A novel nomogram to predict mortality in patients with stroke: a survival analysis based on the MIMIC-III clinical database

Development and validation of an interpretable machine learning model for predicting in-hospital mortality for ischemic stroke patients in ICU

BACKGROUND

Timely and accurate outcome prediction is essential for clinical decision-making for ischemic stroke patients in the intensive care unit (ICU). However, the interpretation and translation of predictive models into clinical applications are equally crucial. This study aims to develop an interpretable machine learning (IML) model that effectively predicts in-hospital mortality for ischemic stroke patients.

METHODS

In this study, an IML model was developed and validated using multicenter cohorts of 3225 ischemic stroke patients admitted to the ICU. Nine machine learning (ML) models, including logistic regression (LR), K-nearest neighbors (KNN), naive Bayes (NB), decision tree (DT), support vector machine (SVM), random forest (RF), XGBoost, LightGBM, and artificial neural network (ANN), were developed to predict in-hospital mortality using data from the MIMIC-IV and externally validated in Shanghai Changhai Hospital. Feature selection was conducted using three algorithms. Model's performance was assessed using area under the receiver operating characteristic (AUROC), accuracy, sensitivity, specificity and F1 score. Calibration curve and Brier score were used to evaluate the degree of calibration of the model, and decision curve analysis were generated to assess the net clinical benefit. Additionally, the SHapley Additive exPlanations (SHAP) method was employed to evaluate the risk of in-hospital mortality among ischemic stroke patients admitted to the ICU.

RESULTS

Mechanical ventilation, age, statins, white blood cell, blood urea nitrogen, hematocrit, warfarin, bicarbonate and systolic blood pressure were selected as the nine most influential variables. The RF model demonstrated the most robust predictive performance, achieving AUROC values of 0.908 and 0.858 in the testing set and external validation set, respectively. Calibration curves also revealed a high consistency between observations and predictions. Decision curve analysis showed that the model had the greatest net benefit rate when the prediction probability threshold is 0.10 ∼ 0.80. SHAP was employed to interpret the RF model. In addition, we have developed an online prediction calculator for ischemic stroke patients.

CONCLUSION

This study develops a machine learning-based calculator to predict the probability of in-hospital mortality among patients with ischemic stroke in ICU. The calculator has the potential to guide clinical decision-making and improve the care of patients with ischemic stroke by identifying patients at a higher risk of in-hospital mortality.

Comprehensive analysis of risk factors and nomogram development for predicting hepatic metastasis following radical resection of adenocarcinoma of the esophagogastric junction

BACKGROUND

Adenocarcinoma of the esophagogastric junction (AEG) often presents with subtle early symptoms and delayed diagnosis, frequently resulting in liver metastasis and a poor prognosis. This study aimed to investigate the primary risk factors influencing postoperative liver metastasis in AEG and to develop a simple predictive model to facilitate clinical risk stratification and individualized follow-up strategies.

METHODS

This retrospective study analyzed data from 524 patients with AEG who underwent radical resection, with patients randomly divided into a training group (368 cases) and a validation group (156 cases). Clinical and pathological information was collected, and independent factors significantly associated with postoperative liver metastasis were identified using univariate and multivariate Cox regression analyses. Based on these findings, a nomogram model was constructed to predict the 1-year and 3-year liver metastasis-free survival rates, and the model's predictive performance and clinical utility were evaluated using the C-index, ROC curves, and calibration curves.

RESULTS

Multivariate analysis revealed that thoracoabdominal surgery, higher N stage (N1 and N2/N3), moderate-to-poor differentiation, the presence of vascular tumor thrombus, intestinal type according to Lauren classification, and P53 status were independent risk factors for postoperative liver metastasis. The nomogram model based on these six indicators demonstrated high predictive accuracy in both the training group (C-index = 0.966) and the validation group (C-index = 0.976), with ROC AUCs for both the 1-year and 3-year predictions exceeding 0.96 and favorable calibration curves, confirming the model's strong predictive efficacy.

CONCLUSIONS

The predictive model developed in this study can effectively assess the risk of postoperative liver metastasis in patients with AEG, thereby providing a scientific basis for postoperative monitoring and individualized treatment, with the potential to improve patient outcomes in clinical practice.

In-Hospital Mortality Prediction among Intensive Care Unit Patients with Acute Ischemic Stroke: A Machine Learning Approach

Background: Acute ischemic stroke is a leading cause of death in the United States. Identifying patients with stroke at high risk of mortality is crucial for timely intervention and optimal resource allocation. This study aims to develop and validate machine learning-based models to predict in-hospital mortality risk for intensive care unit (ICU) patients with acute ischemic stroke and identify important associated factors. Methods: Our data include 3,489 acute ischemic stroke admissions to the ICU for patients not discharged or dead within 48 h from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. Demographic, hospitalization type, procedure, medication, intake (intravenous and oral), laboratory, vital signs, and clinical assessment [e.g., Glasgow Coma Scale Scores (GCS)] during the initial 48 h of admissions were used to predict in-hospital mortality after 48 h of ICU admission. We explored 3 machine learning models (random forests, logistic regression, and XGBoost) and applied Bayesian optimization for hyperparameter tuning. Important features were identified using learned coefficients. Results: Experiments show that XGBoost tuned for area under the receiver operating characteristic curve (AUC ROC) was the best performing model (AUC ROC 0.86, F1 0.52), compared to random forests (AUC ROC 0.85, F1 0.47) and logistic regression (AUC ROC 0.75, F1 0.40). Top features include GCS, blood urea nitrogen, and Richmond RASS score. The model also demonstrates good fairness for males versus females and across racial/ethnic groups. Conclusions: Machine learning has shown great potential in predicting in-hospital mortality risk for people with acute ischemic stroke in the ICU setting. However, more ethical considerations need to be applied to ensure that performance differences across different racial/ethnic groups will not exacerbate existing health disparities and will not harm historically marginalized populations.

Multimodal Metaverse Healthcare: A Collaborative Representation and Adaptive Fusion Approach for Generative Artificial-Intelligence-Driven Diagnosis

The metaverse enables immersive virtual healthcare environments, presenting opportunities for enhanced care delivery. A key challenge lies in effectively combining multimodal healthcare data and generative artificial intelligence abilities within metaverse-based healthcare applications, which is a problem that needs to be addressed. This paper proposes a novel multimodal learning framework for metaverse healthcare, MMLMH, based on collaborative intra- and intersample representation and adaptive fusion. Our framework introduces a collaborative representation learning approach that captures shared and modality-specific features across text, audio, and visual health data. By combining modality-specific and shared encoders with carefully formulated intrasample and intersample collaboration mechanisms, MMLMH achieves superior feature representation for complex health assessments. The framework's adaptive fusion approach, utilizing attention mechanisms and gated neural networks, demonstrates robust performance across varying noise levels and data quality conditions. Experiments on metaverse healthcare datasets demonstrate MMLMH's superior performance over baseline methods across multiple evaluation metrics. Longitudinal studies and visualization further illustrate MMLMH's adaptability to evolving virtual environments and balanced performance across diagnostic accuracy, patient-system interaction efficacy, and data integration complexity. The proposed framework has a unique advantage in that a similar level of performance is maintained across various patient populations and virtual avatars, which could lead to greater personalization of healthcare experiences in the metaverse. MMLMH's successful functioning in such complicated circumstances suggests that it can combine and process information streams from several sources. They can be successfully utilized in next-generation healthcare delivery through virtual reality.

Nomogram and randomized survival forest model for predicting sepsis risk in patients with cerebral infarction in the intensive care unit

BACKGROUND

To construct a nomogram and a Randomized Survival Forest (RSF) model for predicting the occurrence of sepsis in patients with cerebral infarction in intensive care units (ICUs).

METHODS

A total of 1,963 patients were included from the Medical Information Mart for Intensive Care IV database version 2.0 (MIMIC-IV v2.0). Screening features based on Cox regression and Lasso regression for nomogram and RSF modeling.

RESULTS

Patients were randomly split into a training set (1,374 cases) and a validation set (589 cases) at a ratio of 7:3. Risk factors in the nomogram model included atenolol, bicarbonate, calcium, clopidogrel, dipyridamole, heart failure, lymphocyte percent, midazolam, propofol, rhabdomyolysis, vancomycin, white blood cells, and antibiotics. In the training and validation sets, the nomogram predicted sepsis on the 3rd day of admission with an AUC of 0.798 and 0.765 and predicted sepsis on the 7th day with an AUC of 0.808 and 0.736, respectively. In the training and validation sets, the RSF model predicted sepsis on the 3rd day of admission with an AUC of 0.899 and 0.775 and predicted sepsis on the 7th day with an AUC of 0.913 and 0.768, respectively CONCLUSIONS: The two models can reliably predict the probability of sepsis in patients with cerebral infarction in the intensive care unit, which can help clinicians to assess the condition and provide timely medical interventions for patients. The RSF model has better performance.

Nomogram based on burn characteristics and the National Early Warning Score to predict survival in severely burned patients

BACKGROUND

Extensive burns are associated with a high mortality rate. Early prediction and action can reduce mortality. The National Early Warning Score (NEWS) is considered the best early warning score for predicting mortality. However, there has been no assessment conducted on the clinical prognostic significance of NEWS in individuals suffering from severe burns. The objective of this research was to establish a nomogram based on burn characteristics and the NEWS to predict survival in severely burned patients.

METHODS

A retrospective analysis was performed on 335 patients diagnosed with extensive burns from 2005 to 2021 in the Department of Burn Surgery of Changhai Hospital, the First Affiliated Hospital of Naval Medical University. Univariate and multivariate analyses were used to determine independent prognostic factors. A nomogram was developed using these prognostic factors and its internal validity was assessed through bootstrap resampling.

RESULTS

The results of multivariate analysis showed that the independent factors affecting the prognosis of severe burn patients were age, full-thickness burn, creatinine, inhalation tracheotomy, and the NEWS, all of which were identified to create the nomogram. The Akaike Information Criterion and Bayesian Information Criterion values of the nomogram demonstrated superior goodness-of-fit in predicting severe burns compared to NEWS, with lower scores (195.21 vs. 201.24; 221.91 vs. 224.12, respectively). The bootstrap-adjusted concordance index (C-index) of the nomogram yielded a higher value of 0.923(95 % CI 0.892-0.953), compared to NEWS which had a C-index of 0.699 (95 % CI 0.628-0.770). The calibration curves demonstrated excellent agreement between predicted probabilities and observed outcomes in the nomogram analysis. Furthermore, decision curve analysis indicated promising clinical utility for the proposed nomogram model. By applying an appropriate cutoff value derived from receiver operating characteristics curve analysis, it was observed that the high-risk group identified by the nomogram exhibited a significantly higher mortality rate than the low-risk group.

CONCLUSION

This study introduces an innovative nomogram that predicts the survival rate of individuals with severe burn injuries by combining clinical attributes and laboratory examinations, demonstrating superior efficacy compared to conventional NEWS systems.

A comprehensive analysis of stroke risk factors and development of a predictive model using machine learning approaches

Stroke is a leading cause of death and disability globally, particularly in China. Identifying risk factors for stroke at an early stage is critical to improving patient outcomes and reducing the overall disease burden. However, the complexity of stroke risk factors requires advanced approaches for accurate prediction. The objective of this study is to identify key risk factors for stroke and develop a predictive model using machine learning techniques to enhance early detection and improve clinical decision-making. Data from the China Health and Retirement Longitudinal Study (2011-2020) were analyzed, classifying participants based on baseline characteristics. We evaluated correlations among 12 chronic diseases and applied machine learning algorithms to identify stroke-associated parameters. A dose-response relationship between these parameters and stroke was assessed using restricted cubic splines with Cox proportional hazards models. A refined predictive model, incorporating age, sex, and key risk factors, was developed. Stroke patients were significantly older (average age 69.03 years) and had a higher proportion of women (53%) compared to non-stroke individuals. Additionally, stroke patients were more likely to reside in rural areas, be unmarried, smoke, and suffer from various diseases. While the 12 chronic diseases were correlated (p < 0.05), the correlation coefficients were generally weak (r < 0.5). Machine learning identified nine parameters significantly associated with stroke risk: TyG-WC, WHtR, TyG-BMI, TyG, TMO, CysC, CREA, SBP, and HDL-C. Of these, TyG-WC, WHtR, TyG-BMI, TyG, CysC, CREA, and SBP exhibited a positive dose-response relationship with stroke risk. In contrast, TMO and HDL-C were associated with reduced stroke risk. In the fully adjusted model, elevated CysC (HR = 2.606, 95% CI 1.869-3.635), CREA (HR = 1.819, 95% CI 1.240-2.668), and SBP (HR = 1.008, 95% CI 1.003-1.012) were significantly associated with increased stroke risk, while higher HDL-C (HR = 0.989, 95% CI 0.984-0.995) and TMO (HR = 0.99995, 95% CI 0.99994-0.99997) were protective. A nomogram model incorporating age, sex, and the identified parameters demonstrated superior predictive accuracy, with a significantly higher Harrell's C-index compared to individual predictors. This study identifies several significant stroke risk factors and presents a predictive model that can enhance early detection of high-risk individuals. Among them, CREA, CysC, SBP, TyG-BMI, TyG, TyG-WC, and WHtR were positively associated with stroke risk, whereas TMO and HDL-C were opposite. This serves as a valuable decision-support resource for clinicians, facilitating more effective prevention and treatment strategies, ultimately improving patient outcomes.

Construction of a nomogram for predicting the prognosis of patients with cervical spondylotic myelopathy: a retrospective study

Cervical spondylotic myelopathy (CSM) is a neurological disorder characterized by degenerative changes in the spinal cord and compression of the spinal cord and its adjacent structures due to various reasons, such as intervertebral disc herniation. The Japan Orthopaedic Association score is a disease-specific outcome tool that provides quantitative measurements for CSM patients. At present, no scholars have developed a model that can directly predict the prognosis of CSM patients. This study used binary logistic regression analysis to identify clinical independent predictive factors, and then used R language to construct a nomogram. The results showed that Hb (0.71 [0.51, 0.99]), LYM (0.32 [0.22, 0.46]), and ALB (0.32 [0.23, 0.45]) were independent predictive factors. Good prediction performance with modest errors was shown by the nomogram in both the training and validation groups. Hb, LYM, ALB are independent predictors of the prognosis of ACDF in CSM patients. The constructed nomogram can help clinicians assess and treat patients early, benefiting more patients.

Artificial intelligence-based personalized clinical decision-making for patients with localized prostate cancer: surgery versus radiotherapy

BACKGROUND

Surgery and radiotherapy are primary nonconservative treatments for prostate cancer (PCa). However, personalizing treatment options between these treatment modalities is challenging due to unclear criteria. We developed an artificial intelligence (AI)-based model that can identify patients with localized PCa who would benefit more from either radiotherapy or surgery, thereby providing personalized clinical decision-making.

MATERIAL AND METHODS

Data from consecutive patients with localized PCa who received radiotherapy or surgery with complete records of clinicopathological variables and follow-up results in 12 registries of the Surveillance, Epidemiology, and End Results database were analyzed. Patients from 7 registries were randomly assigned to training (TD) and internal validation datasets (IVD) at a 9:1 ratio. The remaining 5 registries constituted the external validation dataset (EVD). TD was divided into training-radiotherapy (TRD) and training-surgery (TSD) datasets, and IVD was divided into internal-radiotherapy (IRD) and internal-surgery (ISD) datasets. Six models for radiotherapy and surgery were trained using TRD and TSD to predict radiotherapy survival probability (RSP) and surgery survival probability (SSP), respectively. The models with the highest concordance index (C-index) on IRD and ISD were chosen to form the final treatment recommendation model (FTR). FTR recommendations were based on the higher value between RSP and SSP. Kaplan-Meier curves were generated for patients receiving recommended (consistent group) and nonrecommended treatments (inconsistent group), which were compared using the log-rank test.

RESULTS

The study included 118 236 patients, categorized into TD (TRD: 44 621; TSD: 41 500), IVD (IRD: 4949; ISD: 4621), and EVD (22 545). Both radiotherapy and surgery models accurately predicted RSP and SSP (C-index: 0.735-0.787 and 0.769-0.797, respectively). The consistent group exhibited higher survival rates than the inconsistent group, particularly among patients not suitable for active surveillance (P < .001).

CONCLUSION

FTR accurately identifies patients with localized PCa who would benefit more from either radiotherapy or surgery, offering clinicians an effective AI tool to make informed choices between these 2 treatments.

Association between platelet count and neonatal acute kidney injury: a cohort study using the medical information mart for intensive care III database

OBJECTIVE

A decrease in platelet count has been reported to be associated with several neonatal inflammatory diseases, including sepsis and necrotizing enterocolitis; while its association with neonatal acute kidney injury (AKI) has not been reported. This study aims to explore the association between platelet count and neonatal AKI.

METHODS

This was a retrospective cohort study based on the Medical Information Mart for Intensive Care III (MIMIC-III) database. Data were extracted based on baseline characteristics, comorbidities, vital signs, laboratory parameters, and intervention measures. Logistic regression analysis was used to assess the association between platelet count and AKI, and results were shown as odds ratios (OR) with 95% confidence intervals (CI).

RESULTS

A total of 1,576 neonates were finally included in the analysis. After adjusting birth weight, sepsis, patent ductus arteriosus, hematocrit, percentage of neutrophils, and vasopressor use, we found that platelet count in the lowest quartile (Q1) was significantly associated with the higher odds of AKI than platelet count in the highest quartile (Q4) (OR = 1.70, 95% CI: 1.01-2.87).

CONCLUSIONS

Low platelet count was associated with the high odds of AKI in the neonatal intensive care unit (NICU), indicating that platelet count might be a biomarker for neonatal AKI. Large-scale multicenter studies should be performed to verify the results.

Construction of Survival Nomogram for Ventilator-Associated Pneumonia Patients: Based on MIMIC Database

Objective: To construct and validate a predictive nomogram model for the survival of patients with ventilator-associated pneumonia (VAP) to enhance prediction of 28-day survival rate in critically ill patients with VAP. Methods: A total of 1,438 intensive care unit (ICU) patients with VAP were screened through Medical Information Mart for Intensive Care (MIMIC)-IV. On the basis of multi-variable Cox regression analysis data, nomogram performance in predicting survival status of patients with VAP at ICU admission for 7, 14, and 28 days was evaluated using the C-index and area under the curve (AUC). Calibration and decision curve analysis curves were generated to assess clinical value and effectiveness of model, and risk stratification was performed for patients with VAP. Result: Through stepwise regression screening of uni-variable and multi-variable Cox regression models, independent prognostic factors for predicting nomogram were determined, including age, race, body temperature, Sequential Organ Failure Assessment score, anion gap, bicarbonate concentration, partial pressure of carbon dioxide, mean corpuscular hemoglobin, and liver disease. The model had C-index values of 0.748 and 0.628 in the train and test sets, respectively. The receiver operating characteristic curve showed that nomogram had better performance in predicting 28-day survival status in the train set (AUC = 0.74), whereas it decreased in the test set (AUC = 0.66). Calibration and decision curve analysis curve results suggested that nomogram had favorable predictive performance and clinical efficacy. Kaplan-Meier curves showed significant differences in survival between low, medium, and high-risk groups in the total set and training set (log-rank p < 0.05), further validating the effectiveness of the model. Conclusion: The VAP patient admission ICU 7, 14, and 28-day survival prediction nomogram was constructed, contributing to risk stratification and decision-making for such patients. The model is expected to play a positive role in supporting personalized treatment and management of VAP.

Construction and verification of the prediction model for risk of sleep disturbance in elderly patients with hypertension: a cross-sectional survey based on NHANES database from 2005 to 2018

OBJECTIVE

To construct and verify a risk prediction model for sleep disturbance in elderly patients with hypertension, aiming to offer guidance for sleep management in this demographic.

METHODS

A cohort of 6,708 elderly hypertensive patients from the NHANES database, spanning 2005 to 2018, met the inclusion criteria and were selected for this study. Participants were randomly assigned to a development group (n = 4,696) and a verification group (n = 2,012) in a 7:3 ratio. The occurrence of sleep disturbance was assessed across the subjects. Independent risk factors for sleep disturbance were analyzed using weighted multivariate logistic regression within the development group. A predictive model for sleep disturbance risk in elderly hypertensive patients was developed and verified using Stata 17.0. The model's predictive accuracy and stability were evaluated using the verification group's data.

RESULTS

Of the 6,708 subjects, 2,014 (30.02%) were identified with sleep disturbance, and the weighted prevalence of sleep disturbance among elderly hypertensive patients was 33.283%. Weighted multivariate logistic regression analysis in the development group revealed that six factors were independently associated with sleep disturbance: higher total depression scores, higher education level, asthma, overweight, arthritis, and work restriction (OR > 1 and P < 0.05). The area under the receiver operating characteristic (ROC) curve (AUC) for the nomogram prediction model was 0.709 in the development group and 0.707 in the verification group, indicating good discrimination ability. Brier scores for the nomogram model were 0.185 in the development group and 0.189 in the verification group, both below 0.25, suggesting good calibration. Decision Curve Analysis (DCA) determined that the nomogram's clinical net benefit was maximized when the threshold probability for sleep disturbance in elderly hypertensive patients was 0.13-0.67 in the development group and 0.14-0.61 in the verification group, highlighting the model's clinical utility.

LIMITATIONS

This study is not without its limitations, including issues with data collection, the absence of external validation, and the non-extrapolation of results.

CONCLUSION

The prevalence of sleep disturbance among elderly hypertensive patients stands at 33.283%. The nomogram model, based on identified risk factors for sleep disturbance in this population, has demonstrated good predictive efficiency and clinical relevance. It serves as a valuable tool to assist healthcare providers in identifying elderly hypertensive patients at high risk for sleep disturbance.

Single-center nomogram model for sepsis complicated by acute lung injury

OBJECTIVE

To construct and validate a nomogram model for predicting sepsis complicated by acute lung injury (ALI).

METHODS

The healthcare records of 193 sepsis patients hospitalized at The Affiliated Tai'an City Central Hospital of Qingdao University from January 2022 to December 2023 were retrospectively reviewed. Among these patients, 69 were in the ALI group and 124 in the non-ALI group. A nomogram prediction model was constructed using logistic regression analysis. Its predictive performance was evaluated through various measures, including the area under the curve (AUC), calibration curve, decision curve, sensitivity, specificity, accuracy, recall rate, and precision rate.

RESULTS

The predictive factors included the neutrophil/lymphocyte ratio (NLR), oxygenation index (PaO2/FiO2), tumor necrosis factor-α (TNF-α), and acute physiology and chronic health evaluation II (APACHE II). The nomogram training set achieved an AUC of 0.959 (95% CI: 0.924-0.995), an accuracy of 92.59%, a recall of 96.70%, and a precision of 92.63%. In the validation set, the AUC was 0.938 (95% CI: 0.880-0.996), with an accuracy of 89.66%, a recall of 93.94%, and a precision of 88.57%. The calibration curve demonstrated that the prediction results were consistent with the actual findings. The decision curve indicated that the model has clinical applicability.

CONCLUSION

NLR, PaO2/FiO2, TNF-α, and APACHE II are closely associated with ALI in sepsis patients. A nomogram model based on these four variables shows strong predictive performance and may be used as a clinical decision-support tool to help physicians better identify high-risk groups.

Predictive nomogram for 28-day mortality risk in mitral valve disorder patients in the intensive care unit: A comprehensive assessment from the MIMIC-III database

BACKGROUND

Mitral valve disorder (MVD) stands as the most prevalent valvular heart disease. Presently, a comprehensive clinical index to predict mortality in MVD remains elusive. The aim of our study is to construct and assess a nomogram for predicting the 28-day mortality risk of MVD patients.

METHODS

Patients diagnosed with MVD were identified via ICD-9 code from the MIMIC-III database. Independent risk factors were identified utilizing the LASSO method and multivariate logistic regression to construct a nomogram model aimed at predicting the 28-day mortality risk. The nomogram's performance was assessed through various metrics including the area under the curve (AUC), calibration curves, Hosmer-Lemeshow test, integrated discriminant improvement (IDI), net reclassification improvement (NRI), and decision curve analysis (DCA).

RESULTS

The study encompassed a total of 2771 patients diagnosed with MVD. Logistic regression analysis identified several independent risk factors: age, anion gap, creatinine, glucose, blood urea nitrogen level (BUN), urine output, systolic blood pressure (SBP), respiratory rate, saturation of peripheral oxygen (SpO2), Glasgow Coma Scale score (GCS), and metastatic cancer. These factors were found to independently influence the 28-day mortality risk among patients with MVD. The calibration curve demonstrated adequate calibration of the nomogram. Furthermore, the nomogram exhibited favorable discrimination in both the training and validation cohorts. The calculations of IDI, NRI, and DCA analyses demonstrate that the nomogram model provides a greater net benefit compared to the Simplified Acute Physiology Score II (SAPSII), Acute Physiology Score III (APSIII), and Sequential Organ Failure Assessment (SOFA) scoring systems.

CONCLUSION

This study successfully identified independent risk factors for 28-day mortality in patients with MVD. Additionally, a nomogram model was developed to predict mortality, offering potential assistance in enhancing the prognosis for MVD patients. It's helpful in persuading patients to receive early interventional catheterization treatment, for example, transcatheter mitral valve replacement (TMVR), transcatheter mitral valve implantation (TMVI).

Development and validation of a risk prediction model for aspiration in patients with acute ischemic stroke

BACKGROUND

Aspiration is a frequently observed complication in individuals diagnosed with acute ischemic stroke, leading to potentially severe consequences. However, the availability of predictive tools for assessing aspiration probabilities remains limited. Hence, our study aimed to develop and validate a nomogram for accurately predicting aspiration probability in patients with acute ischemic stroke.

METHODS

We analyzed 30 potential risk factors associated with aspiration in 359 adult patients diagnosed with acute ischemic stroke. Advanced statistical techniques, such as Least absolute shrinkage and selection operator (LASSO) and Multivariate Logistic regression, were employed to identify independent predictors. Subsequently, we developed a nomogram prediction model based on these predictors, which underwent internal validation through 1000 bootstrap resampling. Two additional cohorts (Cohort A n = 64; Cohort B, n = 105) were included for external validation. The discriminatory power and calibration performance of the nomogram were assessed using rigorous methods, including the Hosmer-Lemeshow test, area under the receiver operating characteristic curve (AUC), calibration curve analyses, and decision curve analyses (DCA).

RESULTS

The nomogram was established based on four variables: sputum suction, brain stem infarction, temporal lobe infarction, and Barthel Index score. The predictive model exhibited satisfactory discriminative ability, with an area under the receiver operating characteristic curve of 0.853 (95 % confidence interval, 0.795-0.910), which remained consistent at 0.852 (95 % confidence interval, 0.794-0.912) during the internal validation. The Hosmer-Lemeshow test (P = 0.394) and calibration curve demonstrated favorable consistency between the predicted and observed outcomes in the development cohort. The AUC was 0.872 (95 % confidence interval, 0.783-0.962) in validation cohort A and 0.877 (95 % confidence interval, 0.764-0.989) in validation cohort B, demonstrating sustained accuracy. DCA showed a good net clinical benefit of the nomogram.

CONCLUSIONS

A nomogram for predicting the probability of aspiration in patients with acute ischemia has been successfully developed and validated.

Risk Factors and Prognosis Analysis of Upper Gastrointestinal Bleeding in Patients With Acute Severe Cerebral Stroke

GOALS

We aim to explore the relationship between the use of proton pump inhibitors (PPIs) and upper gastrointestinal bleeding (UGIB). We develop a nomogram model to predict mortality in critically ill stroke patients.

STUDY

This is a retrospective study based on the MIMIC IV database. We extracted clinical information including demographic data, comorbidities, and laboratory indicators. Univariate and multivariable logistic regressions were used to assess and identify risk factors for the occurrence of UGIB and for the in-hospital mortality of critically ill stroke patients. The resulting model was used to construct a nomogram for predicting in-hospital mortality.

RESULTS

Five thousand seven hundred sixteen patients from the MIMIC-IV database were included in our analysis. UGIB occurred in 109 patients (1.9%), whereas the PPI use rate was as high as 60.6%. Chronic liver disease, sepsis, shock, anemia, and increased level of urea nitrogen were independent risk factors for the occurrence of UGIB in severe stroke patients. We identified age, heart failure, shock, coagulopathy, mechanical ventilation, continuous renal replacement therapy, antiplatelet drugs, anticoagulation, simplified acute physiology score-II, and Glasgow coma score as independent risk factors for in-hospital mortality in severe stroke patients. The C-index for the final nomograms was 0.852 (95% confidence interval: 0.840, 0.864).

CONCLUSIONS

We found that the overall rate of UGIB in severe stroke patients is low, whereas the rate of PPI usage is high. In our study, PPI was not identified as a risk factor for the occurrence of UGIB and UGIB was not associated with all-cause mortality. More clinical trials are needed to evaluate the benefits of using PPI in critically ill stroke patients.

Development and validation of a nomogram for predicting 28-day mortality in patients with ischemic stroke

BACKGROUND/AIM

We aimed to construct a validated nomogram model for predicting short-term (28-day) ischemic stroke mortality among critically ill populations.

MATERIALS AND METHODS

We collected raw data from the Medical Information Mart for Intensive Care IV database, a comprehensive repository renowned for its depth and breadth in critical care information. Subsequently, a rigorous analytical framework was employed, incorporating a 10-fold cross-validation procedure to ensure robustness and reliability. Leveraging advanced statistical methodologies, specifically the least absolute shrinkage and selection operator regression, variables pertinent to 28-day mortality in ischemic stroke were meticulously screened. Next, binary logistic regression was utilized to establish nomogram, then applied concordance index to evaluate discrimination of the prediction models. Predictive performance of the nomogram was assessed by integrated discrimination improvement (IDI) and net reclassification index (NRI). Additionally, we generated calibration curves to assess calibrating ability. Finally, we evaluated the nomogram's net clinical benefit using decision curve analysis (DCA), in comparison with scoring systems clinically applied under common conditions.

RESULTS

A total of 2089 individuals were identified and assigned into training (n = 1443) or validation (n = 646) cohorts. Various identified risk factors, including age, ethnicity, marital status, underlying metastatic solid tumor, Charlson comorbidity index, heart rate, Glasgow coma scale, glucose concentrations, white blood cells, sodium concentrations, potassium concentrations, mechanical ventilation, use of heparin and mannitol, were associated with short-term (28-day) mortality in ischemic stroke individuals. A concordance index of 0.834 was obtained in the training dataset, indicating that our nomogram had good discriminating ability. Results of IDI and NRI in both cohorts proved that our nomogram had positive improvement of predictive performance, compared to other scoring systems. The actual and predicted incidence of mortality showed favorable concordance on calibration curves (P > 0.05). DCA curves revealed that, compared with scoring systems clinically used under common conditions, the constructed nomogram yielded a greater net clinical benefit.

CONCLUSIONS

Utilizing a comprehensive array of fourteen readily accessible variables, a prognostic nomogram was meticulously formulated and rigorously validated to provide precise prognostication of short-term mortality within the ischemic stroke cohort.

Development of m6A/m5C/m1A regulated lncRNA signature for prognostic prediction, personalized immune intervention and drug selection in LUAD

Research indicates that there are links between m6A, m5C and m1A modifications and the development of different types of tumours. However, it is not yet clear if these modifications are involved in the prognosis of LUAD. The TCGA-LUAD dataset was used as for signature training, while the validation cohort was created by amalgamating publicly accessible GEO datasets including GSE29013, GSE30219, GSE31210, GSE37745 and GSE50081. The study focused on 33 genes that are regulated by m6A, m5C or m1A (mRG), which were used to form mRGs clusters and clusters of mRG differentially expressed genes clusters (mRG-DEG clusters). Our subsequent LASSO regression analysis trained the signature of m6A/m5C/m1A-related lncRNA (mRLncSig) using lncRNAs that exhibited differential expression among mRG-DEG clusters and had prognostic value. The model's accuracy underwent validation via Kaplan-Meier analysis, Cox regression, ROC analysis, tAUC evaluation, PCA examination and nomogram predictor validation. In evaluating the immunotherapeutic potential of the signature, we employed multiple bioinformatics algorithms and concepts through various analyses. These included seven newly developed immunoinformatic algorithms, as well as evaluations of TMB, TIDE and immune checkpoints. Additionally, we identified and validated promising agents that target the high-risk mRLncSig in LUAD. To validate the real-world expression pattern of mRLncSig, real-time PCR was carried out on human LUAD tissues. The signature's ability to perform in pan-cancer settings was also evaluated. The study created a 10-lncRNA signature, mRLncSig, which was validated to have prognostic power in the validation cohort. Real-time PCR was applied to verify the actual manifestation of each gene in the signature in the real world. Our immunotherapy analysis revealed an association between mRLncSig and immune status. mRLncSig was found to be closely linked to several checkpoints, such as IL10, IL2, CD40LG, SELP, BTLA and CD28, which could be appropriate immunotherapy targets for LUAD. Among the high-risk patients, our study identified 12 candidate drugs and verified gemcitabine as the most significant one that could target our signature and be effective in treating LUAD. Additionally, we discovered that some of the lncRNAs in mRLncSig could play a crucial role in certain cancer types, and thus, may require further attention in future studies. According to the findings of this study, the use of mRLncSig has the potential to aid in forecasting the prognosis of LUAD and could serve as a potential target for immunotherapy. Moreover, our signature may assist in identifying targets and therapeutic agents more effectively.

Indian Society of Critical Care Medicine and Indian Association of Palliative Care Expert Consensus and Position Statements for End-of-life and Palliative Care in the Intensive Care Unit

End-of-life care (EOLC) exemplifies the joint mission of intensive and palliative care (PC) in their human-centeredness. The explosion of technological advances in medicine must be balanced with the culture of holistic care. Inevitably, it brings together the science and the art of medicine in their full expression. High-quality EOLC in the ICU is grounded in evidence, ethical principles, and professionalism within the framework of the Law. Expert professional statements over the last two decades in India were developed while the law was evolving. Recent landmark Supreme Court judgments have necessitated a review of the clinical pathway for EOLC outlined in the previous statements. Much empirical and interventional evidence has accumulated since the position statement in 2014. This iteration of the joint Indian Society of Critical Care Medicine-Indian Association of Palliative Care (ISCCM-IAPC) Position Statement for EOLC combines contemporary evidence, ethics, and law for decision support by the bedside in Indian ICUs.

HOW TO CITE THIS ARTICLE

Mani RK, Bhatnagar S, Butola S, Gursahani R, Mehta D, Simha S, et al. Indian Society of Critical Care Medicine and Indian Association of Palliative Care Expert Consensus and Position Statements for End-of-life and Palliative Care in the Intensive Care Unit. Indian J Crit Care Med 2024;28(3):200-250.

Constructing a Nomogram Model to Estimate the Risk of Ventilator-Associated Pneumonia for Elderly Patients in the Intensive Care Unit

BACKGROUND

Ventilator-associated pneumonia (VAP) causes heavy losses in terms of finances, hospitalization, and death for elderly patients in the intensive care unit (ICU); however, the risk is difficult to evaluate due to a lack of reliable assessment tools. We aimed to create and validate a nomogram to estimate VAP risk to provide early intervention for high-risk patients.

METHODS

Between January 2016 and March 2021, 293 patients from a tertiary hospital in China were retrospectively reviewed as a training set. Another 84 patients were enrolled for model validation from April 2021 to February 2022. Least absolute shrinkage and selection operator (LASSO) regression and multivariable logistic regression analysis were employed to select predictors, and a nomogram model was constructed. The calibration, discrimination, and clinical utility of the nomogram were verified. Finally, a web-based online scoring system was created to make the model more practical.

RESULTS

The predictors were hypoproteinemia, long-term combined antibiotic use, intubation time, length of mechanical ventilation, and tracheotomy/intubation. The area under the curve (AUC) was 0.937 and 0.925 in the training and validation dataset, respectively, suggesting the model exhibited effective discrimination. The calibration curve demonstrated high consistency with the observed result and the estimated values. Decision curve analysis (DCA) demonstrated that the nomogram was clinically applicable.

CONCLUSIONS

We have created a novel nomogram model that can be utilized to anticipate VAP risk in elderly ICU patients, which is helpful for healthcare professionals to detect patients at high risk early and adopt protective interventions.

Development and Validation of a Dynamic Nomogram for Predicting 3-Month Mortality in Acute Ischemic Stroke Patients with Atrial Fibrillation

Background

Acute ischemic stroke (AIS) in patients with atrial fibrillation (AF) carries a substantial risk of mortality, emphasizing the need for effective risk assessment and timely interventions. This study aimed to develop and validate a practical dynamic nomogram for predicting 3-month mortality in AIS patients with AF.

Methods

AIS patients with AF were enrolled and randomly divided into training and validation cohorts. The nomogram was developed based on independent risk factors identified by multivariate logistic regression analysis. The prediction performance of the nomogram was evaluated using the area under the receiver operating characteristic curve (AUC-ROC), calibration plots, decision curve analysis (DCA), and Kaplan-Meier survival analysis.

Results

A total of 412 patients with AIS and AF entered final analysis, 288 patients in the training cohort and 124 patients in the validation cohort. The nomogram was developed using age, baseline National Institutes of Health Stroke Scale score, early introduction of novel oral anticoagulants, and pneumonia as independent risk factors. The nomogram exhibited good discrimination both in the training cohort (AUC, 0.851; 95% CI, 0.802-0.899) and the validation cohort (AUC, 0.811; 95% CI, 0.706-0.916). The calibration plots, DCA and Kaplan-Meier survival analysis demonstrated that the nomogram was well calibrated and clinically useful, effectively distinguishing the 3-month survival status of patients with AIS and AF, respectively. The dynamic nomogram can be obtained at the website: https://yanxiaodi.shinyapps.io/3-monthmortality/.

Conclusion

The dynamic nomogram represents the first predictive model for 3-month mortality and may contribute to managing the mortality risk of patients with AIS and AF.

Relationship between baseline bicarbonate and 30-day mortality in patients with non-traumatic subarachnoid hemorrhage

Objective

This study aimed to explore the relationship between baseline bicarbonate levels and 30-day mortality in individuals with non-traumatic subarachnoid hemorrhage (SAH).

Methods

Patients with non-traumatic SAH were chosen from the Medical Information Mart for Intensive Care (MIMIC)-IV database. The relationship between baseline bicarbonate and 30-day mortality was examined using Cox regression models. Restricted cubic splines were used to test the hypothesis that there was an association between bicarbonate and mortality. With the use of Kaplan-Meier survival curve analysis, we looked deeper into the validity of these correlations. To find subgroups with differences, interaction tests were utilized.

Results

This retrospective cohort study consisted of 521 participants in total. Bicarbonate had a negative association with death at 30 days (HR = 0.93, 95%CI: 0.88-0.98, p = 0.004). Next, we divided bicarbonate into quartile groups. In comparison to the reference group Q1 (20 mEq/L), groups Q3 (23-25 mEq/L) and Q4 (26 mEq/L) had adjusted HR values of 0.47 (95%CI: 0.27-0.82, p = 0.007) and 0.56 (95%CI: 0.31-0.99, p = 0.047). No definite conclusions can be derived from this study, since there is no obvious curve link between baseline bicarbonate and 30-day mortality. Patients' 30-day mortality increased statistically significantly (p < 0.001, K-M analysis) in patients with low bicarbonate levels. The relationship between bicarbonate and 30-day mortality remained consistent in the stratified analysis, with no observed interactions.

Conclusion

Finally, 30-day mortality was negatively associated with baseline bicarbonate levels. Patients with non-traumatic SAH are more at risk of mortality if their bicarbonate levels are low.

Clinical decision support systems for 3-month mortality in elderly patients admitted to ICU with ischemic stroke using interpretable machine learning

Background

Elderly patients are more likely to suffer from severe ischemic stroke (IS) and have worse outcomes, including death and disability. We aimed to develop and validate predictive models using novel machine learning algorithms for the 3-month mortality in elderly patients with IS admitted to the intensive care unit (ICU).

Methods

We conducted a retrospective cohort study. Data were extracted from Medical Information Mart for Intensive Care (MIMIC)-IV and International Stroke Perfusion Imaging Registry (INSPIRE) database. Ten machine learning algorithms including Categorical Boosting (CatBoost), Random Forest (RF), Support Vector Machine (SVM), Neural Network (NN), Gradient Boosting Machine (GBM), K-Nearest Neighbors (KNNs), Multi-Layer Perceptron (MLP), Naive Bayes (NB), eXtreme Gradient Boosting (XGBoost) and Logistic Regression (LR) were used to build the models. Performance was measured using area under the curve (AUC) and accuracy. Finally, interpretable machine learning (IML) models presenting as Shapley additive explanation (SHAP) values were applied for mortality risk prediction.

Results

A total of 1826 elderly patients with IS admitted to the ICU were included in the analysis, of whom 624 (34.2%) died, and endovascular treatment was performed in 244 patients. After feature selection, a total of eight variables, including minimum Glasgow Coma Scale values, albumin, lactate dehydrogenase, age, alkaline phosphatase, body mass index, platelets, and types of surgery, were finally used for model construction. The AUCs of the CatBoost model were 0.737 in the testing set and 0.709 in the external validation set. The Brier scores in the training set and testing set were 0.12 and 0.21, respectively. The IML of the CatBoost model was performed based on the SHAP value and the Local Interpretable Model-Agnostic Explanations method.

Conclusion

The CatBoost model had the best predictive performance for predicting mortality in elderly patients with IS admitted to the ICU. The IML model would further aid in clinical decision-making and timely healthcare services by the early identification of high-risk patients.

Concept Coverage Analysis of Ophthalmic Infections and Trauma among the Standardized Medical Terminologies SNOMED-CT, ICD-10-CM, and ICD-11

Purpose

Widespread electronic health record adoption has generated a large volume of data and emphasized the need for standardized terminology to describe clinical concepts. Here, we undertook a systematic concept coverage analysis to determine the representation of clinical concepts in ophthalmic infection and ophthalmic trauma among standardized medical terminologies, including the Systematized Nomenclature of Medicine Clinical Terms (SNOMED-CT), the International Classification of Diseases (ICD) version 10 with clinical modifications (ICD-10-CM), and ICD version 11 (ICD-11).

Design

Extraction of concepts related to ophthalmic infection and ophthalmic trauma and structured search in terminology browsers.

Data Sources

The American Academy of Ophthalmology Basic and Clinical Science Course (BCSC), SNOMED-CT, and ICD-10-CM terminologies from the Observational Health Data Sciences and Informatics Athena browser, and the ICD-11 terminology browser.

Methods

Concepts pertaining to ophthalmic infection and ophthalmic trauma were extracted from the 2022 BCSC free text and index terms. We searched terminology browsers to identify corresponding codes and classified the extent of semantic alignment as equal, wide, narrow, or unmatched in each terminology. The overlap of equal concepts in each terminology was represented in a Venn diagram.

Main Outcome Measures

Proportions of clinical concepts with corresponding codes at various levels of semantic alignment.

Results

A total of 443 concepts were identified: 304 concepts related to ophthalmic infection and 139 concepts related to ophthalmic trauma. The SNOMED-CT had the highest proportion of equal coverage, with 82.0% (249 of 304) among concepts related to ophthalmic infection and 82.0% (115 of 139) among concepts related to ophthalmic trauma. Across all concepts, 28% (124 of 443) were classified as equal in ICD-10-CM and 52.8% (234 of 443) were classified as equal in ICD-11.

Conclusions

The SNOMED-CT had significantly better semantic alignment than ICD-10-CM and ICD-11 for ophthalmic infections and ophthalmic trauma. This demonstrates opportunity for continuing advancement of representation of ophthalmic concepts in standardized medical terminologies.

An interpretable ensemble learning model facilitates early risk stratification of ischemic stroke in intensive care unit: Development and external validation of ICU-ISPM

Ischemic stroke (IS) is a common and severe condition that requires intensive care unit (ICU) admission, with high mortality and variable prognosis. Accurate and reliable predictive tools that enable early risk stratification can facilitate interventions to improve patient outcomes; however, such tools are currently lacking. In this study, we developed and validated novel ensemble learning models based on soft voting and stacking methods to predict in-hospital mortality from IS in the ICU using two public databases: MIMIC-IV and eICU-CRD. Additionally, we identified the key predictors of mortality and developed a user-friendly online prediction tool for clinical use. The soft voting ensemble model, named ICU-ISPM, achieved an AUROC of 0.861 (95% CI: 0.829-0.892) and 0.844 (95% CI: 0.819-0.869) in the internal and external test cohorts, respectively. It significantly outperformed the APACHE scoring system and was more robust than individual models. ICU-ISPM obtained the highest performance compared to other models in similar studies. Using the SHAP method, the model was interpretable, revealing that GCS score, age, and intubation were the most important predictors of mortality. This model also provided a risk stratification system that can effectively distinguish between low-, medium-, and high-risk patients. Therefore, the ICU-ISPM is an accurate, reliable, interpretable, and clinically applicable tool, which is expected to assist clinicians in stratifying IS patients by the risk of mortality and rationally allocating medical resources. Based on ICU-ISPM, an online risk prediction tool was further developed, which was freely available at: http://ispm.idrblab.cn/.

Derivation and validation of a screening tool for stroke-associated sepsis

BACKGROUND

Post-stroke infections may cause sepsis, which is associated with poor clinical outcome. Sepsis is defined by life-threatening organ dysfunction that can be identified using the Sequential Organ Failure Assessment (SOFA) score. The applicability of the SOFA score for patients not treated on an intensive care unit (ICU) is limited. The aim of this study was to develop and validate an easier-to-use modification of the SOFA score for stroke patients.

METHODS

Using a registry-based cohort of 212 patients with large vessel occlusion stroke and infection, potential predictors of a poor outcome indicating sepsis were assessed by logistic regression. The derived score was validated on a separate cohort of 391 patients with ischemic stroke and infection admitted to our hospital over a period of 1.5 years.

RESULTS

The derived Stroke-SOFA (S-SOFA) score included the following predictors: National Institutes of Health stroke scale ≥ 14, peripheral oxygen saturation < 90%, mean arterial pressure < 70 mmHg, thrombocyte count < 150 10⁹/l and creatinine ≥ 1.2 mg/dl. The area under the receiver operating curve for the prediction of a poor outcome indicating sepsis was 0.713 [95% confidence interval: 0.665-0.762] for the S-SOFA score, which was comparable to the standard SOFA score (0.750 [0.703-0.798]), but the prespecified criteria for non-inferiority were not met (p = 0.115). However, the S-SOFA score was non-inferior compared to the SOFA score in non-ICU patients (p = 0.013).

CONCLUSIONS

The derived S-SOFA score may be useful to identify non-ICU patients with stroke-associated sepsis who have a high risk of a poor outcome.

Relationship between serum bicarbonate levels and the risk of death within 30 days in ICU patients with acute ischemic stroke

Aim

To explore the relationship between baseline bicarbonate levels and their changes with 30-day mortality in patients with acute ischemic stroke who were admitted to the intensive care unit (ICU).

Methods

This cohort study collected the data of 4,048 participants from the Medical Information Mart for Intensive Care (MIMIC)-III and MIMIC-IV databases. Univariate and multivariable Cox proportional risk models were utilized to explore the relationship between bicarbonate T0 and Δbicarbonate with 30-day mortality in patients with acute ischemic stroke. The Kaplan-Meier curves were plotted to measure the 30-day survival probability of patients with acute ischemic stroke.

Results

The median follow-up time was 30 days. At the end of the follow-up, 3,172 patients survived. Bicarbonate T0 ≤ 21 mEq/L [hazard ratio (HR) = 1.24, a 95% confidence interval (CI): 1.02-1.50] or 21 mEq/L < bicarbonate T0 ≤ 23 mEq/L (HR = 1.29, 95%CI: 1.05-1.58) were associated with an increased risk of 30-day mortality in patients with acute ischemic stroke compared with bicarbonate T0 > 26 mEq/L. -2 mEq/L < Δbicarbonate ≤ 0 mEq/L (HR = 1.40, 95%CI: 1.14-1.71), 0 mEq/L < Δbicarbonate ≤ 2 mEq/L (HR = 1.44, 95%CI: 1.17-1.76), and Δbicarbonate >2 mEq/L (HR = 1.40, 95%CI: 1.15-1.71) were correlated with an elevated risk of 30-day mortality in acute ischemic stroke patients. The 30-day survival probability of acute ischemic stroke patients with 21 mEq/L < bicarbonate T0 ≤ 23 mEq/L, 23 mEq/L < bicarbonate T0 ≤ 26 mEq/L, or bicarbonate T0 >26 mEq/L was higher than that of patients with bicarbonate T0 ≤ 21 mEq/L. The 30-day survival probability was greater for patients in the Δbicarbonate ≤ -2 mEq/L group than for those in the Δbicarbonate >2 mEq/L group.

Conclusion

Low baseline bicarbonate levels and decreased bicarbonate levels during the ICU stay were associated with a high risk of 30-day mortality in acute ischemic stroke patients. Special interventions should be offered to those with low baseline and decreased bicarbonate levels during their ICU stay.

Prediction of long-term mortality in patients with ischemic stroke based on clinical characteristics on the first day of ICU admission: An easy-to-use nomogram

Background

This study aimed to establish and validate an easy-to-use nomogram for predicting long-term mortality among ischemic stroke patients.

Methods

All raw data were obtained from the Medical Information Mart for Intensive Care IV database. Clinical features associated with long-term mortality (1-year mortality) among ischemic stroke patients were identified using least absolute shrinkage and selection operator regression. Then, binary logistic regression was used to construct a nomogram, the discrimination of which was evaluated by the concordance index (C-index), integrated discrimination improvement (IDI), and net reclassification index (NRI). Finally, a calibration curve and decision curve analysis (DCA) were employed to study calibration and net clinical benefit, compared to the Glasgow Coma Scale (GCS) and the commonly used disease severity scoring system.

Results

Patients who were identified with ischemic stroke were randomly assigned into developing (n = 1,443) and verification (n = 646) cohorts. The following factors were associated with 1-year mortality among ischemic stroke patients, including age on ICU admission, marital status, underlying dementia, underlying malignant cancer, underlying metastatic solid tumor, heart rate, respiratory rate, oxygen saturation, white blood cells, anion gap, mannitol injection, invasive mechanical ventilation, and GCS. The construction of the nomogram was based on the abovementioned features. The C-index of the nomogram in the developing and verification cohorts was 0.820 and 0.816, respectively. Compared with GCS and the commonly used disease severity scoring system, the IDI and NRI of the constructed nomogram had a statistically positive improvement in predicting long-term mortality in both developing and verification cohorts (all with p < 0.001). The actual mortality was consistent with the predicted mortality in the developing (p = 0.862) and verification (p = 0.568) cohorts. Our nomogram exhibited greater net clinical benefit than GCS and the commonly used disease severity scoring system.

Conclusion

This proposed nomogram has good performance in predicting long-term mortality among ischemic stroke patients.

The Combination of Whole-Brain Features and Local-Lesion Features in DSC-PWI May Improve Ischemic Stroke Outcome Prediction

Accurate and reliable outcome predictions can help evaluate the functional recovery of ischemic stroke patients and assist in making treatment plans. Given that recovery factors may be hidden in the whole-brain features, this study aims to validate the role of dynamic radiomics features (DRFs) in the whole brain, DRFs in local ischemic lesions, and their combination in predicting functional outcomes of ischemic stroke patients. First, the DRFs in the whole brain and the DRFs in local lesions of dynamic susceptibility contrast-enhanced perfusion-weighted imaging (DSC-PWI) images are calculated. Second, the least absolute shrinkage and selection operator (Lasso) is used to generate four groups of DRFs, including the outstanding DRFs in the whole brain (Lasso (WB)), the outstanding DRFs in local lesions (Lasso (LL)), the combination of them (combined DRFs), and the outstanding DRFs in the combined DRFs (Lasso (combined)). Then, the performance of the four groups of DRFs is evaluated to predict the functional recovery in three months. As a result, Lasso (combined) in the four groups achieves the best AUC score of 0.971, which improves the score by 8.9% compared with Lasso (WB), and by 3.5% compared with Lasso (WB) and combined DRFs. In conclusion, the outstanding combined DRFs generated from the outstanding DRFs in the whole brain and local lesions can predict functional outcomes in ischemic stroke patients better than the single DRFs in the whole brain or local lesions.