Association between the atherogenic index of plasma and coronary collateral circulation in patients with chronic total occlusion.

BACKGROUND: The atherogenic index of plasma (AIP) is considered an independent risk factor for coronary artery disease (CAD). The present study investigated whether AIP correlates with the formation of coronary collateral circulation (CCC) in CAD patients with chronic total occlusion (CTO). METHODS: This retrospective study included 1093 CAD patients with CTO confirmed by coronary angiography from January 2020 to December 2020 at Beijing Anzhen Hospital. Based on the Rentrop scoring system, the patients were divided into the good CCC group and the poor CCC group. AIP was calculated by log (triglyceride/high-density lipoprotein cholesterol). Meanwhile, the study population was further divided into four groups according to the quartiles of AIP. RESULTS: Patients in the poor CCC group exhibited significantly higher AIP compared to those in the good CCC group (0.31 ± 0.27 vs. 0.14 ± 0.24, p < 0.001). Multivariate logistic regression analysis revealed an independent association between AIP and poor CCC, regardless of whether AIP was treated as a continuous or categorical variable (p < 0.001), after adjusting for confounding factors. Besides, this association remained consistent across most subgroups. The incorporation of AIP into the baseline model significantly enhanced the accuracy of identifying poor CCC [area under the curve (AUC): baseline model, 0.661 vs. baseline model + AIP, 0.721, p for comparison < 0.001]. CONCLUSIONS: Elevated AIP is independently associated with an increased risk of poor CCC in CAD patients with CTO, and AIP may improve the ability to identify poor CCC in clinical practice.

Assessment of Myocardial Viability in Ischemic Cardiomyopathy With Reduced Left Ventricular Function Undergoing Coronary Artery Bypass Grafting.

BACKGROUND: We aim to provide a comprehensive review of the current state of knowledge of myocardial viability assessment in patients undergoing coronary artery bypass grafting (CABG), with a focus on the clinical markers of viability for each imaging modality. We also compare mortality between patients with viable myocardium and those without viability who undergo CABG. METHODS: A systematic database search with meta-analysis was conducted of comparative original articles (both observations and randomized controlled studies) of patients undergoing CABG with either viable or nonviable myocardium, in EMBASE, MEDLINE, Cochrane database, and Google Scholar, from inception to 2022. Imaging modalities included were dobutamine stress echocardiography (DSE), cardiac magnetic resonance (CMR), single-photon emission computed tomography (SPECT), and positron emission tomography (PET). RESULTS: A total of 17 studies incorporating a total of 2317 patients were included. Across all imaging modalities, the relative risk of death post-CABG was reduced in patients with versus without viability (random-effects model: odds ratio: 0.42; 95% confidence interval: 0.29-0.61; p < 0.001). Imaging for myocardial viability has significant clinical implications as it can affect the accuracy of the diagnosis, guide treatment decisions, and predict patient outcomes. Generally, based on local availability and expertise, either SPECT or DSE should be considered as the first step in evaluating viability, while PET or CMR would provide further evaluation of transmurality, perfusion metabolism, and extent of scar tissue. CONCLUSION: The assessment of myocardial viability is an essential component of preoperative evaluation in patients with ischemic heart disease undergoing surgical revascularization. Careful patient selection and individualized assessment of viability remain paramount.

Comprehensive analysis of the association between triglyceride-glucose index and coronary artery disease severity across different glucose metabolism states: a large-scale cross-sectional study from an Asian cohort.

BACKGROUND: The triglyceride-glucose (TyG) index is associated with the development and prognosis of coronary artery disease (CAD). However, the impact of the TyG index on CAD severity across different glucose metabolism states exhibits significant disparities in previous research. METHODS: This cross-sectional study comprised 10,433 participants from a prospective cohort. Participants were categorized into four groups based on glucose metabolism state: normal glucose regulation (NGR), prediabetes (pre-DM), diabetes mellitus (DM) without insulin prescribed (Rx), and DM with insulin Rx. The TyG index was determined by the following formula: Ln [TG (mg/dL) × FPG (mg/dL) / 2], where TG is triglycerides and FPG is fasting plasm glucose. Statistical methods such as binary logistic regression, interaction analysis, restricted cubic spline (RCS), and receiver operating characteristic (ROC) were employed to analyze the relationship between the TyG index and CAD severity across the entire population and glucose metabolism subgroups. Mediation analysis was conducted to examine the mediating effects of glycated hemoglobin (HbA1c) on these relationships. Sensitivity analysis was performed to ensure the robustness of the findings. RESULTS: Multivariable logistic regression analysis revealed a significant positive association between the TyG index and multi-vessel CAD in the entire population (OR: 1.34; 95% CI: 1.22-1.47 per 1-unit increment). Subgroup analysis demonstrated consistent positive associations in the NGR, pre-DM, and DM non-insulin Rx groups, with the highest OR observed in the NGR group (OR: 1.67; 95% CI: 1.3-2.14 per 1-unit increment). No correlation was found in the DM with insulin Rx subgroup. RCS analyses indicated the distinct dose-response relationships across different glucose metabolism subgroups. Including the TyG index in the established model slightly improved the predictive accuracy, particularly in the NGR group. Mediation analyses showed varying mediating effects of HbA1c among different glucose metabolism subgroups. Sensitivity analysis confirmed the robustness of the aforementioned relationships in the new-onset CAD population and in individuals not using antilipidemic medications. CONCLUSIONS: The TyG index positively associated with CAD severity across all glucose metabolism states, except for individuals receiving insulin treatment. Moreover, it might serve as a supplementary noninvasive predictor of CAD severity in addition to established factors, especially in NGR patients.

Uric acid to albumin ratio as a novel predictor for coronary slow flow phenomenon in patients with chronic coronary syndrome and non-obstructive coronary arteries.

BACKGROUND: The plasma uric acid to albumin ratio (UAR) is considered as a novel indicator for Inflammation. However, the association between UAR and coronary slow flow phenomenon (CSFP) remains unclear. METHODS: A total of 1328 individuals with chronic coronary syndrome (CCS) receiving coronary angiography (CAG) and found no obvious obstructive stenosis (< 40%) were included in this study. 79 individuals developed CSFP and were divided into CSFP group. The 1:2 age-matched patients with normal coronary blood flow were allocated to the control group (n = 158). The clinical characteristics, laboratory parameters including uric acid, albumin ratio, UAR and the angiographic characteristics were compared between the two groups. RESULTS: Patients with CSFP had a higher level of uric acid (392.3 ± 85.3 vs. 273.8 ± 71.5, P < 0.001), UAR (10.7 ± 2.2 vs. 7.2 ± 1.9, P < 0.001), but a lower level of plasma albumin (36.9 ± 4.2 vs. 38.5 ± 3.6, P = 0.003). Moreover, UAR increased as the numbers of vessels involved in CSFP increased. The logistic regression analysis demonstrated that UAR was independent predictors for CSFP. The Receiver operating characteristic (ROC) curve analysis showed that when UAR was more than 7.9, the AUC was 0.883 (95% CI: 0.840-0.927, p < 0.001), with the sensitivity and specificity were 78.2% and 88.2% respectively. CONCLUSION: Combined uric acid with plasma albumin, UAR could serve as an independent predictor for CSFP.

Lipoprotein(a) and Coronary Plaque in Asymptomatic Individuals: The Miami Heart Study at Baptist Health South Florida.

BACKGROUND: Elevated levels of lipoprotein(a) (Lp(a)) are independently associated with an increased risk of atherosclerotic cardiovascular disease events. However, the mechanisms driving this association are poorly understood. We aimed to evaluate the association between Lp(a) and coronary plaque characteristics in a contemporary US cohort without clinical atherosclerotic cardiovascular disease, undergoing coronary computed tomography angiography, the noninvasive gold standard for the assessment of coronary atherosclerosis. METHODS: We used baseline data from the Miami Heart Study-a community-based, prospective cohort study-which included asymptomatic adults aged 40 to 65 years evaluated using coronary computed tomography angiography. Those taking any lipid-lowering therapies were excluded. Elevated Lp(a) was defined as ≥125 nmol/L. Outcomes included any plaque, coronary artery calcium score >0, maximal stenosis ≥50%, presence of any high-risk plaque feature (positive remodeling, spotty calcification, low-attenuation plaque, napkin ring), and the presence of ≥2 high-risk plaque features. RESULTS: Among 1795 participants (median age, 52 years; 54.3% women; 49.6% Hispanic), 291 (16.2%) had Lp(a) ≥125 nmol/L. In unadjusted analyses, individuals with Lp(a) ≥125 nmol/L had a higher prevalence of all outcomes compared with Lp(a) <125 nmol/L, although differences were only statistically significant for the presence of any coronary plaque and ≥2 high-risk features. In multivariable models, elevated Lp(a) was independently associated with the presence of any coronary plaque (odds ratio, 1.40, [95% CI, 1.05-1.86]) and with ≥2 high-risk features (odds ratio, 3.94, [95% CI, 1.82-8.52]), although only 35 participants had this finding. Among participants with a coronary artery calcium score of 0 (n=1200), those with Lp(a) ≥125 nmol/L had a significantly higher percentage of any plaque compared with those with Lp(a) <125 nmol/L (24.2% versus 14.2%; P<0.001). CONCLUSIONS: In this contemporary analysis, elevated Lp(a) was independently associated with the presence of coronary plaque. Larger studies are needed to confirm the strong association observed with the presence of multiple high-risk coronary plaque features.

Rapid Plaque Progression Is Independently Associated With Hyperglycemia and Low HDL Cholesterol in Patients With Stable Coronary Artery Disease: A PARADIGM Study.

BACKGROUND: We assessed whether combinations of cardiometabolic risk factors independently predict coronary plaque progression (PP) and major adverse cardiovascular events in patients with stable coronary artery disease. METHODS: Patients with known or suspected stable coronary artery disease (60.9±9.3 years, 55.4% male) undergoing serial coronary computed tomography angiographies (≥2 years apart), with clinical characterization and follow-up (N=1200), were analyzed from the PARADIGM study (Progression of Atherosclerotic Plaque Determined by Computed Tomographic Angiography Imaging). Plaque volumes measured in coronary segments (≥2 mm in diameter) were summed to provide whole heart plaque volume (mm3) and percent atheroma volume (plaque volume/vessel volume×100; %) per patient at baseline and follow-up. Rapid PP was defined as a percent atheroma volume increase of ≥1.0%/y. Major adverse cardiovascular events included nonfatal myocardial infarction, death, and unplanned coronary revascularization. RESULTS: In an interscan period of 3.2 years (interquartile range, 1.9), rapid PP occurred in 341 patients (28%). At multivariable analysis, the combination of cardiometabolic risk factors defined as metabolic syndrome predicted rapid PP (odds ratio, 1.51 [95% CI, 1.12-2.03]; P=0.007) together with older age, smoking habits, and baseline percent atheroma volume. Among single cardiometabolic variables, high fasting plasma glucose (diabetes or fasting plasma glucose >100 mg/dL) and low HDL-C (high-density lipoprotein cholesterol; <40 mg/dL in males and <50 mg/dL in females) were independently associated with rapid PP, in particular when combined (odds ratio, 2.37 [95% CI, 1.56-3.61]; P<0.001). In a follow-up of 8.23 years (interquartile range, 5.92-9.53), major adverse cardiovascular events occurred in 201 patients (17%). At multivariable Cox analysis, the combination of high fasting plasma glucose with high systemic blood pressure (treated hypertension or systemic blood pressure >130/85 mm Hg) was an independent predictor of events (hazard ratio, 1.79 [95% CI, 1.10-2.90]; P=0.018) together with family history, baseline percent atheroma volume, and rapid PP. CONCLUSIONS: In patients with stable coronary artery disease, the combination of hyperglycemia with low HDL-C is associated with rapid PP independently of other risk factors, baseline plaque burden, and treatment. The combination of hyperglycemia with high systemic blood pressure independently predicts the worse outcome beyond PP. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifier: NCT02803411.

Self-Report Tool for Identification of Individuals With Coronary Atherosclerosis: The Swedish CardioPulmonary BioImage Study.

BACKGROUND: Coronary atherosclerosis detected by imaging is a marker of elevated cardiovascular risk. However, imaging involves large resources and exposure to radiation. The aim was, therefore, to test whether nonimaging data, specifically data that can be self-reported, could be used to identify individuals with moderate to severe coronary atherosclerosis. METHODS AND RESULTS: We used data from the population-based SCAPIS (Swedish CardioPulmonary BioImage Study) in individuals with coronary computed tomography angiography (n=25 182) and coronary artery calcification score (n=28 701), aged 50 to 64 years without previous ischemic heart disease. We developed a risk prediction tool using variables that could be assessed from home (self-report tool). For comparison, we also developed a tool using variables from laboratory tests, physical examinations, and self-report (clinical tool) and evaluated both models using receiver operating characteristic curve analysis, external validation, and benchmarked against factors in the pooled cohort equation. The self-report tool (n=14 variables) and the clinical tool (n=23 variables) showed high-to-excellent discriminative ability to identify a segment involvement score ≥4 (area under the curve 0.79 and 0.80, respectively) and significantly better than the pooled cohort equation (area under the curve 0.76, P<0.001). The tools showed a larger net benefit in clinical decision-making at relevant threshold probabilities. The self-report tool identified 65% of all individuals with a segment involvement score ≥4 in the top 30% of the highest-risk individuals. Tools developed for coronary artery calcification score ≥100 performed similarly. CONCLUSIONS: We have developed a self-report tool that effectively identifies individuals with moderate to severe coronary atherosclerosis. The self-report tool may serve as prescreening tool toward a cost-effective computed tomography-based screening program for high-risk individuals.

Coronary Artery Disease Is A Stronger Predictor of All-Cause Mortality Than Left Ventricular Ejection Fraction Among Patients With Newly Diagnosed Heart Failure: Insights From the WDHR.

BACKGROUND: In patients with newly diagnosed heart failure (HF) and left ventricular ejection fraction (LVEF) <50%, little is known whether LVEF per se or presence of coronary artery disease (CAD) provides independent prognostic information on all-cause mortality. METHODS AND RESULTS: Using the WDHR (Western Denmark Heart Registry), we identified 3620 patients with newly diagnosed HF and LVEF 10% to 49% referred for first-time coronary angiography as part of general workup of HF. Patients were stratified by LVEF (10%-35% versus 36%-49%) and presence of CAD. We estimated 10-year all-cause mortality risk and calculated hazard ratios adjusted for relevant comorbidities and risk factors (aHRs). CAD was present in 1592 (44%) patients. Lower LVEF was associated with a relative 15% increased 10-year mortality: 37% for LVEF 36% to 49% versus 42% for LVEF 10% to 35% (aHR, 1.15 [95% CI, 0.99-1.34]). This result did not change when stratified into those with CAD (52% versus 56%; aHR, 1.11 [95% CI, 0.91-1.35]) and those without CAD (27% versus 33%; aHR, 1.24 [95% CI, 0.97-1.57]). In comparison, presence and extent of CAD were associated with a relative 43% increased 10-year mortality (CAD versus no CAD, 55.0% versus 31.5%; aHR, 1.43 [95% CI, 1.25-1.64]). Compared with a matched general population, excess mortality risk was higher for patients with HF and CAD (54.7% versus 26.3%; aHR, 2.10 [95% CI, 1.85-2.39]) versus those with HF and no CAD (31.4% versus 17.2%; aHR, 1.76 [95% CI, 1.52-2.02]). CONCLUSIONS: Among newly diagnosed patients with HF and LVEF <50%, presence and extent of CAD are associated with substantial higher all-cause mortality risk than lower LVEF.

High-Density Lipoprotein Particle Concentration and Size Predict Incident Coronary Artery Disease Events in a Cohort With Type 1 Diabetes.

BACKGROUND: The cholesterol efflux capacity of high density lipoprotein (HDL) is negatively associated with cardiovascular risk. Small HDL particles account almost quantitatively for cholesterol efflux capacity, perhaps mediated through efflux of cholesterol and outer leaflet plasma membrane phospholipids by ABCA1 (ATP binding cassette subfamily A member 1). People with type 1 diabetes are at increased coronary artery disease (CAD) risk despite normal HDL-cholesterol concentrations. We therefore tested the hypothesis that small HDL particles (HDL-P)-rather than HDL-cholesterol-predict incident CAD in type 1 diabetes. METHODS AND RESULTS: Incident CAD (CAD death, myocardial infarction, or coronary revascularization) was determined in 550 individuals with childhood-onset type 1 diabetes. HDL-P was quantified by calibrated ion mobility analysis and cholesterol efflux capacity was quantified with validated assays. During a median follow-up of 26 years, 36.5% of the participants developed incident CAD, for an incidence density of 181.3 per 10 000 person-years. In multivariable Cox models, neither HDL-cholesterol nor apolipoprotein A1 concentration was significantly associated with CAD risk. In contrast, higher extra-small HDL-P concentrations were significantly associated with decreased CAD risk (hazard ratio [HR], 0.26 [95% CI, 0.14-0.50]). Weaker associations were observed for total HDL-P (HR, 0.88 [95% CI, 0.83-0.93]), small HDL (HR, 0.83 [95% CI, 0.68-1.02]), medium HDL (HR, 0.79 [95% CI, 0.71-0.89]), and large HDL (HR, 0.72 [95% CI, 0.59-0.89]). Although cholesterol efflux capacity was negatively associated with incident CAD, this association was no longer significant after adjustment for total HDL-P. CONCLUSIONS: Lower concentrations of total HDL-P and HDL subpopulations were positively associated with incident CAD independently of HDL-cholesterol, apolipoprotein A1, and other common CVD risk factors. Extra-small HDL was a much stronger predictor of risk than the other HDLs. Our data are consistent with the proposal that extra-small HDL plays a critical role in cardioprotection in type 1 diabetes, mediated by macrophage cholesterol efflux by the ABCA1 pathway.

Circulating Levels of Calprotectin as a Biomarker in Patients With Coronary Artery Disease: A Systematic Review and Meta-Analysis.

BACKGROUND: Calprotectin, also known as MRP8/14, is generated by immune cells and is altered in several inflammatory diseases. Studies have assessed their levels in patients with coronary artery disease (CAD) and its subtypes (stable CAD and acute coronary syndrome [ACS]). Herein, we aimed to systematically investigate these associations through a systematic review and meta-analysis. METHODS: A systematic search was conducted in four online databases, including PubMed, Scopus, Embase, and the Web of Science. Relevant studies were retrieved, screened, and extracted. Random-effect meta-analysis was performed for the calculation of standardized mean difference (SMD) and 95% confidence interval (CI). Blood calprotectin levels were compared between CAD patients and controls, as well as CAD subtypes. RESULTS: A total of 20 studies were included in the systematic review and meta-analysis, comprising 3300 CAD patients and 1230 controls. Patients with CAD had significantly higher calprotectin levels (SMD 0.81, 95% CI 0.32-1.30, p < 0.01). Similarly, patients with ACS were reported to have higher levels compared to those with stable CAD. However, there was no significant difference in terms of blood calprotectin levels between stable CAD cases and healthy controls. Finally, studies have shown that calprotectin could be used as a diagnostic biomarker of CAD while also predicting major adverse events and mortality in these patients. CONCLUSION: Based on our findings, calprotectin, as an inflammatory marker, could be used as a possible biomarker for patients with CAD and ACS. These suggest the possibility of pathophysiological pathways for this involvement and warrant further research on these associations as well as their clinical utility.

Ischaemia with no obstructive coronary arteries: a review with focus on the Asian population.

ABSTRACT: Ischaemia with no obstructive coronary arteries (INOCA) has been a diagnostic and therapeutic challenge for decades. Several studies have demonstrated that INOCA is associated with an increased risk of death, adverse cardiovascular events, poor quality of life and high healthcare cost. Although there is increasing recognition of this entity in the Western population, in the Asian population, INOCA remains elusive and its prevalence uncertain. Despite its prognostic significance, diagnosis of INOCA is often delayed. In this review, we identified the multiple barriers to its diagnosis and management, and proposed strategies to overcome them.

Lipoprotein(a) as a cardiovascular risk factor among patients with and without diabetes Mellitus: the Mass General Brigham Lp(a) Registry.

BACKGROUND: Diabetes mellitus (DM) and Lp(a) are well-established predictors of coronary artery disease (CAD) outcomes. However, their combined association remains poorly understood. OBJECTIVE: To investigate the relationship between elevated Lp(a) and DM with CAD outcomes. METHODS: Retrospective analysis of the MGB Lp(a) Registry involving patients ≥ 18 years who underwent Lp(a) measurements between 2000 and 2019. Exclusion criteria were severe kidney dysfunction, malignant neoplasms, and prior atherosclerotic cardiovascular disease (ASCVD). The primary outcome was a combination of cardiovascular death or myocardial infarction (MI). Elevated Lp(a) was defined as > 90th percentile (≥ 216 nmol/L). RESULTS: Among 6,238 patients who met the eligibility criteria, the median age was 54, 45% were women, and 12% had DM. Patients with DM were older, more frequently male, and had a higher prevalence of additional cardiovascular risk factors. Over a median follow-up of 12.9 years, patients with either DM or elevated Lp(a) experienced higher rates of the primary outcome. Notably, those with elevated Lp(a) had a higher incidence of the primary outcome regardless of their DM status. The annual event rates were as follows: No-DM and Lp(a) < 90th% - 0.6%; No-DM and Lp(a) > 90th% - 1.3%; DM and Lp(a) < 90th% - 1.9%; DM and Lp(a) > 90th% - 4.7% (p < 0.001). After adjusting for confounders, elevated Lp(a) remained independently associated with the primary outcome among both patients with DM (HR = 2.66 [95%CI: 1.55-4.58], p < 0.001) and those without DM (HR = 2.01 [95%CI: 1.48-2.74], p < 0.001). CONCLUSIONS: Elevated Lp(a) constitutes an independent and incremental risk factor for CAD outcomes in patients with and without DM.

Association of diabetes with cardiovascular calcification and all-cause mortality in end-stage renal disease in the early stages of hemodialysis: a retrospective cohort study.

BACKGROUND: The main goal of this study was to examine how diabetes, cardiovascular calcification characteristics and other risk factors affect mortality in end-stage renal disease (ESRD) patients in the early stages of hemodialysis. METHODS: A total of 285 ESRD patients in the early stages of hemodialysis were enrolled in this research, including 101 patients with diabetes. Survival time was monitored, and general data, biochemical results, cardiac ultrasound calcification of valvular tissue, and thoracic CT calcification of the coronary artery and thoracic aorta were recorded. Subgroup analysis and logistic regression were applied to investigate the association between diabetes and calcification. Cox regression analysis and survival between calcification, diabetes, and all-cause mortality. Additionally, the nomogram model was used to estimate the probability of survival for these individuals, and its performance was evaluated using risk stratification, receiver operating characteristic, decision, and calibration curves. RESULTS: Cardiovascular calcification was found in 81.2% of diabetic patients (82/101) and 33.7% of nondiabetic patients (62/184). Diabetic patients had lower phosphorus, calcium, calcium-phosphorus product, plasma PTH levels and lower albumin levels (p < 0.001). People with diabetes were more likely to have calcification than people without diabetes (OR 5.66, 95% CI 1.96-16.36; p < 0.001). The overall mortality rate was 14.7% (42/285). The risk of death was notably greater in patients with both diabetes and calcification (29.27%, 24/82). Diabetes and calcification, along with other factors, collectively predict the risk of death in these patients. The nomogram model demonstrated excellent discriminatory power (area under the curve (AUC) = 0.975 at 5 years), outstanding calibration at low to high-risk levels and provided the greatest net benefit across a wide range of clinical decision thresholds. CONCLUSIONS: In patients with ESRD during the early period of haemodialysis, diabetes significantly increases the risk of cardiovascular calcification, particularly multisite calcification, which is correlated with a higher mortality rate. The risk scores and nomograms developed in this study can assist clinicians in predicting the risk of death and providing individualised treatment plans to lower mortality rates in the early stages of hemodialysis.

Coronary Artery Revascularization in Patients Undergoing Transcatheter Aortic Valve Replacement.

Patients with concomitant severe aortic stenosis and significant coronary artery disease present a diagnostic and therapeutic challenge in clinical practice. There are no clear-cut guidelines as to the timing of revascularization in these patients who are referred for transcatheter aortic valve replacement (TAVR). This article aims to show that in patients without high-grade proximal coronary artery disease, revascularization after TAVR is safe, feasible, and practical. Additionally, the use of preoperative TAVR computed tomographic angiography might be used in both intermediate and high-risk patients rather than invasive coronary angiography to assess for significant proximal coronary artery disease to help guide the timing of revascularization.

Enhancing the diagnosis of functionally relevant coronary artery disease with machine learning.

Functionally relevant coronary artery disease (fCAD) can result in premature death or nonfatal acute myocardial infarction. Its early detection is a fundamentally important task in medicine. Classical detection approaches suffer from limited diagnostic accuracy or expose patients to possibly harmful radiation. Here we show how machine learning (ML) can outperform cardiologists in predicting the presence of stress-induced fCAD in terms of area under the receiver operating characteristic (AUROC: 0.71 vs. 0.64, p = 4.0E-13). We present two ML approaches, the first using eight static clinical variables, whereas the second leverages electrocardiogram signals from exercise stress testing. At a target post-test probability for fCAD of <15%, ML facilitates a potential reduction of imaging procedures by 15-17% compared to the cardiologist's judgement. Predictive performance is validated on an internal temporal data split as well as externally. We also show that combining clinical judgement with conventional ML and deep learning using logistic regression results in a mean AUROC of 0.74.