Critical appraisal of artificial intelligence-based prediction models for cardiovascular disease

Machine-learning approaches for risk prediction in transcatheter aortic valve implantation: Systematic review and meta-analysis

OBJECTIVES

With the expanding integration of artificial intelligence (AI) and machine learning (ML) into the structural heart domain, numerous ML models have emerged for the prediction of adverse outcomes after transcatheter aortic valve implantation (TAVI). We aim to identify, describe, and critically appraise ML prediction models for adverse outcomes after TAVI. Key objectives consisted in summarizing model performance, evaluating adherence to reporting guidelines, and transparency.

METHODS

We searched PubMed, SCOPUS, and Embase through August 2023. We selected published machine learning models predicting TAVI outcomes. Two reviewers independently screened articles, extracted data, and assessed the study quality according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Outcomes included summary C-statistics and model risk of bias assessed with the Prediction Model Risk of Bias Assessment Tool. C-statistics were pooled using a random-effects model.

RESULTS

Twenty-one studies (118,153 patients) employing various ML algorithms (76 models) were included in the systematic review. Predictive ability of models varied: 11.8% inadequate (C-statistic <0.60), 26.3% adequate (C-statistic 0.60-0.70), 31.6% acceptable (C-statistic 0.70-0.80), and 30.3% demonstrated excellent (C-statistic >0.80) performance. Meta-analyses revealed excellent predictive performance for early mortality (C-statistic: 0.81; 95% confidence interval [CI], 0.65-0.91), acceptable performance for 1-year mortality (C-statistic: 0.76; 95% CI, 0.67-0.84), and acceptable performance for predicting permanent pacemaker implantation (C-statistic: 0.75; 95% CI, 0.51-0.90).

CONCLUSIONS

ML models for TAVI outcomes exhibit adequate-to-excellent performance, suggesting potential clinical utility. We identified concerns in methodology and transparency, emphasizing the need for improved scientific reporting standards.

Preprocedural CT angiography and machine learning for mortality prediction after transcatheter aortic valve replacement

Prediction of outcomes following transcatheter aortic valve replacement (TAVR) is challenging. Considering that in aortic stenosis outcomes are governed by both valve degeneration and myocardial adverse remodeling, we aimed to evaluate machine-learning leveraging pre-procedural computed tomography (CT) for the prediction of 1-year mortality following TAVR. The analysis included data of consecutive patients who underwent TAVR at a high-volume center between January 2017 and January 2022 and was externally validated on unseen data from 3 international sites. Machine learning by extreme gradient boosting was trained and tested using clinical variables, CT-derived volumetric measurements including myocardial mass, and quantitative fibrocalcific aortic valve characteristics measured using standardized software. The EuroScore II and a separate machine learning risk score based exclusively on baseline clinical characteristics served as comparators. The derivation cohort included 631 consecutive patients (48 % men, 80 ± 8 years old, EuroSCORE II 6.5 [4.6-10.3] %). Machine learning was externally validated on data of 596 patients (48 % men, 81 ± 8 years old, EuroSCORE II 5.4 [4.7-8.1] %). In external validation, the machine learning prognostic risk score had an area under the receiver operator curve of 0.79 (0.74-0.84) which was superior to the EuroSCORE 0.59 (0.53-0.66), and the machine learning risk based on clinical data alone 0.64 (0.59-0.69), p < 0.001 for difference. Machine-learning integrating clinical data and CT-derived imaging characteristics was found to predict 1-year all-cause mortality following TAVR significantly better than clinical variables or clinical risk scores alone; and can help identify patients at higher prognostic risk prior to the procedure.

Artificial Intelligence in Cardiology: General Perspectives and Focus on Interventional Cardiology

Artificial intelligence (AI) is being intensively applied to cardiology, particularly in diagnostics, risk prediction, treatment planning, and invasive procedures. While AI-driven advancements have demonstrated promise, their real-world implementation remains constrained by critical challenges. Current AI applications, such as electrocardiogram interpretation and automated imaging analysis, have improved diagnostic accuracy and workflow efficiency, yet generalizability, regulatory hurdles, and integration into existing clinical workflows remain major obstacles. Algorithmic bias and the lack of explainable AI further complicate widespread adoption, potentially leading to disparities in healthcare outcomes. In interventional cardiology, robotic-assisted percutaneous coronary intervention has emerged as a technological innovation, but comparative clinical evidence supporting its superiority (or even non-inferiority) over conventional approaches is still limited. Additionally, AI-based decision support systems in high-risk cardiovascular procedures require rigorous validation to ensure safety and reliability. Ethical considerations, including patient data security and region-specific regulatory frameworks, also pose significant barriers. Addressing these challenges requires interdisciplinary collaboration, robust external validation, and the development of transparent, interpretable AI models. This review provides a critical appraisal of the current role of AI in cardiology, emphasizing both its potential and its limitations, and outlines future directions to facilitate its responsible integration into clinical practice.

Harnessing Electronic Health Records and Artificial Intelligence for Enhanced Cardiovascular Risk Prediction: A Comprehensive Review

Electronic health records (EHR) have revolutionized cardiovascular disease (CVD) research by enabling comprehensive, large-scale, and dynamic data collection. Integrating EHR data with advanced analytical methods, including artificial intelligence (AI), transforms CVD risk prediction and management methodologies. This review examines the advancements and challenges of using EHR in developing CVD prediction models, covering traditional and AI-based approaches. While EHR-based CVD risk prediction has greatly improved, moving from models that integrate real-world data on medication use and imaging, challenges persist regarding data quality, standardization across health care systems, and geographic variability. The complexity of EHR data requires sophisticated computational methods and multidisciplinary approaches for effective CVD risk modeling. AI's deep learning enhances prediction performance but faces limitations in interpretability and the need for validation and recalibration for diverse populations. The future of CVD risk prediction and management increasingly depends on using EHR and AI technologies effectively. Addressing data quality issues and overcoming limitations from retrospective data analysis are critical for improving the reliability and applicability of risk prediction models. Integrating multidimensional data, including environmental, lifestyle, social, and genomic factors, could significantly enhance risk assessment. These models require continuous validation and recalibration to ensure their adaptability to diverse populations and evolving health care environments, providing reassurance about their reliability.

Profiling of Cardiogenic Shock: Incorporating Machine Learning Into Bedside Management

Cardiogenic shock (CS) is a complex clinical syndrome with various etiologies and clinical presentations. Despite advances in therapeutic options, mortality remains high, and clinical trials in the field are complicated in part by the heterogeneity of CS patients. More individualized targeted therapeutic approaches might improve outcomes in CS, but their implementation remains challenging. The present review discusses current and emerging machine learning-based approaches, including unsupervised and supervised learning methods that use real-world clinical data to individualize therapeutic strategies for CS patients. We will discuss the rationale for each approach, potential advantages and disadvantages, and how these strategies can inform clinical trial design and management decisions.

Prediction of sudden cardiac death using artificial intelligence: Current status and future directions

Sudden cardiac death (SCD) remains a pressing health issue, affecting hundreds of thousands each year globally. The heterogeneity among people who suffer a SCD, ranging from individuals with severe heart failure to seemingly healthy individuals, poses a significant challenge for effective risk assessment. Conventional risk stratification, which primarily relies on left ventricular ejection fraction, has resulted in only modest efficacy of implantable cardioverter-defibrillators for SCD prevention. In response, artificial intelligence (AI) holds promise for personalized SCD risk prediction and tailoring preventive strategies to the unique profiles of individual patients. Machine and deep learning algorithms have the capability to learn intricate nonlinear patterns between complex data and defined end points, and leverage these to identify subtle indicators and predictors of SCD that may not be apparent through traditional statistical analysis. However, despite the potential of AI to improve SCD risk stratification, there are important limitations that need to be addressed. We aim to provide an overview of the current state-of-the-art of AI prediction models for SCD, highlight the opportunities for these models in clinical practice, and identify the key challenges hindering widespread adoption.

Machine learning-based scoring system to predict cardiogenic shock in acute coronary syndrome

Aims

Cardiogenic shock (CS) is a severe complication of acute coronary syndrome (ACS) with mortality rates approaching 50%. The ability to identify high-risk patients prior to the development of CS may allow for pre-emptive measures to prevent the development of CS. The objective was to derive and externally validate a simple, machine learning (ML)-based scoring system using variables readily available at first medical contact to predict the risk of developing CS during hospitalization in patients with ACS.

Methods and results

Observational multicentre study on ACS patients hospitalized at intensive care units. Derivation cohort included over 40 000 patients from Beth Israel Deaconess Medical Center, Boston, USA. Validation cohort included 5123 patients from the Sheba Medical Center, Ramat Gan, Israel. The final derivation cohort consisted of 3228 and the final validation cohort of 4904 ACS patients without CS at hospital admission. Development of CS was adjudicated manually based on the patients' reports. From nine ML models based on 13 variables (heart rate, respiratory rate, oxygen saturation, blood glucose level, systolic blood pressure, age, sex, shock index, heart rhythm, type of ACS, history of hypertension, congestive heart failure, and hypercholesterolaemia), logistic regression with elastic net regularization had the highest externally validated predictive performance (c-statistics: 0.844, 95% CI, 0.841-0.847).

Conclusion

STOP SHOCK score is a simple ML-based tool available at first medical contact showing high performance for prediction of developing CS during hospitalization in ACS patients. The web application is available at https://stopshock.org/#calculator.

Deep Learning-Enabled Assessment of Right Ventricular Function Improves Prognostication After Transcatheter Edge-to-Edge Repair for Mitral Regurgitation

BACKGROUND

Right ventricular (RV) function has a well-established prognostic role in patients with severe mitral regurgitation (MR) undergoing transcatheter edge-to-edge repair (TEER) and is typically assessed using echocardiography-measured tricuspid annular plane systolic excursion. Recently, a deep learning model has been proposed that accurately predicts RV ejection fraction (RVEF) from 2-dimensional echocardiographic videos, with similar diagnostic accuracy as 3-dimensional imaging. This study aimed to evaluate the prognostic value of the deep learning-predicted RVEF values in patients with severe MR undergoing TEER.

METHODS

This multicenter registry study analyzed the associations between the predicted RVEF values and 1-year mortality in patients with severe MR undergoing TEER. To predict RVEF, 2-dimensional apical 4-chamber view videos from preprocedural transthoracic echocardiographic studies were exported and processed by a rigorously validated deep learning model.

RESULTS

Good-quality 2-dimensional apical 4-chamber view videos could be retrieved for 1154 patients undergoing TEER between 2017 and 2023. Survival at 1 year after TEER was 84.7%. The predicted RVEF values ranged from 26.6% to 64.0% and correlated only modestly with tricuspid annular plane systolic excursion (Pearson R=0.33; P<0.001). Importantly, predicted RVEF was superior to tricuspid annular plane systolic excursion levels in predicting 1-year mortality after TEER (area under the curve, 0.687 versus 0.625; P=0.029). Furthermore, Kaplan-Meier survival analysis revealed that patients with reduced RV function (n=723; defined as a predicted RVEF of <45%) had significantly worse 1-year survival rates than patients with preserved RV function (n=431; defined as a predicted RVEF of ≥45%; 80.3% [95% CI, 77.4%-83.3%] versus 92.1% [95% CI, 89.5%-94.7%]; hazard ratio for 1-year mortality, 2.67 [95% CI, 1.82-3.90]; P<0.001).

CONCLUSIONS

Deep learning-enabled assessment of RV function using standard 2-dimensional echocardiographic videos can refine the prognostication of patients with severe MR undergoing TEER. Thus, it can be used to screen for patients with RV dysfunction who might benefit from intensified follow-up care.

Fluid status assessment in heart failure patients: pilot validation of the Maastricht Decompensation Questionnaire

BACKGROUND

eHealth products have the potential to enhance heart failure (HF) care by identifying at-risk patients. However, existing risk models perform modestly and require extensive data, limiting their practical application in clinical settings. This study aims to address this gap by validating a more suitable risk model for eHealth integration.

METHODS

We developed the Maastricht Decompensation Questionnaire (MDQ) based on expert opinion to assess HF patients' fluid status using common signs and symptoms. Subsequently, the MDQ was administered to a cohort of HF outpatients at Maastricht University Medical Centre. Patients with ≥ 10 MDQ points were categorised as 'decompensated', patients with < 10 MDQ points as 'not decompensated'. HF nurses, blinded to MDQ scores, served as the gold standard for fluid status assessment. Patients were classified as 'correctly' if MDQ and nurse assessments aligned; otherwise, they were classified as 'incorrectly'.

RESULTS

A total of 103 elderly HF patients were included. The MDQ classified 50 patients as 'decompensated', with 17 of them being correctly classified (34%). Additionally, 53 patients were categorised as 'not decompensated', with 48 of them being correctly classified (90%). The calculated area under the curve was 0.69 (95% confidence interval: 0.57-0.81; p < 0.05). Cronbach's alpha reliability coefficient for the MDQ was 0.85.

CONCLUSIONS

The MDQ helps identify decompensated HF patients through clinical signs and symptoms. Further trials with larger samples are needed to confirm its validity, reliability and applicability. Tailoring the MDQ to individual patient profiles may improve its accuracy.

The power of artificial intelligence for managing pandemics: A primer for public health professionals

Artificial intelligence (AI) applications are complex and rapidly evolving, and thus often poorly understood, but have potentially profound implications for public health. We offer a primer for public health professionals that explains some of the key concepts involved and examines how these applications might be used in the response to a future pandemic. They include early outbreak detection, predictive modelling, healthcare management, risk communication, and health surveillance. Artificial intelligence applications, especially predictive algorithms, have the ability to anticipate outbreaks by integrating diverse datasets such as social media, meteorological data, and mobile phone movement data. Artificial intelligence-powered tools can also optimise healthcare delivery by managing the allocation of resources and reducing healthcare workers' exposure to risks. In resource distribution, they can anticipate demand and optimise logistics, while AI-driven robots can minimise physical contact in healthcare settings. Artificial intelligence also shows promise in supporting public health decision-making by simulating the social and economic impacts of different policy interventions. These simulations help policymakers evaluate complex scenarios such as lockdowns and resource allocation. Additionally, it can enhance public health messaging, with AI-generated health communications shown to be more effective than human-generated messages in some cases. However, there are risks, such as privacy concerns, biases in models, and the potential for 'false confirmations', where AI reinforces incorrect decisions. Despite these challenges, we argue that AI will become increasingly important in public health crises, but only if integrated thoughtfully into existing systems and processes.

Systematic evaluation of machine learning models for postoperative surgical site infection prediction

BACKGROUND

Surgical site infections (SSIs) lead to increased mortality and morbidity, as well as increased healthcare costs. Multiple models for the prediction of this serious surgical complication have been developed, with an increasing use of machine learning (ML) tools.

OBJECTIVE

The aim of this systematic review was to assess the performance as well as the methodological quality of validated ML models for the prediction of SSIs.

METHODS

A systematic search in PubMed, Embase and the Cochrane library was performed from inception until July 2023. Exclusion criteria were the absence of reported model validation, SSIs as part of a composite adverse outcome, and pediatric populations. ML performance measures were evaluated, and ML performances were compared to regression-based methods for studies that reported both methods. Risk of bias (ROB) of the studies was assessed using the Prediction model Risk of Bias Assessment Tool.

RESULTS

Of the 4,377 studies screened, 24 were included in this review, describing 85 ML models. Most models were only internally validated (81%). The C-statistic was the most used performance measure (reported in 96% of the studies) and only two studies reported calibration metrics. A total of 116 different predictors were described, of which age, steroid use, sex, diabetes, and smoking were most frequently (100% to 75%) incorporated. Thirteen studies compared ML models to regression-based models and showed a similar performance of both modelling methods. For all included studies, the overall ROB was high or unclear.

CONCLUSIONS

A multitude of ML models for the prediction of SSIs are available, with large variability in performance. However, most models lacked external validation, performance was reported limitedly, and the risk of bias was high. In studies describing both ML models and regression-based models, one modelling method did not outperform the other.

Comprehensive Analysis of Cardiovascular Diseases: Symptoms, Diagnosis, and AI Innovations

Cardiovascular diseases are some of the underlying reasons contributing to the relentless rise in mortality rates across the globe. In this regard, there is a genuine need to integrate advanced technologies into the medical realm to detect such diseases accurately. Moreover, numerous academic studies have been published using AI-based methodologies because of their enhanced accuracy in detecting heart conditions. This research extensively delineates the different heart conditions, e.g., coronary artery disease, arrhythmia, atherosclerosis, mitral valve prolapse/mitral regurgitation, and myocardial infarction, and their underlying reasons and symptoms and subsequently introduces AI-based detection methodologies for precisely classifying such diseases. The review shows that the incorporation of artificial intelligence in detecting heart diseases exhibits enhanced accuracies along with a plethora of other benefits, like improved diagnostic accuracy, early detection and prevention, reduction in diagnostic errors, faster diagnosis, personalized treatment schedules, optimized monitoring and predictive analysis, improved efficiency, and scalability. Furthermore, the review also indicates the conspicuous disparities between the results generated by previous algorithms and the latest ones, paving the way for medical researchers to ascertain the accuracy of these results through comparative analysis with the practical conditions of patients. In conclusion, AI in heart disease detection holds paramount significance and transformative potential to greatly enhance patient outcomes, mitigate healthcare expenditure, and amplify the speed of diagnosis.

Prediction of incident atrial fibrillation using deep learning, clinical models, and polygenic scores

BACKGROUND AND AIMS

Deep learning applied to electrocardiograms (ECG-AI) is an emerging approach for predicting atrial fibrillation or flutter (AF). This study introduces an ECG-AI model developed and tested at a tertiary cardiac centre, comparing its performance with clinical models and AF polygenic score (PGS).

METHODS

Electrocardiograms in sinus rhythm from the Montreal Heart Institute were analysed, excluding those from patients with pre-existing AF. The primary outcome was incident AF at 5 years. An ECG-AI model was developed by splitting patients into non-overlapping data sets: 70% for training, 10% for validation, and 20% for testing. The performance of ECG-AI, clinical models, and PGS was assessed in the test data set. The ECG-AI model was externally validated in the Medical Information Mart for Intensive Care-IV (MIMIC-IV) hospital data set.

RESULTS

A total of 669 782 ECGs from 145 323 patients were included. Mean age was 61 ± 15 years, and 58% were male. The primary outcome was observed in 15% of patients, and the ECG-AI model showed an area under the receiver operating characteristic (AUC-ROC) curve of .78. In time-to-event analysis including the first ECG, ECG-AI inference of high risk identified 26% of the population with a 4.3-fold increased risk of incident AF (95% confidence interval: 4.02-4.57). In a subgroup analysis of 2301 patients, ECG-AI outperformed CHARGE-AF (AUC-ROC = .62) and PGS (AUC-ROC = .59). Adding PGS and CHARGE-AF to ECG-AI improved goodness of fit (likelihood ratio test P < .001), with minimal changes to the AUC-ROC (.76-.77). In the external validation cohort (mean age 59 ± 18 years, 47% male, median follow-up 1.1 year), ECG-AI model performance remained consistent (AUC-ROC = .77).

CONCLUSIONS

ECG-AI provides an accurate tool to predict new-onset AF in a tertiary cardiac centre, surpassing clinical and PGS.

An efficient cardiovascular disease prediction model through AI-driven IoT technology

Conditions affecting the circulatory system and blood vessels are referred to as cardiovascular diseases that include strokes and heart attacks. Internet of Things (IoT) technologies monitor health metrics, identify irregularities and enable remote patient care, resulting in earlier intervention and more individualized therapy. This research aims to establish an efficient cardiovascular disease prediction model through Artificial intelligence (AI)-driven IoT technology. We propose a novel Shuffled Frog leaping-tuned Iterative Improved Adaptive Boosting (SF-IIAdaboost) algorithm for predicting cardiovascular disease with the implementation of IoT device data. IoT medical sensors and wearable devices will collect the patient's clinical data in our proposed framework. Z-score normalization is used to preprocess the gathered data and optimize its quality. Kernel principal component analysis (Kernel-PCA) extracts the relevant features from the processed data. We obtained a dataset that contains various health data gathered from numerous sensing devices to train our recommended model. Our proposed methodology is implemented using Python software. During the evaluation phase, we assess the effectiveness of our model across different parameters. We conduct comparative analyses against conventional methods to ascertain the superiority of our approach. Experimental findings demonstrate the superior performance of our recognition method over traditional approaches. The proposed SF-IIAdaboost algorithm, integrated with IoT device data, presents a promising avenue for predicting cardiovascular disease. The SF-IIAdaboost model demonstrated notable enhancements, attaining 95.37 % accuracy, 93.51 % precision, 94.3 % sensitivity, 96.31 % specificity, and 95.72 % F-measure. Future developments are predicted to involve computing on the edge, where immediate evaluations can be performed in the edge layer to avoid the basic constraints of the clouds, such as high latency, utilization of bandwidth and performing the growth of IoT data. Edge computing can revolutionize the healthcare industry's efficacy by enabling providers to make flexible decisions, operate quickly, and accurately anticipate diseases. It can improve the average level of service standards.

Machine learning for outcome prediction in patients with non-valvular atrial fibrillation from the GLORIA-AF registry

Clinical risk scores that predict outcomes in patients with atrial fibrillation (AF) have modest predictive value. Machine learning (ML) may achieve greater results when predicting adverse outcomes in patients with recently diagnosed AF. Several ML models were tested and compared with current clinical risk scores on a cohort of 26,183 patients (mean age 70.13 (standard deviation 10.13); 44.8% female) with non-valvular AF. Inputted into the ML models were 23 demographic variables alongside comorbidities and current treatments. For one-year stroke prediction, ML achieved an area under the curve (AUC) of 0.653 (95% confidence interval 0.576-0.730), compared to the CHADS2 and CHA2DS2-VASc scores performance of 0.587 (95% CI 0.559-0.615) and 0.535 (95% CI 0.521-0.550), respectively. Using ML for one-year major bleed prediction increased the AUC from 0.537 (95% CI 0.518-0.557) generated by the HAS-BLED score to 0.677 (95% CI 0.619-0.724). ML was able to predict one-year and three-year all-cause mortality with an AUC of 0.734 (95% CI 0.696-0.771) and 0.742 (95% CI 0.718-0.766). In this study a significant improvement in performance was observed when transitioning from clinical risk scores to machine learning-based approaches across all applications tested. Obtaining precise prediction tools is desirable for increased interventions to reduce event rates.Trial Registry https://www.clinicaltrials.gov ; Unique identifier: NCT01468701, NCT01671007, NCT01937377.

Artificial intelligence in cardiovascular medicine: clinical applications

Clinical medicine requires the integration of various forms of patient data including demographics, symptom characteristics, electrocardiogram findings, laboratory values, biomarker levels, and imaging studies. Decision-making on the optimal management should be based on a high probability that the envisaged treatment is appropriate, provides benefit, and bears no or little potential harm. To that end, personalized risk-benefit considerations should guide the management of individual patients to achieve optimal results. These basic clinical tasks have become more and more challenging with the massively growing data now available; artificial intelligence and machine learning (AI/ML) can provide assistance for clinicians by obtaining and comprehensively preparing the history of patients, analysing face and voice and other clinical features, by integrating laboratory results, biomarkers, and imaging. Furthermore, AI/ML can provide a comprehensive risk assessment as a basis of optimal acute and chronic care. The clinical usefulness of AI/ML algorithms should be carefully assessed, validated with confirmation datasets before clinical use, and repeatedly re-evaluated as patient phenotypes change. This review provides an overview of the current data revolution that has changed and will continue to change the face of clinical medicine radically, if properly used, to the benefit of physicians and patients alike.

A scoping review on advancements in noninvasive wearable technology for heart failure management

Wearables offer a promising solution for enhancing remote monitoring (RM) of heart failure (HF) patients by tracking key physiological parameters. Despite their potential, their clinical integration faces challenges due to the lack of rigorous evaluations. This review aims to summarize the current evidence and assess the readiness of wearables for clinical practice using the Medical Device Readiness Level (MDRL). A systematic search identified 99 studies from 3112 found articles, with only eight being randomized controlled trials. Accelerometery was the most used measurement technique. Consumer-grade wearables, repurposed for HF monitoring, dominated the studies with most of them in the feasibility testing stage (MDRL 6). Only two of the described wearables were specifically designed for HF RM, and received FDA approval. Consequently, the actual impact of wearables on HF management remains uncertain due to limited robust evidence, posing a significant barrier to their integration into HF care.

An Artificial Intelligence Algorithm for Detection of Severe Aortic Stenosis: A Clinical Cohort Study

Background

Identifying individuals with severe aortic stenosis (AS) at high risk of mortality remains challenging using current clinical imaging methods.

Objectives

The purpose of this study was to evaluate an artificial intelligence decision support algorithm (AI-DSA) to augment the detection of severe AS within a well-resourced health care setting.

Methods

Agnostic to clinical information, an AI-DSA trained to identify echocardiographic phenotype associated with an aortic valve area (AVA)<1 cm2 using minimal input data (excluding left ventricular outflow tract measures) was applied to routine transthoracic echocardiograms (TTE) reports from 31,141 U.S. Medicare beneficiaries at an academic medical center (2003-2017).

Results

Performance of AI-DSA to detect the phenotype associated with an AVA<1 cm2 was excellent (sensitivity 82.2%, specificity 98.1%, negative predictive value 9.2%, c-statistic = 0.986). In addition to identifying clinical severe AS cases, AI-DSA identified an additional 1,034 (3.3%) individuals with guideline-defined moderate AS but with a similar clinical and TTE phenotype to those with severe AS with low rates of aortic valve replacement (6.6%). Five-year mortality was 75.9% in those with known severe AS, 73.5% in those with a similar phenotype to severe AS, and 44.6% in those without severe AS. The AI-DSA continued to perform well to identify severe AS among those with a depressed left ventricular ejection fraction. Overall rates of aortic valve replacement remained low, even in those with an AVA<1 cm2 (21.9%).

Conclusions

Without relying on left ventricular outflow tract measurements, an AI-DSA used echocardiographic reports to reliably identify the phenotype of severe AS. These results suggest possible utility for this AI-DSA to enhance detection of severe AS individuals at risk for adverse outcomes.

Don't be misled: 3 misconceptions about external validation of clinical prediction models

Clinical prediction models provide risks of health outcomes that can inform patients and support medical decisions. However, most models never make it to actual implementation in practice. A commonly heard reason for this lack of implementation is that prediction models are often not externally validated. While we generally encourage external validation, we argue that an external validation is often neither sufficient nor required as an essential step before implementation. As such, any available external validation should not be perceived as a license for model implementation. We clarify this argument by discussing 3 common misconceptions about external validation. We argue that there is not one type of recommended validation design, not always a necessity for external validation, and sometimes a need for multiple external validations. The insights from this paper can help readers to consider, design, interpret, and appreciate external validation studies.

Cardiac sympathetic activity and lethal arrhythmic events: insight into bell-shaped relationship between 123I-meta-iodobenzylguanidine activity and event rates

BACKGROUND

123I-meta-iodobenzylguanidine (mIBG) has been applied to patients with chronic heart failure (CHF). However, the relationship between 123I-mIBG activity and lethal arrhythmic events (ArE) is not well defined. This study aimed to determine this relationship in Japanese and European cohorts.

RESULTS

We calculated heart-to-mediastinum (H/M) count ratios and washout rates (WRs) of 827 patients using planar 123I-mIBG imaging. We defined ArEs as sudden cardiac death, arrhythmic death, and potentially lethal events such as sustained ventricular tachycardia, cardiac arrest with resuscitation, and appropriate implantable cardioverter defibrillator (ICD) discharge, either from a single ICD or as part of a cardiac resynchronization therapy device (CRTD). We analyzed the incidence of ArE with respect to H/M ratios, WRs and New York Heart Association (NYHA) functional classes among Japanese (J; n = 581) and European (E; n = 246) cohorts. We also simulated ArE rates versus H/M ratios under specific conditions using a machine-learning model incorporating 13 clinical variables. Consecutive patients with CHF were selected in group J, whereas group E comprised candidates for cardiac electronic devices. Groups J and E mostly comprised patients with NYHA functional classes I/II (95%) and II/III (91%), respectively, and 21% and 72% were respectively implanted with ICD/CRTD devices. The ArE rate increased with lower H/M ratios in group J, but the relationship was bell-shaped, with a high ArE rate within the intermediate H/M range, in group E. This bell-shaped curve was also evident in patients with NYHA classes II/III in the combined J and E groups, particularly in those with a high (> 15%) mIBG WR and with ischemic, but not in those with non-ischemic etiologies. Machine learning-based prediction of ArE risk aligned with these findings, indicating a bell-shaped curve in NYHA class II/III but not in class I.

CONCLUSIONS

The relationship between cardiac 123I-mIBG activity and lethal arrhythmic events is influenced by the background of patients. The bell-shaped relationship in NYHA classes II/III, high WR, and ischemic etiology likely aids in identifying patients at high risk for ArEs.

Artificial Intelligence in Cardiovascular Care-Part 2: Applications: JACC Review Topic of the Week

Recent artificial intelligence (AI) advancements in cardiovascular care offer potential enhancements in effective diagnosis, treatment, and outcomes. More than 600 U.S. Food and Drug Administration-approved clinical AI algorithms now exist, with 10% focusing on cardiovascular applications, highlighting the growing opportunities for AI to augment care. This review discusses the latest advancements in the field of AI, with a particular focus on the utilization of multimodal inputs and the field of generative AI. Further discussions in this review involve an approach to understanding the larger context in which AI-augmented care may exist, and include a discussion of the need for rigorous evaluation, appropriate infrastructure for deployment, ethics and equity assessments, regulatory oversight, and viable business cases for deployment. Embracing this rapidly evolving technology while setting an appropriately high evaluation benchmark with careful and patient-centered implementation will be crucial for cardiology to leverage AI to enhance patient care and the provider experience.

Prognostic models for patients suffering a heart failure with a preserved ejection fraction: a systematic review

The purpose of this study was to systematically review the development, performance, and applicability of prognostic models developed for predicting poor events in patients with heart failure with preserved ejection fraction (HFpEF). Databases including Embase, PubMed, Web of Science Core Collection, the Cochrane Library, China National Knowledge Infrastructure, Wan Fang, Wei Pu, and China Biological Medicine were queried from their respective dates of inception to 1 June 2023, to examine multivariate models for prognostic prediction in HFpEF. Both forward and backward citations of all studies were included in our analysis. Two researchers individually used the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist to extract data and assess the quality of the models using the Predictive Mode Bias Risk Assessment Tool (PROBAST). Among the 6897 studies screened, 16 studies derived and/or validated a total of 39 prognostic models. The sample size ranges for model development, internal validation, and external validation are 119 to 5988, 152 to 1000, and 30 to 5957, respectively. The most frequently employed modelling technique was Cox proportional hazards regression. Six studies (37.50%) conducted internal validation of models; bootstrap and k-fold cross-validation were the commonly used methods for internal validation of models. Ten of these models (25.64%) were validated externally, with reported the c-statistic in the external validation set ranging from 0.70 to 0.96, while the remaining models await external validation. The MEDIA echo score and I-PRESERVE-sudden cardiac death prediction mode have been externally validated using multiple cohorts, and the results consistently show good predictive performance. The most frequently used predictors identified among the models were age, n-terminal pro-brain natriuretic peptide, ejection fraction, albumin, and hospital stay in the last 5 months owing to heart failure. All study predictor domains and outcome domains were at low risk of bias, high or unclear risk of bias of all prognostic models due to underreporting in the area of analysis. All studies did not evaluate the clinical utility of the prognostic models. Predictive models for predicting prognostic outcomes in patients with HFpEF showed good discriminatory ability but their utility and generalization remain uncertain due to the risk of bias, differences in predictors between models, and the lack of clinical application studies. Future studies should improve the methodological quality of model development and conduct external validation of models.

Prediction of diagnosis and diastolic filling pressure by AI-enhanced cardiac MRI: a modelling study of hospital data

BACKGROUND

With increasing numbers of patients and novel drugs for distinct causes of systolic and diastolic heart failure, automated assessment of cardiac function is important. We aimed to provide a non-invasive method to predict diagnosis of patients undergoing cardiac MRI (cMRI) and to obtain left ventricular end-diastolic pressure (LVEDP).

METHODS

For this modelling study, patients who had undergone cardiac catheterisation at University Hospital Heidelberg (Heidelberg, Germany) between July 15, 2004 and March 16, 2023, were identified, as were individual left ventricular pressure measurements. We used existing patient data from routine cardiac diagnostics. From this initial group, we extracted patients who had been diagnosed with ischaemic cardiomyopathy, dilated cardiomyopathy, hypertrophic cardiomyopathy, or amyloidosis, as well as control individuals with no structural phenotype. Data were pseudonymised and only processed within the university hospital's AI infrastructure. We used the data to build different models to predict either demographic (ie, AI-age and AI-sex), diagnostic (ie, AI-coronary artery disease and AI-cardiomyopathy [AI-CMP]), or functional parameters (ie, AI-LVEDP). We randomly divided our datasets via computer into training, validation, and test datasets. AI-CMP was not compared with other models, but was validated in a prospective setting. Benchmarking was also done.

FINDINGS

66 936 patients who had undergone cardiac catheterisation at University Hospital Heidelberg were identified, with more than 183 772 individual left ventricular pressure measurements. We extracted 4390 patients from this initial group, of whom 1131 (25·8%) had been diagnosed with ischaemic cardiomyopathy, 1064 (24·2%) had been diagnosed with dilated cardiomyopathy, 816 (18·6%) had been diagnosed with hypertrophic cardiomyopathy, 202 (4·6%) had been diagnosed with amyloidosis, and 1177 (26·7%) were control individuals with no structural phenotype. The core cohort only included patients with cardiac catherisation and cMRI within 30 days, and emergency cases were excluded. AI-sex was able to predict patient sex with areas under the receiver operating characteristic curves (AUCs) of 0·78 (95% CI 0·77-0·78) and AI-age was able to predict patient age with a mean absolute error of 7·86 years (7·77-7·95), with a Pearson correlation of 0·57 (95% CI 0·56-0·57). The AUCs for the classification tasks ranged between 0·82 (95% CI 0·79-0·84) for ischaemic cardiomyopathy and 0·92 (0·91-0·94) for hypertrophic cardiomyopathy.

INTERPRETATION

Our AI models could be easily integrated into clinical practice and provide added value to the information content of cMRI, allowing for disease classification and prediction of diastolic function.

FUNDING

Informatics for Life initiative of the Klaus-Tschira Foundation, German Center for Cardiovascular Research, eCardiology section of the German Cardiac Society, and AI Health Innovation Cluster Heidelberg.

Machine learning models for assessing risk factors affecting health care costs: 12-month exercise-based cardiac rehabilitation

Introduction

Exercise-based cardiac rehabilitation (ECR) has proven to be effective and cost-effective dominant treatment option in health care. However, the contribution of well-known risk factors for prognosis of coronary artery disease (CAD) to predict health care costs is not well recognized. Since machine learning (ML) applications are rapidly giving new opportunities to assist health care professionals' work, we used selected ML tools to assess the predictive value of defined risk factors for health care costs during 12-month ECR in patients with CAD.

Methods

The data for analysis was available from a total of 71 patients referred to Oulu University Hospital, Finland, due to an acute coronary syndrome (ACS) event (75% men, age 61 ± 12 years, BMI 27 ± 4 kg/m2, ejection fraction 62 ± 8, 89% have beta-blocker medication). Risk factors were assessed at the hospital immediately after the cardiac event, and health care costs for all reasons were collected from patient registers over a year. ECR was programmed in accordance with international guidelines. Risk analysis algorithms (cross-decomposition algorithms) were employed to rank risk factors based on variances in their effects. Regression analysis was used to determine the accounting value of risk factors by entering first the risk factor with the highest degree of explanation into the model. After that, the next most potent risk factor explaining costs was added to the model one by one (13 forecast models in total).

Results

The ECR group used health care services during the year at an average of 1,624 ± 2,139€ per patient. Diabetes exhibited the strongest correlation with health care expenses (r = 0.406), accounting for 16% of the total costs (p < 0.001). When the next two ranked markers (body mass index; r = 0.171 and systolic blood pressure; r = - 0.162, respectively) were added to the model, the predictive value was 18% for the costs (p = 0.004). The depression scale had the weakest independent explanation rate of all 13 risk factors (explanation value 0.1%, r = 0.029, p = 0.811).

Discussion

Presence of diabetes is the primary reason forecasting health care costs in 12-month ECR intervention among ACS patients. The ML tools may help decision-making when planning the optimal allocation of health care resources.

The use of artificial intelligence for predicting postinfarction myocardial viability in echocardiographic images

BACKGROUND

Evaluation of standard echocardiographic examination with artificial intelligence may help in the diagnosis of myocardial viability and function recovery after acute coronary syndrome.

METHODS

Sixty-one consecutive patients with acute coronary syndrome were enrolled in the present study (43 men, mean age 61 ± 9 years). All patients underwent percutaneous coronary intervention (PCI). 533 segments of the heart echo images were used. After 12 ± 1 months of follow-up, patients had an echocardiographic evaluation. After PCI each patient underwent cardiac magnetic resonance (CMR) with late enhancement and low-dose dobutamine echocardiographic examination. For texture analysis, custom software was used (MaZda 5.20, Institute of Electronics).Linear and non-linear (neural network) discriminative analyses were performed to identify the optimal analytic method correlating with CMR regarding the necrosis extent and viability prediction after follow-up. Texture parameters were analyzed using machine learning techniques: Artificial Neural Networks, Namely Multilayer Perceptron, Nonlinear Discriminant Analysis, Support Vector Machine, and Adaboost algorithm.

RESULTS

The mean concordance between the CMR definition of viability and three classification models in Artificial Neural Networks varied from 42% to 76%. Echo-based detection of non-viable tissue was more sensitive in the segments with the highest relative transmural scar thickness: 51-75% and 76-99%. The best results have been obtained for images with contrast for red and grey components (74% of proper classification). In dobutamine echocardiography, the results of appropriate prediction were 67% for monochromatic images.

CONCLUSIONS

Detection and semi-quantification of scar transmurality are feasible in echocardiographic images analyzed with artificial intelligence. Selected analytic methods yielded similar accuracy, and contrast enhancement contributed to the prediction accuracy of myocardial viability after myocardial infarction in 12 months of follow-up.

Advancements in Artificial Intelligence for Precision Diagnosis and Treatment of Myocardial Infarction: A Comprehensive Review of Clinical Trials and Randomized Controlled Trials

Coronary artery disease (CAD) is still a serious global health issue that has a substantial impact on death and illness rates. The goal of primary prevention strategies is to lower the risk of developing CAD. Nevertheless, current methods usually rely on simple risk assessment instruments that might overlook significant individual risk factors. This limitation highlights the need for innovative methods that can accurately assess cardiovascular risk and offer personalized preventive care. Recent advances in machine learning and artificial intelligence (AI) have opened up interesting new avenues for optimizing primary preventive efforts for CAD and improving risk prediction models. By leveraging large-scale databases and advanced computational techniques, AI has the potential to fundamentally alter how cardiovascular risk is evaluated and managed. This review looks at current randomized controlled studies and clinical trials that explore the application of AI and machine learning to improve primary preventive measures for CAD. The emphasis is on their ability to recognize and include a range of risk elements in sophisticated risk assessment models.

Use of Artificial Intelligence in Improving Outcomes in Heart Disease: A Scientific Statement From the American Heart Association

A major focus of academia, industry, and global governmental agencies is to develop and apply artificial intelligence and other advanced analytical tools to transform health care delivery. The American Heart Association supports the creation of tools and services that would further the science and practice of precision medicine by enabling more precise approaches to cardiovascular and stroke research, prevention, and care of individuals and populations. Nevertheless, several challenges exist, and few artificial intelligence tools have been shown to improve cardiovascular and stroke care sufficiently to be widely adopted. This scientific statement outlines the current state of the art on the use of artificial intelligence algorithms and data science in the diagnosis, classification, and treatment of cardiovascular disease. It also sets out to advance this mission, focusing on how digital tools and, in particular, artificial intelligence may provide clinical and mechanistic insights, address bias in clinical studies, and facilitate education and implementation science to improve cardiovascular and stroke outcomes. Last, a key objective of this scientific statement is to further the field by identifying best practices, gaps, and challenges for interested stakeholders.

Transforming the cardiometabolic disease landscape: Multimodal AI-powered approaches in prevention and management

The rise of artificial intelligence (AI) has revolutionized various scientific fields, particularly in medicine, where it has enabled the modeling of complex relationships from massive datasets. Initially, AI algorithms focused on improved interpretation of diagnostic studies such as chest X-rays and electrocardiograms in addition to predicting patient outcomes and future disease onset. However, AI has evolved with the introduction of transformer models, allowing analysis of the diverse, multimodal data sources existing in medicine today. Multimodal AI holds great promise in more accurate disease risk assessment and stratification as well as optimizing the key driving factors in cardiometabolic disease: blood pressure, sleep, stress, glucose control, weight, nutrition, and physical activity. In this article we outline the current state of medical AI in cardiometabolic disease, highlighting the potential of multimodal AI to augment personalized prevention and treatment strategies in cardiometabolic disease.

ECG-based data-driven solutions for diagnosis and prognosis of cardiovascular diseases: A systematic review

Cardiovascular diseases (CVD) are a leading cause of death globally, and result in significant morbidity and reduced quality of life. The electrocardiogram (ECG) plays a crucial role in CVD diagnosis, prognosis, and prevention; however, different challenges still remain, such as an increasing unmet demand for skilled cardiologists capable of accurately interpreting ECG. This leads to higher workload and potential diagnostic inaccuracies. Data-driven approaches, such as machine learning (ML) and deep learning (DL) have emerged to improve existing computer-assisted solutions and enhance physicians' ECG interpretation of the complex mechanisms underlying CVD. However, many ML and DL models used to detect ECG-based CVD suffer from a lack of explainability, bias, as well as ethical, legal, and societal implications (ELSI). Despite the critical importance of these Trustworthy Artificial Intelligence (AI) aspects, there is a lack of comprehensive literature reviews that examine the current trends in ECG-based solutions for CVD diagnosis or prognosis that use ML and DL models and address the Trustworthy AI requirements. This review aims to bridge this knowledge gap by providing a systematic review to undertake a holistic analysis across multiple dimensions of these data-driven models such as type of CVD addressed, dataset characteristics, data input modalities, ML and DL algorithms (with a focus on DL), and aspects of Trustworthy AI like explainability, bias and ethical considerations. Additionally, within the analyzed dimensions, various challenges are identified. To these, we provide concrete recommendations, equipping other researchers with valuable insights to understand the current state of the field comprehensively.

A comprehensive review on efficient artificial intelligence models for classification of abnormal cardiac rhythms using electrocardiograms

Deep learning has made many advances in data classification using electrocardiogram (ECG) waveforms. Over the past decade, data science research has focused on developing artificial intelligence (AI) based models that can analyze ECG waveforms to identify and classify abnormal cardiac rhythms accurately. However, the primary drawback of the current AI models is that most of these models are heavy, computationally intensive, and inefficient in terms of cost for real-time implementation. In this review, we first discuss the current state-of-the-art AI models utilized for ECG-based cardiac rhythm classification. Next, we present some of the upcoming modeling methodologies which have the potential to perform real-time implementation of AI-based heart rhythm diagnosis. These models hold significant promise in being lightweight and computationally efficient without compromising the accuracy. Contemporary models predominantly utilize 12-lead ECG for cardiac rhythm classification and cardiovascular status prediction, increasing the computational burden and making real-time implementation challenging. We also summarize research studies evaluating the potential of efficient data setups to reduce the number of ECG leads without affecting classification accuracy. Lastly, we present future perspectives on AI's utility in precision medicine by providing opportunities for accurate prediction and diagnostics of cardiovascular status in patients.

Deep learning-based prediction of major arrhythmic events in dilated cardiomyopathy: A proof of concept study

Prediction of major arrhythmic events (MAEs) in dilated cardiomyopathy represents an unmet clinical goal. Computational models and artificial intelligence (AI) are new technological tools that could offer a significant improvement in our ability to predict MAEs. In this proof-of-concept study, we propose a deep learning (DL)-based model, which we termed Deep ARrhythmic Prevention in dilated cardiomyopathy (DARP-D), built using multidimensional cardiac magnetic resonance data (cine videos and hypervideos and LGE images and hyperimages) and clinical covariates, aimed at predicting and tracking an individual patient's risk curve of MAEs (including sudden cardiac death, cardiac arrest due to ventricular fibrillation, sustained ventricular tachycardia lasting ≥30 s or causing haemodynamic collapse in <30 s, appropriate implantable cardiac defibrillator intervention) over time. The model was trained and validated in 70% of a sample of 154 patients with dilated cardiomyopathy and tested in the remaining 30%. DARP-D achieved a 95% CI in Harrell's C concordance indices of 0.12-0.68 on the test set. We demonstrate that our DL approach is feasible and represents a novelty in the field of arrhythmic risk prediction in dilated cardiomyopathy, able to analyze cardiac motion, tissue characteristics, and baseline covariates to predict an individual patient's risk curve of major arrhythmic events. However, the low number of patients, MAEs and epoch of training make the model a promising prototype but not ready for clinical usage. Further research is needed to improve, stabilize and validate the performance of the DARP-D to convert it from an AI experiment to a daily used tool.

Artificial intelligence in the risk prediction models of cardiovascular disease and development of an independent validation screening tool: a systematic review

BACKGROUND

A comprehensive overview of artificial intelligence (AI) for cardiovascular disease (CVD) prediction and a screening tool of AI models (AI-Ms) for independent external validation are lacking. This systematic review aims to identify, describe, and appraise AI-Ms of CVD prediction in the general and special populations and develop a new independent validation score (IVS) for AI-Ms replicability evaluation.

METHODS

PubMed, Web of Science, Embase, and IEEE library were searched up to July 2021. Data extraction and analysis were performed for the populations, distribution, predictors, algorithms, etc. The risk of bias was evaluated with the prediction risk of bias assessment tool (PROBAST). Subsequently, we designed IVS for model replicability evaluation with five steps in five items, including transparency of algorithms, performance of models, feasibility of reproduction, risk of reproduction, and clinical implication, respectively. The review is registered in PROSPERO (No. CRD42021271789).

RESULTS

In 20,887 screened references, 79 articles (82.5% in 2017-2021) were included, which contained 114 datasets (67 in Europe and North America, but 0 in Africa). We identified 486 AI-Ms, of which the majority were in development (n = 380), but none of them had undergone independent external validation. A total of 66 idiographic algorithms were found; however, 36.4% were used only once and only 39.4% over three times. A large number of different predictors (range 5-52,000, median 21) and large-span sample size (range 80-3,660,000, median 4466) were observed. All models were at high risk of bias according to PROBAST, primarily due to the incorrect use of statistical methods. IVS analysis confirmed only 10 models as "recommended"; however, 281 and 187 were "not recommended" and "warning," respectively.

CONCLUSION

AI has led the digital revolution in the field of CVD prediction, but is still in the early stage of development as the defects of research design, report, and evaluation systems. The IVS we developed may contribute to independent external validation and the development of this field.

Artificial intelligence in detecting left atrial appendage thrombus by transthoracic echocardiography and clinical features: the Left Atrial Thrombus on Transoesophageal Echocardiography (LATTEE) registry

AIMS

Transoesophageal echocardiography (TOE) is often performed before catheter ablation or cardioversion to rule out the presence of left atrial appendage thrombus (LAT) in patients on chronic oral anticoagulation (OAC), despite associated discomfort. A machine learning model [LAT-artificial intelligence (AI)] was developed to predict the presence of LAT based on clinical and transthoracic echocardiography (TTE) features.

METHODS AND RESULTS

Data from a 13-site prospective registry of patients who underwent TOE before cardioversion or catheter ablation were used. LAT-AI was trained to predict LAT using data from 12 sites (n = 2827) and tested externally in patients on chronic OAC from two sites (n = 1284). Areas under the receiver operating characteristic curve (AUC) of LAT-AI were compared with that of left ventricular ejection fraction (LVEF) and CHA2DS2-VASc score. A decision threshold allowing for a 99% negative predictive value was defined in the development cohort. A protocol where TOE in patients on chronic OAC is performed depending on the LAT-AI score was validated in the external cohort. In the external testing cohort, LAT was found in 5.5% of patients. LAT-AI achieved an AUC of 0.85 [95% confidence interval (CI): 0.82-0.89], outperforming LVEF (0.81, 95% CI 0.76-0.86, P < .0001) and CHA2DS2-VASc score (0.69, 95% CI: 0.63-0.7, P < .0001) in the entire external cohort. Based on the proposed protocol, 40% of patients on chronic OAC from the external cohort would safely avoid TOE.

CONCLUSION

LAT-AI allows accurate prediction of LAT. A LAT-AI-based protocol could be used to guide the decision to perform TOE despite chronic OAC.

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

BACKGROUND

The utilization of artificial intelligence and machine learning as diagnostic and predictive tools in perioperative medicine holds great promise. Indeed, many studies have been performed in recent years to explore the potential. The purpose of this systematic review is to assess the current state of machine learning in perioperative medicine, its utility in prediction of complications and prognostication, and limitations related to bias and validation.

METHODS

A multidisciplinary team of clinicians and engineers conducted a systematic review using the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) protocol. Multiple databases were searched, including Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Library, PubMed, Medline, Embase, and Web of Science. The systematic review focused on study design, type of machine learning model used, validation techniques applied, and reported model performance on prediction of complications and prognostication. This review further classified outcomes and machine learning applications using an ad hoc classification system. The Prediction model Risk Of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability of the studies.

RESULTS

A total of 103 studies were identified. The models reported in the literature were primarily based on single-center validations (75%), with only 13% being externally validated across multiple centers. Most of the mortality models demonstrated a limited ability to discriminate and classify effectively. The PROBAST assessment indicated a high risk of systematic errors in predicted outcomes and artificial intelligence or machine learning applications.

CONCLUSIONS

The findings indicate that the development of this field is still in its early stages. This systematic review indicates that application of machine learning in perioperative medicine is still at an early stage. While many studies suggest potential utility, several key challenges must be first overcome before their introduction into clinical practice.

EDITOR’S PERSPECTIVE

Discriminative fusion of moments-aligned latent representation of multimodality medical data

Fusion of multimodal medical data provides multifaceted, disease-relevant information for diagnosis or prognosis prediction modeling. Traditional fusion strategies such as feature concatenation often fail to learn hidden complementary and discriminative manifestations from high-dimensional multimodal data. To this end, we proposed a methodology for the integration of multimodality medical data by matching their moments in a latent space, where the hidden, shared information of multimodal data is gradually learned by optimization with multiple feature collinearity and correlation constrains. We first obtained the multimodal hidden representations by learning mappings between the original domain and shared latent space. Within this shared space, we utilized several relational regularizations, including data attribute preservation, feature collinearity and feature-task correlation, to encourage learning of the underlying associations inherent in multimodal data. The fused multimodal latent features were finally fed to a logistic regression classifier for diagnostic prediction. Extensive evaluations on three independent clinical datasets have demonstrated the effectiveness of the proposed method in fusing multimodal data for medical prediction modeling.

Five critical quality criteria for artificial intelligence-based prediction models

To raise the quality of clinical artificial intelligence (AI) prediction modelling studies in the cardiovascular health domain and thereby improve their impact and relevancy, the editors for digital health, innovation, and quality standards of the European Heart Journal propose five minimal quality criteria for AI-based prediction model development and validation studies: complete reporting, carefully defined intended use of the model, rigorous validation, large enough sample size, and openness of code and software.

Mapping the Heartbeat of America with ChatGPT-4: Unpacking the Interplay of Social Vulnerability, Digital Literacy, and Cardiovascular Mortality in County Residency Choices

Cardiovascular disease remains a leading cause of morbidity and mortality in the United States (US). Although high-quality data are accessible in the US for cardiovascular research, digital literacy (DL) has not been explored as a potential factor influencing cardiovascular mortality, although the Social Vulnerability Index (SVI) has been used previously as a variable in predictive modeling. Utilizing a large language model, ChatGPT4, we investigated the variability in CVD-specific mortality that could be explained by DL and SVI using regression modeling. We fitted two models to calculate the crude and adjusted CVD mortality rates. Mortality data using ICD-10 codes were retrieved from CDC WONDER, and the geographic level data was retrieved from the US Department of Agriculture. Both datasets were merged using the Federal Information Processing Standards code. The initial exploration involved data from 1999 through 2020 (n = 65,791; 99.98% complete for all US Counties) for crude cardiovascular mortality (CCM). Age-adjusted cardiovascular mortality (ACM) had data for 2020 (n = 3118 rows; 99% complete for all US Counties), with the inclusion of SVI and DL in the model (a composite of literacy and internet access). By leveraging on the advanced capabilities of ChatGPT4 and linear regression, we successfully highlighted the importance of incorporating the SVI and DL in predicting adjusted cardiovascular mortality. Our findings imply that just incorporating internet availability in the regression model may not be sufficient without incorporating significant variables, such as DL and SVI, to predict ACM. Further, our approach could enable future researchers to consider DL and SVI as key variables to study other health outcomes of public-health importance, which could inform future clinical practices and policies.

Personalized diagnosis in suspected myocardial infarction

BACKGROUND

In suspected myocardial infarction (MI), guidelines recommend using high-sensitivity cardiac troponin (hs-cTn)-based approaches. These require fixed assay-specific thresholds and timepoints, without directly integrating clinical information. Using machine-learning techniques including hs-cTn and clinical routine variables, we aimed to build a digital tool to directly estimate the individual probability of MI, allowing for numerous hs-cTn assays.

METHODS

In 2,575 patients presenting to the emergency department with suspected MI, two ensembles of machine-learning models using single or serial concentrations of six different hs-cTn assays were derived to estimate the individual MI probability (ARTEMIS model). Discriminative performance of the models was assessed using area under the receiver operating characteristic curve (AUC) and logLoss. Model performance was validated in an external cohort with 1688 patients and tested for global generalizability in 13 international cohorts with 23,411 patients.

RESULTS

Eleven routinely available variables including age, sex, cardiovascular risk factors, electrocardiography, and hs-cTn were included in the ARTEMIS models. In the validation and generalization cohorts, excellent discriminative performance was confirmed, superior to hs-cTn only. For the serial hs-cTn measurement model, AUC ranged from 0.92 to 0.98. Good calibration was observed. Using a single hs-cTn measurement, the ARTEMIS model allowed direct rule-out of MI with very high and similar safety but up to tripled efficiency compared to the guideline-recommended strategy.

CONCLUSION

We developed and validated diagnostic models to accurately estimate the individual probability of MI, which allow for variable hs-cTn use and flexible timing of resampling. Their digital application may provide rapid, safe and efficient personalized patient care.

TRIAL REGISTRATION NUMBERS

Data of following cohorts were used for this project: BACC ( www.

CLINICALTRIALS

gov ; NCT02355457), stenoCardia ( www.

CLINICALTRIALS

gov ; NCT03227159), ADAPT-BSN ( www.australianclinicaltrials.gov.au ; ACTRN12611001069943), IMPACT ( www.australianclinicaltrials.gov.au , ACTRN12611000206921), ADAPT-RCT ( www.anzctr.org.au ; ANZCTR12610000766011), EDACS-RCT ( www.anzctr.org.au ; ANZCTR12613000745741); DROP-ACS ( https://www.umin.ac.jp , UMIN000030668); High-STEACS ( www.

CLINICALTRIALS

gov ; NCT01852123), LUND ( www.

CLINICALTRIALS

gov ; NCT05484544), RAPID-CPU ( www.

CLINICALTRIALS

gov ; NCT03111862), ROMI ( www.

CLINICALTRIALS

gov ; NCT01994577), SAMIE ( https://anzctr.org.au ; ACTRN12621000053820), SEIGE and SAFETY ( www.

CLINICALTRIALS

gov ; NCT04772157), STOP-CP ( www.

CLINICALTRIALS

gov ; NCT02984436), UTROPIA ( www.

CLINICALTRIALS

gov ; NCT02060760).

Machine learning-powered, device-embedded heart sound measurement can optimize AV delay in patients with CRT

BACKGROUND

Continuous optimization of atrioventricular (AV) delay for cardiac resynchronization therapy (CRT) is mainly performed by electrical means.

OBJECTIVE

The purpose of this study was to develop an estimation model of cardiac function that uses a piezoelectric microphone embedded in a pulse generator to guide CRT optimization.

METHODS

Electrocardiogram, left ventricular pressure (LVP), and heart sounds were simultaneously collected during CRT device implantation procedures. A piezoelectric alarm transducer embedded in a modified CRT device facilitated recording of heart sounds in patients undergoing a pacing protocol with different AV delays. Machine learning (ML) was used to produce a decision-tree ensemble model capable of estimating absolute maximal LVP (LVPmax) and maximal rise of LVP (LVdP/dtmax) using 3 heart sound-based features. To gauge the applicability of ML in AV delay optimization, polynomial curves were fitted to measured and estimated values.

RESULTS

In the data set of ∼30,000 heartbeats, ML indicated S1 amplitude, S2 amplitude, and S1 integral (S1 energy for LVdP/dtmax) as most prominent features for AV delay optimization. ML resulted in single-beat estimation precision for absolute values of LVPmax and LVdP/dtmax of 67% and 64%, respectively. For 20-30 beat averages, cross-correlation between measured and estimated LVPmax and LVdP/dtmax was 0.999 for both. The estimated optimal AV delays were not significantly different from those measured using invasive LVP (difference -5.6 ± 17.1 ms for LVPmax and +5.1 ± 6.7 ms for LVdP/dtmax). The difference in function at estimated and measured optimal AV delays was not statiscally significant (1 ± 3 mm Hg for LVPmax and 9 ± 57 mm Hg/s for LVdP/dtmax).

CONCLUSION

Heart sound sensors embedded in a CRT device, powered by a ML algorithm, provide a reliable assessment of optimal AV delays and absolute LVPmax and LVdP/dtmax.

Consolidated Reporting Guidelines for Prognostic and Diagnostic Machine Learning Modeling Studies: Development and Validation

BACKGROUND

The reporting of machine learning (ML) prognostic and diagnostic modeling studies is often inadequate, making it difficult to understand and replicate such studies. To address this issue, multiple consensus and expert reporting guidelines for ML studies have been published. However, these guidelines cover different parts of the analytics lifecycle, and individually, none of them provide a complete set of reporting requirements.

OBJECTIVE

We aimed to consolidate the ML reporting guidelines and checklists in the literature to provide reporting items for prognostic and diagnostic ML in in-silico and shadow mode studies.

METHODS

We conducted a literature search that identified 192 unique peer-reviewed English articles that provide guidance and checklists for reporting ML studies. The articles were screened by their title and abstract against a set of 9 inclusion and exclusion criteria. Articles that were filtered through had their quality evaluated by 2 raters using a 9-point checklist constructed from guideline development good practices. The average κ was 0.71 across all quality criteria. The resulting 17 high-quality source papers were defined as having a quality score equal to or higher than the median. The reporting items in these 17 articles were consolidated and screened against a set of 6 inclusion and exclusion criteria. The resulting reporting items were sent to an external group of 11 ML experts for review and updated accordingly. The updated checklist was used to assess the reporting in 6 recent modeling papers in JMIR AI. Feedback from the external review and initial validation efforts was used to improve the reporting items.

RESULTS

In total, 37 reporting items were identified and grouped into 5 categories based on the stage of the ML project: defining the study details, defining and collecting the data, modeling methodology, model evaluation, and explainability. None of the 17 source articles covered all the reporting items. The study details and data description reporting items were the most common in the source literature, with explainability and methodology guidance (ie, data preparation and model training) having the least coverage. For instance, a median of 75% of the data description reporting items appeared in each of the 17 high-quality source guidelines, but only a median of 33% of the data explainability reporting items appeared. The highest-quality source articles tended to have more items on reporting study details. Other categories of reporting items were not related to the source article quality. We converted the reporting items into a checklist to support more complete reporting.

CONCLUSIONS

Our findings supported the need for a set of consolidated reporting items, given that existing high-quality guidelines and checklists do not individually provide complete coverage. The consolidated set of reporting items is expected to improve the quality and reproducibility of ML modeling studies.

Artificial Intelligence in Thyroidology: A Narrative Review of the Current Applications, Associated Challenges, and Future Directions

Background: The use of artificial intelligence (AI) in health care has grown exponentially with the promise of facilitating biomedical research and enhancing diagnosis, treatment, monitoring, disease prevention, and health care delivery. We aim to examine the current state, limitations, and future directions of AI in thyroidology. Summary: AI has been explored in thyroidology since the 1990s, and currently, there is an increasing interest in applying AI to improve the care of patients with thyroid nodules (TNODs), thyroid cancer, and functional or autoimmune thyroid disease. These applications aim to automate processes, improve the accuracy and consistency of diagnosis, personalize treatment, decrease the burden for health care professionals, improve access to specialized care in areas lacking expertise, deepen the understanding of subtle pathophysiologic patterns, and accelerate the learning curve of less experienced clinicians. There are promising results for many of these applications. Yet, most are in the validation or early clinical evaluation stages. Only a few are currently adopted for risk stratification of TNODs by ultrasound and determination of the malignant nature of indeterminate TNODs by molecular testing. Challenges of the currently available AI applications include the lack of prospective and multicenter validations and utility studies, small and low diversity of training data sets, differences in data sources, lack of explainability, unclear clinical impact, inadequate stakeholder engagement, and inability to use outside of the research setting, which might limit the value of their future adoption. Conclusions: AI has the potential to improve many aspects of thyroidology; however, addressing the limitations affecting the suitability of AI interventions in thyroidology is a prerequisite to ensure that AI provides added value for patients with thyroid disease.

Prediction of short-term atrial fibrillation risk using primary care electronic health records

OBJECTIVE

Atrial fibrillation (AF) screening by age achieves a low yield and misses younger individuals. We aimed to develop an algorithm in nationwide routinely collected primary care data to predict the risk of incident AF within 6 months (Future Innovations in Novel Detection of Atrial Fibrillation (FIND-AF)).

METHODS

We used primary care electronic health record data from individuals aged ≥30 years without known AF in the UK Clinical Practice Research Datalink-GOLD dataset between 2 January 1998 and 30 November 2018, randomly divided into training (80%) and testing (20%) datasets. We trained a random forest classifier using age, sex, ethnicity and comorbidities. Prediction performance was evaluated in the testing dataset with internal bootstrap validation with 200 samples, and compared against the CHA2DS2-VASc (Congestive heart failure, Hypertension, Age >75 (2 points), Stroke/transient ischaemic attack/thromboembolism (2 points), Vascular disease, Age 65-74, Sex category) and C2HEST (Coronary artery disease/Chronic obstructive pulmonary disease (1 point each), Hypertension, Elderly (age ≥75, 2 points), Systolic heart failure, Thyroid disease (hyperthyroidism)) scores. Cox proportional hazard models with competing risk of death were fit for incident longer-term AF between higher and lower FIND-AF-predicted risk.

RESULTS

Of 2 081 139 individuals in the cohort, 7386 developed AF within 6 months. FIND-AF could be applied to all records. In the testing dataset (n=416 228), discrimination performance was strongest for FIND-AF (area under the receiver operating characteristic curve 0.824, 95% CI 0.814 to 0.834) compared with CHA2DS2-VASc (0.784, 0.773 to 0.794) and C2HEST (0.757, 0.744 to 0.770), and robust by sex and ethnic group. The higher predicted risk cohort, compared with lower predicted risk, had a 20-fold higher 6-month incidence rate for AF and higher long-term hazard for AF (HR 8.75, 95% CI 8.44 to 9.06).

CONCLUSIONS

FIND-AF, a machine learning algorithm applicable at scale in routinely collected primary care data, identifies people at higher risk of short-term AF.

Deep-Learning for Epicardial Adipose Tissue Assessment With Computed Tomography: Implications for Cardiovascular Risk Prediction

BACKGROUND

Epicardial adipose tissue (EAT) volume is a marker of visceral obesity that can be measured in coronary computed tomography angiograms (CCTA). The clinical value of integrating this measurement in routine CCTA interpretation has not been documented.

OBJECTIVES

This study sought to develop a deep-learning network for automated quantification of EAT volume from CCTA, test it in patients who are technically challenging, and validate its prognostic value in routine clinical care.

METHODS

The deep-learning network was trained and validated to autosegment EAT volume in 3,720 CCTA scans from the ORFAN (Oxford Risk Factors and Noninvasive Imaging Study) cohort. The model was tested in patients with challenging anatomy and scan artifacts and applied to a longitudinal cohort of 253 patients post-cardiac surgery and 1,558 patients from the SCOT-HEART (Scottish Computed Tomography of the Heart) Trial, to investigate its prognostic value.

RESULTS

External validation of the deep-learning network yielded a concordance correlation coefficient of 0.970 for machine vs human. EAT volume was associated with coronary artery disease (odds ratio [OR] per SD increase in EAT volume: 1.13 [95% CI: 1.04-1.30]; P = 0.01), and atrial fibrillation (OR: 1.25 [95% CI: 1.08-1.40]; P = 0.03), after correction for risk factors (including body mass index). EAT volume predicted all-cause mortality (HR per SD: 1.28 [95% CI: 1.10-1.37]; P = 0.02), myocardial infarction (HR: 1.26 [95% CI:1.09-1.38]; P = 0.001), and stroke (HR: 1.20 [95% CI: 1.09-1.38]; P = 0.02) independently of risk factors in SCOT-HEART (5-year follow-up). It also predicted in-hospital (HR: 2.67 [95% CI: 1.26-3.73]; P ≤ 0.01) and long-term post-cardiac surgery atrial fibrillation (7-year follow-up; HR: 2.14 [95% CI: 1.19-2.97]; P ≤ 0.01).

CONCLUSIONS

Automated assessment of EAT volume is possible in CCTA, including in patients who are technically challenging; it forms a powerful marker of metabolically unhealthy visceral obesity, which could be used for cardiovascular risk stratification.

Continuous Data-Driven Monitoring in Critical Congenital Heart Disease: Clinical Deterioration Model Development

BACKGROUND

Critical congenital heart disease (cCHD)-requiring cardiac intervention in the first year of life for survival-occurs globally in 2-3 of every 1000 live births. In the critical perioperative period, intensive multimodal monitoring at a pediatric intensive care unit (PICU) is warranted, as their organs-especially the brain-may be severely injured due to hemodynamic and respiratory events. These 24/7 clinical data streams yield large quantities of high-frequency data, which are challenging in terms of interpretation due to the varying and dynamic physiology innate to cCHD. Through advanced data science algorithms, these dynamic data can be condensed into comprehensible information, reducing the cognitive load on the medical team and providing data-driven monitoring support through automated detection of clinical deterioration, which may facilitate timely intervention.

OBJECTIVE

This study aimed to develop a clinical deterioration detection algorithm for PICU patients with cCHD.

METHODS

Retrospectively, synchronous per-second data of cerebral regional oxygen saturation (rSO₂) and 4 vital parameters (respiratory rate, heart rate, oxygen saturation, and invasive mean blood pressure) in neonates with cCHD admitted to the University Medical Center Utrecht, the Netherlands, between 2002 and 2018 were extracted. Patients were stratified based on mean oxygen saturation during admission to account for physiological differences between acyanotic and cyanotic cCHD. Each subset was used to train our algorithm in classifying data as either stable, unstable, or sensor dysfunction. The algorithm was designed to detect combinations of parameters abnormal to the stratified subpopulation and significant deviations from the patient's unique baseline, which were further analyzed to distinguish clinical improvement from deterioration. Novel data were used for testing, visualized in detail, and internally validated by pediatric intensivists.

RESULTS

A retrospective query yielded 4600 hours and 209 hours of per-second data in 78 and 10 neonates for, respectively, training and testing purposes. During testing, stable episodes occurred 153 times, of which 134 (88%) were correctly detected. Unstable episodes were correctly noted in 46 of 57 (81%) observed episodes. Twelve expert-confirmed unstable episodes were missed in testing. Time-percentual accuracy was 93% and 77% for, respectively, stable and unstable episodes. A total of 138 sensorial dysfunctions were detected, of which 130 (94%) were correct.

CONCLUSIONS

In this proof-of-concept study, a clinical deterioration detection algorithm was developed and retrospectively evaluated to classify clinical stability and instability, achieving reasonable performance considering the heterogeneous population of neonates with cCHD. Combined analysis of baseline (ie, patient-specific) deviations and simultaneous parameter-shifting (ie, population-specific) proofs would be promising with respect to enhancing applicability to heterogeneous critically ill pediatric populations. After prospective validation, the current-and comparable-models may, in the future, be used in the automated detection of clinical deterioration and eventually provide data-driven monitoring support to the medical team, allowing for timely intervention.

Artificial intelligence in medical device software and high-risk medical devices - a review of definitions, expert recommendations and regulatory initiatives

INTRODUCTION

Artificial intelligence (AI) encompasses a wide range of algorithms with risks when used to support decisions about diagnosis or treatment, so professional and regulatory bodies are recommending how they should be managed.

AREAS COVERED

AI systems may qualify as standalone medical device software (MDSW) or be embedded within a medical device. Within the European Union (EU) AI software must undergo a conformity assessment procedure to be approved as a medical device. The draft EU Regulation on AI proposes rules that will apply across industry sectors, while for devices the Medical Device Regulation also applies. In the CORE-MD project (Coordinating Research and Evidence for Medical Devices), we have surveyed definitions and summarize initiatives made by professional consensus groups, regulators, and standardization bodies.

EXPERT OPINION

The level of clinical evidence required should be determined according to each application and to legal and methodological factors that contribute to risk, including accountability, transparency, and interpretability. EU guidance for MDSW based on international recommendations does not yet describe the clinical evidence needed for medical AI software. Regulators, notified bodies, manufacturers, clinicians and patients would all benefit from common standards for the clinical evaluation of high-risk AI applications and transparency of their evidence and performance.