Machine Learning for Assessment of Coronary Artery Disease in Cardiac CT: A Survey.
Ontology highlight
ABSTRACT:
SUBMITTER: Hampe N
PROVIDER: S-EPMC6988816 | biostudies-literature | 2019
REPOSITORIES: biostudies-literature
Publications
Machine Learning for Assessment of Coronary Artery Disease in Cardiac CT: A Survey.
Hampe Nils N Wolterink Jelmer M JM van Velzen Sanne G M SGM Leiner Tim T Išgum Ivana I
Frontiers in cardiovascular medicine 20191126
Cardiac computed tomography (CT) allows rapid visualization of the heart and coronary arteries with high spatial resolution. However, analysis of cardiac CT scans for manifestation of coronary artery disease is time-consuming and challenging. Machine learning (ML) approaches have the potential to address these challenges with high accuracy and consistent performance. In this mini review, we present a survey of the literature on ML-based analysis of coronary artery disease in cardiac CT. We summa ...[more]
Similar Datasets
Project description:The aim of this review is to provide an overview of different functional cardiac CT techniques which can be used to supplement assessment of the coronary arteries to establish the significance of coronary artery stenoses. We focus on cine-CT, CT-FFR, CT-myocardial perfusion and how developments in machine learning can supplement these techniques.
| S-EPMC7465866 | biostudies-literature
Project description:BackgroundArtificial intelligence (AI) techniques are increasingly applied to cardiovascular (CV) medicine in arenas ranging from genomics to cardiac imaging analysis. Cardiac Phase Space Tomography Analysis (cPSTA), employing machine-learned linear models from an elastic net method optimized by a genetic algorithm, analyzes thoracic phase signals to identify unique mathematical and tomographic features associated with the presence of flow-limiting coronary artery disease (CAD). This novel approach does not require radiation, contrast media, exercise, or pharmacological stress. The objective of this trial was to determine the diagnostic performance of cPSTA in assessing CAD in patients presenting with chest pain who had been referred by their physician for coronary angiography.MethodsThis prospective, multicenter, non-significant risk study was designed to: 1) develop machine-learned algorithms to assess the presence of CAD (defined as one or more ? 70% stenosis, or fractional flow reserve ? 0.80) and 2) test the accuracy of these algorithms prospectively in a naïve verification cohort. This report is an analysis of phase signals acquired from 606 subjects at rest just prior to angiography. From the collective phase signal data, features were extracted and paired with the known angiographic results. A development set, consisting of signals from 512 subjects, was used for machine learning to determine an algorithm that correlated with significant CAD. Verification testing of the algorithm was performed utilizing previously untested phase signals from 94 subjects.ResultsThe machine-learned algorithm had a sensitivity of 92% (95% CI: 74%-100%) and specificity of 62% (95% CI: 51%-74%) on blind testing in the verification cohort. The negative predictive value (NPV) was 96% (95% CI: 85%-100%).ConclusionsThese initial multicenter results suggest that resting cPSTA may have comparable diagnostic utility to functional tests currently used to assess CAD without requiring cardiac stress (exercise or pharmacological) or exposure of the patient to radioactivity.
| S-EPMC6082503 | biostudies-literature
Project description:Coronary artery disease (CAD) remains the leading cause of death and morbidity worldwide. To date, diagnostic evaluation of patients with suspected CAD has relied upon the use of physiologic non-invasive testing by stress electrocardiography, echocardiography, myocardial perfusion imaging (MPI) and magnetic resonance imaging. Indeed, the importance of physiologic evaluation of CAD has been highlighted by large-scale randomized trials that demonstrate the propitious benefit of an integrated anatomic-physiologic evaluation method by performing lesion-specific ischemia assessment by fractional flow reserve (FFR)-widely considered the "gold" standard for ischemia assessment-at the time of invasive angiography. Coronary CT angiography (CCTA) has emerged as an attractive non-invasive test for anatomic illustration of the coronary arteries and atherosclerotic plaque. In a series of prospective multicenter trials, CCTA has been proven as having high diagnostic performance for stenosis detection as compared to invasive angiography. Nevertheless, CCTA evaluation of obstructive stenoses is prone to overestimation of severity and further, detection of stenoses by CCTA does not reliably determine the hemodynamic significance of the visualized lesions. Recently, a series of technological innovations have advanced the possibility of CCTA to enable physiologic evaluation of CAD, thereby creating the potential of this test to provide an integrated anatomic-physiologic assessment of CAD. These advances include rest-stress MPI by CCTA as well as the use of computational fluid dynamics to non-invasively calculate FFR from a typically acquired CCTA. The purpose of this review is to summarize the most recent data addressing these 2 physiologic methods of CAD evaluation by CCTA.
| S-EPMC3744730 | biostudies-literature
Project description:Positron emission tomography (PET) is increasingly being applied in the evaluation of myocardial perfusion. Cardiac PET can be performed with an increasing variety of cyclotron- and generator-produced radiotracers. Compared with single photon emission computed tomography, PET offers lower radiation exposure, fewer artifacts, improved spatial resolution, and, most important, improved diagnostic performance. With its capacity to quantify rest-peak stress left ventricular systolic function as well as coronary flow reserve, PET is superior to other methods for the detection of multivessel coronary artery disease and, potentially, for risk stratification. Coronary artery calcium scoring may be included for further risk stratification in patients with normal perfusion imaging findings. Furthermore, PET allows quantification of absolute myocardial perfusion, which also carries substantial prognostic value. Hybrid PET-computed tomography scanners allow functional evaluation of myocardial perfusion combined with anatomic characterization of the epicardial coronary arteries, thereby offering great potential for both diagnosis and management. Additional studies to further validate the prognostic value and cost effectiveness of PET are warranted.
| S-EPMC3173713 | biostudies-other
Project description:BACKGROUNDCeramides are sphingolipids that play causative roles in diabetes and heart disease, with their serum levels measured clinically as biomarkers of cardiovascular disease (CVD).METHODSWe performed targeted lipidomics on serum samples from individuals with familial coronary artery disease (CAD) (n = 462) and population-based controls (n = 212) to explore the relationship between serum sphingolipids and CAD, using unbiased machine learning to identify sphingolipid species positively associated with CAD.RESULTSNearly every sphingolipid measured (n = 30 of 32) was significantly elevated in subjects with CAD compared with measurements in population controls. We generated a novel sphingolipid-inclusive CAD risk score, termed SIC, that demarcates patients with CAD independently and more effectively than conventional clinical CVD biomarkers including serum LDL cholesterol and triglycerides. This new metric comprises several minor lipids that likely serve as measures of flux through the ceramide biosynthesis pathway rather than the abundant deleterious ceramide species that are included in other ceramide-based scores.CONCLUSIONThis study validates serum ceramides as candidate biomarkers of CVD and suggests that comprehensive sphingolipid panels should be considered as measures of CVD.FUNDINGThe NIH (DK112826, DK108833, DK115824, DK116888, and DK116450); the Juvenile Diabetes Research Foundation (JDRF 3-SRA-2019-768-A-B); the American Diabetes Association; the American Heart Association; the Margolis Foundation; the National Cancer Institute, NIH (5R00CA218694-03); and the Huntsman Cancer Institute Cancer Center Support Grant (P30CA040214).
| S-EPMC7269567 | biostudies-literature
Project description:Coronary artery disease (CAD) is a major and common disease that poses a threat to human health. Recent studies suggested that epicardial fat may have an important role in the pathogenesis of CAD. Therefore, the association between epicardial fat volume (EFV) and left ventricular function with CAD was investigated in the present study. A total of 61 patients with suspected CAD who underwent CT scanning were enrolled. Baseline data, parameters of left heart function and EFV of the subjects were collected and analyzed. The degree of coronary artery lesions was assessed using the Gensini score. Pearson's correlation analysis and a logistic regression model were applied to assess the association between EFV and risk factors for CAD, the Gensini score and left ventricular function index. A total of 29 female and 32 male subjects with a median age of 63 years were enrolled. The median body mass index (BMI) of the subjects was 23.37 kg/m2 and the median EFV was 86.41 cm3. It was revealed that risk factors of CAD, specially hypertension, diabetes mellitus, dyslipidemia, history of myocardial infarction and smoking, had no significant association with the EFV (P>0.05); however, the EFV was significantly positively correlated with the BMI (r=0.479, P<0.0001), interventricular septal thickness (r=0.436, P=0.004), left ventricular posterior wall thickness (r=0.350, P=0.0058), left ventricular end diastolic diameter (r=0.265, P=0.0388), left ventricular mass (r=0.445, P=0.0003) and left ventricular mass index (r=0.371, P=0.0035). However, no correlation was identified between the EFV and the Gensini score (r=0.131, P=0.3137). In conclusion, the EFV measured by cardiac CT scanning was positively correlated with the BMI and left ventricular function, but was not associated with the presence of CAD according to the Gensini scores.
| S-EPMC7444410 | biostudies-literature
Project description:RationaleIndividuals with HIV are at increased risk for coronary artery disease (CAD). Early detection of subclinical CAD by assessment of coronary artery calcium (CAC) may help risk stratify and prevent CAD events in these individuals. However, the current standard to quantify CAC i.e. Agatston scoring requires EKG-gated cardiac CT imaging.ObjectiveTo determine if the assessment of CAC using non-EKG-gated chest CT and the Weston scoring system is a useful surrogate for Agatston scores in HIV-infected and HIV-uninfected individuals.Methods and measurementsCAC was assessed by both the Weston and Agatston score in 108 men enrolled in the Multicenter AIDS Cohort Study.ResultsParticipants were 55.2 (IQR 50.4; 59.9) years old and 62 (57.4%) were seropositive for HIV. Inter-observer agreement (rs = 0.94, ? = 90.0%, p<0.001, n = 21) and intra-observer agreement (rs = 0.95, ? = 95.2%, p<0.001, n = 97) for category of Weston score were excellent. Weston scores were associated with similar CAD risk factors as Agatston scores (age, race, HDL cholesterol level, all p<0.05) in our cohort. There was excellent correlation (rs = 0.92, p<0.001) and agreement (?w = 0.77, p<0.001) between Weston and Agatston scores.ConclusionsThis study is the first to examine calcium scoring using chest CT in HIV-infected individuals and to independently validate the Weston score as a surrogate for the Agatston score. In clinical or research settings where EKG-gated cardiac CT is not feasible for the assessment of coronary calcium, Weston scoring by using chest CT should be considered.
| S-EPMC5409142 | biostudies-literature
Project description:MotivationSelecting the optimal machine learning (ML) model for a given dataset is often challenging. Automated ML (AutoML) has emerged as a powerful tool for enabling the automatic selection of ML methods and parameter settings for the prediction of biomedical endpoints. Here, we apply the tree-based pipeline optimization tool (TPOT) to predict angiographic diagnoses of coronary artery disease (CAD). With TPOT, ML models are represented as expression trees and optimal pipelines discovered using a stochastic search method called genetic programing. We provide some guidelines for TPOT-based ML pipeline selection and optimization-based on various clinical phenotypes and high-throughput metabolic profiles in the Angiography and Genes Study (ANGES).ResultsWe analyzed nuclear magnetic resonance-derived lipoprotein and metabolite profiles in the ANGES cohort with a goal to identify the role of non-obstructive CAD patients in CAD diagnostics. We performed a comparative analysis of TPOT-generated ML pipelines with selected ML classifiers, optimized with a grid search approach, applied to two phenotypic CAD profiles. As a result, TPOT-generated ML pipelines that outperformed grid search optimized models across multiple performance metrics including balanced accuracy and area under the precision-recall curve. With the selected models, we demonstrated that the phenotypic profile that distinguishes non-obstructive CAD patients from no CAD patients is associated with higher precision, suggesting a discrepancy in the underlying processes between these phenotypes.Availability and implementationTPOT is freely available via http://epistasislab.github.io/tpot/.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC7703753 | biostudies-literature
Project description:BackgroundCoronary inflammation induces dynamic changes in the balance between water and lipid content in perivascular adipose tissue (PVAT), as captured by perivascular Fat Attenuation Index (FAI) in standard coronary CT angiography (CCTA). However, inflammation is not the only process involved in atherogenesis and we hypothesized that additional radiomic signatures of adverse fibrotic and microvascular PVAT remodelling, may further improve cardiac risk prediction.Methods and resultsWe present a new artificial intelligence-powered method to predict cardiac risk by analysing the radiomic profile of coronary PVAT, developed and validated in patient cohorts acquired in three different studies. In Study 1, adipose tissue biopsies were obtained from 167 patients undergoing cardiac surgery, and the expression of genes representing inflammation, fibrosis and vascularity was linked with the radiomic features extracted from tissue CT images. Adipose tissue wavelet-transformed mean attenuation (captured by FAI) was the most sensitive radiomic feature in describing tissue inflammation (TNFA expression), while features of radiomic texture were related to adipose tissue fibrosis (COL1A1 expression) and vascularity (CD31 expression). In Study 2, we analysed 1391 coronary PVAT radiomic features in 101 patients who experienced major adverse cardiac events (MACE) within 5 years of having a CCTA and 101 matched controls, training and validating a machine learning (random forest) algorithm (fat radiomic profile, FRP) to discriminate cases from controls (C-statistic 0.77 [95%CI: 0.62-0.93] in the external validation set). The coronary FRP signature was then tested in 1575 consecutive eligible participants in the SCOT-HEART trial, where it significantly improved MACE prediction beyond traditional risk stratification that included risk factors, coronary calcium score, coronary stenosis, and high-risk plaque features on CCTA (Δ[C-statistic] = 0.126, P < 0.001). In Study 3, FRP was significantly higher in 44 patients presenting with acute myocardial infarction compared with 44 matched controls, but unlike FAI, remained unchanged 6 months after the index event, confirming that FRP detects persistent PVAT changes not captured by FAI.ConclusionThe CCTA-based radiomic profiling of coronary artery PVAT detects perivascular structural remodelling associated with coronary artery disease, beyond inflammation. A new artificial intelligence (AI)-powered imaging biomarker (FRP) leads to a striking improvement of cardiac risk prediction over and above the current state-of-the-art.
| S-EPMC6855141 | biostudies-literature
Project description:Machine learning (ML) algorithms have been applied to predicting coronary artery disease (CAD). Our purpose was to utilize autoantibody isotypes against four different unmodified and malondialdehyde (MDA)-modified peptides among Taiwanese with CAD and healthy controls (HCs) for CAD prediction. In this study, levels of MDA, MDA-modified protein (MDA-protein) adducts, and autoantibody isotypes against unmodified peptides and MDA-modified peptides were measured with enzyme-linked immunosorbent assay (ELISA). To improve the performance of ML, we used decision tree (DT), random forest (RF), and support vector machine (SVM) coupled with five-fold cross validation and parameters optimization. Levels of plasma MDA and MDA-protein adducts were higher in CAD patients than in HCs. IgM anti-IGKC76-99 MDA and IgM anti-A1AT284-298 MDA decreased the most in patients with CAD compared to HCs. In the experimental results of CAD prediction, the decision tree classifier achieved an area under the curve (AUC) of 0.81; the random forest classifier achieved an AUC of 0.94; the support vector machine achieved an AUC of 0.65 for differentiating between CAD patients with stenosis rates of 70% and HCs. In this study, we demonstrated that autoantibody isotypes imported into machine learning algorithms can lead to accurate models for clinical use.
| S-EPMC8229983 | biostudies-literature