scholarly journals Population Bias in Polygenic Risk Prediction Models for Coronary Artery Disease

2020 ◽  
Vol 13 (6) ◽  
Author(s):  
Damian Gola ◽  
Jeanette Erdmann ◽  
Kristi Läll ◽  
Reedik Mägi ◽  
Bertram Müller-Myhsok ◽  
...  
Keyword(s):  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Risk Prediction ◽  
Prediction Models ◽  
Data Sets ◽  
Risk Prediction Models ◽  
Polygenic Risk ◽  
Testing Data ◽  
Artery Disease ◽  
European Populations

Background: Individual risk prediction based on genome-wide polygenic risk scores (PRSs) using millions of genetic variants has attracted much attention. It is under debate whether PRS models can be applied—without loss of precision—to populations of similar ethnic but different geographic background than the one the scores were trained on. Here, we examine how PRS trained in population-specific but European data sets perform in other European subpopulations in distinguishing between coronary artery disease patients and healthy individuals. Methods: We use data from UK and Estonian biobanks (UKB, EB) as well as case-control data from the German population (DE) to develop and evaluate PRS in the same and different populations. Results: PRSs have the highest performance in their corresponding population testing data sets, whereas their performance significantly drops if applied to testing data sets from different European populations. Models trained on DE data revealed area under the curves in independent testing sets in DE: 0.6752, EB: 0.6156, and UKB: 0.5989; trained on EB and tested on EB: 0.6565, DE: 0.5407, and UKB: 0.6043; trained on UKB and tested on UKB: 0.6133, DE: 0.5143, and EB: 0.6049. Conclusions: This result has a direct impact on the clinical usability of PRS for risk prediction models using PRS: a population effect must be kept in mind when applying risk estimation models, which are based on additional genetic information even for individuals from different European populations of the same ethnicity.

scholarly journals Utility of family history in disease prediction in the era of polygenic scores

2021 ◽  
Author(s):  
Brooke N Wolford ◽  
Ida Surakka ◽  
Sarah E Graham ◽  
Jonas B Nielsen ◽  
Wei Zhou ◽  
...  
Keyword(s):  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Family History ◽  
Risk Prediction ◽  
Prediction Models ◽  
Preventive Treatment ◽  
Health Study ◽  
Polygenic Scores ◽  
Artery Disease ◽  
The Uk

Clinicians have historically used family history and other risk prediction algorithms to guide patient care and preventive treatment such as statin therapeutics for coronary artery disease. As polygenic scores move towards clinical use, we have begun to consider the interplay of these scores with other predictors for optimal second generation risk prediction. Here, we assess the use of family history and polygenic scores as independent predictors of coronary artery disease and type 2 diabetes. We highlight considerations for use of family history as a predictor of these two diseases after evaluating their effectiveness in the Trøndelag Health Study and the UK Biobank. From these, we advocate for collection of high resolution family history variables in biobanks for future prediction models.

Probabilistic Graphical Modeling for Estimating Risk of Coronary Artery Disease: Applications of a Flexible Machine-Learning Method

2019 ◽  
Vol 39 (8) ◽  
pp. 1032-1044 ◽  
Author(s):  
Alind Gupta ◽  
Justin J. Slater ◽  
Devon Boyne ◽  
Nicholas Mitsakakis ◽  
Audrey Béliveau ◽  
...  
Keyword(s):  
Machine Learning ◽  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Risk Prediction ◽  
Prediction Models ◽  
The United States ◽  
Individual Risk ◽  
Support Vector ◽  
Data Set ◽  
Artery Disease

Objectives. Coronary artery disease (CAD) is the leading cause of death and disease burden worldwide, causing 1 in 7 deaths in the United States alone. Risk prediction models that can learn the complex causal relationships that give rise to CAD from data, instead of merely predicting the risk of disease, have the potential to improve transparency and efficacy of personalized CAD diagnosis and therapy selection for physicians, patients, and other decision makers. Methods. We use Bayesian networks (BNs) to model the risk of CAD using the Z-Alizadehsani data set—a published real-world observational data set of 303 Iranian patients at risk for CAD. We also describe how BNs can be used for incorporation of background knowledge, individual risk prediction, handling missing observations, and adaptive decision making under uncertainty. Results. BNs performed on par with machine-learning classifiers at predicting CAD and showed better probability calibration. They achieved a mean 10-fold area under the receiver-operating characteristic curve (AUC) of 0.93 ± 0.04, which was comparable with the performance of logistic regression with L1 or L2 regularization (AUC: 0.92 ± 0.06), support vector machine (AUC: 0.92 ± 0.06), and artificial neural network (AUC: 0.91 ± 0.05). We describe the use of BNs to predict with missing data and to adaptively calculate prognostic values of individual variables under uncertainty. Conclusion. BNs are powerful and versatile tools for risk prediction and health outcomes research that can complement traditional statistical techniques and are particularly useful in domains in which information is uncertain or incomplete and in which interpretability is important, such as medicine.

Markers of eosinophilic inflammation and risk prediction in patients with coronary artery disease

2006 ◽  
Vol 36 (4) ◽  
pp. 211-217 ◽  
Author(s):  
C. Falcone ◽  
P. Minoretti ◽  
A. D'Angelo ◽  
M. P. Buzzi ◽  
E. Coen ◽  
...  
Keyword(s):  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Risk Prediction ◽  
Eosinophilic Inflammation ◽  
Artery Disease

scholarly journals The role of coronary artery calcification testing in incident coronary artery disease risk prediction in type 1 diabetes

Diabetologia ◽  
2018 ◽  
Vol 62 (2) ◽  
pp. 259-268 ◽  
Author(s):  
Jingchuan Guo ◽  
Sebhat A. Erqou ◽  
Rachel G. Miller ◽  
Daniel Edmundowicz ◽  
Trevor J. Orchard ◽  
...  
Keyword(s):  
Coronary Artery Disease ◽  
Type 1 Diabetes ◽  
Coronary Artery ◽  
Risk Prediction ◽  
Coronary Artery Calcification ◽  
Disease Risk ◽  
Coronary Artery Disease Risk ◽  
Artery Disease

scholarly journals Proteomic biomarker profiles and vascular dysfunction in angina with no obstructive coronary artery disease

2021 ◽  
Vol 42 (Supplement_1) ◽  
Author(s):  
A Ejupi ◽  
A Aziz ◽  
P Ong ◽  
B H Shafi ◽  
T Lange ◽  
...  
Keyword(s):  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Prediction Model ◽  
Prediction Models ◽  
Vascular Dysfunction ◽  
Obstructive Coronary Artery Disease ◽  
Protein Biomarkers ◽  
Age And Gender ◽  
Artery Disease ◽  
And Gender

Abstract Background Coronary vascular dysfunction is a common cause of symptoms in patients with angina and no obstructed coronary arteries (ANOCA). Several endotypes have been defined but there are big gaps in our understanding of the underlying pathophysiology. Proteomic analyses may improve the understanding of the pathophysiology. Purpose Exploratory approach to 1) compare the proteomic biomarker profile across different types of vascular dysfunction in ANOCA and 2) assess the value of prediction models with protein biomarkers for vascular dysfunction in ANOCA. Methods We included 107 angina patients without previous coronary artery disease, left ventricular ejection fraction >45% and no obstructive coronary artery disease (CAD) (<50% stenosis of epicardial vessels) on coronary angiography. Three types of vascular dysfunction were assessed: 1) Vasomotor dysfunction (VMD) defined as epicardial or microvascular vasospasm on acetylcholine provocation, 2) Coronary microvascular dysfunction (CMD) defined as coronary flow velocity reserve (CFVR) ≤2.5 on echocardiography of the LAD on adenosine stimulation and 3) Reactive Hyperaemia Index (RHI) ≤1.67 as a measure of peripheral endothelial dysfunction. Blood samples were analysed for 184 protein biomarkers related to cardiovascular disease. Correlations between biomarkers and results of vascular function assessments were analysed with Pearson's correlation coefficient and visualized with volcano plots. Significantly correlated biomarkers (p<0.05) were tested in prediction models for their incremental value over age and gender with C-statistics. Results CFVR was correlated to 24 biomarkers before (figure 1a) and 2 biomarkers after adjustment for age and gender. The basic prediction model had AUC of 0.68 and was not significantly improved by adding biomarkers (figure 2a). RHI was correlated to 27 biomarkers before (figure 1b) and 10 biomarkers after adjustment for age and gender. The clinical prediction model was significantly improved (p=0.037) by adding TRAIL R2 and IL-18, in addition to age and gender, with an AUC of 84.4 (figure 2b). VMD was correlated to 14 biomarkers before (figure 1c) and 6 biomarkers after adjustment for age and gender. The prediction model was significantly improved (p=0.011) by adding HSP-27, RARRES-2 and SERPINA-12 in addition to age and gender in prediction of VMD with an AUC of 85.4 (figure 2c). Conclusion Several biomarkers were associated with vascular dysfunction in ANOCA patients with little overlap between different endotypes. We identified biomarkers that may contribute to the understanding of the underlying pathophysiology and have applications for screening. Results need to be confirmed in larger studies. FUNDunding Acknowledgement Type of funding sources: Public hospital(s). Main funding source(s): Department of Cardiology, Bispebjerg and Frederiksberg Hospital, University of Copenhagen, Denmark.Department of Cardiology and Angiology, Robert Bosch Krankenhaus, Stuttgart, Germany

scholarly journals Patient-Reported Outcomes Improves the Prediction of In-patient and Emergency Department Readmission Risks in Coronary Artery Disease

2018 ◽  
Vol 3 (4) ◽  
Author(s):  
Danielle Southern ◽  
Colleen Norris ◽  
Hude Quan ◽  
Maria Santana ◽  
Matthew James ◽  
...  
Keyword(s):  
Emergency Department ◽  
Risk Prediction ◽  
Prediction Models ◽  
Policy Decision ◽  
Patient Reported Outcome Measures ◽  
Baseline Survey ◽  
Post Discharge ◽  
Risk Prediction Models ◽  
Patient Reported ◽  
Artery Disease

IntroductionCoronary Artery Disease (CAD) patients are known to report higher healthcare resource use, such as inpatient [IP] and emergency department [ED] readmissions, than the general population. We investigate if the patient reported outcome measures (PROMs) improve the accuracy of readmissions risk prediction models in CAD. Objectives and ApproachPatients enrolled in the Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease (APPROACH) registry between 1995 and 2014 who received catheterization (CATH) and completed baseline PROMs were linked to discharge abstract data and national ambulatory data. Logistic regression (LR) was used to develop 30-day and 1-year readmissions risk prediction models adjusting for patients’ demographic, clinical, and self-reported characteristics. PROM was measured using the 19-item Seattle Angina Questionnaire (SAQ). The discriminatory performance of each prediction model was assessed using the Harrel’s c-statistic for LR. ResultsOf the 13,264 patients who completed baseline SAQ, 59 (0.3%) had IP readmissions or ED visits within 30 days, and up to 356 (1.9%) within 1 year of baseline survey. The C-statistics for one-year readmissions risk prediction models that only adjusted for demographic and clinical variables only ranged between 56.4% and 61.2%. The prognostic improvement in the discrimination of these models ranged between 2% to 10% when patient-reported SAQ was included as predictor. The addition of SAQ improves the model discrimination in all types of admission. Conclusion/ImplicationsThe addition of PROMs improves the moderate accuracy of readmissions risk prediction models. These findings highlight the need for routine collection of PROMs in clinical settings and their potential use for aiding clinical and policy decision-making and post-discharge outcomes monitoring in the management of cardiovascular diseases.

Sign in / Sign up

Export Citation Format

Share Document