Development of a machine learning-based risk prediction model for cerebral infarction and comparison with nomogram model

Predicting cerebral infarction and transient ischemic attack in healthy individuals and those with dysmetabolism: a machine learning approach combined with routine blood tests

Ischemic cerebral infarction is the most prevalent type of stroke, causing significant disability and death worldwide. Transient ischemic attack (TIA) is a strong predictor of subsequent stroke. Individuals with dysmetabolism, such as hypertension, hypercholesterolemia, and diabetes, are at increased risk for cerebral infarction (CI) and TIA. In resource-limited settings, diagnosing CI and TIA can be particularly difficult due to a lack of advanced imaging and specialized expertise. Therefore, we aim to develop a simple, convenient, blood-based approach that could assist clinicians in diagnosing CI and TIA, especially in regions where advanced imaging or stroke-specific expertise is limited. All study subjects were patients admitted to the First Hospital of Xiamen University and healthy check-up populations between January 2018 and September 2023. This study employed five machine learning methods alongside 21 blood routine indicators, 30 blood biochemical indicators, age, and gender to construct predictive models for CI and TIA in both healthy individuals and those with dysmetabolism. The Area Under the Receiver Operating Characteristic (ROC) Curve (AUC) served as the metric to assess the comprehensive predictive capability of the models. Subsequently, the SHAP package was employed for model interpretation. Extreme Gradient Boosting (XGBoost) outperforms other models in all predictive models. In the models predicting CI and TIA among healthy, the AUC is 0.9958 (0.9947-0.9969) and 0.9928 (0.9899-0.9951), respectively. Among the nine shared key features of the two models are indicators of glucose metabolism, lipid metabolism, and liver metabolism. In the models for predicting CI and TIA among patients with hypertension, hypercholesterolemia, diabetes, and those with all three metabolic disorders combined, the AUCs ranged from 0.6990 to 0.8591. We found that the indicators K significantly contributed to predict CI and TIA from those with dysmetabolism. Additionally, metabolic-related indicators, such as glucose (GLU) and high-density lipoprotein cholesterol (HDL-C), are ranked highly among the top ten contributing features. XGBoost performed the best in all models. It can effectively differentiate CI and TIA from healthy and dysmetabolic patients by combining blood routine and blood biochemical indicators. Moreover, it can also differentiate CI from TIA. Although any suspicious findings from this model would still require confirmatory imaging, the simplicity and low cost of blood-based testing may offer a practical adjunct for clinicians-particularly in areas lacking advanced imaging or extensive stroke expertise-and could facilitate earlier diagnostic decision-making.

Machine learning to predict stroke risk from routine hospital data: A systematic review

PURPOSE

Stroke remains a leading cause of morbidity and mortality. Despite this, current risk stratification tools such as CHA2DS2-VASc and QRISK3 are of limited accuracy, particularly in those without a diagnosis of atrial-fibrillation. Hence, there is a need for more accurate stroke risk prediction models. Machine-learning (ML) may provide a solution to this by leveraging existing routine hospital databases to build accurate stroke risk prediction models and identify novel risk factors for stroke.

AIMS

In this systematic review we appraise current research using ML to predict stroke risk from routine hospital data. Based on these findings we then highlight common methodological limitations and recommendations for future research.

METHODS

In this review we identify 49 original research (38 in the general population and 11 in AF specific populations) articles from the PUBMED database from January-2013 to December-2024 using ML and routine hospital data to predict the risk of stroke.

RESULTS

ML models were able to accurately predict stroke risk in both AF specific and general populations, with AUCs ranging from 0.64 to 0.99. Where tested, ML also consistently outperformed traditional risk stratification tool, such as CHA2DS2-VASc. ML also appeared useful in identifying several novel risk factors from electrocardiogram, laboratory test and echocardiography data. However, the quality of datasets were often limited, there was a high suspicion of overfitting and models often lacked calibration, external validation and explainability analysis.

CONCLUSION

Whilst ML has shown great potential in stroke prediction and identifying novel risk factors for stroke, improvements in study methodology is required prior to integration of ML into routine healthcare. Future research should adhere to the EQUATOR guidance on prediction models and encourage interdisciplinary collaboration between computer scientists and clinicians. Further prospective RCTs are also required to validate models in the clinical setting and the identify barriers of integrating ML into routine healthcare.

Development and validation of a support vector machine-based nomogram for diagnosis of obstetric antiphospholipid syndrome

BACKGROUND

Antiphospholipid Syndrome (APS) is a systemic autoimmune disorder characterized by arterial or venous thrombosis and/or pregnancy complications. This study aims to develop a diagnostic model for Obstetric APS (OAPS) using the Support Vector Machine (SVM) algorithm.

METHODS

Data were retrospectively collected from 102 patients with OAPS and 80 healthy controls (HC). Utilizing random sampling, patients were randomly allocated into a training set and a validation set. The training set comprised 72 OAPS patients and 52 HCs, while the validation set included 30 OAPS patients and 24 HCs. Univariate logistic regression analysis and the LASSO method were employed to screen feature variables. Subsequently, the selected feature variables were used to construct a diagnostic model based on the SVM algorithm, which was then validated within the training set.

RESULTS

An optimal subset comprising 12 clinical features was curated. This ensemble of clinical features exhibited formidable predictive efficacy within both the training and validation datasets, as evidenced by Area Under the Curve (AUC) values of 0.969 and 0.942, sensitivities of 0.875 and 0.867, and specificities of 0.929 and 0.875, respectively. Furthermore, the nomogram generated a Concordance Index (C-index) of 0.851 across the entire dataset. Decision curve analysis demonstrates that the combined nomogram and TAT nomogram offer greater net benefit compared to nomograms based on other individual clinical indicators within the dataset.

CONCLUSION

The SVM-based model can effectively diagnose patients with OAPS.

Machine learning and explainable artificial intelligence to predict pathologic stage in men with localized prostate cancer

BACKGROUND

Though several nomograms exist, machine learning (ML) approaches might improve prediction of pathologic stage in patients with prostate cancer. To develop ML models to predict pathologic stage that outperform existing nomograms that use readily available clinicopathologic variables.

METHODS

Patients with prostate adenocarcinoma who underwent surgery were identified in the National Cancer Database. Seven ML models were trained to predict organ-confined (OC) disease, extracapsular extension, seminal vesicle invasion (SVI), and lymph node involvement (LNI). Model performance was measured using area under the curve (AUC) on a holdout testing data set. Clinical utility was evaluated using decision curve analysis (DCA). Performance metrics were confirmed on an external validation data set.

RESULTS

The ML-based extreme gradient boosted trees model achieved the best performance with an AUC of 0.744, 0.749, 0.816, 0.811 for the OC, ECE, SVI, and LNI models, respectively. The MSK nomograms achieved an AUC of 0.708, 0.742, 0.806, 0.802 for the OC, ECE, SVI, and LNI models, respectively. These models also performed the best on DCA. Findings were consistent on both a holdout internal validation data set as well as an external validation data set.

CONCLUSIONS

Our ML models better predicted pathologic stage relative to existing nomograms at predicting pathologic stage. Accurate prediction of pathologic stage can help oncologists and patients determine optimal definitive treatment options for patients with prostate cancer.

Development of machine learning-based models for predicting risk factors in acute cerebral infarction patients: a clinical retrospective study

OBJECTIVES

The aim of this study was to develop machine learning-based models for predicting acute cerebral infarction (ACI) in patients.

METHODS

We extracted the data of ACI patients and non-ACI patients (as control) from two hospitals. The Lasso algorithm was employed to select the most crucial features associated with ACI. Five machine learning algorithms-based models were trained, which was performed with 10-fold cross-validation. Then, the area under the receiver operating characteristic curve (AUC), accuracy, and F1-score were calculated in the training models. Accordingly, the training models with excellent performance was selected as the final predictive model. The relative importance of variables was analyzed and ranked.

RESULTS

A total of 150 patients were diagnosed with ACI (50.00%), with a higher proportion of males (70.67% vs. 44.00%) compared to the non-ACI patients. The logistic regression model exhibited a good performance in predicting ACI in the training set, as evidenced by its highest AUC, accuracy, sensitivity, and F1-score. Furthermore, feature importance analysis showed that blood glucose, gender, smoking history, serum homocysteine, folic acid, and C-reactive protein were the top six crucial variables of the logistic regression.

CONCLUSIONS

In our work, the ACI risk prediction model developed by the logistic regression exhibited excellent performance. This could contribute to the identification of risk variables for ACI patients and enables clinicians timely and effective interventions.

Application of interpretable machine learning algorithms to predict acute kidney injury in patients with cerebral infarction in ICU

BACKGROUND

Acute kidney injury (AKI) is not only a complication but also a serious threat to patients with cerebral infarction (CI). This study aimed to explore the application of interpretable machine learning algorithms in predicting AKI in patients with cerebral infarction.

METHODS

The study included 3920 patients with CI admitted to the Intensive Care Unit and Emergency Medicine of the Central Hospital of Lishui City, Zhejiang Province. Nine machine learning techniques, including XGBoost, logistics, LightGBM, random forest (RF), AdaBoost, GaussianNB (GNB), Multi-Layer Perceptron (MLP), support vector machine (SVM), and k-nearest neighbors (KNN) classification, were used to develop a predictive model for AKI in these patients. SHapley Additive exPlanations (SHAP) analysis provided visual explanations for each patient. Finally, model effectiveness was assessed using metrics such as average precision (AP), sensitivity, specificity, accuracy, F1 score, precision-recall (PR) curve, calibration plot, and decision curve analysis (DCA).

RESULTS

The XGBoost model performed better in the internal validation set and the external validation set, with an AUC of 0.940 and 0.887, respectively. The five most important variables in the model were, in order, glomerular filtration rate, low-density lipoprotein, total cholesterol, hemiplegia and serum kalium.

CONCLUSION

This study demonstrates the potential of interpretable machine learning algorithms in predicting CI patients with AKI.

Development and comparison of machine learning-based models for predicting heart failure after acute myocardial infarction

AIMS

Heart failure (HF) is one of the common adverse cardiovascular events after acute myocardial infarction (AMI), but the predictive efficacy of numerous machine learning (ML) built models is unclear. This study aimed to build an optimal model to predict the occurrence of HF in AMI patients by comparing seven ML algorithms.

METHODS

Cohort 1 included AMI patients from 2018 to 2019 divided into HF and control groups. All first routine test data of the study subjects were collected as the features to be selected for the model, and seven ML algorithms with screenable features were evaluated. Cohort 2 contains AMI patients from 2020 to 2021 to establish an early warning model with external validation. ROC curve and DCA curve to analyze the diagnostic efficacy and clinical benefit of the model respectively.

RESULTS

The best performer among the seven ML algorithms was XgBoost, and the features of XgBoost algorithm for troponin I, triglycerides, urine red blood cell count, γ-glutamyl transpeptidase, glucose, urine specific gravity, prothrombin time, prealbumin, and urea were ranked high in importance. The AUC of the HF-Lab9 prediction model built by the XgBoost algorithm was 0.966 and had good clinical benefits.

CONCLUSIONS

This study screened the optimal ML algorithm as XgBoost and developed the model HF-Lab9 will improve the accuracy of clinicians in assessing the occurrence of HF after AMI and provide a reference for the selection of subsequent model-building algorithms.

Validation and comparison of the stroke prognosis instrument (SPI-II) and the essen stroke risk score (ESRS) in predicting stroke recurrence in Asian population

BACKGROUND

Currently, there are limited data on the accuracy of available risk scores to predict stroke recurrence in the Asian population.

METHOD

A single-center, retrospective cohort study was conducted among patients with acute ischemic stroke during January 2014 - December 2018. Longitudinal data with three years of follow-up among these patients were collected and validated through both electronic and manual chart review. The area under the receiver-operating curve (AUROC) method or C-statistic and calibration plot were used to evaluate and compare the Stroke Prognosis Instrument II (SPI-II) and the Essen Stroke Risk Score (ESRS) in predicting the long-term risk of recurrent ischemic stroke. The predictive performances of the two scores were compared using DeLong's method.

RESULTS

The study cohort consisted of 543 patients, including 181 and 362 patients with and without recurrent events. There were no significant differences in mean age and gender between the two groups. Recurrence cases tended to have significant more risk factors compared to those without events. Among cases with recurrent events, 134 (74.03%) and 65.74% (119) cases were classified as high-risk based on SPI-II and ESRS, respectively. The AUROC curve of the SPI-II and ESRS score was 0.646 (95% CI, 0.594-0.697) and 0.614 (95%CI, 0.563-0.665), respectively (p = 0.394). Based on the calibration plot, the SPI-II and ESRS scores showed similar moderate predictive performance on recurrence stroke with a C statistic (95% CI) of 0.655 (95% CI: 0.603-0.707) and 0.631 (95% CI 0.579-0.684), respectively.

CONCLUSION

Both ESRS and SPI-II scores had moderate predictive performance in Thai population.

Machine learning predicts the risk of hemorrhagic transformation of acute cerebral infarction and in-hospital death

BACKGROUND

The incidence of hemorrhagic transformation (HT) during thrombolysis after acute cerebral infarction (ACI) is very high. We aimed to develop a model to predict the occurrence of HT after ACI and the risk of death after HT.

METHODS

Cohort 1 is divided into HT and non-HT groups, to train the model and perform internal validation. All first laboratory test results of study subjects were used as features to be selected for machine learning, and the models built by four machine learning algorithms were compared to screen the best algorithm and model. Following that, the HT group was divided into death and non-death for subgroup analysis. Receiver operating characteristic (ROC) curves etc. to evaluate the model. ACI patients in cohort 2 for external validation.

RESULTS

In cohort 1, the HT risk prediction model HT-Lab10 built by the XgBoost algorithm performed the best with AUCROC=0.95 (95% CI, 0.93-0.96). Ten features were included in the model, namely B-type natriuretic peptide precursor, ultrasensitive C-reactive protein, glucose, absolute neutrophil value, myoglobin, uric acid, creatinine, Ca2+, Thrombin time, and carbon dioxide combining power. The model also had the ability to predict death after HT with AUCROC=0.85 (95% CI, 0.78-0.91). The ability of HT-Lab10 to predict the occurrence of HT as well as death after HT was validated in cohort 2.

CONCLUSIONS

The model HT-Lab10 built using the XgBoost algorithm showed excellent predictive ability in both the occurrence of HT and the risk of HT death, achieving a model with multiple uses.

AutoScore-Ordinal: an interpretable machine learning framework for generating scoring models for ordinal outcomes

Background

Risk prediction models are useful tools in clinical decision-making which help with risk stratification and resource allocations and may lead to a better health care for patients. AutoScore is a machine learning–based automatic clinical score generator for binary outcomes. This study aims to expand the AutoScore framework to provide a tool for interpretable risk prediction for ordinal outcomes.

Methods

The AutoScore-Ordinal framework is generated using the same 6 modules of the original AutoScore algorithm including variable ranking, variable transformation, score derivation (from proportional odds models), model selection, score fine-tuning, and model evaluation. To illustrate the AutoScore-Ordinal performance, the method was conducted on electronic health records data from the emergency department at Singapore General Hospital over 2008 to 2017. The model was trained on 70% of the data, validated on 10% and tested on the remaining 20%.

Results

This study included 445,989 inpatient cases, where the distribution of the ordinal outcome was 80.7% alive without 30-day readmission, 12.5% alive with 30-day readmission, and 6.8% died inpatient or by day 30 post discharge. Two point-based risk prediction models were developed using two sets of 8 predictor variables identified by the flexible variable selection procedure. The two models indicated reasonably good performance measured by mean area under the receiver operating characteristic curve (0.758 and 0.793) and generalized c-index (0.737 and 0.760), which were comparable to alternative models.

Conclusion

AutoScore-Ordinal provides an automated and easy-to-use framework for development and validation of risk prediction models for ordinal outcomes, which can systematically identify potential predictors from high-dimensional data.