Introduction
Cardiovascular diseases are part of complex or multifactorial diseases, in which numerous causes converge, not all of which are easily recognizable.1
Numerous advances have been made in the last 50 years in cardiovascular disease prevention; however, cardiovascular disease is still a significant cause of morbidity and mortality.2
Over the past few decades, major cardiovascular disease risk factors have been identified and effective and well tolerated risk factor treatments have been developed. However, the prevalence of unhealthy lifestyles is still high, and cardiovascular risk factors are often poorly treated2; consequently, the reduction of cardiovascular risk and the prediction of the development of cardiovascular disease remain two of the main challenges of our time.
Accordingly, the research on cardiovascular prevention and risk stratification focuses, in addition to large genomic datasets, on biomarkers,3 metabolic phenotypes, imaging,4,5 both ultrasound and radiological, and on their integration, favoured by artificial intelligence applications.1,6
Accordingly, a basic knowledge of this science is essential; therefore, in this review, we have reported several applications of artificial intelligence in cardiovascular prevention, without neglecting the limitations of this new scientific approach.
Artificial intelligence and machine learning techniques
Data analysis with artificial intelligence methods is an approach based on information systems capable of learning and adapting without following explicit instructions, using algorithms and statistical models to analyse and derive pattern inferences in the data. Specifically, artificial intelligence is a branch of information technology capable of analysing (receiving, processing and interpreting) complex medical data using algorithms and complex mathematical calculations, artificially simulating what happens in the process of learning and processing of the human mind.
Artificial intelligence focuses on machine learning techniques, which can establish composite relationships among data, behavioural patterns, rules governing a system, classification schemes and more.7
Moreover, machine learning can manage and incorporate different data resources (clinical measurements and observations, biological -omics, experimental results, environmental information, wearable devices) into models for describing and predicting human diseases (Fig. 1).8
;)
Fig. 1: The typical machine learning workflow in healthcare research.
Generally, relations are discovered by using standard machine learning strategies, typically grouped into two types: supervised and unsupervised learning.
Supervised learning involves teaching the model with a collection of input data with the correct output already associated with it.9 Specifically, standard supervised learning algorithms include linear regression, logistic regression, random forest, decision tree and support vector machine (SVM).
Instead, in unsupervised learning techniques, there is no information on the feature to be predicted; indeed, these techniques must learn from the relationships among the elements of a dataset and classify them without basing them on categories or labels.10 For this task, the most frequently used approaches are deep learning algorithms, tensor factorization and topological data analysis.8
Using these methods, artificial intelligence has started a profound reform in the field of cardiovascular prevention, opening new essential ways for the improvement of patient healthcare and cost-effectiveness.
Artificial intelligence in cardiovascular prevention
There are several areas of opportunity for artificial intelligence in cardiovascular prevention, and it is essential to see the development of artificial intelligence-guided medicine as part of the battle against chronic diseases worldwide.7 Specifically, artificial intelligence has a fundamental position in novel therapeutic agent discovery, precision cardiovascular disease stratification, integration of multiomic data, extension of physician efficiency and efficacy, continuous remote monitoring and diagnostics, and optimized resource allocation in the field of cardiovascular prevention.
Management of cardiovascular risk factors (arterial hypertension, diabetes and dyslipidaemia)
Among the most recent applications of artificial intelligence is the possibility to help physicians and patients in the management of the main cardiovascular risk factors, with particular regard to hypertension, diabetes and dyslipidaemia. From a practical point of view, the use of artificial intelligence might be helpful for clinicians to make the more accurate decision and predict patient outcomes leading to a more personalized approach. From the patients’ point of view, the use of artificial intelligence will help to collect and analyse a long medical history rapidly and connect electronic medical record systems with received patient information to screen for the essential message to clinicians. In the cardiovascular prevention analysis of big databases, using artificial intelligence is helping the scientific community to develop accurate prediction analyses for cardiovascular death. This has been tested in particular in the field of coronary artery disease (CAD).
In managing patients with hypertension, artificial intelligence has been used to predict the incidence of future hypertension in population-based studies.7 Using the machine learning approach with information from more than 18000 patients, Japanese researchers have developed a very sensitive algorithm for predicting new-onset hypertension, which showed higher accuracy than the usual logistic regression model, reaching an AUC close to 0.99.11 Similar results have been obtained in a more extensive study including more than 8 000 000 individuals from East Asia using an open-source platform, with potential large-scale applicability.12 Secondary hypertension is potentially treatable, but its identification is challenging. Recently, medical records analysis using the machine learning approach led to the development of precise predictive modelling for secondary hypertension. This tool using artificial intelligence may potentially facilitate the clinical diagnosis and decision-making in managing hypertension.13 Successful treatment of hypertension is sometimes not achieved, leading to incident cardiovascular morbidity and mortality.14 A different aspect may impact the achievement of blood pressure control.15–18 Recently an analysis of big data using the machine learning approach with artificial intelligence identified possible factors associated with uncontrolled hypertension using electronic charts of more than 2 million Israelis.19 While exemplifying another avenue of interrogating data using the machine learning approach, defining the clinical value of these results still requires evidence from randomized trials to resolve possible confounding.
Similarly, diabetes is a leading cause of death and its management as another chronic disease requires a novel approach to improving patients’ prognosis. In this field, the use of artificial intelligence has been tested to enhance the prediction and diagnosis of significant complications of diabetes leading to adverse cardiovascular events.20 This is of particular interest in the field of gestational diabetes and recent research has identified the possibility of using a dedicated machine learning approach to improve its prediction.21
In the management of dyslipidaemia, different potential applications of artificial intelligence have been tested so far, starting from the diagnosis to the management and prognosis related to the disease. Recently machine learning modelling has been applied to big datasets to obtain more accurate predictive models for incident dyslipidaemia, taking into account monogenic or polygenic variants.22,23 Recently, estimation of LDL cholesterol has been achieved with better accuracy using machine learning outcome giving the opportunity to better assess cardiovascular risk.24
Current evidence supports the possibility of the development of a new prediction model based on artificial intelligence that outperforms traditional risk prediction tools. Data from the MESA study have demonstrated that calculating cardiovascular risk using the machine learning approach is more accurate than the American Heart Calculator risk.25 Similar results were obtained in a recent study involving participants in UK Biobank. The study demonstrated that an approach using artificial intelligence could analyse and includes a large number of informative variables (more than 400), which in part explains the increased accuracy with the use of artificial intelligence.26
Risk stratification and prediction of development of cardiovascular diseases (atrial fibrillation, coronary artery disease and heart failure)
Atrial fibrillation
Early atrial fibrillation identification is essential for early treatment and preventing atrial fibrillation-related strokes, complications and disability. However, a cost-effective strategy for population screening for atrial fibrillation is yet to be available. Machine learning and artificial intelligence can be applied to atrial fibrillation screening to support atrial fibrillation-risk prediction or improve the automated atrial fibrillation diagnosis using different rhythm monitoring modalities.
Several clinical risk score models have been generated to predict the likelihood of atrial fibrillation,27–36 mainly developed using data from specific population cohorts providing a 5-year or 10-year prediction risk of atrial fibrillation. However, as risk factors for atrial fibrillation vary among population cohorts (i.e. median age, ethnicity and prevailing cardiovascular risk factors), it has been repeatedly found that the performance of a model is not as precise as when performed on the original cohort.37,38
In recent decades, technologies for heart rhythm monitoring are improving [i.e. handheld or smartphone or smartwatch single-lead ECG recorders, ECG patch monitors, photoplethysmography (PPG) smartwatches or smartphones, blood pressure monitors, and external Holter monitors], providing more opportunities for ambulatory monitoring, and increasing the likelihood of detecting atrial fibrillation.39
The machine learning approach to predict future atrial fibrillation uses 12-lead ECGs, electronic health records, or sinus rhythm ambulatory ECGs. These approaches help maximize screening efficiency and the number of new atrial fibrillation cases detected.40–42
Initially, machine learning was applied to sinus rhythm segments of ambulatory ECGs to predict paroxysmal atrial fibrillation based mainly on premature atrial contractions (PACs). More recently, another approach based on heart rate variability features, including times and frequency domains, is used with an accuracy above 0.9.43–46
Combining machine learning and clinical risk scoring shows potential benefits for reducing the number needed to screen and improving the effectiveness of atrial fibrillation screening.47–49 Although results are encouraging, these have yet to be tested within prospective studies.
Several rhythm monitoring modalities, including ECGs, are used for atrial fibrillation detection. Machine learning could be applied in atrial fibrillation screening to improve the accuracy of automated diagnosis using other rhythm monitoring modalities, potentially increasing the number of individuals screened and allowing quicker population or mass screening. Studies have developed machine learning models to detect atrial fibrillation from a 12-lead ECG.50,51 Neural networks are typically used in the decision process to determine if the ECG demonstrates atrial fibrillation. Neural networks warrant high accuracy values for models differentiating between atrial fibrillation and sinus rhythm; however, their performance decreased when the models were applied to determine other rhythms too.
Photoplethysmography (PPG) allows reading of the pulse pressure signals resulting from the propagation of blood pressure pulses along arteries.52 Detection of irregularities in pulse pressure signals measured peripherally supports the diagnosis of atrial fibrillation. Multiple studies have tested different machine learning algorithms to detect atrial fibrillation from PPG signals.53–57 Notably, when compared, neural networks were superior, with accuracy, specificity and sensitivity of above 0.90.
Many studies involving machine learning algorithms to identify atrial fibrillation from single-lead or ambulatory ECGs have been performed.58–76 As an extensive database of single-lead ECGs is easily accessible, many machine learning models have been created compared with 12-lead ECGs. Overall, accuracy, sensitivity and specificity tended to be above 0.85.
Great efforts were brought to the use of machine learning to improve atrial fibrillation screening. Machine learning algorithms help detect individuals at the highest risk of atrial fibrillation to target the screening better or to facilitate more precise automated atrial fibrillation diagnosis during the screening process. Unfortunately, most machine learning algorithms have been developed and tested on retrospective data. Consequently, prospective studies assessing their impact on actual world data are eagerly awaited.
Coronary artery disease
Current guidelines emphasize the importance of early detection and risk stratification in CAD to implement goal-directed medical therapies, which can alter the natural CAD trajectory.77
Specifically, Dogan et al. built an ensemble model of eight random-forest classifiers to predict the risk of symptomatic CAD using genetic and epigenetic variables along with clinical risk factors: the model predicted symptomatic CAD with an accuracy, sensitivity and specificity of 0.78, 0.75 and 0.80, respectively, in the internal validation cohort (n = 142).78
Moreover, using clinical and demographic features, machine learning models have been employed to estimate the pretest probability (PTP) of CAD.79–81
Finally, various machine learning algorithms based on stress imaging, particularly single-photon emission computed tomography (SPECT), have been devised to facilitate the prediction of CAD.82–86
Risk scores, such as the SYNTAX score87 and the Global Registry of Acute Coronary Events (GRACE),88 are currently used to predict major adverse cardiovascular events (MACEs) in patients with previous or current acute coronary syndrome (ACS); nevertheless, these tools lack accuracy.89 Nowadays, by using machine learning techniques, identifying patients with higher morbidity and mortality following ACS might be possible90–95: these new models appear to perform better than the commonly used score.
Specifically, the PRAISE score95 showed accurate discriminative capabilities for predicting all-cause death, myocardial infarction and major bleeding over a time period of 1 year.
Finally, artificial intelligence may allow identifying who will benefit more, less or will not benefit from any specific treatment after an ACS.89,96,97
Heart failure
The natural history of heart failure alternates phases of stability with exacerbation phases, with a progressive decline in functional capacity.98 Machine learning is currently attempting to simplify heart failure diagnosis, classification, severity estimation, and re-hospitalization prediction. Aljaaf et al.99 suggested a predictive model of heart failure; specifically, they proposed a decision-tree model trained on the Cleveland Clinic Heart Disease Data Set with 13 clinical variables, which stratified patients into one of five risk classes for the development of heart failure.
Moreover, Shah et al.100 emphasized the distinction between heart failure with preserved ejection fraction (HFpEF) subtypes; precisely, the authors used ‘phenomapping’, an innovative technique that employs machine learning to define clusters of patients based on dense phenotypic data, thereby supplying an unbiased way to classify heterogeneous clinical syndromes. Briefly, this study showed three distinct phenogroups in terms of clinical characteristics, cardiac structure and function, haemodynamics and outcomes. Another field of application of artificial intelligence in heart failure is the prediction of mortality. The GWTG-HF101 and MAGGIC102 scores are well validated conventional models for risk stratification of AHF patients; however, they have several limitations103; accordingly, Kwon et al.104 validated a deep learning-based artificial intelligence algorithm for predicting mortality of acute heart failure (DAHF) and showed that this algorithm predicted the in-hospital and long-term mortality of AHF patients with more accuracy than the existing risk scores and other machine learning models.
The prediction of re-hospitalization is particularly challenging. In particular, Lorenzoni et al.105 aimed to compare the performance of eight machine learning techniques [logistic regression, generalized linear model net (GLMN), classification and regression tree, random forest, adaboost, logitboost, SVM and ANN] in the forecast of hospitalization among heart failure patients, using data from the Gestione Integrata dello Scompenso Cardiaco (GISC). Specifically, the GLMN showed better performance in predicting hospitalization than the other machine learning techniques.
Finally, numerous invasive and noninvasive devices have been developed to predict heart failure exacerbations.106–112
Specifically, the HeartLogic algorithm (Boston Scientific, St. Paul, Minnesota, USA), based on several parameters (first and third heart sounds, intrathoracic impedance, respiration rate, the ratio of respiration rate to tidal volume, night heart rate and patient activity) collected from a cardiac resynchronization therapy defibrillator, showed a 70% sensitivity for detection of subsequent hospitalization for worsening of heart failure.110 Precisely, the device uses a machine learning technique to combine these multiple data into a model specific to each patient.
Accordingly, in the LINK-HF study, machine learning combined multiple physiological parameter streams into a model specific to each individual through an adhesive patch.106
All these devices use artificial intelligence algorithms and open the way to personalized medicine.
Future prospective
Cardiologists plan primary and secondary prevention strategies from data and they tend to have access to a larger volume of patient data than other specialties.2 Despite some potential pitfalls, it is emerging that the use of techniques based on artificial intelligence will be the best approach for making data-based decisions in the future.108,112
The increased application of artificial intelligence, also supported by increased funding from public and/or private organizations, depends on the rapid growth of clinical databases, medical applications for smartphones and wearable devices. In this sense, preventive cardiology represents one of the largest fields of investment as it covers not only patients but also the broader general population.
The study of CAD has been a model for modern medicine as it has taught how to identify a number of associated conditions, called risk factors, which can be incorporated into general risk stratification models.113 On this subject, artificial intelligence has the potential to analyse an immense range of variables, identify nonlinear associations and help identify new nontraditional risk factors.
One example of these applications was provided in a study conducted on a 13-year follow-up data set from the Multi-Ethnic Study of Atherosclerosis of 6459 participants who were free from atherosclerotic cardiovascular disease at baseline. The American College of Cardiology/American Heart Association (ACC/AHA) Pooled Cohort Equations Risk Calculator is one of the mostly used tools for primary risk stratification of CAD and atherosclerotic cardiovascular disease but underestimates the risk of atherosclerotic cardiovascular disease in women and certain ethnic groups.114–116 Using the same nine traditional risk factors, the machine learning risk calculator significantly improved risk stratification, identifying 13% more high-risk patients and recommending 25% less statin therapy in low-risk subjects, in part by identifying nonlinear relationships.25
Ambale-Venkatesh and colleagues by using random survival forests, a machine learning technique, aimed at evaluating the prediction of cardiovascular adverse events in comparison to standard cardiovascular risk scores. They obtained 735 variables from imaging and noninvasive tests, questionnaires, and biomarker panels, and identified the 20 best predictors for each outcome of interest. They found that the random survival forests technique performed better than established risk scores with increased prediction accuracy.117 This finding emphasizes the importance of nontraditional risk factors and of considering more unsupervised learning methods instead of selecting the variables based on biologic plausibility.
Another example of artificial intelligence applications comes from the analysis of digital data. Voice analysis through smartphone recording with a machine learning algorithm revealed two voice features that were independently associated with CAD, particularly when patients were asked to describe an emotionally significant experience.118 This result may have clinical implications for telemedicine, when clinical healthcare is provided at a distance, and represents an additional example of how digital information and machine learning processing may change some accepted paradigms of cardiovascular research.
Although machine learning is currently the prerogative of highly trained professionals, the use of these techniques will be progressively easier and more accessible to the clinical cardiologist. Knowledge of pathophysiology and clinical pictures based on experience cannot be replaced. Rather, they will become central to the decision of when to apply these techniques, how to interpret the results of models and translate them into clinical practice.119
How coronavirus disease 2019 affected cardiovascular prevention: development of a new predictive model through artificial intelligence
Coronavirus disease 2019 (COVID-19) may deteriorate the clinical status of subjects with underlying cardiovascular diseases and may cause several de novo cardiovascular complications, including heart failure, myocardial infarction, myocarditis, takotsubo syndrome, life-threatening arrhythmias and thromboembolic events.120–124 However, the interrelationship of COVID-19 and CVD is still not completely understood, with consequences for optimal management of such patients.
Some authors have developed artificial intelligence-based models, including clinical and imaging characteristics for better diagnosis and management of patients with COVID-19.125
Machine learning has been proposed for deeper cardiovascular risk assessment in these patients, including those who are asymptomatic. Specific applications cover patient cardiovascular monitoring using surrogate noninvasive markers, telemedicine-based care using the machine learning approach and universal biobanks in the context of artificial intelligence platform developments.126
Artificial intelligence could become an integral part of the COVID-19 management system, from cardiovascular risk assessment to long-term follow-up after recovery.
Limitations of applying artificial intelligence in cardiovascular prevention
Artificial intelligence and machine learning are finding huge applications in modern medicine, particularly cardiology. Artificial intelligence is used to improve work efficiency, detect patterns behind observed data and play a role in predicting future events and prognosis with solid potential to become an integral part in cardiovascular disease prevention and rehabilitation. All this imposes many challenges.
Several regulatory authorities in healthcare, such as the United States Food and Drug Administration (FDA), have approved various algorithms, but there are still no universal approval guidelines. Added to the fact that the people who create the algorithms are not always doctors treating patients, the symbolists may need to learn more about healthcare, and clinicians must find out what tasks the algorithms can and cannot do.
Artificial intelligence prediction models over existing and often simpler prediction models should, therefore, be based on solid evidence coming, for example, from careful model and prediction comparisons.127 For example, an artificial intelligence-based prediction model aiming to predict the 10-year risk of cardiovascular events might be compared with a canonical model like the Framingham risk score.127 It is essential to determine the acceptable threshold for artificial intelligence-induced errors and what to do if those mistakes are ultimately fatal.128
It must be clearly understood that both artificial intelligence and machine learning should act to support and not replace the physician and his clinical skills.129,130 Machine learning algorithms can identify patterns based on the huge volume of data and often tend to identify the average characteristics of a patient and ignore the outliers.130 However, in medicine, every individual case is different, and artificial intelligence might be error prone in these cases, especially if they are outliers.130 Artificial intelligence as a tool in cardiology should act as an aid in decision-making and not actually make those decisions.130,131
Among the biggest problems that arise with the use of are the possibility of greater inequalities and the legal and ethical concerns when large volumes of data are processed.
To implement new technology in some fields, some economic and social gaps may need to be addressed. What cannot be overlooked is that the digital health devices and artificial intelligence algorithms carry the additional challenge of digital literacy, which shows the ability to appraise and apply information or knowledge gained from electronic sources.132,133 Engaging new technologies like artificial intelligence might lead to health services being inaccessible for those patients with limited digital literacy.133
The ethical issues of artificial intelligence remain sticky when handled and should be focused upon to be careful of them and further reviewed to develop the regulatory system.133
Data protection
Artificial intelligence needs an amount of data for building up and training models, which may contain personal or clinical data of patients.7,133 How to defend the security of patients’ data urgently needs to be considered prudently to preserve the human rights of privacy.133 Data provenance and permission for use are mainly important.133 Adopting measures to avoid the unethical use of patients’ data is essential, such as ensuring the rights of patients to be informed, have access and be allowed rectification.10,133 The General Data Protection Regulation (GDPR) endorsed by the European Union aims to regulate and standardize personal data utilization, strengthening and unifying the data protection for all people within the European Union.133,134 Specific informed consent requirements for using and granting data, and numerous rights that must be respected in data processing, were set up in this law.133,134
Transparency
Transparency and fairness of algorithms are the most prevalent issues reported.133 The highly automatic decision-making ability enables artificial intelligence to make decisions without human interference.133,135 This utility not only brings more efficient data management but also makes algorithms harder to explain.133 Nevertheless, unmatured artificial intelligence models might achieve results that reflect preexisting bias in the real world if models were trained by unrepresentative or poor datasets.30,133,136 This kind of algorithm bias could entrench or exacerbate health discrepancies.133,137 Regardless of the notion of bias being very complex, it is possible and ethically necessary to design artificial intelligence systems to support the offset of human biases to try to lead to outcomes closer to fairness.133,138 For now, to achieve better transparency, it is suggested to improve the interpretability and auditability of artificial intelligence.133,139
Responsibility
Accountability is required in artificial intelligence algorithms to elucidate who should be liable for decisions made with algorithmic support.133 The accountability component includes multiple professionals, which is considered the most challenging part of implementation.133,137 Although responsible artificial intelligence has garnered widespread attention, whether we should consider artificial intelligence algorithms as a subject of responsibility or humans as the only actors who are responsible for algorithm-based decisions remains debatable.133 Statements about the specific actors accountable for artificial intelligence's decisions are also diverse.133 In fact, the responsibility may be problematic to determine. Without transparency, it is difficult to enforce accountability.133,135
Trustworthiness
Not all clinicians and patients are interested in adapting to new technologies.133 Some clinicians do not trust artificial intelligence because of the difficulties associated with understanding how it works, which may affect their willingness to use artificial intelligence-based products.133,140 This situation seems more serious in patients. Patients worry about losing connection with clinicians and are also concerned about the misuse of artificial intelligence that can cause additional damage to their health, which may be caused by inadequate follow-up and insufficient education.133,141
One of the key issues is the absence of standardization across institutions and interoperability.142
If we take the example of cardiac telerehabilitation, it can be obtained by applying artificial intelligence in wearable monitoring and support systems in addition to the use of various devices and mobile applications. Artificial intelligence combined with cardiac telerehabilitation support systems could analyse observed indicators in real time and then rank the results, which allows systems to provide timely feedback and more targeted recommendations at varying degrees as preliminary interventions.133 The selection of the algorithm and its optimization are important factors for the success of a rehabilitation treatment worthy of further research and development work.
Conclusions
The different artificial intelligence applications allow the detection of a patient's health trajectory leading to personalized medicine and tailored treatment. Specifically, artificial intelligence-based systems support cardiologists in daily medical activities, improving disease detection and treatment; however, the ethical issues and limits of artificial intelligence should not be overlooked.
Conflicts of interest
There are no conflicts of interest.
References
1. Boccanelli A, Botta G. New prevention scenarios: polygenic risk.
Eur Heart J Suppl 2021; 23: (Suppl E): E33–E35.
2. Visseren FLJ, Mach F, Smulders YM, et al. ESC National Cardiac Societies, ESC Scientific Document Group. 2021 ESC Guidelines on cardiovascular disease prevention in clinical practice.
Eur Heart J 2021; 42:3227–3337.
3. Frary CE, Blicher MK, Olesen TB, et al. Circulating biomarkers for long-term cardiovascular risk stratification in apparently healthy individuals from the MONICA 10 cohort.
Eur J Prev Cardiol 2020; 27:570–578.
4. Faggiano P, Dasseni N, Gaibazzi N, Rossi A, Henein M, Pressman G. Cardiac calcification as a marker of subclinical atherosclerosis and predictor of cardiovascular events: a review of the evidence.
Eur J Prev Cardiol 2019; 26:1191–1204.
5. Kozakova M, Palombo C. Imaging subclinical atherosclerosis in cardiovascular risk stratification.
Eur J Prev Cardiol 2020; 98:177–184.
6. Tamarappoo BK, Lin A, Commandeur F, et al. Machine learning integration of circulating and imaging biomarkers for explainable patient-specific prediction of cardiac events: a prospective study.
Atherosclerosis 2021; 318:76–82.
7. Visco V, Ferruzzi GJ, Nicastro F, et al. Artificial intelligence as a business partner in cardiovascular precision medicine: an emerging approach for disease detection and treatment optimization.
Curr Med Chem 2021; 28:6569–6590.
8. Johnson KW, Torres Soto J, Glicksberg BS, et al. Artificial intelligence in cardiology.
J Am Coll Cardiol 2018; 71:2668–2679.
9. Nichols JA, Herbert Chan HW, Baker MAB. Machine learning: applications of artificial intelligence to imaging and diagnosis.
Biophys Rev 2019; 11:111–118.
10. Dorado-Diaz PI, Sampedro-Gomez J, Vicente-Palacios V, Sanchez PL. Applications of artificial intelligence in cardiology: the future is already here.
Rev Esp Cardiol (Engl Ed) 2019; 72:1065–1075.
11. Kanegae H, Suzuki K, Fukatani K, Ito T, Harada N, Kario K. Highly precise risk prediction model for new-onset hypertension using artificial intelligence techniques.
J Clin Hypertens (Greenwich) 2020; 22:445–450.
12. Islam SMS, Talukder A, Awal MA, et al. Machine learning approaches for predicting hypertension and its associated factors using population-level data from three South Asian countries.
Front Cardiovasc Med 2022; 9:839379.
13. Diao X, Huo Y, Yan Z, et al. An application of machine learning to etiological diagnosis of secondary hypertension: retrospective study using electronic medical records.
JMIR Med Inform 2021; 9:e19739.
14. Mancusi C, Bisogni V, Maloberti A, et al. Accuracy of home blood pressure measurement: the ACCURAPRESS study: a proposal of Young Investigator Group of the Italian Hypertension Society (Societa Italiana dell’Ipertensione Arteriosa).
Blood Press 2022; 31:297–304.
15. Visco V, Finelli R, Pascale AV, et al. Larger blood pressure reduction by fixed-dose compared to free dose combination therapy of ACE inhibitor and calcium antagonist in hypertensive patients.
Transl Med UniSa 2017; 16:17–23.
16. Visco V, Finelli R, Pascale AV, et al. Difficult-to-control hypertension: identification of clinical predictors and use of ICT-based integrated care to facilitate blood pressure control.
J Hum Hypertens 2018; 32:467–476.
17. De Marco M, de Simone G, Izzo R, et al. Classes of antihypertensive medications and blood pressure control in relation to metabolic risk factors.
J Hypertens 2012; 30:188–193.
18. Mancusi C, Manzi MV, de Simone G, et al. Carotid atherosclerosis predicts blood pressure control in patients with hypertension: the Campania Salute Network Registry.
J Am Heart Assoc 2022; 11:e022345.
19. Koren G, Nordon G, Radinsky K, Shalev V. Machine learning of big data in gaining insight into successful treatment of hypertension.
Pharmacol Res Perspect 2018; 6:e00396.
20. Huang J, Yeung AM, Armstrong DG, et al. Artificial intelligence for predicting and diagnosing complications of diabetes.
J Diabetes Sci Technol 2022; 19322968221124583.
21. Liao LD, Ferrara A, Greenberg MB, et al. Development and validation of prediction models for gestational diabetes treatment modality using supervised machine learning: a population-based cohort study.
BMC Med 2022; 20:307.
22. Wu J, Qin S, Wang J, et al. Develop and evaluate a new and effective approach for predicting dyslipidemia in steel workers.
Front Bioeng Biotechnol 2020; 8:839.
23. Correia M, Kagenaar E, van Schalkwijk DB, Bourbon M, Gama-Carvalho M. Machine learning modelling of blood lipid biomarkers in familial hypercholesterolaemia versus polygenic/environmental dyslipidaemia.
Sci Rep 2021; 11:3801.
24. Oh GC, Ko T, Kim JH, et al. Estimation of low-density lipoprotein cholesterol levels using machine learning.
Int J Cardiol 2022; 352:144–149.
25. Kakadiaris IA, Vrigkas M, Yen AA, Kuznetsova T, Budoff M, Naghavi M. Machine learning outperforms ACC/AHA CVD risk calculator in MESA.
J Am Heart Assoc 2018; 7:e009476.
26. Alaa AM, Bolton T, Di Angelantonio E, Rudd JHF, van der Schaar M. Cardiovascular disease risk prediction using automated machine learning: a prospective study of 423,604 UK Biobank participants.
PLoS One 2019; 14:e0213653.
27. de Vos CB, Pisters R, Nieuwlaat R, et al. Progression from paroxysmal to persistent atrial fibrillation clinical correlates and prognosis.
J Am Coll Cardiol 2010; 55:725–731.
28. Hulme OL, Khurshid S, Weng LC, et al. Development and validation of a prediction model for atrial fibrillation using electronic health records.
JACC Clin Electrophysiol 2019; 5:1331–1341.
29. Linker DT, Murphy TB, Mokdad AH. Selective screening for atrial fibrillation using multivariable risk models.
Heart 2018; 104:1492–1499.
30. Cho MK. Rising to the challenge of bias in healthcare AI.
Nat Med 2021; 27:2079–2081.
31. Li C, Sun D, Liu J, et al. A prediction model of essential hypertension based on genetic and environmental risk factors in northern Han Chinese.
Int J Med Sci 2019; 16:793–799.
32. Kokubo Y, Watanabe M, Higashiyama A, Nakao YM, Kusano K, Miyamoto Y. Development of a basic risk score for incident atrial fibrillation in a Japanese general population: the Suita Study.
Circ J 2017; 81:1580–1588.
33. Hamada R, Muto S. Simple risk model and score for predicting of incident atrial fibrillation in Japanese.
J Cardiol 2019; 73:65–72.
34. Alonso A, Krijthe BP, Aspelund T, et al. Simple risk model predicts incidence of atrial fibrillation in a racially and geographically diverse population: the CHARGE-AF consortium.
J Am Heart Assoc 2013; 2:e000102.
35. Chamberlain AM, Agarwal SK, Folsom AR, et al. A clinical risk score for atrial fibrillation in a biracial prospective cohort [from the Atherosclerosis Risk in Communities (ARIC) study].
Am J Cardiol 2011; 107:85–91.
36. Schnabel RB, Sullivan LM, Levy D, et al. Development of a risk score for atrial fibrillation (Framingham Heart Study): a community-based cohort study.
Lancet 2009; 373:739–745.
37. Pfister R, Bragelmann J, Michels G, Wareham NJ, Luben R, Khaw KT. Performance of the CHARGE-AF risk model for incident atrial fibrillation in the EPIC Norfolk cohort.
Eur J Prev Cardiol 2015; 22:932–939.
38. Shulman E, Kargoli F, Aagaard P, et al. Validation of the Framingham Heart Study and CHARGE-AF risk scores for atrial fibrillation in Hispanics, African-Americans, and Non-Hispanic Whites.
Am J Cardiol 2016; 117:76–83.
39. Hindricks G, Potpara T, Dagres N, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): the Task Force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC.
Eur Heart J 2021; 42:373–498.
40. Raghunath S, Pfeifer JM, Ulloa-Cerna AE, et al. Deep neural networks can predict new-onset atrial fibrillation from the 12-lead ECG and help identify those at risk of atrial fibrillation-related stroke.
Circulation 2021; 143:1287–1298.
41. Hirota N, Suzuki S, Arita T, et al. Prediction of current and new development of atrial fibrillation on electrocardiogram with sinus rhythm in patients without structural heart disease.
Int J Cardiol 2021; 327:93–99.
42. Attia ZI, Noseworthy PA, Lopez-Jimenez F, et al. An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction.
Lancet 2019; 394:861–867.
43. Parsi A, Glavin M, Jones E, Byrne D. Prediction of paroxysmal atrial fibrillation using new heart rate variability features.
Comput Biol Med 2021; 133:104367.
44. Wang LH, Yan ZH, Yang YT, et al. A classification and prediction hybrid model construction with the IQPSO-SVM algorithm for atrial fibrillation arrhythmia.
Sensors (Basel) 2021; 21:5222.
45. Bashar SK, Ding EY, Walkey AJ, McManus DD, Chon KH. Atrial fibrillation prediction from critically ill sepsis patients.
Biosensors (Basel) 2021; 11:269.
46. Ganapathy N, Baumgartel D, Deserno TM. Automatic detection of atrial fibrillation in ECG using co-occurrence patterns of dynamic symbol assignment and machine learning.
Sensors (Basel) 2021; 21:3542.
47. Grout RW, Hui SL, Imler TD, et al. Development, validation, and proof-of-concept implementation of a two-year risk prediction model for undiagnosed atrial fibrillation using common electronic health data (UNAFIED).
BMC Med Inform Decis Mak 2021; 21:112.
48. Suzuki R, Katada J, Ramagopalan S, McDonald L. Potential of machine learning methods to identify patients with nonvalvular atrial fibrillation.
Future Cardiol 2020; 16:43–52.
49. Hill NR, Ayoubkhani D, McEwan P, et al. Predicting atrial fibrillation in primary care using machine learning.
PLoS One 2019; 14:e0224582.
50. Jo YY, Cho Y, Lee SY, et al. Explainable artificial intelligence to detect atrial fibrillation using electrocardiogram.
Int J Cardiol 2021; 328:104–110.
51. Ribeiro AH, Ribeiro MH, Paixao GMM, et al. Automatic diagnosis of the 12-lead ECG using a deep neural network.
Nat Commun 2020; 11:1760.
52. Pereira T, Tran N, Gadhoumi K, et al. Photoplethysmography based atrial fibrillation detection: a review.
NPJ Digit Med 2020; 3:3.
53. Wasserlauf J, You C, Patel R, Valys A, Albert D, Passman R. Smartwatch performance for the detection and quantification of atrial fibrillation.
Circ Arrhythm Electrophysiol 2019; 12:e006834.
54. Kwon S, Hong J, Choi EK, et al. Detection of atrial fibrillation using a ring-type wearable device (CardioTracker) and deep learning analysis of photoplethysmography signals: prospective observational proof-of-concept study.
J Med Internet Res 2020; 22:e16443.
55. Eerikainen LM, Bonomi AG, Schipper F, et al. Detecting atrial fibrillation and atrial flutter in daily life using photoplethysmography data.
IEEE J Biomed Health Inform 2020; 24:1610–1618.
56. Jacobsen M, Dembek TA, Ziakos AP, et al. Reliable detection of atrial fibrillation with a medical wearable during inpatient conditions.
Sensors (Basel) 2020; 20:5517.
57. Valiaho ES, Kuoppa P, Lipponen JA, et al. Wrist band photoplethysmography autocorrelation analysis enables detection of atrial fibrillation without pulse detection.
Front Physiol 2021; 12:654555.
58. Bashar SK, Hossain MB, Lazaro J, et al. Feasibility of atrial fibrillation detection from a novel wearable armband device.
Cardiovasc Digit Health J 2021; 2:179–191.
59. Sadr N, Jayawardhana M, Pham TT, Tang R, Balaei AT, de Chazal P. A low-complexity algorithm for detection of atrial fibrillation using an ECG.
Physiol Meas 2018; 39:064003.
60. Parvaneh S, Rubin J, Rahman A, Conroy B, Babaeizadeh S. Analyzing single-lead short ECG recordings using dense convolutional neural networks and feature-based postprocessing to detect atrial fibrillation.
Physiol Meas 2018; 39:084003.
61. He R, Wang K, Zhao N, et al. Automatic detection of atrial fibrillation based on continuous wavelet transform and 2D convolutional neural networks.
Front Physiol 2018; 9:1206.
62. Shao M, Bin G, Wu S, Bin G, Huang J, Zhou Z. Detection of atrial fibrillation from ECG recordings using decision tree ensemble with multilevel features.
Physiol Meas 2018; 39:094008.
63. Hannun AY, Rajpurkar P, Haghpanahi M, et al. Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network.
Nat Med 2019; 25:65–69.
64. Shao M, Zhou Z, Bin G, Bai Y, Wu S. A wearable electrocardiogram telemonitoring system for atrial fibrillation detection.
Sensors (Basel) 2020; 20:606.
65. Oster J, Hopewell JC, Ziberna K, et al. Identification of patients with atrial fibrillation: a big data exploratory analysis of the UK Biobank.
Physiol Meas 2020; 41:025001.
66. Shi H, Wang H, Qin C, Zhao L, Liu C. An incremental learning system for atrial fibrillation detection based on transfer learning and active learning.
Comput Methods Programs Biomed 2020; 187:105219.
67. Hsieh CH, Li YS, Hwang BJ, Hsiao CH. Detection of atrial fibrillation using 1D convolutional neural network.
Sensors (Basel) 2020; 20:606.
68. Huang ML, Wu YS. Classification of atrial fibrillation and normal sinus rhythm based on convolutional neural network.
Biomed Eng Lett 2020; 10:183–193.
69. Marinucci D, Sbrollini A, Marcantoni I, Morettini M, Swenne CA, Burattini L. Artificial neural network for atrial fibrillation identification in portable devices.
Sensors (Basel) 2020; 20:
70. Lown M, Brown M, Brown C, et al. Machine learning detection of atrial fibrillation using wearable technology.
PLoS One 2020; 15:e0227401.
71. Yue Y, Chen C, Liu P, Xing Y, Zhou X. Automatic detection of short-term atrial fibrillation segments based on frequency slice wavelet transform and machine learning techniques.
Sensors (Basel) 2021; 21:3570.
72. Chen X, Cheng Z, Wang S, et al. Atrial fibrillation detection based on multifeature extraction and convolutional neural network for processing ECG signals.
Comput Methods Programs Biomed 2021; 202:106009.
73. Michel P, Ngo N, Pons JF, Delliaux S, Giorgi R. A filter approach for feature selection in classification: application to automatic atrial fibrillation detection in electrocardiogram recordings.
BMC Med Inform Decis Mak 2021; 21: (Suppl 4): 130.
74. Rouhi R, Clausel M, Oster J, Lauer F. An interpretable hand-crafted feature-based model for atrial fibrillation detection.
Front Physiol 2021; 12:657304.
75. Ilbeigipour S, Albadvi A, Akhondzadeh Noughabi E. Real-time heart arrhythmia detection using apache spark structured streaming.
J Healthc Eng 2021; 2021:6624829.
76. Taniguchi H, Takata T, Takechi M, et al. Explainable artificial intelligence model for diagnosis of atrial fibrillation using Holter electrocardiogram waveforms.
Int Heart J 2021; 62:534–539.
77. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines.
Circulation 2019; 140:e596–e646.
78. Dogan MV, Grumbach IM, Michaelson JJ, Philibert RA. Integrated genetic and epigenetic prediction of coronary heart disease in the Framingham Heart Study.
PLoS One 2018; 13:e0190549.
79. Velusamy D, Ramasamy K. Ensemble of heterogeneous classifiers for diagnosis and prediction of coronary artery disease with reduced feature subset.
Comput Methods Programs Biomed 2021; 198:105770.
80. Muhammad LJ, Al-Shourbaji I, Haruna AA, Mohammed IA, Ahmad A, Jibrin MB. Machine learning predictive models for coronary artery disease.
SN Comput Sci 2021; 2:350.
81. Li D, Xiong G, Zeng H, Zhou Q, Jiang J, Guo X. Machine learning-aided risk stratification system for the prediction of coronary artery disease.
Int J Cardiol 2021; 326:30–34.
82. Betancur J, Commandeur F, Motlagh M, et al. Deep learning for prediction of obstructive disease from fast myocardial perfusion SPECT: a multicenter study.
JACC Cardiovasc Imaging 2018; 11:1654–1663.
83. Rahmani R, Niazi P, Naseri M, et al. Improved diagnostic accuracy for myocardial perfusion imaging using artificial neural networks on different input variables including clinical and quantification data.
Rev Esp Med Nucl Imagen Mol (Engl Ed) 2019; 38:275–279.
84. Guner LA, Karabacak NI, Akdemir OU, et al. An open-source framework of neural networks for diagnosis of coronary artery disease from myocardial perfusion SPECT.
J Nucl Cardiol 2010; 17:405–413.
85. Betancur J, Hu LH, Commandeur F, et al. Deep learning analysis of upright-supine high-efficiency SPECT myocardial perfusion imaging for prediction of obstructive coronary artery disease: a multicenter study.
J Nucl Med 2019; 60:664–670.
86. Arsanjani R, Xu Y, Dey D, et al. Improved accuracy of myocardial perfusion SPECT for the detection of coronary artery disease using a support vector machine algorithm.
J Nucl Med 2013; 54:549–555.
87. Bundhun PK, Sookharee Y, Bholee A, Huang F. Application of the SYNTAX score in interventional cardiology: a systematic review and meta-analysis.
Medicine (Baltimore) 2017; 96:e7410.
88. Tang EW, Wong CK, Herbison P. Global Registry of Acute Coronary Events (GRACE) hospital discharge risk score accurately predicts long-term mortality postacute coronary syndrome.
Am Heart J 2007; 153:29–35.
89. Wang H, Zu Q, Chen J, Yang Z, Ahmed MA. Application of artificial intelligence in acute coronary syndrome: a brief literature review.
Adv Ther 2021; 38:5078–5086.
90. Sherazi SWA, Bae JW, Lee JY. A soft voting ensemble classifier for early prediction and diagnosis of occurrences of major adverse cardiovascular events for STEMI and NSTEMI during 2-year follow-up in patients with acute coronary syndrome.
PLoS One 2021; 16:e0249338.
91. Hadanny A, Shouval R, Wu J, et al. Predicting 30-day mortality after ST elevation myocardial infarction: machine learning- based random forest and its external validation using two independent nationwide datasets.
J Cardiol 2021; 78:439–446.
92. Hernesniemi JA, Mahdiani S, Tynkkynen JA, et al. Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome: the MADDEC study.
Ann Med 2019; 51:156–163.
93. Zack CJ, Senecal C, Kinar Y, et al. Leveraging machine learning techniques to forecast patient prognosis after percutaneous coronary intervention.
JACC Cardiovasc Interv 2019; 12:1304–1311.
94. de Carvalho LSF, Gioppato S, Fernandez MD, et al. Machine learning improves the identification of individuals with higher morbidity and avoidable health costs after acute coronary syndromes.
Value Health 2020; 23:1570–1579.
95. D’Ascenzo F, De Filippo O, Gallone G, et al. Machine learning-based prediction of adverse events following an acute coronary syndrome (PRAISE): a modelling study of pooled datasets.
Lancet 2021; 397:199–207.
96. Xue Y, Hu Z, Jing Y, et al. Efficacy assessment of ticagrelor versus clopidogrel in Chinese patients with acute coronary syndrome undergoing percutaneous coronary intervention by data mining and machine-learning decision tree approaches.
J Clin Pharm Ther 2020; 45:1076–1086.
97. Chu J, Dong W, Wang J, He K, Huang Z. Treatment effect prediction with adversarial deep learning using electronic health records.
BMC Med Inform Decis Mak 2020; 20: (Suppl 4): 139.
98. Visco V, Radano I, Campanile A, et al. Predictors of sacubitril/valsartan high dose tolerability in a real world population with HFrEF.
ESC Heart Fail 2022; 9:2909–2917.
99. Aljaaf AJ, Al-Jumeily D, Hussain AJ, Dawson T, Fergus P, Al-Jumaily M. Predicting the likelihood of heart failure with a multi level risk assessment using decision tree. 2015 Third International Conference on Technological Advances in Electrical, Electronics and Computer Engineering (TAEECE) 2015. pp. 101–106.
100. Shah SJ, Katz DH, Selvaraj S, et al. Phenomapping for novel classification of heart failure with preserved ejection fraction.
Circulation 2015; 131:269–279.
101. Shiraishi Y, Kohsaka S, Abe T, et al. West Tokyo Heart Failure Registry Investigators. Validation of the Get With The Guideline-Heart Failure risk score in Japanese patients and the potential improvement of its discrimination ability by the inclusion of B-type natriuretic peptide level.
Am Heart J 2016; 171:33–39.
102. Sartipy U, Dahlstrom U, Edner M, Lund LH. Predicting survival in heart failure: validation of the MAGGIC heart failure risk score in 51,043 patients from the Swedish heart failure registry.
Eur J Heart Fail 2014; 16:173–179.
103. Bo X, Zhang Y, Liu Y, Kharbuja N, Chen L. Performance of the heart failure risk scores in predicting 1 year mortality and short-term readmission of patients [published online ahead of print, 2022 Nov 3].
ESC Heart Fail 2022; 10.1002/ehf2.14208.
104. Kwon JM, Kim KH, Jeon KH, et al. Artificial intelligence algorithm for predicting mortality of patients with acute heart failure.
PLoS One 2019; 14:e0219302.
105. Lorenzoni G, Sabato SS, Lanera C, et al. Comparison of machine learning techniques for prediction of hospitalization in heart failure patients.
J Clin Med 2019; 8:1298.
106. Stehlik J, Schmalfuss C, Bozkurt B, et al. Continuous wearable monitoring analytics predict heart failure hospitalization: the LINK-HF Multicenter Study.
Circ Heart Fail 2020; 13:e006513.
107. Anand IS, Tang WH, Greenberg BH, et al. Design and performance of a multisensor heart failure monitoring algorithm: results from the multisensor monitoring in congestive heart failure (MUSIC) study.
J Card Fail 2012; 18:289–295.
108. Visco V, Esposito C, Vitillo P, Vecchione C, Ciccarelli M. It is easy to see, but it is better to foresee: a case report on the favourable alliance between CardioMEMS and levosimendan.
Eur Heart J Case Rep 2020; 4:1–5.
109. Yu CM, Wang L, Chau E, et al. Intrathoracic impedance monitoring in patients with heart failure: correlation with fluid status and feasibility of early warning preceding hospitalization.
Circulation 2005; 112:841–848.
110. Boehmer JP, Hariharan R, Devecchi FG, et al. A multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE Study.
JACC Heart Fail 2017; 5:216–225.
111. Heywood JT, Zalawadiya S, Bourge RC, et al. Sustained reduction in pulmonary artery pressures and hospitalizations during 2 years of ambulatory monitoring [published online ahead of print, 2022 Nov 2].
J Card Fail 2022; S1071-9164(22)01170-8.
112. Visco V, Esposito C, Manzo M, et al. A multistep approach to deal with advanced heart failure: a case report on the positive effect of cardiac contractility modulation therapy on pulmonary pressure measured by CardioMEMS.
Front Cardiovasc Med 2022; 9:874433.
113. Kannel WB. Some lessons in cardiovascular epidemiology from Framingham.
Am J Cardiol 1976; 37:269–282.
114. DeFilippis AP, Young R, McEvoy JW, et al. Risk score overestimation: the impact of individual cardiovascular risk factors and preventive therapies on the performance of the American Heart Association-American College of Cardiology-Atherosclerotic Cardiovascular Disease risk score in a modern multiethnic cohort.
Eur Heart J 2017; 38:598–608.
115. Kavousi M, Leening MJ, Nanchen D, et al. Comparison of application of the ACC/AHA guidelines, Adult Treatment Panel III guidelines, and European Society of Cardiology guidelines for cardiovascular disease prevention in a European cohort.
JAMA 2014; 311:1416–1423.
116. Muntner P, Colantonio LD, Cushman M, et al. Validation of the atherosclerotic cardiovascular disease pooled cohort risk equations.
JAMA 2014; 311:1406–1415.
117. Ambale-Venkatesh B, Yang X, Wu CO, et al. Cardiovascular event prediction by machine learning: the Multi-Ethnic Study of Atherosclerosis.
Circ Res 2017; 121:1092–1101.
118. Maor E, Sara JD, Orbelo DM, Lerman LO, Levanon Y, Lerman A. Voice signal characteristics are independently associated with coronary artery disease.
Mayo Clin Proc 2018; 93:840–847.
119. Silverio A, Cavallo P, De Rosa R, Galasso G. Big health data and cardiovascular diseases: a challenge for research, an opportunity for clinical care.
Front Med (Lausanne) 2019; 6:36.
120. Izzo C, Visco V, Gambardella J,
et al. Cardiovascular implications of miRNAs in COVID-19 [published online ahead of print, 2022 Jul 2].
J Pharmacol Exp Ther. 2022;JPET-MR-2022-001210.
121. Citro R, Pontone G, Bellino M, et al. Role of multimodality imaging in evaluation of cardiovascular involvement in COVID-19.
Trends Cardiovasc Med 2021; 31:8–16.
122. Visco V, Vitale C, Rispoli A, et al. Post-COVID-19 syndrome: involvement and interactions between respiratory, cardiovascular and nervous systems.
J Clin Med 2022; 11:524.
123. Silverio A, Di Maio M, Citro R, et al. Cardiovascular risk factors and mortality in hospitalized patients with COVID-19: systematic review and meta-analysis of 45 studies and 18,300 patients.
BMC Cardiovasc Disord 2021; 21:23.
124. Polito MV, Silverio A, Bellino M, et al. Cardiovascular involvement in COVID-19: what sequelae should we expect?
Cardiol Ther 2021; 10:377–396.
125. Dong D, Tang Z, Wang S, et al. The role of imaging in the detection and management of COVID-19: a review.
IEEE Rev Biomed Eng 2021; 14:16–29.
126. Suri JS, Puvvula A, Majhail M, et al. Integration of cardiovascular risk assessment with COVID-19 using artificial intelligence.
Rev Cardiovasc Med 2020; 21:541–560.
127. van Smeden M, Heinze G, Van Calster B, et al. Critical appraisal of artificial intelligence-based prediction models for cardiovascular disease.
Eur Heart J 2022; 43:2921–2930.
128. Krajcer Z. Artificial intelligence in cardiovascular medicine: historical overview, current status, and future directions.
Tex Heart Inst J 2022; 49:e207527.
129. Yan Y, Zhang JW, Zang GY, Pu J. The primary use of artificial intelligence in cardiovascular diseases: what kind of potential role does artificial intelligence play in future medicine?
J Geriatr Cardiol 2019; 16:585–591.
130. Gupta MD, Kunal S, Girish MP, Gupta A, Yadav R. Artificial intelligence in cardiology: the past, present and future.
Indian Heart J 2022; 74:265–269.
131. Sermesant M, Delingette H, Cochet H, Jais P, Ayache N. Applications of artificial intelligence in cardiovascular imaging.
Nat Rev Cardiol 2021; 18:600–609.
132. Smith B, Magnani JW. New technologies, new disparities: the intersection of electronic health and digital health literacy.
Int J Cardiol 2019; 292:280–282.
133. Su J, Zhang Y, Ke QQ, Su JK, Yang QH. Mobilizing artificial intelligence to cardiac telerehabilitation.
Rev Cardiovasc Med 2022; 23:45.
134. McCall B. What does the GDPR mean for the medical community?
Lancet 2018; 391:1249–1250.
135. Olhede SC, Wolfe PJ. The growing ubiquity of algorithms in society: implications, impacts and innovations.
Philos Trans A Math Phys Eng Sci 2018; 376:
136. Vayena E, Blasimme A, Cohen IG. Machine learning in medicine: addressing ethical challenges.
PLoS Med 2018; 15:e1002689.
137. Reddy S, Allan S, Coghlan S, Cooper P. A governance model for the application of AI in healthcare.
J Am Med Inform Assoc 2020; 27:491–497.
138. Char DS, Shah NH, Magnus D. Implementing machine learning in healthcare: addressing ethical challenges.
N Engl J Med 2018; 378:981–983.
139. Char DS, Abramoff MD, Feudtner C. Identifying ethical considerations for machine learning healthcare applications.
Am J Bioeth 2020; 20:7–17.
140. Asan O, Bayrak AE, Choudhury A. Artificial intelligence and human trust in healthcare: focus on clinicians.
J Med Internet Res 2020; 22:e15154.
141. Duncker D, Ding WY, Etheridge S, et al. Smart wearables for cardiac monitoring-real-world use beyond atrial fibrillation.
Sensors (Basel) 2021; 21:2539.
142. Seetharam K, Balla S, Bianco C, et al. Applications of machine learning in cardiology.
Cardiol Ther 2022; 11:355–368.
