. Similarly, age, serum creatinine level, AKI stage, estimated glomerular filtration rate, and comorbid conditions were the common predictive features for kidney function recovery.

Limitations

Many studies developed prediction models within specific hospital settings without broad validation, restricting their generalizability and clinical application.

Conclusions

Machine learning models outperformed traditional approaches in predicting mortality for patients with AKI, although they are less accurate in predicting kidney function recovery. Overall, these models demonstrate significant potential to help physicians improve clinical decision making and patient outcomes.

Registration

CRD42024535965.

Machine learning for the prediction of mortality in patients with sepsis-associated acute kidney injury: a systematic review and meta-analysis

BACKGROUND

Predicting mortality in sepsis-related acute kidney injury facilitates early data-driven treatment decisions. Machine learning is predicting mortality in S-AKI in a growing number of studies. Therefore, we conducted this systematic review and meta-analysis to investigate the predictive value of machine learning for mortality in patients with septic acute kidney injury.

METHODS

The PubMed, Web of Science, Cochrane Library and Embase databases were searched up to 20 July 2024 This was supplemented by a manual search of study references and review articles. Data were analysed using STATA 14.0 software. The risk of bias in the prediction model was assessed using the Predictive Model Risk of Bias Assessment Tool.

RESULTS

A total of 8 studies were included, with a total of 53 predictive models and 17 machine learning algorithms used. Meta-analysis using a random effects model showed that the overall C index in the training set was 0.81 (95% CI: 0.78-0.84), sensitivity was 0.39 (0.32-0.47), and specificity was 0.92 (95% CI: 0.89-0.95). The overall C-index in the validation set was 0.73 (95% CI: 0.71-0.74), sensitivity was 0.54 (95% CI: 0.48-0.60) and specificity was 0.90 (95% CI: 0.88-0.91). The results showed that the machine learning algorithms had a good performance in predicting sepsis-related acute kidney injury death prediction.

CONCLUSION

Machine learning has been shown to be an effective tool for predicting sepsis-associated acute kidney injury deaths, which has important implications for enhancing risk assessment and clinical decision-making to improve sepsis patient care. It is also eagerly anticipated that future research efforts will incorporate larger sample sizes and multi-centre studies to more intensively examine the external validation of these models in different patient populations, allowing for a more in-depth exploration of sepsis-associated acute kidney injury in terms of accurate diagnostic efficacy across a diverse range of model and predictor types.

TRIAL REGISTRATION

This study was registered with PROSPERO (CRD42024569420).

Identification of a Susceptible and High-Risk Population for Postoperative Systemic Inflammatory Response Syndrome in Older Adults: Machine Learning-Based Predictive Model

BACKGROUND

Systemic inflammatory response syndrome (SIRS) is a serious postoperative complication among older adult surgical patients that frequently develops into sepsis or even death. Notably, the incidences of SIRS and sepsis steadily increase with age. It is important to identify the risk of postoperative SIRS for older adult patients at a sufficiently early stage, which would allow preemptive individualized enhanced therapy to be conducted to improve the prognosis of older adult patients. In recent years, machine learning (ML) models have been deployed by researchers for many tasks, including disease prediction and risk stratification, exhibiting good application potential.

OBJECTIVE

We aimed to develop and validate an individualized predictive model to identify susceptible and high-risk populations for SIRS in older adult patients to instruct appropriate early interventions.

METHODS

Data for surgical patients aged ≥65 years from September 2015 to September 2020 in 3 independent medical centers were retrieved and analyzed. The eligible patient cohort in the Third Affiliated Hospital of Sun Yat-sen University was randomly separated into an 80% training set (2882 patients) and a 20% internal validation set (720 patients). We developed 4 ML models to predict postoperative SIRS. The area under the receiver operating curve (AUC), F1 score, Brier score, and calibration curve were used to evaluate the model performance. The model with the best performance was further validated in the other 2 independent data sets involving 844 and 307 cases, respectively.

RESULTS

The incidences of SIRS in the 3 medical centers were 24.3% (876/3602), 29.6% (250/844), and 6.5% (20/307), respectively. We identified 15 variables that were significantly associated with postoperative SIRS and used in 4 ML models to predict postoperative SIRS. A balanced cutoff between sensitivity and specificity was chosen to ensure as high a true positive as possible. The random forest classifier (RF) model showed the best overall performance to predict postoperative SIRS, with an AUC of 0.751 (95% CI 0.709-0.793), sensitivity of 0.682, specificity of 0.681, and F1 score of 0.508 in the internal validation set and higher AUCs in the external validation-1 set (0.759, 95% CI 0.723-0.795) and external validation-2 set (0.804, 95% CI 0.746-0.863).

CONCLUSIONS

We developed and validated a generalizable RF model to predict postoperative SIRS in older adult patients, enabling clinicians to screen susceptible and high-risk patients and implement early individualized interventions. An online risk calculator to make the RF model accessible to anesthesiologists and peers around the world was developed.

Artificial intelligence and predictive models for early detection of acute kidney injury: transforming clinical practice

Acute kidney injury (AKI) presents a significant clinical challenge due to its rapid progression to kidney failure, resulting in serious complications such as electrolyte imbalances, fluid overload, and the potential need for renal replacement therapy. Early detection and prediction of AKI can improve patient outcomes through timely interventions. This review was conducted as a narrative literature review, aiming to explore state-of-the-art models for early detection and prediction of AKI. We conducted a comprehensive review of findings from various studies, highlighting their strengths, limitations, and practical considerations for implementation in healthcare settings. We highlight the potential benefits and challenges of their integration into routine clinical care and emphasize the importance of establishing robust early-detection systems before the introduction of artificial intelligence (AI)-assisted prediction models. Advances in AI for AKI detection and prediction are examined, addressing their clinical applicability, challenges, and opportunities for routine implementation.

Aurantiamide mitigates acute kidney injury by suppressing renal necroptosis and inflammation via GRPR-dependent mechanism

Acute kidney injury (AKI) manifests as a clinical syndrome characterised by the rapid accumulation of metabolic wastes, such as blood creatinine and urea nitrogen, leading to a sudden decline in renal function. Currently, there is a lack of specific therapeutic drugs for AKI. Previously, we identified gastrin-releasing peptide receptor (GRPR) as a pathogenic factor in AKI. In this study, we investigated the therapeutic potential of a novel Chinese medicine monomer, aurantiamide (AA), which exhibits structural similarities to our previously reported GRPR antagonist, RH-1402. We compared the therapeutic efficacy of AA with RH-1402 both in vitro and in vivo using various AKI models. Our results demonstrated that, in vitro, AA attenuated injury, necroptosis, and inflammatory responses in human renal tubular epithelial cells subjected to repeated hypoxia/reoxygenation and lipopolysaccharide stimulation. In vivo, AA ameliorated renal tubular injury and inflammation in mouse models of ischemia/reperfusion and cecum ligation puncture-induced AKI, surpassing the efficacy of RH-1402. Furthermore, molecular docking and cellular thermal shift assay confirmed GRPR as a direct target of AA, which was further validated in primary cells. Notably, in GRPR-silenced HK-2 cells and GRPR systemic knockout mice, AA failed to mitigate renal inflammation and injury, underscoring the importance of GRPR in AA's mechanism of action. In conclusion, our study has demonstrated that AA serve as a novel antagonist of GRPR and a promising clinical candidate for AKI treatment.

Construction and validation of predictive models for intravenous immunoglobulin-resistant Kawasaki disease using an interpretable machine learning approach

BACKGROUND

Intravenous immunoglobulin (IVIG)-resistant Kawasaki disease is associated with coronary artery lesion development.

PURPOSE

This study aimed to explore the factors associated with IVIG-resistance and construct and validate an interpretable machine learning (ML) prediction model in clinical practice.

METHODS

Between December 2014 and November 2022, 602 patients were screened and risk factors for IVIG-resistance investigated. Five ML models are used to establish an optimal prediction model. The SHapley Additive exPlanations (SHAP) method was used to interpret the ML model.

RESULTS

Na+, hemoglobin (Hb), C-reactive protein (CRP), and globulin were independent risk factors for IVIG-resistance. A nonlinear relationship was identified between globulin level and IVIG-resistance. The XGBoost model exhibited excellent performance, with an area under the receiver operating characteristic curve of 0.821, accuracy of 0.748, sensitivity of 0.889, and specificity of 0.683 in the testing set. The XGBoost model was interpreted globally and locally using the SHAP method.

CONCLUSION

Na+, Hb, CRP, and globulin levels were independently associated with IVIG-resistance. Our findings demonstrate that ML models can reliably predict IVIG-resistance. Moreover, use of the SHAP method to interpret the established XGBoost model's findings would provide evidence of IVIG-resistance and guide the individualized treatment of Kawasaki disease.

Effects of environmental phenols on eGFR: machine learning modeling methods applied to cross-sectional studies

Purpose

Limited investigation is available on the correlation between environmental phenols' exposure and estimated glomerular filtration rate (eGFR). Our target is established a robust and explainable machine learning (ML) model that associates environmental phenols' exposure with eGFR.

Methods

Our datasets for constructing the associations between environmental phenols' and eGFR were collected from the National Health and Nutrition Examination Survey (NHANES, 2013-2016). Five ML models were contained and fine-tuned to eGFR regression by phenols' exposure. Regression evaluation metrics were used to extract the limitation of the models. The most effective model was then utilized for regression, with interpretation of its features carried out using shapley additive explanations (SHAP) and the game theory python package to represent the model's regression capacity.

Results

The study identified the top-performing random forest (RF) regressor with a mean absolute error of 0.621 and a coefficient of determination of 0.998 among 3,371 participants. Six environmental phenols with eGFR in linear regression models revealed that the concentrations of triclosan (TCS) and bisphenol S (BPS) in urine were positively correlated with eGFR, and the correlation coefficients were β = 0.010 (p = 0.026) and β = 0.007 (p = 0.004) respectively. SHAP values indicate that BPS (1.38), bisphenol F (BPF) (0.97), 2,5-dichlorophenol (0.87), TCS (0.78), BP3 (0.60), bisphenol A (BPA) (0.59) and 2,4-dichlorophenol (0.47) in urinary contributed to the model.

Conclusion

The RF model was efficient in identifying a correlation between phenols' exposure and eGFR among United States NHANES 2013-2016 participants. The findings indicate that BPA, BPF, and BPS are inversely associated with eGFR.

Development and validation of machine learning models to predict MDRO colonization or infection on ICU admission by using electronic health record data

BACKGROUND

Multidrug-resistant organisms (MDRO) pose a significant threat to public health. Intensive Care Units (ICU), characterized by the extensive use of antimicrobial agents and a high prevalence of bacterial resistance, are hotspots for MDRO proliferation. Timely identification of patients at high risk for MDRO can aid in curbing transmission, enhancing patient outcomes, and maintaining the cleanliness of the ICU environment. This study focused on developing a machine learning (ML) model to identify patients at risk of MDRO during the initial phase of their ICU stay.

METHODS

Utilizing patient data from the First Medical Center of the People's Liberation Army General Hospital (PLAGH-ICU) and the Medical Information Mart for Intensive Care (MIMIC-IV), the study analyzed variables within 24 h of ICU admission. Machine learning algorithms were applied to these datasets, emphasizing the early detection of MDRO colonization or infection. Model efficacy was evaluated by the area under the receiver operating characteristics curve (AUROC), alongside internal and external validation sets.

RESULTS

The study evaluated 3,536 patients in PLAGH-ICU and 34,923 in MIMIC-IV, revealing MDRO prevalence of 11.96% and 8.81%, respectively. Significant differences in ICU and hospital stays, along with mortality rates, were observed between MDRO positive and negative patients. In the temporal validation, the PLAGH-ICU model achieved an AUROC of 0.786 [0.748, 0.825], while the MIMIC-IV model reached 0.744 [0.723, 0.766]. External validation demonstrated reduced model performance across different datasets. Key predictors included biochemical markers and the duration of pre-ICU hospital stay.

CONCLUSIONS

The ML models developed in this study demonstrated their capability in early identification of MDRO risks in ICU patients. Continuous refinement and validation in varied clinical contexts remain essential for future applications.

Artificial intelligence and machine learning's role in sepsis-associated acute kidney injury

Sepsis-associated acute kidney injury (SA-AKI) is a serious complication in critically ill patients, resulting in higher mortality, morbidity, and cost. The intricate pathophysiology of SA-AKI requires vigilant clinical monitoring and appropriate, prompt intervention. While traditional statistical analyses have identified severe risk factors for SA-AKI, the results have been inconsistent across studies. This has led to growing interest in leveraging artificial intelligence (AI) and machine learning (ML) to predict SA-AKI better. ML can uncover complex patterns beyond human discernment by analyzing vast datasets. Supervised learning models like XGBoost and RNN-LSTM have proven remarkably accurate at predicting SA-AKI onset and subsequent mortality, often surpassing traditional risk scores. Meanwhile, unsupervised learning reveals clinically relevant sub-phenotypes among diverse SA-AKI patients, enabling more tailored care. In addition, it potentially optimizes sepsis treatment to prevent SA-AKI through continual refinement based on patient outcomes. However, utilizing AI/ML presents ethical and practical challenges regarding data privacy, algorithmic biases, and regulatory compliance. AI/ML allows early risk detection, personalized management, optimal treatment strategies, and collaborative learning for SA-AKI management. Future directions include real-time patient monitoring, simulated data generation, and predictive algorithms for timely interventions. However, a smooth transition to clinical practice demands continuous model enhancements and rigorous regulatory oversight. In this article, we outlined the conventional methods used to address SA-AKI and explore how AI and ML can be applied to diagnose and manage SA-AKI, highlighting their potential to revolutionize SA-AKI care.

An explainable machine learning-based model to predict intensive care unit admission among patients with community-acquired pneumonia and connective tissue disease

BACKGROUND

There is no individualized prediction model for intensive care unit (ICU) admission on patients with community-acquired pneumonia (CAP) and connective tissue disease (CTD) so far. In this study, we aimed to establish a machine learning-based model for predicting the need for ICU admission among those patients.

METHODS

This was a retrospective study on patients admitted into a University Hospital in China between November 2008 and November 2021. Patients were included if they were diagnosed with CAP and CTD during admission and hospitalization. Data related to demographics, CTD types, comorbidities, vital signs and laboratory results during the first 24 h of hospitalization were collected. The baseline variables were screened to identify potential predictors via three methods, including univariate analysis, least absolute shrinkage and selection operator (Lasso) regression and Boruta algorithm. Nine supervised machine learning algorithms were used to build prediction models. We evaluated the performances of differentiation, calibration, and clinical utility of all models to determine the optimal model. The Shapley Additive Explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) techniques were performed to interpret the optimal model.

RESULTS

The included patients were randomly divided into the training set (1070 patients) and the testing set (459 patients) at a ratio of 70:30. The intersection results of three feature selection approaches yielded 16 predictors. The eXtreme gradient boosting (XGBoost) model achieved the highest area under the receiver operating characteristic curve (AUC) (0.941) and accuracy (0.913) among various models. The calibration curve and decision curve analysis (DCA) both suggested that the XGBoost model outperformed other models. The SHAP summary plots illustrated the top 6 features with the greatest importance, including higher N-terminal pro-B-type natriuretic peptide (NT-proBNP) and C-reactive protein (CRP), lower level of CD4 + T cell, lymphocyte and serum sodium, and positive serum (1,3)-β-D-glucan test (G test).

CONCLUSION

We successfully developed, evaluated and explained a machine learning-based model for predicting ICU admission in patients with CAP and CTD. The XGBoost model could be clinical referenced after external validation and improvement.

Explainable Boosting Machine approach identifies risk factors for acute renal failure

BACKGROUND

Risk stratification and outcome prediction are crucial for intensive care resource planning. In addressing the large data sets of intensive care unit (ICU) patients, we employed the Explainable Boosting Machine (EBM), a novel machine learning model, to identify determinants of acute kidney injury (AKI) in these patients. AKI significantly impacts outcomes in the critically ill.

METHODS

An analysis of 3572 ICU patients was conducted. Variables such as average central venous pressure (CVP), mean arterial pressure (MAP), age, gender, and comorbidities were examined. This analysis combined traditional statistical methods with the EBM to gain a detailed understanding of AKI risk factors.

RESULTS

Our analysis revealed chronic kidney disease, heart failure, arrhythmias, liver disease, and anemia as significant comorbidities influencing AKI risk, with liver disease and anemia being particularly impactful. Surgical factors were also key; lower GI surgery heightened AKI risk, while neurosurgery was associated with a reduced risk. EBM identified four crucial variables affecting AKI prediction: anemia, liver disease, and average CVP increased AKI risk, whereas neurosurgery decreased it. Age was a progressive risk factor, with risk escalating after the age of 50 years. Hemodynamic instability, marked by a MAP below 65 mmHg, was strongly linked to AKI, showcasing a threshold effect at 60 mmHg. Intriguingly, average CVP was a significant predictor, with a critical threshold at 10.7 mmHg.

CONCLUSION

Using an Explainable Boosting Machine enhance the precision in AKI risk factors in ICU patients, providing a more nuanced understanding of known AKI risks. This approach allows for refined predictive modeling of AKI, effectively overcoming the limitations of traditional statistical models.

Country-specific determinants for COVID-19 case fatality rate and response strategies from a global perspective: an interpretable machine learning framework

BACKGROUND

There are significant geographic inequities in COVID-19 case fatality rates (CFRs), and comprehensive understanding its country-level determinants in a global perspective is necessary. This study aims to quantify the country-specific risk of COVID-19 CFR and propose tailored response strategies, including vaccination strategies, in 156 countries.

METHODS

Cross-temporal and cross-country variations in COVID-19 CFR was identified using extreme gradient boosting (XGBoost) including 35 factors from seven dimensions in 156 countries from 28 January, 2020 to 31 January, 2022. SHapley Additive exPlanations (SHAP) was used to further clarify the clustering of countries by the key factors driving CFR and the effect of concurrent risk factors for each country. Increases in vaccination rates was simulated to illustrate the reduction of CFR in different classes of countries.

FINDINGS

Overall COVID-19 CFRs varied across countries from 28 Jan 2020 to 31 Jan 31 2022, ranging from 68 to 6373 per 100,000 population. During the COVID-19 pandemic, the determinants of CFRs first changed from health conditions to universal health coverage, and then to a multifactorial mixed effect dominated by vaccination. In the Omicron period, countries were divided into five classes according to risk determinants. Low vaccination-driven class (70 countries) mainly distributed in sub-Saharan Africa and Latin America, and include the majority of low-income countries (95.7%) with many concurrent risk factors. Aging-driven class (26 countries) mainly distributed in high-income European countries. High disease burden-driven class (32 countries) mainly distributed in Asia and North America. Low GDP-driven class (14 countries) are scattered across continents. Simulating a 5% increase in vaccination rate resulted in CFR reductions of 31.2% and 15.0% for the low vaccination-driven class and the high disease burden-driven class, respectively, with greater CFR reductions for countries with high overall risk (SHAP value > 0.1), but only 3.1% for the ageing-driven class.

CONCLUSIONS

Evidence from this study suggests that geographic inequities in COVID-19 CFR is jointly determined by key and concurrent risks, and achieving a decreasing COVID-19 CFR requires more than increasing vaccination coverage, but rather targeted intervention strategies based on country-specific risks.

INTERPRETABLE MACHINE LEARNING FOR PREDICTING RISK OF INVASIVE FUNGAL INFECTION IN CRITICALLY ILL PATIENTS IN THE INTENSIVE CARE UNIT: A RETROSPECTIVE COHORT STUDY BASED ON MIMIC-IV DATABASE

ABSTRACT

The delayed diagnosis of invasive fungal infection (IFI) is highly correlated with poor prognosis in patients. Early identification of high-risk patients with invasive fungal infections and timely implementation of targeted measures is beneficial for patients. The objective of this study was to develop a machine learning-based predictive model for invasive fungal infection in patients during their intensive care unit (ICU) stay. Retrospective data was extracted from adult patients in the MIMIC-IV database who spent a minimum of 48 h in the ICU. Feature selection was performed using LASSO regression, and the dataset was balanced using the BL-SMOTE approach. Predictive models were built using six machine learning algorithms. The Shapley additive explanation algorithm was used to assess the impact of various clinical features in the optimal model, enhancing interpretability. The study included 26,346 ICU patients, of whom 379 (1.44%) were diagnosed with invasive fungal infection. The predictive model was developed using 20 risk factors, and the dataset was balanced using the borderline-SMOTE (BL-SMOTE) algorithm. The BL-SMOTE random forest model demonstrated the highest predictive performance (area under curve = 0.88, 95% CI = 0.84-0.91). Shapley additive explanation analysis revealed that the three most influential clinical features in the BL-SMOTE random forest model were dialysis treatment, APSIII scores, and liver disease. The machine learning model provides a reliable tool for predicting the occurrence of IFI in ICU patients. The BL-SMOTE random forest model, based on 20 risk factors, exhibited superior predictive performance and can assist clinicians in early assessment of IFI occurrence in ICU patients. Importance: Invasive fungal infections are characterized by high incidence and high mortality rates characteristics. In this study, we developed a clinical prediction model for invasive fungal infections in critically ill patients based on machine learning algorithms. The results show that the machine learning model based on 20 clinical features has good predictive value.

Interpretable machine learning framework to predict gout associated with dietary fiber and triglyceride-glucose index

BACKGROUND

Gout prediction is essential for the development of individualized prevention and treatment plans. Our objective was to develop an efficient and interpretable machine learning (ML) model using the SHapley Additive exPlanation (SHAP) to link dietary fiber and triglyceride-glucose (TyG) index to predict gout.

METHODS

Using datasets from the National Health and Nutrition Examination Survey (NHANES) (2005-2018) population to study dietary fiber, the TyG index was used to predict gout. After evaluating the performance of six ML models and selecting the Light Gradient Boosting Machine (LGBM) as the optimal algorithm, we interpret the LGBM model for predicting gout using SHAP and reveal the decision-making process of the model.

RESULTS

An initial survey of 70,190 participants was conducted, and after a gradual exclusion process, 12,645 cases were finally included in the study. Selection of the best performing LGBM model for prediction of gout associated with dietary fiber and TyG index (Area under the ROC curve (AUC): 0.823, 95% confidence interval (CI): 0.798-0.848, Accuracy: 95.3%, Brier score: 0.077). The feature importance of SHAP values indicated that age was the most important feature affecting the model output, followed by uric acid (UA). The SHAP values showed that lower dietary fiber values had a more pronounced effect on the positive prediction of the model, while higher values of the TyG index had a more pronounced effect on the positive prediction of the model.

CONCLUSION

The interpretable LGBM model associated with dietary fiber and TyG index showed high accuracy, efficiency, and robustness in predicting gout. Increasing dietary fiber intake and lowering the TyG index are beneficial in reducing the potential risk of gout.

DEVELOPMENT AND VALIDATION OF A NOMOGRAM FOR PREDICTING 28-DAY IN-HOSPITAL MORTALITY IN SEPSIS PATIENTS BASED ON AN OPTIMIZED ACUTE PHYSIOLOGY AND CHRONIC HEALTH EVALUATION II SCORE

ABSTRACT

Purpose : The objective of this study is to establish a nomogram that correlates optimized Acute Physiology and Chronic Health Evaluation II (APACHE II) score with sepsis-related indicators, aiming to provide a robust model for early prediction of sepsis prognosis in clinical practice and serve as a valuable reference for improved diagnosis and treatment strategies. Methods : This retrospective study extracted sepsis patients meeting the inclusion criteria from the MIMIC-IV database to form the training group. An optimized APACHE II score integrated with relevant indicators was developed using a nomogram for predicting the prognosis of sepsis patients. External validation was conducted using data from the intensive care unit at Lanzhou University Second Hospital. Results : The study enrolled 1805 patients in the training cohort and 203 patients in the validation cohort. A multifactor analysis was conducted to identify factors affecting patient mortality within 28 days, resulting in the development of an optimized score by simplifying evaluation indicators from APACHE II score. The results showed that the optimized score (area under the ROC curve [AUC] = 0.715) had a higher area under receiver operating characteristic curve than Sequential Organ Failure Assessment score (AUC = 0.637) but slightly lower than APACHE II score (AUC = 0.720). Significant indicators identified through multifactor analysis included platelet count, total bilirubin level, albumin level, prothrombin time, activated partial thromboplastin time, mechanical ventilation use and renal replacement therapy use. These seven indicators were combined with optimized score to construct a nomogram based on these seven indicators. The nomogram demonstrated good clinical predictive value in both training cohort (AUC = 0.803) and validation cohort (AUC = 0.750). Calibration curves and decision curve analyses also confirmed its good predictive ability, surpassing the APACHE II score and Sequential Organ Failure Assessment score in identifying high-risk patients. Conclusions : The nomogram was established in this study using the MIMIC-IV database and validated with external data, demonstrating its robust discriminability, calibration, and clinical practicability for predicting 28-day mortality in sepsis patients. These findings aim to provide substantial support for clinicians' decision making.

Machine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study

BACKGROUND

Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes.

OBJECTIVE

We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps.

METHODS

Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps.

RESULTS

For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699-0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model's top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2).

CONCLUSIONS

We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.

Diagnosing autism severity associated with physical fitness and gray matter volume in children with autism spectrum disorder: Explainable machine learning method

PURPOSE

This study aimed to investigate the relationship between physical fitness, gray matter volume (GMV), and autism severity in children with autism spectrum disorder (ASD). Besides, we sought to diagnose autism severity associated with physical fitness and GMV using machine learning methods.

METHODS

Ninety children diagnosed with ASD underwent physical fitness tests, magnetic resonance imaging scans, and autism severity assessments. Diagnosis models were established using extreme gradient boosting (XGB), random forest (RF), support vector machine (SVM), and decision tree (DT) algorithms. Hyperparameters were optimized through the grid search cross-validation method. The shapley additive explanation (SHAP) method was employed to explain the diagnosis results.

RESULTS

Our study revealed associations between muscular strength in physical fitness and GMV in specific brain regions (left paracentral lobule, bilateral thalamus, left inferior temporal gyrus, and cerebellar vermis I-II) with autism severity in children with ASD. The accuracy (95 % confidence interval) of the XGB, RF, SVM, and DT models were 77.9 % (77.3, 78.6 %), 72.4 % (71.7, 73.2 %), 71.9 % (71.1, 72.6 %), and 66.9 % (66.2, 67.7 %), respectively. SHAP analysis revealed that muscular strength and thalamic GMV significantly influenced the decision-making process of the XGB model.

CONCLUSION

Machine learning methods can effectively diagnose autism severity associated with physical fitness and GMV in children with ASD. In this respect, the XGB model demonstrated excellent performance across various indicators, suggesting its potential for diagnosing autism severity.