The Comparison and Interpretation of Machine-Learning Models in Post-Stroke Functional Outcome Prediction

Association Between Multiple Plasma Toxic Metal and Metalloid Exposures and Hypertension in Elderly Chinese Adults

Although environmental exposure to toxic metals and metalloids is linked with the risk of cardiovascular diseases, the evidence is limited in the elderly. We evaluated the associations between 12 plasma metal levels including aluminum (Al), titanium (Ti), strontium (Sr), lead (Pb), vanadium (V), chromium (Cr), cobalt (Co), nickel (Ni), cuprum (Cu), zinc (Zn), arsenic (As), and selenium (Se) with prevalence of hypertension in the elderly Chinese population. In this study, stratified cluster sampling was conducted among elderly residents in three communities in Gansu province from June to July 2023, with a total of 330 participants included. The concentrations of metals in whole plasma were measured using inductively coupled plasma mass spectrometry (ICP-MS). Logistic regression and restricted cubic spline analyses were used to evaluate the dose-response relationship between plasma metal levels and hypertension, with all metal concentrations log-transformed. We applied quantile g-computation (QG-comp) and Bayesian kernel machine regression (BKMR) models to examine the associations of both individual metals and metal mixtures with hypertension. After multivariable adjustments, the odds ratios (ORs) and 95% confidence intervals (CIs) for hypertension associated with the highest quartile of metal concentrations compared to the lowest quartile were as follows: 4.20 (1.36, 12.98) for Sr, 3.95 (1.30, 12.03) for V, 3.43 (1.09, 10.78) for Cr, 3.28 (1.16, 9.28) for Cu, 3.28 (1.13, 9.52) for Zn, and 2.87 (0.94, 8.74) for As. Using BKMR and restricted cubic spline analysis, we found that exposure to metal mixtures was positively associated with an increased risk of hypertension, with Ni, Cr, As, and V being the primary contributing factors. In addition, Zn, Ni, and Sr were significantly and positively correlated with hypertension, while plasma titanium levels were negatively associated with hypertension development. These results suggest a complex interaction between various metals and the risk of hypertension in the elderly. Exposure to metal mixtures was positively associated with hypertension risk in elderly Chinese adults, with Ni, Cr, As, and V as key contributors. In addition, Zn, Ni, and Sr are significantly associated with an increased risk of hypertension, while Ti was positively associated with its development.

Predictive modelling of hospital-acquired infection in acute ischemic stroke using machine learning

Hospital-acquired infections (HAIs) are serious complication for patients with acute ischemic stroke (AIS), often resulting in poor functional outcomes. However, no existing model can specifically predict HAI in AIS patients. Therefore, we employed the Gradient Boosting matching learning algorithm to establish predictive models for HAI occurrence in AIS patients and poor 30-day functional outcomes (modified Rankin Scale > 2) in AIS patients with HAI by analyzing electronic health records from 6560 AIS patients. Model performance was evaluated through internal cross-validation and external validation using an independent cohort of 3521 AIS patients. The established models demonstrated robust predictive performance for HAI in AIS patients, achieving area under the receiver operating characteristic curves (AUROCs) of 0.857 ± 0.008 during internal validation and 0.825 ± 0.002 during external validation. For AIS patients with HAI, the second model effectively predict poor 30-day functional outcomes, with AUROCs of 0.905 ± 0.009 during internal validation and 0.907 ± 0.002 during external validation. In conclusion, machine learning models effectively identify the HAI occurrence and predict poor 30-day functional outcomes in AIS patients with HAI. Future prospective studies are crucial for validating and refining these models for clinical application, as well as for developing an accessible flowchart or scoring system to enhance clinical practices.

Impact of Interprofessional Collaborative Practice on Functional Improvements Among Post-Acute Stroke Survivors: A Retrospective Cross-Sectional Study

Background

Stroke survivors in post-acute care frequently experience physiological dysfunction and reduced quality of life. This study aims to assess the impact of the Post-Acute Care Interprofessional Collaborative Practice (PAC-IPCP) program across different care settings, and to identify sensitive tools for assessing physiological functions among post-acute stroke survivors.

Methods

This retrospective study involved 210 stroke survivors in Taiwan. Participants who self-selection for their preferred between hospital care setting and home care setting under PAC-IPCP. Multiple assessment tools were utilized, including the Barthel Index (BI), Functional Oral Intake Scale (FOIS), Mini Nutritional Assessment (MNA), EQ-5D-3L, and Instrumental Activities of Daily Living (IADL). The logistic regression was used to estimate the odds ratios of various functional assessment tools between hospital and home care settings. Additionally, the area under the ROC curves was used to determine which functional assessment tools had higher accuracy in measuring the association between care settings.

Results

Of the study population, 138 stroke survivors (65.71%) selection hospital care setting and 72 stroke survivors (34.29%) selection home care setting. The PAC-IPCP program was equally effective in both care settings for physical function status and quality of life improvements. Specifically, the BI emerged as the most sensitive tool for assessing care settings, with an adjusted OR of 1.04 (95% CI:1.02-1.07, p < 0.0001; AUC = 0.7557). IPCP-based hospital and home care models are equally effective in facilitating improved functional outcomes in post-acute stroke survivors.

Conclusion

The PAC-IPCP program is versatile and effective across care settings. The BI stands out as a robust assessment tool for physiological functions, endorsing its broader clinical application. Future studies should also consider swallowing and nutritional status for a more holistic approach to rehabilitation.

Predicting Language Function Post-Stroke: A Model-Based Structural Connectivity Approach

BACKGROUND

The prediction of post-stroke language function is essential for the development of individualized treatment plans based on the personal recovery potential of aphasic stroke patients.

OBJECTIVE

To establish a framework for integrating information on connectivity disruption of the language network based on routinely collected clinical magnetic resonance (MR) images into Random Forest modeling to predict post-stroke language function.

METHODS

Language function was assessed in 76 stroke patients from the Non-Invasive Repeated Therapeutic Stimulation for Aphasia Recovery trial, using the Token Test (TT), Boston Naming Test (BNT), and Semantic Verbal Fluency (sVF) Test as primary outcome measures. Individual infarct masks were superimposed onto a diffusion tensor imaging tractogram reference set to calculate Change in Connectivity scores of language-relevant gray matter regions as estimates of structural connectivity disruption. Multivariable Random Forest models were derived to predict language function.

RESULTS

Random Forest models explained moderate to high amount of variance at baseline and follow-up for the TT (62.7% and 76.2%), BNT (47.0% and 84.3%), and sVF (52.2% and 61.1%). Initial language function and non-verbal cognitive ability were the most important variables to predict language function. Connectivity disruption explained additional variance, resulting in a prediction error increase of up to 12.8% with variable omission. Left middle temporal gyrus (12.8%) and supramarginal gyrus (9.8%) were identified as among the most important network nodes.

CONCLUSION

Connectivity disruption of the language network adds predictive value beyond lesion volume, initial language function, and non-verbal cognitive ability. Obtaining information on connectivity disruption based on routine clinical MR images constitutes a significant advancement toward practical clinical application.

Estimating the volume of penumbra in rodents using DTI and stack-based ensemble machine learning framework

BACKGROUND

This study investigates the potential of diffusion tensor imaging (DTI) in identifying penumbral volume (PV) compared to the standard gadolinium-required perfusion-diffusion mismatch (PDM), utilizing a stack-based ensemble machine learning (ML) approach with enhanced explainability.

METHODS

Sixteen male rats were subjected to middle cerebral artery occlusion. The penumbra was identified using PDM at 30 and 90 min after occlusion. We used 11 DTI-derived metrics and 14 distance-based features to train five voxel-wise ML models. The model predictions were integrated using stack-based ensemble techniques. ML-estimated and PDM-defined PVs were compared to evaluate model performance through volume similarity assessment, the Pearson correlation analysis, and Bland-Altman analysis. Feature importance was determined for explainability.

RESULTS

In the test rats, the ML-estimated median PV was 106.4 mL (interquartile range 44.6-157.3 mL), whereas the PDM-defined median PV was 102.0 mL (52.1-144.9 mL). These PVs had a volume similarity of 0.88 (0.79-0.96), a Pearson correlation coefficient of 0.93 (p < 0.001), and a Bland-Altman bias of 2.5 mL (2.4% of the mean PDM-defined PV), with 95% limits of agreement ranging from -44.9 to 49.9 mL. Among the features used for PV prediction, the mean diffusivity was the most important feature.

CONCLUSIONS

Our study confirmed that PV can be estimated using DTI metrics with a stack-based ensemble ML approach, yielding results comparable to the volume defined by the standard PDM. The model explainability enhanced its clinical relevance. Human studies are warranted to validate our findings.

RELEVANCE STATEMENT

The proposed DTI-based ML model can estimate PV without the need for contrast agent administration, offering a valuable option for patients with kidney dysfunction. It also can serve as an alternative if perfusion map interpretation fails in the clinical setting.

KEY POINTS

• Penumbral volume can be estimated by DTI combined with stack-based ensemble ML. • Mean diffusivity was the most important feature used for predicting penumbral volume. • The proposed approach can be beneficial for patients with kidney dysfunction.

Clinical machine learning predicting best stroke rehabilitation responders to exoskeletal robotic gait rehabilitation

BACKGROUND

Although clinical machine learning (ML) algorithms offer promising potential in forecasting optimal stroke rehabilitation outcomes, their specific capacity to ascertain favorable outcomes and identify responders to robotic-assisted gait training (RAGT) in individuals with hemiparetic stroke undergoing such intervention remains unexplored.

OBJECTIVE

We aimed to determine the best predictive model based on the international classification of functioning impairment domain features (Fugl- Meyer assessment (FMA), Modified Barthel index related-gait scale (MBI), Berg balance scale (BBS)) and reveal their responsiveness to robotic assisted gait training (RAGT) in patients with subacute stroke.

METHODS

Data from 187 people with subacute stroke who underwent a 12-week Walkbot RAGT intervention were obtained and analyzed. Overall, 18 potential predictors encompassed demographic characteristics and the baseline score of functional and structural features. Five predictive ML models, including decision tree, random forest, eXtreme Gradient Boosting, light gradient boosting machine, and categorical boosting, were used.

RESULTS

The initial and final BBS, initial BBS, final Modified Ashworth scale, and initial MBI scores were important features, predicting functional improvements. eXtreme Gradient Boosting demonstrated superior performance compared to other models in predicting functional recovery after RAGT in patients with subacute stroke.

CONCLUSION

eXtreme Gradient Boosting may be an invaluable prognostic tool, providing clinicians and caregivers with a robust framework to make precise clinical decisions regarding the identification of optimal responders and effectively pinpoint those who are most likely to derive maximum benefits from RAGT interventions.

Prediction of enhanced drug solubility related to clathrate compositions and operating conditions: Machine learning study

Although complexation technique has been documented as a promising strategy to enhance the dissolution rate and bioavailability of water-insoluble drugs, prediction of the enhanced drug solubility related to clathrate compositions and operating conditions is still a challenge. Herein, clathrate compositions (drug content (DC), drug molecular weight (M) and molar ratio (Ratio)), operating conditions (drug concentration (C), pH, pressure (P), temperature (T) and dissolution time (t)) under the different excipients (PEG, PVP, HPMC and cyclodextrin) as main solubilizers of the clathrates condition as input parameters were used to predict two indexes (drug dissolved percentage and dissolution efficiency) simultaneously through machine learning methodfor the first time. The results show that PVP as the main solubilizer of clathrates had higher prediction accuracy to the drug dissolved percentage, and HPMC as the main solubilizer of clathrates had higher prediction accuracy to the drug dissolution efficiency. In addition, the influence of various factors and interactions on the target variables were analyzed. This study affords achievable hints to the quantitative prediction of the drug solubility affected by various compositions and different operating conditions.

Prediction of posttraumatic functional recovery in middle-aged and older patients through dynamic ensemble selection modeling

Introduction

Age-specific risk factors may delay posttraumatic functional recovery; complex interactions exist between these factors. In this study, we investigated the prediction ability of machine learning models for posttraumatic (6 months) functional recovery in middle-aged and older patients on the basis of their preexisting health conditions.

Methods

Data obtained from injured patients aged ≥45 years were divided into training-validation (n = 368) and test (n = 159) data sets. The input features were the sociodemographic characteristics and baseline health conditions of the patients. The output feature was functional status 6 months after injury; this was assessed using the Barthel Index (BI). On the basis of their BI scores, the patients were categorized into functionally independent (BI >60) and functionally dependent (BI ≤60) groups. The permutation feature importance method was used for feature selection. Six algorithms were validated through cross-validation with hyperparameter optimization. The algorithms exhibiting satisfactory performance were subjected to bagging to construct stacking, voting, and dynamic ensemble selection models. The best model was evaluated on the test data set. Partial dependence (PD) and individual conditional expectation (ICE) plots were created.

Results

In total, nineteen of twenty-seven features were selected. Logistic regression, linear discrimination analysis, and Gaussian Naive Bayes algorithms exhibited satisfactory performances and were, therefore, used to construct ensemble models. The k-Nearest Oracle Elimination model outperformed the other models when evaluated on the training-validation data set (sensitivity: 0.732, 95% CI: 0.702-0.761; specificity: 0.813, 95% CI: 0.805-0.822); it exhibited compatible performance on the test data set (sensitivity: 0.779, 95% CI: 0.559-0.950; specificity: 0.859, 95% CI: 0.799-0.912). The PD and ICE plots showed consistent patterns with practical tendencies.

Conclusion

Preexisting health conditions can predict long-term functional outcomes in injured middle-aged and older patients, thus predicting prognosis and facilitating clinical decision-making.

Predictive model for early functional outcomes following acute care after traumatic brain injuries: A machine learning-based development and validation study

INTRODUCTION

Few studies on early functional outcomes following acute care after traumatic brain injury (TBI) are available. The aim of this study was to develop and validate a predictive model for functional outcomes at discharge for TBI patients using machine learning methods.

PATIENTS AND METHODS

In this retrospective study, data from 5281 TBI patients admitted for acute care who were identified in the Beijing hospital discharge abstract database were analysed. Data from 4181 patients in 52 tertiary hospitals were used for model derivation and internal validation. Data from 1100 patients in 21 secondary hospitals were used for external validation. A poor outcome was defined as a Barthel Index (BI) score ≤ 60 at discharge. Logistic regression, XGBoost, random forest, decision tree, and back propagation neural network models were used to fit classification models. Performance was evaluated by the area under the receiver operating characteristic curve (AUC), the area under the precision-recall curve (AP), calibration plots, sensitivity/recall, specificity, positive predictive value (PPV)/precision, negative predictive value (NPV) and F1-score.

RESULTS

Compared to the other models, the random forest model demonstrated superior performance in internal validation (AUC of 0.856, AP of 0.786, and F1-score of 0.724) and external validation (AUC of 0.779, AP of 0.630, and F1-score of 0.604). The sensitivity/recall, specificity, PPV/precision, and NPV of the model were 71.8%, 69.2%, 52.2%, and 84.0%, respectively, in external validation. The BI score at admission, age, use of nonsurgical treatment, neurosurgery status, and modified Charlson Comorbidity Index were identified as the top 5 predictors for functional outcome at discharge.

CONCLUSIONS

We established a random forest model that performed well in predicting early functional outcomes following acute care after TBI. The model has utility for informing decision-making regarding patient management and discharge planning and for facilitating health care quality assessment and resource allocation for TBI treatment.

Predicting Arm Nonuse in Individuals with Good Arm Motor Function after Stroke Rehabilitation: A Machine Learning Study

Many stroke survivors demonstrate arm nonuse despite good arm motor function. This retrospective secondary analysis aims to identify predictors of arm nonusers with good arm motor function after stroke rehabilitation. A total of 78 participants were categorized into 2 groups using the Fugl-Meyer Assessment Upper Extremity Scale (FMA-UE) and the Motor Activity Log Amount of Use (MAL-AOU). Group 1 comprised participants with good motor function (FMA-UE ≥ 31) and low daily upper limb use (MAL-AOU ≤ 2.5), and group 2 comprised all other participants. Feature selection analysis was performed on 20 potential predictors to identify the 5 most important predictors for group membership. Predictive models were built with the five most important predictors using four algorithms. The most important predictors were preintervention scores on the FMA-UE, MAL-Quality of Movement, Wolf Motor Function Test-Quality, MAL-AOU, and Stroke Self-Efficacy Questionnaire. Predictive models classified the participants with accuracies ranging from 0.75 to 0.94 and areas under the receiver operating characteristic curve ranging from 0.77 to 0.97. The result indicates that measures of arm motor function, arm use in activities of daily living, and self-efficacy could predict postintervention arm nonuse despite good arm motor function in stroke. These assessments should be prioritized in the evaluation process to facilitate the design of individualized stroke rehabilitation programs to reduce arm nonuse.

The use of machine learning and deep learning techniques to assess proprioceptive impairments of the upper limb after stroke

BACKGROUND

Robots can generate rich kinematic datasets that have the potential to provide far more insight into impairments than standard clinical ordinal scales. Determining how to define the presence or absence of impairment in individuals using kinematic data, however, can be challenging. Machine learning techniques offer a potential solution to this problem. In the present manuscript we examine proprioception in stroke survivors using a robotic arm position matching task. Proprioception is impaired in 50-60% of stroke survivors and has been associated with poorer motor recovery and longer lengths of hospital stay. We present a simple cut-off score technique for individual kinematic parameters and an overall task score to determine impairment. We then compare the ability of different machine learning (ML) techniques and the above-mentioned task score to correctly classify individuals with or without stroke based on kinematic data.

METHODS

Participants performed an Arm Position Matching (APM) task in an exoskeleton robot. The task produced 12 kinematic parameters that quantify multiple attributes of position sense. We first quantified impairment in individual parameters and an overall task score by determining if participants with stroke fell outside of the 95% cut-off score of control (normative) values. Then, we applied five machine learning algorithms (i.e., Logistic Regression, Decision Tree, Random Forest, Random Forest with Hyperparameters Tuning, and Support Vector Machine), and a deep learning algorithm (i.e., Deep Neural Network) to classify individual participants as to whether or not they had a stroke based only on kinematic parameters using a tenfold cross-validation approach.

RESULTS

We recruited 429 participants with neuroimaging-confirmed stroke (< 35 days post-stroke) and 465 healthy controls. Depending on the APM parameter, we observed that 10.9-48.4% of stroke participants were impaired, while 44% were impaired based on their overall task score. The mean performance metrics of machine learning and deep learning models were: accuracy 82.4%, precision 85.6%, recall 76.5%, and F1 score 80.6%. All machine learning and deep learning models displayed similar classification accuracy; however, the Random Forest model had the highest numerical accuracy (83%). Our models showed higher sensitivity and specificity (AUC = 0.89) in classifying individual participants than the overall task score (AUC = 0.85) based on their performance in the APM task. We also found that variability was the most important feature in classifying performance in the APM task.

CONCLUSION

Our ML models displayed similar classification performance. ML models were able to integrate more kinematic information and relationships between variables into decision making and displayed better classification performance than the overall task score. ML may help to provide insight into individual kinematic features that have previously been overlooked with respect to clinical importance.

Novel Prediction Method Applied to Wound Age Estimation: Developing a Stacking Ensemble Model to Improve Predictive Performance Based on Multi-mRNA

(1) Background: Accurate diagnosis of wound age is crucial for investigating violent cases in forensic practice. However, effective biomarkers and forecast methods are lacking. (2) Methods: Samples were collected from rats divided randomly into control and contusion groups at 0, 4, 8, 12, 16, 20, and 24 h post-injury. The characteristics of concern were nine mRNA expression levels. Internal validation data were used to train different machine learning algorithms, namely random forest (RF), support vector machine (SVM), multilayer perceptron (MLP), gradient boosting (GB), and stochastic gradient descent (SGD), to predict wound age. These models were considered the base learners, which were then applied to developing 26 stacking ensemble models combining two, three, four, or five base learners. The best-performing stacking model and base learner were evaluated through external validation data. (3) Results: The best results were obtained using a stacking model of RF + SVM + MLP (accuracy = 92.85%, area under the receiver operating characteristic curve (AUROC) = 0.93, root-mean-square-error (RMSE) = 1.06 h). The wound age prediction performance of the stacking models was also confirmed for another independent dataset. (4) Conclusions: We illustrate that machine learning techniques, especially ensemble algorithms, have a high potential to be used to predict wound age. According to the results, the strategy can be applied to other types of forensic forecasts.