Machine Learning to Predict the Likelihood of Acute Myocardial Infarction

Risk stratification of patients who present with chest pain and have normal troponins using a machine learning model

BACKGROUND

Risk stratification tools exist for patients presenting with chest pain to the emergency room and have achieved the recommended negative predictive value (NPV) of 99%. However, due to low positive predictive value (PPV), current stratification tools result in unwarranted investigations such as serial laboratory tests and cardiac stress tests (CSTs).

AIM

To create a machine learning model (MLM) for risk stratification of chest pain with a better PPV.

METHODS

This retrospective cohort study used de-identified hospital data from January 2016 until November 2021. Inclusion criteria were patients aged > 21 years who presented to the ER, had at least two serum troponins measured, were subsequently admitted to the hospital, and had a CST within 4 d of presentation. Exclusion criteria were elevated troponin value (> 0.05 ng/mL) and missing values for body mass index. The primary outcome was abnormal CST. Demographics, coronary artery disease (CAD) history, hypertension, hyperlipidemia, diabetes mellitus, chronic kidney disease, obesity, and smoking were evaluated as potential risk factors for abnormal CST. Patients were also categorized into a high-risk group (CAD history or more than two risk factors) and a low-risk group (all other patients) for comparison. Bivariate analysis was performed using a χ² test or Fisher’s exact test. Age was compared by t test. Binomial regression (BR), random forest, and XGBoost MLMs were used for prediction. Bootstrapping was used for the internal validation of prediction models. BR was also used for inference. Alpha criterion was set at 0.05 for all statistical tests. R software was used for statistical analysis.

RESULTS

The final cohort of the study included 2328 patients, of which 245 (10.52%) patients had abnormal CST. When adjusted for covariates in the BR model, male sex [risk ratio (RR) = 1.52, 95% confidence interval (CI): 1.2-1.94, P < 0.001)], CAD history (RR = 4.46, 95%CI: 3.08-6.72, P < 0.001), and hyperlipidemia (RR = 3.87, 95%CI: 2.12-8.12, P < 0.001) remained statistically significant. Incidence of abnormal CST was 12.2% in the high-risk group and 2.3% in the low-risk group (RR = 5.31, 95%CI: 2.75-10.24, P < 0.001). The XGBoost model had the best PPV of 24.33%, with an NPV of 91.34% for abnormal CST.

CONCLUSION

The XGBoost MLM achieved a PPV of 24.33% for an abnormal CST, which is better than current stratification tools (13.00%-17.50%). This highlights the beneficial potential of MLMs in clinical decision-making.

Risk Stratification of COVID-19 Using Routine Laboratory Tests: A Machine Learning Approach

The COVID-19 pandemic placed significant stress on an already overburdened health system. The diagnosis was based on detection of a positive RT-PCR test, which may be delayed when there is peak demand for testing. Rapid risk stratification of high-risk patients allows for the prioritization of resources for patient care. The study aims were to classify patients as severe or not severe based on outcomes using machine learning on routine laboratory tests. Data were extracted for all individuals who had at least one SARS-CoV-2 PCR test conducted via the NHLS between the periods of 1 March 2020 to 7 July 2020. Exclusion criteria: those 18 years, and those with indeterminate PCR tests. Results for 15437 patients (3301 positive and 12,136 negative) were used to fit six machine learning models, namely the logistic regression (LR) (the base model), decision trees (DT), random forest (RF), extreme gradient boosting (XGB), convolutional neural network (CNN) and self-normalising neural network (SNN). Model development was carried out by splitting the data into training and testing set of a ratio 70:30, together with a 10-fold cross-validation re-sampling technique. For risk stratification, admission to high care or ICU was the outcome for severe disease. Performance of the models varied: sensitivity was best for RF at 75% and accuracy of 75% for CNN. The area under the curve ranged from 57% for CNN to 75% for RF. RF and SNN were the best-performing models. Machine Learning (ML) can be incorporated into the laboratory information system and offers promise for early identification and risk stratification of COVID-19 patients, particularly in areas of resource-poor settings.

A Predictive Machine Learning Tool for Asthma Exacerbations: Results from a 12-Week, Open-Label Study Using an Electronic Multi-Dose Dry Powder Inhaler with Integrated Sensors

PURPOSE

Machine learning models informed by sensor data inputs have the potential to provide individualized predictions of asthma deterioration. This study aimed to determine if data from an integrated digital inhaler could be used to develop a machine learning model capable of predicting impending exacerbations.

PATIENTS AND METHODS

Adult patients with poorly controlled asthma were enrolled in a 12-week, open-label study using ProAir® Digihaler®, an electronic multi-dose dry powder inhaler (eMDPI) with integrated sensors, as reliever medication (albuterol, 90 µg/dose; 1-2 inhalations every 4 hours, as needed). Throughout the study, the eMDPI recorded inhaler use, peak inspiratory flow (PIF), inhalation volume, inhalation duration, and time to PIF. A model predictive of impending exacerbations was generated by applying machine learning techniques to data downloaded from the inhalers, together with clinical and demographic information. The generated model was evaluated by receiver operating characteristic area under curve (ROC AUC) analysis.

RESULTS

Of 360 patients included in the predictive analysis, 64 experienced a total of 78 exacerbations. Increased albuterol use preceded exacerbations; the mean number of inhalations in the 24-hours preceding an exacerbation was 7.3 (standard deviation 17.3). The machine learning model, using gradient-boosting trees with data from the eMDPI and baseline patient characteristics, predicted an impending exacerbation over the following 5 days with an ROC AUC of 0.83 (95% confidence interval: 0.77-0.90). The feature of the model with the highest weight was the mean number of daily inhalations during the 4 days prior to the day the prediction was made.

CONCLUSION

A machine learning model to predict impending asthma exacerbations using data from the eMDPI was successfully developed. This approach may support a shift from reactive care to proactive, preventative, and personalized management of chronic respiratory diseases.

Influence of Age on the Diagnosis of Myocardial Infarction

BACKGROUND

The 99th centile of cardiac troponin, derived from a healthy reference population, is recommended as the diagnostic threshold for myocardial infarction, but troponin concentrations are strongly influenced by age. Our aim was to assess the diagnostic performance of cardiac troponin in older patients presenting with suspected myocardial infarction.

METHODS

In a secondary analysis of a multicenter trial of consecutive patients with suspected myocardial infarction, we assessed the diagnostic accuracy of high-sensitivity cardiac troponin I at presentation for the diagnosis of type 1, type 2, or type 4b myocardial infarction across 3 age groups (<50, 50-74, and ≥75 years) using guideline-recommended sex-specific and age-adjusted 99th centile thresholds.

RESULTS

In 46 435 consecutive patients aged 18 to 108 years (mean, 61±17 years), 5216 (11%) had a diagnosis of myocardial infarction. In patients <50 (n=12 379), 50 to 74 (n=22 380), and ≥75 (n=11 676) years, the sensitivity of the guideline-recommended threshold was similar at 79.2% (95% CI, 75.5-82.9), 80.6% (95% CI, 79.2-82.1), and 81.6% (95% CI, 79.8-83.2), respectively. The specificity decreased with advancing age from 98.3% (95% CI, 98.1-98.5) to 95.5% (95% CI, 95.2-95.8), and 82.6% (95% CI, 81.9-83.4). The use of age-adjusted 99th centile thresholds improved the specificity (91.3% [90.8%-91.9%] versus 82.6% [95% CI, 81.9%-83.4%]) and positive predictive value (59.3% [57.0%-61.5%] versus 51.5% [49.9%-53.3%]) for myocardial infarction in patients ≥75 years but failed to prevent the decrease in either parameter with increasing age and resulted in a marked reduction in sensitivity compared with the use of the guideline-recommended threshold (55.9% [53.6%-57.9%] versus 81.6% [79.8%-83.3%].

CONCLUSIONS

Age alters the diagnostic performance of cardiac troponin, with reduced specificity and positive predictive value in older patients when applying the guideline-recommended or age-adjusted 99th centiles. Individualized diagnostic approaches rather than the adjustment of binary thresholds are needed in an aging population.

Support vector machine deep mining of electronic medical records to predict the prognosis of severe acute myocardial infarction

Cardiovascular disease is currently one of the most important diseases causing death in China and the world, and acute myocardial infarction is a major cause of cardiovascular disease. This study provides an analytical technique for predicting the prognosis of patients with severe acute myocardial infarction using a support vector machine (SVM) technique based on information gleaned from electronic medical records in the Medical Information Marketplace for Intensive Care (MIMIC)-III database. The MIMIC-III database provided 4785 electronic medical records data for inclusion in the model development after screening 7070 electronic medical records of patients admitted to the intensive care unit for treatment of acute myocardial infarction. Adopting the APS-III score as the criterion for identifying anticipated risk, the dimensions of data information incorporated into the mathematical model design were found using correlation coefficient matrix heatmaps and ordered logistic analysis. An automated prognostic risk-prediction model was developed using SVM, and the fit was evaluated by 5× cross-validation. We used a grid search method to further optimize the parameters and improve the model fit. The excellent generalization ability of SVM was fully verified by calculating the 95% confidence interval of the area under the receiver operating characteristic curve (AUC) for six algorithms (linear discriminant, tree, Kernel Naive Bayes, RUSBoost, KNN, and SVM). Compared to the remaining five models, its confidence interval was the narrowest with higher fitting accuracy and better performance. The patient prognostic risk prediction model constructed using SVM had a relatively impressive accuracy (92.2%) and AUC value (0.98). In this study, a model was designed for fitting that can maximize the potential information to be gleaned in the electronic medical records data. It was demonstrated that SVM models based on electronic medical records data can offer an effective solution for clinical disease prognostic risk assessment and improved clinical outcomes and have great potential for clinical application in the clinical treatment of myocardial infarction.

Predicting acute kidney injury risk in acute myocardial infarction patients: An artificial intelligence model using medical information mart for intensive care databases

Background

Predictive models based on machine learning have been widely used in clinical practice. Patients with acute myocardial infarction (AMI) are prone to the risk of acute kidney injury (AKI), which results in a poor prognosis for the patient. The aim of this study was to develop a machine learning predictive model for the identification of AKI in AMI patients.

Methods

Patients with AMI who had been registered in the Medical Information Mart for Intensive Care (MIMIC) III and IV database were enrolled. The primary outcome was the occurrence of AKI during hospitalization. We developed Random Forests (RF) model, Naive Bayes (NB) model, Support Vector Machine (SVM) model, eXtreme Gradient Boosting (xGBoost) model, Decision Trees (DT) model, and Logistic Regression (LR) models with AMI patients in MIMIC-IV database. The importance ranking of all variables was obtained by the SHapley Additive exPlanations (SHAP) method. AMI patients in MIMIC-III databases were used for model evaluation. The area under the receiver operating characteristic curve (AUC) was used to compare the performance of each model.

Results

A total of 3,882 subjects with AMI were enrolled through screening of the MIMIC database, of which 1,098 patients (28.2%) developed AKI. We randomly assigned 70% of the patients in the MIMIC-IV data to the training cohort, which is used to develop models in the training cohort. The remaining 30% is allocated to the testing cohort. Meanwhile, MIMIC-III patient data performs the external validation function of the model. 3,882 patients and 37 predictors were included in the analysis for model construction. The top 5 predictors were serum creatinine, activated partial prothrombin time, blood glucose concentration, platelets, and atrial fibrillation, (SHAP values are 0.670, 0.444, 0.398, 0.389, and 0.381, respectively). In the testing cohort, using top 20 important features, the models of RF, NB, SVM, xGBoost, DT model, and LR obtained AUC of 0.733, 0.739, 0.687, 0.689, 0.663, and 0.677, respectively. Placing RF models of number of different variables on the external validation cohort yielded their AUC of 0.711, 0.754, 0.778, 0.781, and 0.777, respectively.

Conclusion

Machine learning algorithms, particularly the random forest algorithm, have improved the accuracy of risk stratification for AKI in AMI patients and are applied to accurately identify the risk of AKI in AMI patients.

Effect of Enalapril Combined with Bisoprolol on Cardiac Function and Inflammatory Indexes in Patients with Acute Myocardial Infarction

Objective

The use of enalapril in combination with bisoprolol in patients with acute myocardial infarction (AMI) was studied for its effect on cardiac function and inflammatory parameters.

Methods

Sixty-two cases of AMI patients admitted to our clinic from November 2019 to November 2021 were selected for the study and grouped according to the random number table method, those enrolled were given conventional treatment such as oxygenation, absolute bed rest, and sedation, and administered low molecular heparin, aspirin, atorvastatin calcium tablets, clopidogrel, and nitrates. The control group (31 cases) was treated with enalapril maleate folic acid tablets, and the treatment group (31 cases) was treated with bisoprolol fumarate tablets on top of the control group, and the efficacy, adverse effects, cardiac function, inflammatory indexes, and oxidative stress indexes of the two arms were contrasted.

Results

The incidence of adverse reactions in the therapy cohort was 12.90% higher than that in the controlled arm, but the discrepancy was not medically relevant (P < 0.05). The SOD level was larger than the concentration in the corresponding drug therapy group, and the MDA level was lower than the concentration in the respective test cases (P < 0.05); the incidence of 12.90% adverse reactions in the treatment period was lower than that of 16.13% in the specific drug therapy group, but the variance was not scientifically evident (P > 0.05).

Conclusion

Enalapril application combined with bisoprolol in AMI patients is beneficial to boost the efficacy, promote the improvement of cardiac function, reduce the inflammatory response, and improve the oxidative stress with fewer adverse effects, which can ensure the therapeutic security.

Application of artificial intelligence techniques for automated detection of myocardial infarction: a review

Myocardial infarction (MI) results in heart muscle injury due to receiving insufficient blood flow. MI is the most common cause of mortality in middle-aged and elderly individuals worldwide. To diagnose MI, clinicians need to interpret electrocardiography (ECG) signals, which requires expertise and is subject to observer bias. Artificial intelligence-based methods can be utilized to screen for or diagnose MI automatically using ECG signals. In this work, we conducted a comprehensive assessment of artificial intelligence-based approaches for MI detection based on ECG and some other biophysical signals, including machine learning (ML) and deep learning (DL) models. The performance of traditional ML methods relies on handcrafted features and manual selection of ECG signals, whereas DL models can automate these tasks. The review observed that deep convolutional neural networks (DCNNs) yielded excellent classification performance for MI diagnosis, which explains why they have become prevalent in recent years. To our knowledge, this is the first comprehensive survey of artificial intelligence techniques employed for MI diagnosis using ECG and some other biophysical signals.

An Explainable AI Approach for the Rapid Diagnosis of COVID-19 Using Ensemble Learning Algorithms

Background

Artificial intelligence-based disease prediction models have a greater potential to screen COVID-19 patients than conventional methods. However, their application has been restricted because of their underlying black-box nature.

Objective

To addressed this issue, an explainable artificial intelligence (XAI) approach was developed to screen patients for COVID-19.

Methods

A retrospective study consisting of 1,737 participants (759 COVID-19 patients and 978 controls) admitted to San Raphael Hospital (OSR) from February to May 2020 was used to construct a diagnosis model. Finally, 32 key blood test indices from 1,374 participants were used for screening patients for COVID-19. Four ensemble learning algorithms were used: random forest (RF), adaptive boosting (AdaBoost), gradient boosting decision tree (GBDT), and extreme gradient boosting (XGBoost). Feature importance from the perspective of the clinical domain and visualized interpretations were illustrated by using local interpretable model-agnostic explanations (LIME) plots.

Results

The GBDT model [area under the curve (AUC): 86.4%; 95% confidence interval (CI) 0.821–0.907] outperformed the RF model (AUC: 85.7%; 95% CI 0.813–0.902), AdaBoost model (AUC: 85.4%; 95% CI 0.810–0.899), and XGBoost model (AUC: 84.9%; 95% CI 0.803–0.894) in distinguishing patients with COVID-19 from those without. The cumulative feature importance of lactate dehydrogenase, white blood cells, and eosinophil counts was 0.145, 0.130, and 0.128, respectively.

Conclusions

Ensemble machining learning (ML) approaches, mainly GBDT and LIME plots, are efficient for screening patients with COVID-19 and might serve as a potential tool in the auxiliary diagnosis of COVID-19. Patients with higher WBC count, higher LDH level, or higher EOT count, were more likely to have COVID-19.

Development and validation of a decision support tool for the diagnosis of acute heart failure: systematic review, meta-analysis, and modelling study

OBJECTIVES

To evaluate the diagnostic performance of N-terminal pro-B-type natriuretic peptide (NT-proBNP) thresholds for acute heart failure and to develop and validate a decision support tool that combines NT-proBNP concentrations with clinical characteristics.

DESIGN

Individual patient level data meta-analysis and modelling study.

SETTING

Fourteen studies from 13 countries, including randomised controlled trials and prospective observational studies.

PARTICIPANTS

Individual patient level data for 10 369 patients with suspected acute heart failure were pooled for the meta-analysis to evaluate NT-proBNP thresholds. A decision support tool (Collaboration for the Diagnosis and Evaluation of Heart Failure (CoDE-HF)) that combines NT-proBNP with clinical variables to report the probability of acute heart failure for an individual patient was developed and validated.

MAIN OUTCOME MEASURE

Adjudicated diagnosis of acute heart failure.

RESULTS

Overall, 43.9% (4549/10 369) of patients had an adjudicated diagnosis of acute heart failure (73.3% (2286/3119) and 29.0% (1802/6208) in those with and without previous heart failure, respectively). The negative predictive value of the guideline recommended rule-out threshold of 300 pg/mL was 94.6% (95% confidence interval 91.9% to 96.4%); despite use of age specific rule-in thresholds, the positive predictive value varied at 61.0% (55.3% to 66.4%), 73.5% (62.3% to 82.3%), and 80.2% (70.9% to 87.1%), in patients aged <50 years, 50-75 years, and >75 years, respectively. Performance varied in most subgroups, particularly patients with obesity, renal impairment, or previous heart failure. CoDE-HF was well calibrated, with excellent discrimination in patients with and without previous heart failure (area under the receiver operator curve 0.846 (0.830 to 0.862) and 0.925 (0.919 to 0.932) and Brier scores of 0.130 and 0.099, respectively). In patients without previous heart failure, the diagnostic performance was consistent across all subgroups, with 40.3% (2502/6208) identified at low probability (negative predictive value of 98.6%, 97.8% to 99.1%) and 28.0% (1737/6208) at high probability (positive predictive value of 75.0%, 65.7% to 82.5%) of having acute heart failure.

CONCLUSIONS

In an international, collaborative evaluation of the diagnostic performance of NT-proBNP, guideline recommended thresholds to diagnose acute heart failure varied substantially in important patient subgroups. The CoDE-HF decision support tool incorporating NT-proBNP as a continuous measure and other clinical variables provides a more consistent, accurate, and individualised approach.

STUDY REGISTRATION

PROSPERO CRD42019159407.

Utilizing longitudinal data in assessing all-cause mortality in patients hospitalized with heart failure

AIMS

Risk stratification in patients with a new onset or worsened heart failure (HF) is essential for clinical decision making. We have utilized a novel approach to enrich patient level prognostication using longitudinally gathered data to develop ML-based algorithms predicting all-cause 30, 90, 180, 360, and 720 day mortality.

METHODS AND RESULTS

In a cohort of 2449 HF patients hospitalized between 1 January 2011 and 31 December 2017, we utilized 422 parameters derived from 151 451 patient exams. They included clinical phenotyping, ECG, laboratory, echocardiography, catheterization data or percutaneous and surgical interventions reflecting the standard of care as captured in individual electronic records. The development of predictive models consisted of 101 iterations of repeated random subsampling splits into balanced training and validation sets. ML models yielded area under the receiver operating characteristic curve (AUC-ROC) performance ranging from 0.83 to 0.89 on the outcome-balanced validation set in predicting all-cause mortality at aforementioned time-limits. The 1 year mortality prediction model recorded an AUC of 0.85. We observed stable model performance across all HF phenotypes: HFpEF 0.83 AUC, HFmrEF 0.85 AUC, and HFrEF 0.86 AUC, respectively. Model performance improved when utilizing data from more hospital contacts compared with only data collected at baseline.

CONCLUSIONS

Our findings present a novel, patient-level, comprehensive ML-based algorithm for predicting all-cause mortality in new or worsened heart failure. Its robust performance across phenotypes throughout the longitudinal patient follow-up suggests its potential in point-of-care clinical risk stratification.

Serial cardiac biomarkers for risk stratification of patients with COVID-19

OBJECTIVES

Several studies have demonstrated an association between elevated cardiac biomarkers and adverse outcomes in patients with COVID-19. However, the prognostic and predictive capability of a multimarker panel in a prospectively collected, diverse "all-comers" COVID-19 population has not been fully elucidated.

DESIGN & METHODS

We prospectively assessed high sensitivity cardiac troponin I (hsTnI), NT-pro B-type Natriuretic Peptide (NT-proBNP), Galectin-3 (Gal-3), and procalcitonin (PCT) in 4,282 serial samples from 358 patients admitted with symptomatic, RT-PCR confirmed SARS-CoV-2 infection. Outcomes examined were 30-day in-hospital mortality and requirement for intubation within 10 days.

RESULTS

Baseline hsTnI had the highest AUC for predicting 30-day mortality (0.81; 95% CI, 0.73-0.88), followed by NT-proBNP (0.80; 0.74-0.86), PCT (0.77; 0.70-0.84), and Gal-3 (0.68; 0.60-0.76). HsTnI < 3.5 ng/L at baseline identified patients at low risk for in-hospital mortality (NPV 95.9%, sensitivity 97.3%) and 10-day intubation (NPV 90.4%, sensitivity 88.5%). Continuous, log-2 increases in troponin concentration were associated with reduced survival (p < 0.001) on Kaplan-Meier curves and increased risk of 30-day mortality: HR 1.26 (1.16-1.37) in univariate and 1.19 (1.03-1.4) in multivariate models. Time-varying doubling of concentrations of hsTnI and Gal-3 were associated with increased risk of 30-day mortality (adjusted HR 1.21, 1.06-1.4, and 1.92, 1.40-2.6).

CONCLUSION

HsTnI, NT-proBNP, Gal-3, and PCT are elevated at baseline in patients that have worse outcomes from COVID-19. HsTnI was the only independent predictor of 30-day mortality and intubation. Time-varying, doubling in hsTnI and Gal-3 further aided in prognostication of adverse outcomes. These results support the use of hsTnI for triaging patients with COVID-19.

Machine Learning to Assess for Acute Myocardial Infarction Within 30 Minutes

Variations in high-sensitivity cardiac troponin I by age and sex along with various sampling times can make the evaluation for acute myocardial infarction (AMI) challenging. Machine learning integrates these variables to allow a more accurate evaluation for possible AMI. The goal was to test the diagnostic and prognostic utility of a machine learning algorithm in the evaluation of possible AMI. We applied a machine learning algorithm (myocardial-ischemic-injury-index [MI3]) that incorporates age, sex, and high-sensitivity cardiac troponin I levels at time 0 and 30 minutes in 529 patients evaluated for possible AMI in a single urban emergency department. MI3 generates an index value from 0 to 100 reflecting the likelihood of AMI. Patients were followed at 30-45 days for major adverse cardiac events (MACEs). There were 42 (7.9%) patients that had an AMI. Patients were divided into 3 groups by the MI3 score: low-risk (≤ 3.13), intermediate-risk (> 3.13-51.0), and high-risk (> 51.0). The sensitivity for AMI was 100% with a MI3 value ≤ 3.13 and 353 (67%) ruled-out for AMI at 30 minutes. At 30-45 days, there were 2 (0.6%) MACEs (2 noncardiac deaths) in the low-risk group, in the intermediate-risk group 4 (3.0%) MACEs (3 AMIs, 1 cardiac death), and in the high-risk group 4 (9.1%) MACEs (4 AMIs, 2 cardiac deaths). The MI3 algorithm had 100% sensitivity for AMI at 30 minutes and identified a low-risk cohort who may be considered for early discharge.

Construction of Novel Gene Signature-Based Predictive Model for the Diagnosis of Acute Myocardial Infarction by Combining Random Forest With Artificial Neural Network

Background

Acute myocardial infarction (AMI) is one of the most common causes of mortality around the world. Early diagnosis of AMI contributes to improving prognosis. In our study, we aimed to construct a novel predictive model for the diagnosis of AMI using an artificial neural network (ANN), and we verified its diagnostic value via constructing the receiver operating characteristic (ROC).

Methods

We downloaded three publicly available datasets (training sets GSE48060, GSE60993, and GSE66360) from Gene Expression Omnibus (GEO) database, and differentially expressed genes (DEGs) were identified between 87 AMI and 78 control samples. We applied the random forest (RF) and ANN algorithms to further identify novel gene signatures and construct a model to predict the possibility of AMI. Besides, the diagnostic value of our model was further validated in the validation sets GSE61144 (7 AMI patients and 10 controls), GSE34198 (49 AMI patients and 48 controls), and GSE97320 (3 AMI patients and 3 controls).

Results

A total of 71 DEGs were identified, of which 68 were upregulated and 3 were downregulated. Firstly, 11 key genes in 71 DEGs were screened with RF classifier for the classification of AMI and control samples. Then, we calculated the weight of each key gene using ANN. Furthermore, the diagnostic model was constructed and named neuralAMI, with significant predictive power (area under the curve [AUC] = 0.980). Finally, our model was validated with the independent datasets GSE61144 (AUC = 0.900), GSE34198 (AUC = 0.882), and GSE97320 (AUC = 1.00).

Conclusion

Machine learning was used to develop a reliable predictive model for the diagnosis of AMI. The results of our study provide potential gene biomarkers for early disease screening.

Validation of the myocardial-ischaemic-injury-index machine learning algorithm to guide the diagnosis of myocardial infarction in a heterogenous population: a prespecified exploratory analysis

BACKGROUND

Diagnostic pathways for myocardial infarction rely on fixed troponin thresholds, which do not recognise that troponin varies by age, sex, and time within individuals. To overcome this limitation, we recently introduced a machine learning algorithm that predicts the likelihood of myocardial infarction. Our aim was to evaluate whether this algorithm performs well in routine clinical practice and predicts subsequent events.

METHODS

The myocardial-ischaemic-injury-index (MI3) algorithm was validated in a prespecified exploratory analysis using data from a multi-centre randomised trial done in Scotland, UK that included consecutive patients with suspected acute coronary syndrome undergoing serial high-sensitivity cardiac troponin I measurement. Patients with ST-segment elevation myocardial infarction were excluded. MI3 incorporates age, sex, and two troponin measurements to compute a value (0-100) reflecting an individual's likelihood of myocardial infarction during the index visit and estimates diagnostic performance metrics (including area under the receiver-operating-characteristic curve, and the sensitivity, specificity, negative predictive value, and positive predictive value) at the computed score. Model performance for an index diagnosis of myocardial infarction (type 1 or type 4b), and for subsequent myocardial infarction or cardiovascular death at 1 year was determined using the previously defined low-probability threshold (1·6) and high-probability MI3 threshold (49·7). The trial is registered with ClinicalTrials.gov, NCT01852123.

FINDINGS

In total, 20 761 patients (64 years [SD 16], 9597 [46%] women) enrolled between June 10, 2013, and March 3, 2016, were included from the High-STEACS trial cohort, of whom 3272 (15·8%) had myocardial infarction. MI3 had an area under the receiver-operating-characteristic curve of 0·949 (95% CI 0·946-0·952) identifying 12 983 (62·5%) patients as low-probability for myocardial infarction at the pre-specified threshold (MI3 score <1·6; sensitivity 99·3% [95% CI 99·0-99·6], negative predictive value 99·8% [99·8-99·9]), and 2961 (14·3%) as high-probability at the pre-specified threshold (MI3 score ≥49·7; specificity 95·0% [94·6-95·3], positive predictive value 70·4% [68·7-72·0]). At 1 year, subsequent myocardial infarction or cardiovascular death occurred more often in high-probability patients than low-probability patients (520 [17·6%] of 2961 vs 197 [1·5%] of 12 983], p<0·0001).

INTERPRETATION

In consecutive patients undergoing serial cardiac troponin measurement for suspected acute coronary syndrome, the MI3 algorithm accurately estimated the likelihood of myocardial infarction and predicted subsequent adverse cardiovascular events. By providing individual probabilities the MI3 algorithm could improve the diagnosis and assessment of risk in patients with suspected acute coronary syndrome.

FUNDING

Medical Research Council, British Heart Foundation, National Institute for Health Research, and NHSX.

Machine learning models for prediction of adverse events after percutaneous coronary intervention

An accurate prediction of major adverse events after percutaneous coronary intervention (PCI) improves clinical decisions and specific interventions. To determine whether machine learning (ML) techniques predict peri-PCI adverse events [acute kidney injury (AKI), bleeding, and in-hospital mortality] with better discrimination or calibration than the National Cardiovascular Data Registry (NCDR-CathPCI) risk scores, we developed logistic regression and gradient descent boosting (XGBoost) models for each outcome using data from a prospective, all-comer, multicenter registry that enrolled consecutive coronary artery disease patients undergoing PCI in Japan between 2008 and 2020. The NCDR-CathPCI risk scores demonstrated good discrimination for each outcome (C-statistics of 0.82, 0.76, and 0.95 for AKI, bleeding, and in-hospital mortality) with considerable calibration. Compared with the NCDR-CathPCI risk scores, the XGBoost models modestly improved discrimination for AKI and bleeding (C-statistics of 0.84 in AKI, and 0.79 in bleeding) but not for in-hospital mortality (C-statistics of 0.96). The calibration plot demonstrated that the XGBoost model overestimated the risk for in-hospital mortality in low-risk patients. All of the original NCDR-CathPCI risk scores for adverse periprocedural events showed adequate discrimination and calibration within our cohort. When using the ML-based technique, however, the improvement in the overall risk prediction was minimal.

Prediction of Myocardial Infarction From Patient Features With Machine Learning

This study proposes machine learning-based models to automatically evaluate the severity of myocardial infarction (MI) from physiological, clinical, and paraclinical features. Two types of machine learning models are investigated for the MI assessment: the classification models classify the presence of the infarct and the persistent microvascular obstruction (PMO), and the regression models quantify the Percentage of Infarcted Myocardium (PIM) of patients suspected of having an acute MI during their reception in the emergency department. The ground truth labels for these supervised models are derived from the corresponding Delayed Enhancement MRI (DE-MRI) exams and manual annotations of the myocardium and scar tissues. Experiments were conducted on 150 cases and evaluated with cross-validation. Results showed that for the MI (PMO inclusive) and the PMO (infarct exclusive), the best models obtained respectively a mean error of 0.056 and 0.012 for the quantification, and 88.67 and 77.33% for the classification accuracy of the state of the myocardium. The study of the features' importance also revealed that the troponin value had the strongest correlation to the severity of the MI among the 12 selected features. For the proposal's translational perspective, in cardiac emergencies, qualitative and quantitative analysis can be obtained prior to the achievement of MRI by relying only on conventional tests and patient features, thus, providing an objective reference for further treatment by physicians.

Biomarker Development in Cardiology: Reviewing the Past to Inform the Future

Cardiac biomarkers have become pivotal to the clinical practice of cardiology, but there remains much to discover that could benefit cardiology patients. We review the discovery of key protein biomarkers in the fields of acute coronary syndrome, heart failure, and atherosclerosis, giving an overview of the populations they were studied in and the statistics that were used to validate them. We review statistical approaches that are currently in use to assess new biomarkers and overview a framework for biomarker discovery and evaluation that could be incorporated into clinical trials to evaluate cardiovascular outcomes in the future.

XGBoost, A Novel Explainable AI Technique, in the Prediction of Myocardial Infarction: A UK Biobank Cohort Study

We wanted to assess if "Explainable AI" in the form of extreme gradient boosting (XGBoost) could outperform traditional logistic regression in predicting myocardial infarction (MI) in a large cohort. Two machine learning methods, XGBoost and logistic regression, were compared in predicting risk of MI. The UK Biobank is a population-based prospective cohort including 502 506 volunteers with active consent, aged 40 to 69 years at recruitment from 2006 to 2010. These subjects were followed until end of 2019 and the primary outcome was myocardial infarction. Both models were trained using 90% of the cohort. The remaining 10% was used as a test set. Both models were equally precise, but the regression model classified more of the healthy class correctly. XGBoost was more accurate in identifying individuals who later suffered a myocardial infarction. Receiver operator characteristic (ROC) scores are class size invariant. In this metric XGBoost outperformed the logistic regression model, with ROC scores of 0.86 (accuracy 0.75 (CI ±0.00379) and 0.77 (accuracy 0.77 (CI ± 0.00369) respectively. Secondly, we demonstrate how SHAPley values can be used to visualize and interpret the predictions made by XGBoost models, both for the cohort test set and for individuals. The XGBoost machine learning model shows very promising results in evaluating risk of MI in a large and diverse population. This model can be used, and visualized, both for individual assessments and in larger cohorts. The predictions made by the XGBoost models, points toward a future where "Explainable AI" may help to bridge the gap between medicine and data science.

Sex-Specific Absolute Delta Thresholds for High-Sensitivity Cardiac Troponin T

BACKGROUND

Sex differences in high-sensitivity cardiac troponin (hs-cTn) concentrations from healthy populations have led to the establishment of sex-specific upper reference limits for hs-cTn assays. This study assessed the performance of sex-specific delta (i.e., changes in concentrations) thresholds for the hs-cTnT assay for ruling in acute myocardial infarction (AMI) in different emergency department (ED) populations.

METHODS

This retrospective study consisted of 2 cohorts (Cohort 1 derivation and Cohort 2 validation). Cohort 1 consisted of 18 056 ED patients who had serial hs-cTnT measured using a 0-h/3-h algorithm at a US medical center, with Cohort 2 consisting of 1137 ED patients with 0-h/3-h sampling at a Canadian medical center. The primary outcome was AMI diagnosis with sex-specific deltas derived based on the Youden index and specificity estimates (i.e., ≥90%) in Cohort 1 and validated in Cohort 2.

RESULTS

In Cohort 1, 42% of all patients had 0-h hs-cTnT above the sex-specific 99th percentile. Males had higher 0-h hs-cTnT (median 17 ng/L) and absolute deltas (median 2 ng/L) than females (0-h median 11 ng/L, P < 0.0001; deltas median 1 ng/L, P < 0.0001) in non-AMI patients but not in patients with AMI. For ruling in AMI, the sex-specific delta thresholds based on 90% specificity (14 ng/L for males, 11 ng/L for females) performed best and resulted in 91% diagnostic accuracy in both males and females. The sex-specific delta thresholds yielding high specificity estimates were confirmed in the validation data set.

CONCLUSIONS

Sex-specific absolute delta thresholds can be used to rule in AMI and are robust across different study populations.

Retrospective validation of a machine learning clinical decision support tool for myocardial infarction risk stratification

Diagnosis and appropriate intervention for myocardial infarction (MI) are time-sensitive but rely on clinical measures that can be progressive and initially inconclusive, underscoring the need for an accurate and early predictor of MI to support diagnostic and clinical management decisions. The objective of this study was to develop a machine learning algorithm (MLA) to predict MI diagnosis based on electronic health record data (EHR) readily available during Emergency Department assessment. An MLA was developed using retrospective patient data. The MLA used patient data as they became available in the first 3 h of care to predict MI diagnosis (defined by International Classification of Diseases, 10th revision code) at any time during the encounter. The MLA obtained an area under the receiver operating characteristic curve of 0.87, sensitivity of 87% and specificity of 70%, outperforming the comparator scoring systems TIMI and GRACE on all metrics. An MLA can synthesize complex EHR data to serve as a clinically relevant risk stratification tool for MI.

A Machine Learning-Based Prediction Model for Cardiovascular Risk in Women With Preeclampsia

Objective: Preeclampsia affects 2–8% of women and doubles the risk of cardiovascular disease in women after preeclampsia. This study aimed to develop a model based on machine learning to predict postpartum cardiovascular risk in preeclamptic women.

Methods: Collecting demographic characteristics and clinical serum markers associated with preeclampsia during pregnancy of 907 preeclamptic women retrospectively, we predicted the cardiovascular risk (ischemic heart disease, ischemic cerebrovascular disease, peripheral vascular disease, chronic kidney disease, metabolic system disease or arterial hypertension). The study samples were divided into training sets and test sets randomly in the ratio of 8:2. The prediction model was developed by 5 different machine learning algorithms, including Random Forest. 10-fold cross-validation was performed on the training set, and the performance of the model was evaluated on the test set.

Results: Cardiovascular disease risk occurred in 186 (20.5%) of these women. By weighing area under the curve (AUC), the Random Forest algorithm presented the best performance (AUC = 0.711[95%CI: 0.697–0.726]) and was adopted in the feature selection and the establishment of the prediction model. The most important variables in Random Forest algorithm included the systolic blood pressure, Urea nitrogen, neutrophil count, glucose, and D-Dimer. Random Forest algorithm was well calibrated (Brier score = 0.133) in the test group, and obtained the highest net benefit in the decision curve analysis.

Conclusion: Based on the general situation of patients and clinical variables, a new machine learning algorithm was developed and verified for the individualized prediction of cardiovascular risk in post-preeclamptic women.

Application of machine learning to predict the occurrence of arrhythmia after acute myocardial infarction

Background

Early identification of the occurrence of arrhythmia in patients with acute myocardial infarction plays an essential role in clinical decision-making. The present study attempted to use machine learning (ML) methods to build predictive models of arrhythmia after acute myocardial infarction (AMI).

Methods

A total of 2084 patients with acute myocardial infarction were enrolled in this study. (All data is available on Github: https://github.com/wangsuhuai/AMI-database1.git). The primary outcome is whether tachyarrhythmia occurred during admission containing atrial arrhythmia, ventricular arrhythmia, and supraventricular tachycardia. All data is randomly divided into a training set (80%) and an internal testing set (20%). Apply three machine learning algorithms: decision tree, random forest (RF), and artificial neural network (ANN) to learn the training set to build a model, then use the testing set to evaluate the prediction performance, and compare it with the model built by the Global Registry of Acute Coronary Events (GRACE) risk variable set.

Results

Three ML models predict the occurrence of tachyarrhythmias after AMI. After variable selection, the artificial neural network (ANN) model has reached the highest accuracy rate, which is better than the model constructed using the Grace variable set. After applying SHapley Additive exPlanations (SHAP) to make the model interpretable, the most important features are abnormal wall motion, lesion location, bundle branch block, age, and heart rate. Among them, RBBB (odds ratio [OR]: 4.21; 95% confidence interval [CI]: 2.42–7.02), ≥ 2 ventricular walls motion abnormal (OR: 3.26; 95% CI: 2.01–4.36) and right coronary artery occlusion (OR: 3.00; 95% CI: 1.98–4.56) are significant factors related to arrhythmia after AMI.

Conclusions

We used advanced machine learning methods to build prediction models for tachyarrhythmia after AMI for the first time (especially the ANN model that has the best performance). The current study can supplement the current AMI risk score, provide a reliable evaluation method for the clinic, and broaden the new horizons of ML and clinical research.

Trial registration Clinical Trial Registry No.: ChiCTR2100041960.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12911-021-01667-8.

Machine learning with D-dimer in the risk stratification for pulmonary embolism: a derivation and internal validation study

AIM

To develop a machine learning model to predict the diagnosis of pulmonary embolism (PE).

METHODS AND RESULTS

We undertook a derivation and internal validation study to develop a risk prediction model for use in patients being investigated for possible PE. The machine learning technique, generalized logistic regression using elastic net, was chosen following an assessment of seven machine learning techniques and on the basis that it optimized the area under the receiver operator characteristic curve (AUC) and Brier score. Models were developed both with and without the addition of D-dimer. A total of 3347 patients were included in the study of whom, 219 (6.5%) had PE. Four clinical variables (O2 saturation, previous deep venous thrombosis or PE, immobilization or surgery, and alternative diagnosis equal or more likely than PE) plus D-dimer contributed to the machine learning models. The addition of D-dimer improved the AUC by 0.16 (95% confidence interval 0.13-0.19), from 0.73 to 0.89 (0.87-0.91) and decreased the Brier score by 14% (10-18%). More could be ruled out with a higher positive likelihood ratio than by the Wells score combined with D-dimer, revised Geneva score combined with D-dimer, or the Pulmonary Embolism Rule-out Criteria score. Machine learning with D-dimer maintained a low-false-negative rate at a true-negative rate of nearly 53%, which was better performance than any of the other alternatives.

CONCLUSION

A machine learning model outperformed traditional risk scores for the risk stratification of PE in the emergency department. However, external validation is needed.

A prehospital diagnostic algorithm for strokes using machine learning: a prospective observational study

High precision is optimal in prehospital diagnostic algorithms for strokes and large vessel occlusions. We hypothesized that prehospital diagnostic algorithms for strokes and their subcategories using machine learning could have high predictive value. Consecutive adult patients with suspected stroke as per emergency medical service personnel were enrolled in a prospective multicenter observational study in 12 hospitals in Japan. Five diagnostic algorithms using machine learning, including logistic regression, random forest, support vector machine, and eXtreme Gradient Boosting, were evaluated for stroke and subcategories including acute ischemic stroke with/without large vessel occlusions, intracranial hemorrhage, and subarachnoid hemorrhage. Of the 1446 patients in the analysis, 1156 (80%) were randomly included in the training (derivation) cohort and cohorts, and 290 (20%) were included in the test (validation) cohort. In the diagnostic algorithms for strokes using eXtreme Gradient Boosting had the highest diagnostic value (test data, area under the receiver operating curve 0.980). In the diagnostic algorithms for the subcategories using eXtreme Gradient Boosting had a high predictive value (test data, area under the receiver operating curve, acute ischemic stroke with/without large vessel occlusions 0.898/0.882, intracranial hemorrhage 0.866, subarachnoid hemorrhage 0.926). Prehospital diagnostic algorithms using machine learning had high predictive value for strokes and their subcategories.

Application of Artificial Intelligence in Acute Coronary Syndrome: A Brief Literature Review

Artificial intelligence (AI) is defined as a set of algorithms and intelligence to try to imitate human intelligence. Machine learning is one of them, and deep learning is one of those machine learning techniques. The application of AI in healthcare systems including hospitals and clinics has many possible advantages and future prospects. Applications of AI in cardiovascular medicine are machine learning techniques for diagnostic procedures including imaging modalities and biomarkers and predictive analytics for personalized therapies and improved outcomes. In cardiovascular medicine, AI-based systems have found new applications in risk prediction for cardiovascular diseases, in cardiovascular imaging, in predicting outcomes after revascularization procedures, and in newer drug targets. AI such as machine learning has partially resolved and provided possible solutions to unmet requirements in interventional cardiology. Predicting economically vital endpoints, predictive models with a wide range of health factors including comorbidities, socioeconomic factors, and angiographic factors comprising of the size of stents, the volume of contrast agent which was infused during angiography, stent malposition, and so on have been possible owing to machine learning and AI. Nowadays, machine learning techniques might possibly help in the identification of patients at risk, with higher morbidity and mortality following acute coronary syndrome (ACS). AI through machine learning has shown several potential benefits in patients with ACS. From diagnosis to treatment effects to predicting adverse events and mortality in patients with ACS, machine learning should find an essential place in clinical medicine and in interventional cardiology for the treatment and management of patients with ACS. This paper is a review of the literature which will focus on the application of AI in ACS.

Using Machine Learning Algorithms to Predict Hospital Acquired Thrombocytopenia after Operation in the Intensive Care Unit: A Retrospective Cohort Study

Hospital acquired thrombocytopenia (HAT) is a common hematological complication after surgery. This research aimed to develop and compare the performance of seven machine learning (ML) algorithms for predicting patients that are at risk of HAT after surgery. We conducted a retrospective cohort study which enrolled adult patients transferred to the intensive care unit (ICU) after surgery in West China Hospital of Sichuan University from January 2016 to December 2018. All subjects were randomly divided into a derivation set (70%) and test set (30%). ten-fold cross-validation was used to estimate the hyperparameters of ML algorithms during the training process in the derivation set. After ML models were developed, the sensitivity, specificity, area under the curve (AUC), and net benefit (decision analysis curve, DCA) were calculated to evaluate the performances of ML models in the test set. A total of 10,369 patients were included and in 1354 (13.1%) HAT occurred. The AUC of all seven ML models exceeded 0.7, the two highest were Gradient Boosting (GB) (0.834, 0.814-0.853, p < 0.001) and Random Forest (RF) (0.828, 0.807-0.848, p < 0.001). There was no difference between GB and RF (0.834 vs. 0.828, p = 0.293); however, these two were better than the remaining five models (p < 0.001). The DCA revealed that all ML models had high net benefits with a threshold probability approximately less than 0.6. In conclusion, we found that ML models constructed by multiple preoperative variables can predict HAT in patients transferred to ICU after surgery, which can improve risk stratification and guide management in clinical practice.

Applications of machine learning to undifferentiated chest pain in the emergency department: A systematic review

BACKGROUND

Chest pain is amongst the most common reason for presentation to the emergency department (ED). There are many causes of chest pain, and it is important for the emergency physician to quickly and accurately diagnose life threatening causes such as acute myocardial infarction (AMI). Multiple clinical decision tools have been developed to assist clinicians in risk stratifying patients with chest. There is growing recognition that machine learning (ML) will have a significant impact on the practice of medicine in the near future and may assist with diagnosis and risk stratification. This systematic review aims to evaluate how ML has been applied to adults presenting to the ED with undifferentiated chest pain and assess if ML models show improved performance when compared to physicians or current risk stratification techniques.

METHODS AND FINDINGS

We conducted a systematic review of journal articles that applied a ML technique to an adult patient presenting to an emergency department with undifferentiated chest pain. Multiple databases were searched from inception through to November 2020. In total, 3361 articles were screened, and 23 articles were included. We did not conduct a metanalysis due to a high level of heterogeneity between studies in both their methods, and reporting. The most common primary outcomes assessed were diagnosis of acute myocardial infarction (AMI) (12 studies), and prognosis of major adverse cardiovascular event (MACE) (6 studies). There were 14 retrospective studies and 5 prospective studies. Four studies reported the development of a machine learning model retrospectively then tested it prospectively. The most common machine learning methods used were artificial neural networks (14 studies), random forest (6 studies), support vector machine (5 studies), and gradient boosting (2 studies). Multiple studies achieved high accuracy in both the diagnosis of AMI in the ED setting, and in predicting mortality and composite outcomes over various timeframes. ML outperformed existing risk stratification scores in all cases, and physicians in three out of four cases. The majority of studies were single centre, retrospective, and without prospective or external validation. There were only 3 studies that were considered low risk of bias and had low applicability concerns. Two studies reported integrating the ML model into clinical practice.

CONCLUSIONS

Research on applications of ML for undifferentiated chest pain in the ED has been ongoing for decades. ML has been reported to outperform emergency physicians and current risk stratification tools to diagnose AMI and predict MACE but has rarely been integrated into practice. Many studies assessing the use of ML in undifferentiated chest pain in the ED have a high risk of bias. It is important that future studies make use of recently developed standardised ML reporting guidelines, register their protocols, and share their datasets and code. Future work is required to assess the impact of ML model implementation on clinical decision making, patient orientated outcomes, and patient and physician acceptability.

TRIAL REGISTRATION

International Prospective Register of Systematic Reviews registration number: CRD42020184977.

Cardiac Troponin Thresholds and Kinetics to Differentiate Myocardial Injury and Myocardial Infarction

BACKGROUND

Although the 99th percentile is the recommended diagnostic threshold for myocardial infarction, some guidelines also advocate the use of higher troponin thresholds to rule in myocardial infarction at presentation. It is unclear whether the magnitude or change in troponin concentration can differentiate causes of myocardial injury and infarction in practice.

METHODS

In a secondary analysis of a multicenter randomized controlled trial, we identified 46 092 consecutive patients presenting with suspected acute coronary syndrome without ST-segment-elevation myocardial infarction. High-sensitivity cardiac troponin I concentrations at presentation and on serial testing were compared between patients with myocardial injury and infarction. The positive predictive value and specificity were determined at the sex-specific 99th percentile upper reference limit and rule-in thresholds of 64 ng/L and 5-fold of the upper reference limit for a diagnosis of type 1 myocardial infarction.

RESULTS

Troponin was above the 99th percentile in 8188 patients (18%). The diagnosis was type 1 or type 2 myocardial infarction in 50% and 14% and acute or chronic myocardial injury in 20% and 16%, respectively. Troponin concentrations were similar at presentation in type 1 (median [25th-75th percentile] 91 [30-493] ng/L) and type 2 (50 [22-147] ng/L) myocardial infarction and in acute (50 [26-134] ng/L) and chronic (51 [31-130] ng/L) myocardial injury. The 99th percentile and rule-in thresholds of 64 ng/L and 5-fold upper reference limit gave a positive predictive value of 57% (95% CI, 56%-58%), 59% (58%-61%), and 62% (60%-64%) and a specificity of 96% (96%-96%), 96% (96%-96%), and 98% (97%-98%), respectively. The absolute, relative, and rate of change in troponin concentration were highest in patients with type 1 myocardial infarction (P<0.001 for all). Discrimination improved when troponin concentration and change in troponin were combined compared with troponin concentration at presentation alone (area under the curve, 0.661 [0.642-0.680] versus 0.613 [0.594-0.633]).

CONCLUSIONS

Although we observed important differences in the kinetics, cardiac troponin concentrations at presentation are insufficient to distinguish type 1 myocardial infarction from other causes of myocardial injury or infarction in practice and should not guide management decisions in isolation. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT01852123.

Diagnostic Performance of Serial High-Sensitivity Cardiac Troponin Measurements in the Emergency Setting

Serial high-sensitivity cardiac troponin (hsTn) testing in the emergency department (ED) and the intensive cardiac care unit may assist physicians in ruling out or ruling in acute myocardial infarction (MI). There are three major algorithms proposed for high-sensitivity cardiac troponin I (hsTnI) using serial measurements while incorporating absolute concentration changes for MI or death following ED presentation. We sought to determine the diagnostic estimates of these three algorithms and if one was superior in two different Canadian ED patient cohorts with serial hsTnI measurements. An undifferentiated ED population (Cohort-1) and an ED population with symptoms suggestive of acute coronary syndrome (ACS; Cohort-2) were clinically managed with non-hsTn testing with the hsTnI testing performed in real-time with physicians blinded to these results (i.e., hsTnI not reported). The three algorithms evaluated were the European Society of Cardiology (ESC), the High-STEACS pathway, and the COMPASS-MI algorithm. The diagnostic estimates were derived for each algorithm for the 30-day MI/death outcome for the rule-out and rule-in arm in each cohort and compared to proposed diagnostic benchmarks (i.e., sensitivity ≥ 99.0% and specificity ≥ 90.0%) with 95% confidence intervals (CI). In Cohort-1 (n = 2966 patients, 15.3% had outcome) and Cohort-2 (n = 935 patients, 15.6% had outcome), the algorithm that obtained the highest sensitivity (97.8%; 95% CI: 96.0-98.9 and 98.6%; 95% CI: 95.1-99.8, respectively) in both cohorts was COMPASS-MI. Only Cohort-2 with both the ESC and COMPASS-MI algorithms exceeded the specificity benchmark (97.0%; 95% CI: 95.5-98.0 and 96.7%; 95% CI: 95.2-97.8, respectively). Patient selection for serial hsTnI testing will affect specificity estimates, with no algorithm achieving a sensitivity ≥ 99% for 30-day MI or death.

Novel Criteria for the Observe-Zone of the ESC 0/1h-hs-cTnT Algorithm

Background: The non-ST elevation myocardial infarction (NSTEMI) guidelines of the European Society of Cardiology (ESC) recommend a 3h cardiac troponin determination in patients triaged to the observe-zone of the ESC 0/1h-algorithm; however, no specific cut-off for further triage is endorsed. Recently, a specific cut-off for 0/3h high-sensitivity cardiac troponin T (hs-cTnT) change (7ng/L) was proposed warranting external validation. Methods: Patients presenting with acute chest discomfort to the emergency department were prospectively enrolled into an international multicenter diagnostic study. Final diagnoses were centrally adjudicated by two independent cardiologists applying the 4th universal definition of MI, based on complete cardiac work-up, cardiac imaging and serial hs-cTnT. Hs-cTnT concentrations were measured at presentation, after 1h and 3h. The objective was to externally validate the proposed cut-off, and if necessary, derive and internally as well as externally validate novel 0/3h-criteria for the observe-zone of the ESC 0/1h-hs-cTnT-algorithm in an independent multicenter cohort. Results: Among 2076 eligible patients, application of the ESC 0/1h-hs-cTnT-algorithm triaged 1512 patients (72.8%) to either rule-out or rule-in of NSTEMI, remaining 564 patients (27.2%) in the observe-zone (adjudicated NSTEMI prevalence 120/564 patients, 21.3%). The suggested 0/3h-hs-cTnT-change of <7ng/L triaged 517 patients (91.7%) towards rule-out, resulting in a sensitivity of 33.3% (95%CI 25.5-42.2), missing 80 patients with NSTEMI, and ≥7ng/L triaged 47 patients towards rule-in (8.3%), resulting in a specificity of 98.4% (95%CI 96.8-99.2). Novel derived 0/3h-criteria for the observe-zone patients ruled-out NSTEMI with a 3h hs-cTnT concentration <15 ng/L and a 0/3h-hs-cTnT absolute change <4 ng/L, triaging 138 patients (25%) towards rule-out, resulting in a sensitivity of 99.2% (95%CI 96.0-99.9), missing 1 patient with NSTEMI. A 0/3h-hs-cTnT absolute change ≥6 ng/L triaged 63 patients (11.2%) towards rule-in, resulting in a specificity of 98% (95%CI 96.2-98.9) Thereby, the novel 0/3h-criteria reduced the number of patients in the observe zone by 36%, and the number of T1MI by 50%. Findings were confirmed in both internal and external validation. Conclusions: A combination of a 3h hs-cTnT concentration (<15 ng/L) and a 0/3h absolute change (<4 ng/L) is necessary to safely rule-out NSTEMI in patients remaining in the observe-zone of the ESC 0/1h-hs-cTnT-algorithm. Clinical Trial Registration: URL: https://clinicaltrials.gov Unique Identifier: NCT00470587.

A review of risk prediction models in cardiovascular disease: conventional approach vs. artificial intelligent approach

Cardiovascular disease (CVD) is the leading cause of death worldwide and is a global health issue. Traditionally, statistical models are used commonly in the risk prediction and assessment of CVD. However, the adoption of artificial intelligent (AI) approach is rapidly taking hold in the current era of technology to evaluate patient risks and predict the outcome of CVD. In this review, we outline various conventional risk scores and prediction models and do a comparison with the AI approach. The strengths and limitations of both conventional and AI approaches are discussed. Besides that, biomarker discovery related to CVD are also elucidated as the biomarkers can be used in the risk stratification as well as early detection of the disease. Moreover, problems and challenges involved in current CVD studies are explored. Lastly, future prospects of CVD risk prediction and assessment in the multi-modality of big data integrative approaches are proposed.

Routine Hematological Parameters May Be Predictors of COVID-19 Severity

To date, coronavirus disease 2019 (COVID-19) has affected over 100 million people globally. COVID-19 can present with a variety of different symptoms leading to manifestation of disease ranging from mild cases to a life-threatening condition requiring critical care-level support. At present, a rapid prediction of disease severity and critical care requirement in COVID-19 patients, in early stages of disease, remains an unmet challenge. Therefore, we assessed whether parameters from a routine clinical hematology workup, at the time of hospital admission, can be valuable predictors of COVID-19 severity and the requirement for critical care. Hematological data from the day of hospital admission (day of positive COVID-19 test) for patients with severe COVID-19 disease (requiring critical care during illness) and patients with non-severe disease (not requiring critical care) were acquired. The data were amalgamated and cleaned and modeling was performed. Using a decision tree model, we demonstrated that routine clinical hematology parameters are important predictors of COVID-19 severity. This proof-of-concept study shows that a combination of activated partial thromboplastin time, white cell count-to-neutrophil ratio, and platelet count can predict subsequent severity of COVID-19 with high sensitivity and specificity (area under ROC 0.9956) at the time of the patient's hospital admission. These data, pending further validation, indicate that a decision tree model with hematological parameters could potentially form the basis for a rapid risk stratification tool that predicts COVID-19 severity in hospitalized patients.

Early Rule-Out Strategies in the Emergency Department Utilizing High-Sensitivity Cardiac Troponin Assays

BACKGROUND

Over the past decade, intense collaboration between academic investigators and the diagnostic industry have allowed the integration of high-sensitivity cardiac troponin (hs-cTn) assays into clinical practice worldwide. The hs-cTn assays, with their increased diagnostic accuracy for acute myocardial infarction (AMI), have facilitated the maturation of early rule-out strategies. The first iteration was complex and required the combination of a biomarker panel, the electrocardiogram, and a clinical risk score and allowed the safe rule-out of AMI in only 10% of patients with acute chest pain. In contrast, the latest iterations, including the European Society of Cardiology (ESC) 0/1-h algorithm, are simple. They are based on hs-cTn concentrations only and allow the safe rule-out or rule-in of AMI in up to 75% of patients.

CONTENT

The purposes of this minireview are (a) to describe the best validated hs-cTn-based strategies for early rule-out of AMI, (b) to discuss the advantages and limitations of the different strategies, (c) to identify patient subgroups requiring particular attention, (d) to recognize challenges for widespread clinical implementation, and (e) to provide guidance on strategies for their safe and effective clinical implementation.

SUMMARY

Physicians and institutions may choose among several well-validated rule-out algorithms. The ESC 0/1-h algorithm for hs-cTnT or hs-cTnI seems to be the most attractive option today. It best balances safety and efficacy, and it has been derived and validated for all currently available hs-cTnT/I assays, facilitating widespread clinical implementation.

High-Sensitivity Cardiac Troponin on Presentation to Rule Out Myocardial Infarction: A Stepped-Wedge Cluster Randomized Controlled Trial

BACKGROUND

High-sensitivity cardiac troponin assays enable myocardial infarction to be ruled out earlier, but the safety and efficacy of this approach is uncertain. We investigated whether an early rule-out pathway is safe and effective for patients with suspected acute coronary syndrome.

METHODS

We performed a stepped-wedge cluster randomized controlled trial in the emergency departments of 7 acute care hospitals in Scotland. Consecutive patients presenting with suspected acute coronary syndrome between December 2014 and December 2016 were included. Sites were randomized to implement an early rule-out pathway where myocardial infarction was excluded if high-sensitivity cardiac troponin I concentrations were <5 ng/L at presentation. During a previous validation phase, myocardial infarction was ruled out when troponin concentrations were <99th percentile at 6 to 12 hours after symptom onset. The coprimary outcome was length of stay (efficacy) and myocardial infarction or cardiac death after discharge at 30 days (safety). Patients were followed for 1 year to evaluate safety and other secondary outcomes.

RESULTS

We enrolled 31 492 patients (59±17 years of age [mean±SD]; 45% women) with troponin concentrations <99th percentile at presentation. Length of stay was reduced from 10.1±4.1 to 6.8±3.9 hours (adjusted geometric mean ratio, 0.78 [95% CI, 0.73-0.83]; P<0.001) after implementation and the proportion of patients discharged increased from 50% to 71% (adjusted odds ratio, 1.59 [95% CI, 1.45-1.75]). Noninferiority was not demonstrated for the 30-day safety outcome (upper limit of 1-sided 95% CI for adjusted risk difference, 0.70% [noninferiority margin 0.50%]; P=0.068), but the observed differences favored the early rule-out pathway (0.4% [57/14 700] versus 0.3% [56/16 792]). At 1 year, the safety outcome occurred in 2.7% (396/14 700) and 1.8% (307/16 792) of patients before and after implementation (adjusted odds ratio, 1.02 [95% CI, 0.74-1.40]; P=0.894), and there were no differences in hospital reattendance or all-cause mortality.

CONCLUSIONS

Implementation of an early rule-out pathway for myocardial infarction reduced length of stay and hospital admission. Although noninferiority for the safety outcome was not demonstrated at 30 days, there was no increase in cardiac events at 1 year. Adoption of this pathway would have major benefits for patients and health care providers. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03005158.

A Review on Joint Carotid Intima-Media Thickness and Plaque Area Measurement in Ultrasound for Cardiovascular/Stroke Risk Monitoring: Artificial Intelligence Framework

Cardiovascular diseases (CVDs) are the top ten leading causes of death worldwide. Atherosclerosis disease in the arteries is the main cause of the CVD, leading to myocardial infarction and stroke. The two primary image-based phenotypes used for monitoring the atherosclerosis burden is carotid intima-media thickness (cIMT) and plaque area (PA). Earlier segmentation and measurement methods were based on ad hoc conventional and semi-automated digital imaging solutions, which are unreliable, tedious, slow, and not robust. This study reviews the modern and automated methods such as artificial intelligence (AI)-based. Machine learning (ML) and deep learning (DL) can provide automated techniques in the detection and measurement of cIMT and PA from carotid vascular images. Both ML and DL techniques are examples of supervised learning, i.e., learn from "ground truth" images and transformation of test images that are not part of the training. This review summarizes (1) the evolution and impact of the fast-changing AI technology on cIMT/PA measurement, (2) the mathematical representations of ML/DL methods, and (3) segmentation approaches for cIMT/PA regions in carotid scans based for (a) region-of-interest detection and (b) lumen-intima and media-adventitia interface detection using ML/DL frameworks. AI-based methods for cIMT/PA segmentation have emerged for CVD/stroke risk monitoring and may expand to the recommended parameters for atherosclerosis assessment by carotid ultrasound.

Late Outcomes of the RAPID-TnT Randomized Controlled Trial: 0/1-Hour High-Sensitivity Troponin T Protocol in Suspected ACS

BACKGROUND

High-sensitivity troponin assays are increasingly being adopted to expedite evaluation of patients with suspected acute coronary syndromes. Few direct comparisons have examined whether the enhanced performance of these assays at low concentrations leads to changes in care that improves longer-term outcomes. This study evaluated late outcomes of participants managed under an unmasked 0/1-hour high-sensitivity cardiac troponin T (hs-cTnT) protocol compared with a 0/3-hour masked hs-cTnT protocol.

METHODS

We conducted a multicenter prospective patient-level randomized comparison of care informed by unmasked 0/1-hour hs-cTnT protocol (reported to <5 ng/L) versus standard practice masked hs-cTnT testing (reported to ≤29 ng/L) assessed at 0/3 hours and followed participants for 12 months. Participants included were those presenting to metropolitan emergency departments with suspected acute coronary syndromes, without ECG evidence of coronary ischemia. The primary end point was time to all-cause death or myocardial infarction using Cox proportional hazards models adjusted for clustering within hospitals.

RESULTS

Between August 2015 and April 2019, we randomized 3378 participants, of whom 108 withdrew, resulting in 12-month follow-up for 3270 participants (masked: 1632; unmasked: 1638). Among these, 2993 (91.5%) had an initial troponin concentration of ≤29 ng/L. Deployment of the 0/1-hour hs-cTnT protocol was associated with reductions in functional testing. Over 12-month follow-up, there was no difference in invasive coronary angiography (0/1-hour unmasked: 232/1638 [14.2%]; 0/3-hour masked: 202/1632 [12.4%]; P=0.13), although an increase was seen among patients with hs-cTnT levels within the masked range (0/1-hour unmasked arm: 168/1507 [11.2%]; 0/3-hour masked arm: 124/1486 [8.3%]; P=0.010). By 12 months, all-cause death and myocardial infarction did not differ between study arms overall (0/1-hour: 82/1638 [5.0%] versus 0/3-hour: 62/1632 [3.8%]; hazard ratio, 1.32 [95% CI, 0.95-1.83]; P=0.10). Among participants with initial troponin T concentrations ≤29 ng/L, unmasked hs-cTnT reporting was associated with an increase in death or myocardial infarction (0/1-hour: 55/1507 [3.7%] versus 0/3-hour: 34/1486 [2.3%]; hazard ratio, 1.60 [95% CI, 1.05-2.46]; P=0.030).

CONCLUSIONS

Unmasked hs-cTnT reporting deployed within a 0/1-hour protocol did not reduce ischemic events over 12-month follow-up. Changes in practice associated with the implementation of this protocol may be associated with an increase in death and myocardial infarction among those with newly identified troponin elevations. Registration: URL: https://www.anzctr.org.au; Unique identifier: ACTRN12615001379505.

Clinical History and Detectable Troponin Concentrations below the 99th Percentile for Risk Stratification of Patients with Chest Pain and First Normal Troponin

Decision-making is challenging in patients with chest pain and normal high-sensitivity cardiac troponin T (hs-cTnT; <99th percentile; <14 ng/L) at hospital arrival. Most of these patients might be discharged early. We investigated clinical data and hs-cTnT concentrations for risk stratification. This is a retrospective study including 4476 consecutive patients presenting to the emergency department with chest pain and first normal hs-cTnT. The primary endpoint was one-year death or acute myocardial infarction, and the secondary endpoint added urgent revascularization. The number of primary and secondary endpoints was 173 (3.9%) and 252 (5.6%). Mean hs-cTnT concentrations were 6.9 ± 2.5 ng/L. Undetectable (<5 ng/L) hs-cTnT (n = 1847, 41%) had optimal negative predictive value (99.1%) but suboptimal sensitivity (90.2%) and discrimination accuracy (AUC = 0.664) for the primary endpoint. Multivariable analysis was used to identify the predictive clinical variables. The clinical model showed good discrimination accuracy (AUC = 0.810). The addition of undetectable hs-cTnT (≥ or <5 ng/L; HR, hazard ratio = 3.80; 95% CI, confidence interval 2.27–6.35; p = 0.00001) outperformed the clinical model alone (AUC = 0.836, p = 0.002 compared to the clinical model). Measurable hs-cTnT concentrations (between detection limit and 99th percentile; per 0.1 ng/L, HR = 1.13; CI 1.06–1.20; p = 0.0001) provided further predictive information (AUC = 0.844; p = 0.05 compared to the clinical plus undetectable hs-cTnT model). The results were reproducible for the secondary endpoint and 30-day events. Clinical assessment, undetectable hs-cTnT and measurable hs-cTnT concentrations must be considered for decision-making after a single negative hs-cTnT result in patients presenting to the emergency department with acute chest pain.

Machine Learning and Syncope Management in the ED: The Future Is Coming

In recent years, machine learning (ML) has been promisingly applied in many fields of clinical medicine, both for diagnosis and prognosis prediction. Aims of this narrative review were to summarize the basic concepts of ML applied to clinical medicine and explore its main applications in the emergency department (ED) setting, with a particular focus on syncope management. Through an extensive literature search in PubMed and Embase, we found increasing evidence suggesting that the use of ML algorithms can improve ED triage, diagnosis, and risk stratification of many diseases. However, the lacks of external validation and reliable diagnostic standards currently limit their implementation in clinical practice. Syncope represents a challenging problem for the emergency physician both because its diagnosis is not supported by specific tests and the available prognostic tools proved to be inefficient. ML algorithms have the potential to overcome these limitations and, in the future, they could support the clinician in managing syncope patients more efficiently. However, at present only few studies have addressed this issue, albeit with encouraging results.

Machine learning compared with rule-in/rule-out algorithms and logistic regression to predict acute myocardial infarction based on troponin T concentrations

OBJECTIVE

Computerized decision-support tools may improve diagnosis of acute myocardial infarction (AMI) among patients presenting with chest pain at the emergency department (ED). The primary aim was to assess the predictive accuracy of machine learning algorithms based on paired high-sensitivity cardiac troponin T (hs-cTnT) concentrations with varying sampling times, age, and sex in order to rule in or out AMI.

METHODS

In this register-based, cross-sectional diagnostic study conducted retrospectively based on 5695 chest pain patients at 2 hospitals in Sweden 2013-2014 we used 5-fold cross-validation 200 times in order to compare the performance of an artificial neural network (ANN) with European guideline-recommended 0/1- and 0/3-hour algorithms for hs-cTnT and with logistic regression without interaction terms. Primary outcome was the size of the intermediate risk group where AMI could not be ruled in or out, while holding the sensitivity (rule-out) and specificity (rule-in) constant across models.

RESULTS

ANN and logistic regression had similar (95%) areas under the receiver operating characteristics curve. In patients (n = 4171) where the timing requirements (0/1 or 0/3 hour) for the sampling were met, using ANN led to a relative decrease of 9.2% (95% confidence interval 4.4% to 13.8%; from 24.5% to 22.2% of all tested patients) in the size of the intermediate group compared to the recommended algorithms. By contrast, using logistic regression did not substantially decrease the size of the intermediate group.

CONCLUSION

Machine learning algorithms allow for flexibility in sampling and have the potential to improve risk assessment among chest pain patients at the ED.

Discordance of High-Sensitivity Troponin Assays in Patients With Suspected Acute Coronary Syndromes

BACKGROUND

High-sensitivity cardiac troponin (hs-cTn) assays have different analytic characteristics.

OBJECTIVES

The goal of this study was to quantify differences between assays for common analytical benchmarks and to determine whether they may result in differences in the management of patients with suspected acute coronary syndrome (ACS).

METHODS

The authors included patients with suspected ACS enrolled in the ROMICAT (Rule Out Myocardial Infarction/Ischemia Using Computer Assisted Tomography) I and II trials, with blood samples taken at emergency department presentation (ROMICAT-I and -II) or at 2 and 4 h thereafter (ROMICAT-II). hs-cTn concentrations were measured using 3 assays (Roche Diagnostics, Elecsys 2010 platform; Abbott Diagnostics, ARCHITECT i2000SR; Siemens Diagnostics, HsVista). Per blood sample, we determined concordance across analytic benchmarks (99th percentile). Per-patient, the authors determined concordance of management recommendations (rule-out/observe/rule-in) per the 0/2-h algorithm, and their association with diagnostic test findings (coronary artery stenosis >50% on coronary computed tomography angiography or inducible ischemia on perfusion imaging) and ACS.

RESULTS

Among 1,027 samples from 624 patients (52.8 ± 10.0 years; 39.4% women), samples were classified as 99th percentile (7.2% vs. 6.0% vs. 6.2%) by Roche, Abbott, and Siemens, respectively. A total of 37.4% (n = 384 of 1,027) of blood samples were classified into the same analytical benchmark category, with low concordance across benchmarks (99th percentile 43.6%). Serial samples were available in 242 patients (40.1% women; mean age: 52.8 ± 8.0 years). The concordance of management recommendations across assays was 74.8% (n = 181 of 242) considering serial hs-cTn measurements. Of patients who were recommended to discharge, 19.6% to 21.1% had positive diagnostic test findings and 2.8% to 4.3% had ACS at presentation.

CONCLUSIONS

Caregivers should be aware that there are significant differences between hs-cTn assays in stratifying individual samples and patients with intermediate likelihood of ACS according to analytical benchmarks that may result in different management recommendations. (Rule Out Myocardial Infarction by Computer Assisted Tomography [ROMICAT]; NCT00990262) (Multicenter Study to Rule Out Myocardial Infarction by Cardiac Computed Tomography [ROMICAT-II]; NCT01084239).

99th Percentile Upper-Reference Limit of Cardiac Troponin and the Diagnosis of Acute Myocardial Infarction

BACKGROUND

Concerns exist regarding how the 99th percentile upper reference limit (URL) of cardiac troponin (cTn) is determined and whether it should be derived from normal healthy individuals.

CONTENT

The 99th percentile URL of cTn is an important criterion to standardize the diagnosis of myocardial infarction (MI) for clinical, research, and regulatory purposes. Statistical heterogeneity in its calculation exists but recommendations have been proposed. Some negativity has resulted from the fact that with some high-sensitivity (hs) cTn assays, a greater number of increases above the 99th percentile are observed when transitioning from a contemporary assay. Increases reflect acute or chronic myocardial injury and provide valuable diagnostic and prognostic information. The etiology of increases can sometimes be difficult to determine, making a specific treatment approach challenging. For those reasons, some advocate higher cutoff concentrations. This approach can contribute to missed diagnoses. Contrary to claims, neither clinical or laboratory guidelines have shifted away from the 99th percentile. To support the diagnosis of acute MI, the 99th percentile URL remains the best-established approach given the absence of cTn assay standardization. Importantly, risk stratification algorithms using hs-cTn assays predict the possibility of MI diagnoses established using the 99th percentile.

SUMMARY

The 99th percentile of cTn remains the best-established criterion for the diagnosis of acute MI. While not perfect, it is analytically and clinically evidence-based. Until there are robust data to suggest some other approach, staying with the 99th percentile, a threshold that has served the field well for the past 20 years, appears prudent.

External validation of the clinical chemistry score

BACKGROUND

Combining high-sensitivity cardiac troponin (hs-cTn) with estimated glomerular filtration rate and glucose within the Clinical Chemistry Score (CCS) could help in the assessment of patients with suspected acute myocardial infarction (AMI).

METHODS

In patients presenting with suspected AMI to the emergency department, we aimed to externally validate the performance of the CCS in a prospective international multicenter study and to directly compare the diagnostic and prognostic performance of the CCS with hs-cTnT and hs-cTnI baseline levels alone using a single cut-off approach. The diagnostic endpoint was diagnostic accuracy for AMI as centrally adjudicated by two independent cardiologists including cardiac imaging and serial hs-cTnT/I measurements. The prognostic endpoint was 30-day AMI or death.

RESULTS

AMI was the final diagnosis in 620/3827 patients (16.2%) adjudicated with hs-cTnT and 599 patients (15.7%) adjudicated with hs-cTnI. The CCS resulted in high diagnostic accuracy for AMI and prognostic accuracy for 30-days AMI/death as quantified by the area under the receiver-operating characteristic curve (AUC), using hs-cTnT 0.90 (95%CI 0.89-0.91) and 0.89 (95%CI 0.88-0.90), using hs-cTnI 0.91 (95%Cl 0.90-0.92) and 0.90 (95%CI 0.89-0.91) respectively. E.g. a CCS of 0 points resulted in a sensitivity of 99.8% (95%CI 99.1-100%) for rule-out of index AMI and 99.5% (95%CI 98.5-100%) for AMI/death at 30 days for hs-cTnT and 99.8% (95%CI 98.9-100%) and 99.6% (95%CI 98.6-100%) using hs-cTnI. Overall, the single hs-cTnT/I measurement approach provided comparable diagnostic (sensitivity 99.5-99.7%) and prognostic (sensitivity 98.9-99.5%) performance versus the CCS.

INTERPRETATION

The CCS provided high diagnostic and prognostic performance also in this independent large validation cohort. A single hs-cTnT/I measurement approach for rule-out MI yielded similar estimates.

Mobile application to optimize care for ST-segment elevation myocardial infarction patients in a large healthcare system, STEMIcathAID: rationale and design

Aims

Technological advancements have transformed healthcare. System delays in transferring patients with ST-segment elevation myocardial infarction (STEMI) to a primary percutaneous coronary intervention (PCI) centre are associated with worse clinical outcomes. Our aim was to design and develop a secure mobile application, STEMIcathAID, streamlining communication, and coordination between the STEMI care teams to reduce ischaemia time and improve patient outcomes.

Methods and results

The app was designed for transfer of patients with STEMI to a cardiac catheterization laboratory (CCL) from an emergency department (ED) of either a PCI capable or a non-PCI capable hospital. When a suspected STEMI arrives to a non-PCI hospital ED, the ED physician uploads the electrocardiogram and relevant patient information. An instant notification is simultaneously sent to the on-call CCL attending and transfer centre. The attending reviews the information, makes a video call and decides to either accept or reject the transfer. If accepted, on-call CCL team members receive an immediate push notification and begin communicating with the ED team via a HIPAA compliant chat. The app provides live GPS tracking of the ambulance and frequent clinical status updates of the patient. In addition, it allows for screening of STEMI patients in cardiogenic shock. Prior to discharge, important data elements have to be entered to close the case.

Conclusion

We developed a novel mobile app to optimize care for STEMI patients and facilitate electronic extraction of relevant performance metrics to improve allocation of resources and reduction of costs.