Interpretable machine learning for prediction of clinical outcomes in acute ischemic stroke

Dynamic, Interpretable, Machine Learning-Based Outcome Prediction as a New Emerging Opportunity in Acute Ischemic Stroke Patient Care: A Proof-of-Concept Study

Introduction: While the machine learning (ML) model's black-box nature presents a significant barrier to effective clinical application, the dynamic nature of stroke patients' recovery further undermines the reliability of established predictive scores and models, making them less suitable for accurate prediction and appropriate patient care. This research is aimed at building and evaluating an interpretable ML-based model, which would perform outcome prediction at different time points of patients' recovery, giving more secure and understandable output through interpretable packages. Materials and Methods: A retrospective analysis was conducted on acute ischemic stroke (AIS) patients treated with alteplase at the Neurology Clinic of the University Clinical Center of Vojvodina (Novi Sad, Serbia), for 14 years. Clinical data were grouped into four categories based on collection time-baseline, 2-h, 24-h, and discharge features-serving as inputs for three different classifiers-support vector machine (SVM), logistic regression (LR), and random forest (RF). The 90-day modified Rankin scale (mRS) was used as the outcome measure, distinguishing between favorable (mRS ≤ 2) and unfavorable outcomes (mRS ≥ 3). Results: The sample was described with 49 features and included 355 patients, with a median age of 67 years (interquartile range (IQR) 60-74 years), 66% being male. The models achieved strong discrimination in the testing set, with area under the curve (AUC) values ranging from 0.80 to 0.96. Additionally, they were compared with a model based on the DRAGON score, which showed an AUC of 0.760 (95% confidence interval (CI), 0.640-0.862). The decision-making process was more thoroughly understood using interpretable packages: Shapley additive explanation (SHAP) and local interpretable model-agnostic explanation (LIME). They revealed the most significant features at both the group and individual patient levels. Conclusions and Clinical Implications: This study demonstrated the moderate to strong efficacy of interpretable ML-based models in predicting the functional outcomes of alteplase-treated AIS patients. In all constructed models, age, onset-to-treatment time, and platelet count were recognized as the important predictors, followed by clinical parameters measured at different time points, such as the National Institutes of Health Stroke Scale (NIHSS) and systolic and diastolic blood pressure values. The dynamic approach, coupled with interpretable models, can aid in providing insights into the potential factors that could be modified and thus contribute to a better outcome.

Construction of predictive model for the risk of acute lactic acidosis in patients with ischemic stroke during the ICU stay: A study based on the medical information Mart for intensive care database

BACKGROUND

This study aims to identify the factors influencing the risk of lactic acidosis (LA) in patients with ischemic stroke (IS) and to develop a predictive model for assessing the risk of LA in IS patients during their stay in the intensive care unit (ICU).

METHODS

A retrospective cohort design was employed, with data collected from the Medical Information Mart for Intensive Care (MIMIC)-III and MIMIC-IV databases spanning from 2001 to 2019. LA was defined as pH < 7.35 and lactate ≥ 2 mmol/L. The total sample was randomly divided into a training set and a testing set at a 7:3 ratio. Predictive variables were selected using bidirectional stepwise regression to build the final model. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curves.

RESULTS

The study included 531 patients, of whom 50 (13.47 %) developed LA. The predictive factors included in the model were hypertension, weight, heart rate, Charlson comorbidity index (CCI), Sequential Organ Failure Assessment (SOFA) score, white blood cell (WBC) count, insulin use, sodium bicarbonate administration, and renal replacement therapy (RRT).. The model demonstrated an area under the ROC curve (AUC) of 0.785 [95 % confidence interval (CI): 0.717-0.854] for the training dataset, and 0.721 (95 % CI: 0.615-0.826) for the testing dataset.

CONCLUSION

The predictive model developed for assessing the risk of LA in IS patients demonstrates encouraging predictive performance. It can play a crucial role in managing acid-base balance during ICU stays and assist in the prevention and management of LA in these patients.

A Claims-Based Machine Learning Classifier of Modified Rankin Scale in Acute Ischemic Stroke

Background

We developed a classifier to infer acute ischemic stroke (AIS) severity from Medicare claims using the Modified Rankin Scale (mRS) at discharge. The classifier can be utilized to improve stroke outcomes research and support the development of national surveillance tools.

Methods

This was a multistate study included all participating centers in the Paul Coverdell National Acute Stroke Program (PCNASP) database from nine U.S. states. PCNASP was linked to Medicare data sets for patients hospitalized with AIS, employing demographics, admission details, and diagnosis codes to create unique patient matches. We included Medicare beneficiaries aged 65 and older who were hospitalized for an initial AIS from January 2018 to December 2020. Using Lasso-penalized logistic regression, we developed and validated a binary classifier for mRS outcomes and as a secondary analysis we used ordinal regression to model the full mRS scale. Performance was evaluated on held-out test data using ROC AUC, ROC Precision-Recall, sensitivity, and specificity.

Results

We analyzed data from 68,636 eligible patients. The mean age was 79.5 years old. 77.5% of beneficiaries were White, 14% were Black, 2.6% were Asian, and 2% were Hispanic. The classifier achieved an ROC AUC score of 0.85 (95%CI: 0.85-0.86), sensitivity of 0.81 (95%CI: 0.80-0.81), specificity of 0.73 (0.72 - 0.74), and Precision-Recall AUC of 0.90 (95%CI: 0.90-0.91) on the test set.

Conclusion

Among Medicare beneficiaries hospitalized for AIS, the claims-based classifier demonstrated excellent performance in ROC AUC, Precision-Recall AUC, sensitivity, and acceptable specificity for mRS classification.

Machine learning-based predictive model for the development of thrombolysis resistance in patients with acute ischemic stroke

BACKGROUND

The objective of this study was to establish a predictive model utilizing machine learning techniques to anticipate the likelihood of thrombolysis resistance (TR) in acute ischaemic stroke (AIS) patients undergoing recombinant tissue plasminogen activator (rt-PA) intravenous thrombolysis, given that nearly half of such patients exhibit poor clinical outcomes.

METHODS

Retrospective clinical data were collected from AIS patients who underwent intravenous thrombolysis with rt-PA at the First Affiliated Hospital of Bengbu Medical University. Thrombolysis resistance was defined as ([National Institutes of Health Stroke Scale (NIHSS) at admission - 24-hour NIHSS] × 100％/ NIHSS at admission) ≤ 30%. In this study, we developed five machine learning models: logistic regression (LR), extreme gradient boosting (XGBoost), support vector machine (SVM), the least absolute shrinkage and selection operator (LASSO), and random forest (RF). We assessed the model's performance by using receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA), and presented the results through a nomogram.

RESULTS

This study included a total of 218 patients with AIS who were treated with intravenous thrombolysis, 88 patients experienced TR. Among the five machine learning models, the LASSO model performed the best. The area under the curve (AUC) on the testing group was 0.765 (sensitivity: 0.767, specificity: 0.694, accuracy: 0.727). The apparent curve in the calibration curve was similar to the ideal curve, and DCA showed a positive net benefit. Key features associated with TR included NIHSS at admission, blood glucose, white blood cell count, neutrophil count, and blood urea nitrogen.

CONCLUSION

Machine learning methods with multiple clinical variables can help in early screening of patients at high risk of thrombolysis resistance, particularly in contexts where healthcare resources are limited.

Leveraging multivariate analysis and adjusted mutual information to improve stroke prediction and interpretability

OBJECTIVES

To develop a machine learning model to accurately predict stroke risk based on demographic and clinical data. It also sought to identify the most significant stroke risk factors and determine the optimal machine learning algorithm for stroke prediction.

METHODS

This cross-sectional study analyzed data on 438,693 adults from the 2021 Behavioral Risk Factor Surveillance System. Features encompassed demographics and clinical factors. Descriptive analysis profiled the dataset. Logistic regression quantified risk relationships. Adjusted mutual information evaluated feature importance. Multiple machine learning models were built and evaluated on metrics like accuracy, AUC ROC, and F1 score.

RESULTS

Key factors significantly associated with higher stroke odds included older age, diabetes, hypertension, high cholesterol, and history of myocardial infarction or angina. Random forest model achieved the best performance with accuracy of 72.46%, AUC ROC of 0.72, and F1 score of 0.74. Cross-validation confirmed its reliability. Top features were hypertension, myocardial infarction history, angina, age, diabetes status, and cholesterol.

CONCLUSION

The random forest model robustly predicted stroke risk using demographic and clinical variables. Feature importance highlighted priorities like hypertension and diabetes for clinical monitoring and intervention. This could help enable data-driven stroke prevention strategies.

Magnetic resonance imaging-based deep learning imaging biomarker for predicting functional outcomes after acute ischemic stroke

PURPOSE

Clinical risk scores are essential for predicting outcomes in stroke patients. The advancements in deep learning (DL) techniques provide opportunities to develop prediction applications using magnetic resonance (MR) images. We aimed to develop an MR-based DL imaging biomarker for predicting outcomes in acute ischemic stroke (AIS) and evaluate its additional benefit to current risk scores.

METHOD

This study included 3338 AIS patients. We trained a DL model using deep neural network architectures on MR images and radiomics to predict poor functional outcomes at three months post-stroke. The DL model generated a DL score, which served as the DL imaging biomarker. We compared the predictive performance of this biomarker to five risk scores on a holdout test set. Additionally, we assessed whether incorporating the imaging biomarker into the risk scores improved the predictive performance.

RESULTS

The DL imaging biomarker achieved an area under the receiver operating characteristic curve (AUC) of 0.788. The AUCs of the five studied risk scores were 0.789, 0.793, 0.804, 0.810, and 0.826, respectively. The imaging biomarker's predictive performance was comparable to four of the risk scores but inferior to one (p = 0.038). Adding the imaging biomarker to the risk scores improved the AUCs (p-values) to 0.831 (0.003), 0.825 (0.001), 0.834 (0.003), 0.836 (0.003), and 0.839 (0.177), respectively. The net reclassification improvement and integrated discrimination improvement indices also showed significant improvements (all p < 0.001).

CONCLUSIONS

Using DL techniques to create an MR-based imaging biomarker is feasible and enhances the predictive ability of current risk scores.

A comprehensive review for artificial intelligence on neuroimaging in rehabilitation of ischemic stroke

Stroke is the second leading cause of death worldwide, with ischemic stroke accounting for a significant proportion of morbidity and mortality among stroke patients. Ischemic stroke often causes disability and cognitive impairment in patients, which seriously affects the quality of life of patients. Therefore, how to predict the recovery of patients can provide support for clinical intervention in advance and improve the enthusiasm of patients for rehabilitation treatment. With the popularization of imaging technology, the diagnosis and treatment of ischemic stroke patients are often accompanied by a large number of imaging data. Through machine learning and Deep Learning, information from imaging data can be used more effectively. In this review, we discuss recent advances in neuroimaging, machine learning, and Deep Learning in the rehabilitation of ischemic stroke.