scholarly journals Machine learning for subtype definition and risk prediction in heart failure, acute coronary syndromes and atrial fibrillation: systematic review of validity and clinical utility

BMC Medicine ◽  
2021 ◽  
Vol 19 (1) ◽  
Author(s):  
Amitava Banerjee ◽  
Suliang Chen ◽  
Ghazaleh Fatemifar ◽  
Mohamad Zeina ◽  
R. Thomas Lumbers ◽  
...  
Keyword(s):  
Machine Learning ◽  
Heart Failure ◽  
Atrial Fibrillation ◽  
Systematic Review ◽  
Cardiovascular Diseases ◽  
Risk Prediction ◽  
Clinical Setting ◽  
Acute Coronary Syndromes ◽  
Clinical Utility ◽  
Coronary Syndromes

Abstract Background Machine learning (ML) is increasingly used in research for subtype definition and risk prediction, particularly in cardiovascular diseases. No existing ML models are routinely used for cardiovascular disease management, and their phase of clinical utility is unknown, partly due to a lack of clear criteria. We evaluated ML for subtype definition and risk prediction in heart failure (HF), acute coronary syndromes (ACS) and atrial fibrillation (AF). Methods For ML studies of subtype definition and risk prediction, we conducted a systematic review in HF, ACS and AF, using PubMed, MEDLINE and Web of Science from January 2000 until December 2019. By adapting published criteria for diagnostic and prognostic studies, we developed a seven-domain, ML-specific checklist. Results Of 5918 studies identified, 97 were included. Across studies for subtype definition (n = 40) and risk prediction (n = 57), there was variation in data source, population size (median 606 and median 6769), clinical setting (outpatient, inpatient, different departments), number of covariates (median 19 and median 48) and ML methods. All studies were single disease, most were North American (n = 61/97) and only 14 studies combined definition and risk prediction. Subtype definition and risk prediction studies respectively had limitations in development (e.g. 15.0% and 78.9% of studies related to patient benefit; 15.0% and 15.8% had low patient selection bias), validation (12.5% and 5.3% externally validated) and impact (32.5% and 91.2% improved outcome prediction; no effectiveness or cost-effectiveness evaluations). Conclusions Studies of ML in HF, ACS and AF are limited by number and type of included covariates, ML methods, population size, country, clinical setting and focus on single diseases, not overlap or multimorbidity. Clinical utility and implementation rely on improvements in development, validation and impact, facilitated by simple checklists. We provide clear steps prior to safe implementation of machine learning in clinical practice for cardiovascular diseases and other disease areas.

P5705Machine learning for phenotyping and risk prediction in cardiovascular diseases: a systematic review

2019 ◽  
Vol 40 (Supplement_1) ◽  
Author(s):  
A Banerjee ◽  
S Chen ◽  
G Fatemifar ◽  
H Hemingway ◽  
T Lumbers ◽  
...  
Keyword(s):  
Machine Learning ◽  
Systematic Review ◽  
Cardiovascular Diseases ◽  
North America ◽  
Risk Prediction ◽  
English Language ◽  
Support Vector ◽  
Positive Finding ◽  
Disease Clustering ◽  
The Mean

Abstract Introduction Heart failure (HF), acute coronary syndromes (ACS) and atrial fibrillation (AF) are among the commonest cardiovascular diseases (CVD), frequently co-exist and share pathophysiology. Definitions of diagnosis and prognosis are suboptimal. Machine learning (ML) is increasingly used in subtype definition and risk prediction, but the design, methods and results of studies have not been appraised. Purpose To conduct a systematic review of ML for discovery of new subtypes and risk prediction in HF, ACS and AF. Methods PubMed, MEDLINE, and Web of Science databases were searched (January 2000-August 2018) for English language publications with agreed search terms pertaining to machine learning, clustering, CVD, subtype and risk prediction. The baseline characteristics of the study population, the method of ML, covariates and results were extracted for each study. Results Of 5012 identified studies, 43 met inclusion criteria. Of the 33 studies of unsupervised ML for disease clustering (mean n=2354; min 117, max 44886), there were 22 in HF, 9 in ACS and 2 in AF. 22/33 studies involved <1000 individuals and 24 were based in North America. Across diseases, 27 studies were in outpatients, and 5 used trial data. The mean number of covariates used was 26; most commonly demographic and symptom variables. The ML methods used were partitional (n=12), hierarchical (n=4), self-organising map (n=1) and hidden Markov model (n=1). Most studies used only one ML method (n=25). Only 15 studies validated or replicated findings. 20/33 studies found 2 or 3 disease clusters, Most studies found 2–3 clusters (20/33) and most clusters were based on physical or physiological characteristics (30/33). Of the 10 studies of supervised ML for risk prediction (mean n=43003; min 228, max 378256), 4 were in HF, 5 in ACS and 1 in AF. 2/11 studies involved <1000 individuals and most were from North America (n=6). All studies had an observational design, used at least 2 ML methods and validated or replicated findings. The setting was varied: primary care (n=2), emergency department (n=2), inpatient (n=4) and mixed (n=2). The mean number of covariates was 102. The commonest ML methods were neural networks (n=5), random forest (n=4) and support vector machine (n=4). All studies showed positive finding, i.e. ML approaches improved risk prediction. Conclusions Studies to-date of ML in HF, ACS and AF have focused on North America (68.2%), and 50% included less than 1000 individuals. Moreover, there is heterogeneity in clinical setting, study designs for data collection and ML methods used. Comparison between methods of ML and validation are common to studies of risk prediction but not disease clustering. There is likely to be a publication bias of ML studies in HF, AF and ACS. ML may improve data-driven characterisation of CVD but consensus guidelines for reporting of research using ML are urgently needed to ensure the internal and external validity and applicability of study findings. Acknowledgement/Funding Innovative Medicines Initiative (European Union)

scholarly journals Effect of alirocumab on incidence of atrial fibrillation after acute coronary syndromes: insights from ODYSSEY OUTCOMES

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
R Lopes ◽  
P.G Steg ◽  
D.L Bhatt ◽  
V.A Bittner ◽  
A Dauchy ◽  
...  
Keyword(s):  
Heart Failure ◽  
Atrial Fibrillation ◽  
Acute Coronary Syndromes ◽  
Strong Predictor ◽  
Potential Effect ◽  
Major Adverse Cardiovascular Events ◽  
Funding Source ◽  
Private Company ◽  
History Of ◽  
Coronary Syndromes

Abstract Background Atrial fibrillation (AF) is a marker of risk in patients presenting with acute coronary syndromes (ACS). The potential effect of inhibiting proprotein convertase subtilisin/kexin type 9 (PCSK9) on the incidence of AF is unknown. Methods The ODYSSEY OUTCOMES trial compared randomized treatment with the PCSK9 inhibitor alirocumab or placebo in patients with recent ACS and residual dyslipidaemia despite optimal statin therapy. The current analysis determined: 1) whether alirocumab treatment influenced incident AF; 2) whether a history of AF influenced the risk of major adverse cardiovascular events (MACE); and 3) whether there was interaction between AF at baseline and randomized treatment on MACE. AF was determined from the medical history and investigator reports of adverse events. Results Of 18,924 participants, 662 (3.5%) had a history of AF at randomization and 18,262 (96.5%) had no history of AF. Of the latter category, 499 (2.7%) had incident AF. Older age, randomization in South America or Eastern Europe, history of heart failure or myocardial infarction, and higher body mass index were factors associated with incident AF. Treatment with alirocumab or placebo did not influence incident AF (2.2% vs 2.6%, respectively; hazard ratio 0.90, 95% confidence interval 0.75–1.08; Figure). Patients with a history of AF had a greater burden of comorbidities, including cerebrovascular disease, peripheral artery disease, hypertension and heart failure; they also had higher rates of MACE (Table). There was no significant interaction between AF and randomized treatment on risk of MACE (P interaction=0.78) Conclusions Although treatment with alirocumab did not significantly modify the risk of incident AF after ACS in this analysis, future studies with more sensitive and systematic methods of ascertainment may be warranted. History of AF is a strong predictor of risk of recurrent MACE after ACS. Funding Acknowledgement Type of funding source: Private company. Main funding source(s): Sanofi, Regeneron Pharmaceuticals, Inc

Prognostic utility of routine assessment of left ventricular global longitudinal strain and peak atrial reservoir strain in stable heart disease patients

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
M.I Bolog ◽  
M Dumitrescu ◽  
E Pacuraru ◽  
F Romanoschi ◽  
A Rapa
Keyword(s):  
Heart Failure ◽  
Atrial Fibrillation ◽  
Heart Disease ◽  
Acute Coronary Syndromes ◽  
Longitudinal Strain ◽  
Global Longitudinal Strain ◽  
Left Ventricular ◽  
Cardiac Events ◽  
Coronary Syndromes ◽  
Routine Assessment

Abstract Background Left ventricular ejection fraction above 40% is a poor predictor of cardiac outcomes. Strain imaging seems to add incremental prognostic value in various heart conditions. Purpose The aim of the study is to investigate whether assessment of both left ventricular global longitudinal strain (LVGLS) and peak atrial reservoir strain (PARS) in routine daily practice is useful in predicting cardiac events. Methods We prospectively enrolled 300 patients (212 men, mean age 64.8±10.9 years old) with stable cardiac disease, referred for echocardiographic examination and eligible for strain imaging. Conventional and speckle tracking 2D rest echocardiography were performed and clinical variables were recorded. We excluded patients with acute coronary syndromes, severe valvular disease, cardiomyopathies, arrhythmia and class IV NYHA. Results During a median follow up of 30 months, there were 111 cardiac events (CE) recorded including: 7 cardiac deaths (CD), 25 acute coronary syndromes (ACS), 45 hospitalisations for worsening heart failure (WHF), 23 episodes of atrial fibrillation (AF), 11 stroke (S). Average EF, LVGLS and PARS in patients with CE were significantly lower than in patients without CE (55.8±8.4%, −17.3±3.9% and 16.3±9.8% vs. 60.4±8.4%, −20.1±3.7% and 28.4±8.2%; p&lt;0.05, p&lt;0.01, p&lt;0.01). In univariate analysis, lower LVGLS, respectively lower PARS were associated with a higher risk of cardiac events [Hazard Ratio (HR)1.26; 95% CI (confidence interval): 1.08–1.34; p&lt;0.01 per 1% decrease, respectively HR 1.38; 95% CI: 1.16–1.42; p&lt;0.01 per 1% decrease]. On multivariate analysis this association was independent after adjustment for age, gender, hypertension, diabetes, ejection fraction, left atrial indexed volume. Lower LVGLS was a better predictor of a composite of ACS and CD (HR 1.38 per 1% decrease 95% CI: 1.16–1.48; p&lt;0.01). Lower PARS had a stronger association with AF, S and WHF (HR 1.49 per 1% decrease 95% CI: 1.19–1.58; p&lt;0.01). In a model defined by depressed LVGLS (more positive than −18%) adding lower PALS, with a cut off point of 20%, significantly improved prediction of CE (C-statistic increased from 0.68 to 0.83, p&lt;0.001). Conclusions Left ventricular global longitudinal strain and peak atrial reservoir strain are independent predictors of cardiac events patients with stable heart disease. Acute coronary syndromes and death were better predicted by depressed LVGLS and onset of atrial fibrillation, stroke and worsening heart failure were better predicted by lower PARS. Routine assessment of both parameters improves significantly prediction of cardiac events and helps clinical decision making. Funding Acknowledgement Type of funding source: None

Sign in / Sign up

Export Citation Format

Share Document