Coronary flow reserve, or CFR, is one of the most informative functional measures in cardiology. It captures the ratio between blood flow through the coronary circulation when the heart is working hard and blood flow at rest. In a healthy person, the small resistance vessels that feed the heart muscle can dilate dramatically during exertion, multiplying resting flow several-fold. When atherosclerosis narrows the epicardial arteries, or when the microvasculature itself becomes dysfunctional, that headroom shrinks. A reduced coronary flow reserve is a powerful warning sign, associated with increased risk of myocardial infarction, heart failure, and cardiovascular death even in people whose coronary arteries look relatively clean on an angiogram.

The functional test behind the new study relies on positron emission tomography, or PET, imaging of myocardial perfusion. In a cardiac perfusion PET scan, a radioactive tracer is injected into the bloodstream and tracked as it moves through the heart muscle. By imaging the heart at rest and again during pharmacological stress, when a vasodilating agent forces the coronary vessels to open as far as they can, clinicians and researchers can quantify absolute myocardial blood flow in milliliters per minute per gram of tissue. The ratio of stress flow to rest flow yields the coronary flow reserve for each patient. Because PET provides quantitative, noninvasive measurements of flow deep within the myocardium, it offers a degree of precision that indirect measures of ischemia cannot match, and it has increasingly been used in large clinical cohorts as a phenotyping tool.

Turning that quantitative phenotype into genetic insight required assembling a sufficiently large group of individuals who had all undergone the same rigorous imaging protocol. The research team carried out a genome-wide association study, or GWAS, scanning hundreds of thousands to millions of genetic variants across the genomes of study participants and asking which variants travel together with unusually high or unusually low coronary flow reserve. GWAS is a hypothesis-free approach: rather than testing candidate genes chosen in advance, it surveys the entire genome, letting the data reveal which regions of DNA influence the trait. The statistical burden is considerable, because with millions of comparisons the threshold for significance must be set extremely high to avoid being fooled by chance, but when a signal does clear that bar it represents a genuine and reproducible association between a genetic locus and the measured trait.

The study’s central result is that coronary flow reserve is a heritable trait shaped by many variants of small effect scattered across the genome, and that among the loci and pathways implicated, genes connected to the NF-κB signaling pathway stand out. NF-κB, short for nuclear factor kappa-light-chain-enhancer of activated B cells, is one of the most intensively studied transcription factor systems in biology. It functions as a master switch for inflammation, sitting inactive in the cell’s cytoplasm until stimuli such as cytokines, bacterial products, or oxidative stress trigger its release and translocation into the nucleus, where it switches on hundreds of target genes involved in immune responses, cell survival, and proliferation. In the vasculature, NF-κB activity is activated by disturbed blood flow patterns, oxidized lipids, and other atherosclerosis-promoting conditions, and it drives expression of adhesion molecules and inflammatory genes within the endothelial cells that line the arteries.

The connection between this inflammatory pathway and the ability of coronary vessels to dilate under stress makes biological sense in several ways. Endothelial function depends on a delicate balance between vasodilating signals, most notably nitric oxide, and vasoconstricting and inflammatory forces. Chronic low-grade inflammation, mediated in part by NF-κB, impairs nitric oxide bioavailability, promotes endothelial dysfunction, and encourages the recruitment of immune cells into the vessel wall, all of which erode the microvascular and macrovascular responses that together produce coronary flow reserve. A genetic propensity for heightened or dysregulated NF-κB activity could therefore translate, over decades of life, into measurably poorer flow reserve long before a person develops overt symptoms or even significant angiographic stenoses.

What makes the GWAS approach powerful here is that genetics can help distinguish correlation from causation. People with low coronary flow reserve tend to have many other traits, including hypertension, diabetes, high cholesterol, and smoking exposure, and it can be difficult to know which factor drives which. Genetic variants, by contrast, are fixed at conception and are not themselves changed by disease. When genetic data point to a pathway, that pathway can be prioritized as a plausible causal contributor rather than a mere correlate. The identification of the NF-κB pathway in this study thus provides a form of evidence that observational studies of inflammation and heart disease have struggled to deliver on their own, complementing decades of work linking inflammation to atherosclerosis, including the landmark clinical trials that showed benefit from anti-inflammatory therapies in patients with residual cardiovascular risk.

The clinical implications cut in several directions. First, the results reinforce the idea that coronary microvascular dysfunction, the condition in which the small vessels of the heart fail to dilate properly, is not simply the end stage of visible plaque buildup but a distinct process with its own biology, some of which is inflammatory. Patients, particularly women, who experience symptoms of ischemia without obstructive coronary artery disease often have reduced flow reserve driven by microvascular problems, and this study’s findings suggest that inherited inflammatory tendencies may contribute to that burden. Second, the identified genetic architecture could eventually inform risk stratification. If the variants that influence coronary flow reserve can be combined into a polygenic score, clinicians might one day identify individuals whose coronary vasodilator capacity is genetically limited and who would benefit from earlier or more aggressive preventive therapy. Third, and perhaps most provocatively, the pathway-level findings highlight NF-κB and its upstream and downstream partners as potential therapeutic targets for preserving or restoring coronary vascular function.

The science of genomics has repeatedly shown that large, well-phenotyped cohorts are the engine of discovery, and the new study is a case study in that principle. Cardiac PET imaging is resource-intensive, requiring cyclotron-produced or generator-produced tracers, dedicated scanners, and trained personnel, which has historically limited the size of imaging-based genetic studies compared with those using simple measures such as height or blood pressure. By demonstrating that genome-wide association is feasible for a sophisticated functional imaging phenotype, the work helps pave the way for larger meta-analyses that combine cohorts across institutions and countries. As sample sizes grow, statistical power will increase, allowing researchers to resolve individual genes within the implicated pathways, to separate signals that reflect epicardial disease from those reflecting microvascular function, and to test whether the same genetic architecture governs flow reserve in different populations.

There are also important questions about how the genetic findings translate across ancestries and clinical contexts. GWAS signals are population-dependent in part because patterns of genetic variation, called linkage disequilibrium, differ between ancestral groups, and a variant flagged in one population may not tag the same causal mutation in another. Extending this work to diverse cohorts will be essential both for scientific completeness and for ensuring that any future risk prediction tools work equitably. Similarly, the relationship between genetically influenced flow reserve and hard clinical outcomes such as heart attack and death needs to be mapped in longitudinal follow-up, so that the full chain from DNA sequence to vascular physiology to clinical event can be traced end to end.

The broader significance of the study lies in how it reframes coronary artery disease. For much of the modern era, the disease has been understood primarily through its structural lesions, the plaques that narrow arteries and rupture to cause heart attacks. Over the past two decades, that picture has expanded to include inflammation as a fundamental driver, from the discovery that inflammatory cells populate plaques to the demonstration that lowering inflammation reduces cardiovascular events. The new genetic evidence adds another layer by suggesting that the same inflammatory machinery helps set the functional ceiling on coronary blood delivery throughout life. In this view, atherosclerosis and coronary microvascular dysfunction are twin manifestations of vascular biology gone awry, and the genes that shape that biology act decades before the first symptom appears.

For now, the study stands as a milestone in cardiovascular genomics: a demonstration that one of cardiology’s most precise functional measurements can be connected to specific biological pathways through the power of population genetics. The NF-κB pathway, long a central figure in immunology and vascular biology, now has a documented genetic foothold in the physiology of coronary blood flow. As the researchers and their colleagues build on this foundation, the hope is that understanding the inherited determinants of coronary flow reserve will move from the pages of journals into the clinic, first as refined risk prediction and, eventually, as guidance for therapies that keep the heart’s vital blood supply flowing freely under the stresses of daily life.

The NHCS, alongside the ASTAR Institute for Infocomm Research (ASTAR I²R), is spearheading the implementation of SENSE, a project that will be deployed in three major healthcare institutions: NHCS, the National University Hospital, and Tan Tock Seng Hospital, by the third quarter of 2025. This deployment is a testament to the commitment to improving healthcare access and efficiency, particularly in a demographic where coronary artery disease is a primary cause of mortality.

SENSE serves as a testament to how artificial intelligence can revolutionize complex medical procedures. Traditionally, specialists would require a duration of two to four hours to interpret cardiac scans, but with SENSE, results are expected to be delivered to clinicians within ten minutes. This represents a forty-fold improvement in efficiency, potentially leading to quicker diagnoses and interventions in patients identified at risk of cardiac events. It effectively allows healthcare providers to act sooner, which is critical in addressing CAD, a condition associated with one-third of cardiovascular-related deaths in Singapore.

The foundation of SENSE is built upon cutting-edge AI technologies developed in the NHCS CardioVascular Systems Imaging and Artificial Intelligence (CVS.AI) Research Laboratory. This purpose-built facility, expanding to 164 square meters, is equipped with high-performance GPUs and advanced machine learning software capable of processing vast amounts of patient data. The enhanced infrastructure not only facilitates real-time data analysis but also improves the accuracy of predictive models used in diagnosing heart conditions. The ability to generate immediate insights through AI innovations places NHCS at the forefront of cardiovascular health research.

The initial introduction of AI in the NHCS landscape traces back to the APOLLO project instituted in 2021, which aimed to establish a robust platform for analyzing CT coronary angiography through AI integration. Collaboratively developed with A*STAR’s Bioinformatics Institute and other institutions, APOLLO laid important groundwork for evaluating and diagnosing coronary artery disease through a comprehensive database of cardiac scans. Now, with SENSE, these advanced methods of AI interpretation are being refined and implemented at a larger scale, ensuring a more practical application in everyday clinical settings.

Building upon the APOLLO framework, SENSE focuses on four key determinants of coronary artery disease: coronary calcium scores, epicardial adipose tissue, stenosis levels, and plaque characterization. Each of these factors can be automatically analyzed by the AI modules embedded within the system, delivering comprehensive reports that clinicians can rely on for making informed decisions about patient care. The shift towards AI-powered diagnostics not only enhances efficiency but also introduces a new era of accuracy, as data gleaned from patient scans now enter clinical practice faster than ever before.

As pressures mount on healthcare systems globally, the adoption of initiatives like SENSE illustrates a proactive approach to addressing the inefficiencies inherent in traditional diagnostic methods. Given that coronary artery disease is among the leading causes of mortality worldwide, improving the identification and treatment of this disease is paramount. By revolutionizing the speed and precision of how healthcare professionals assess cardiac images, SENSE assists in closing the gap between medical advancements and patient care.

The National Heart Centre Singapore is dedicated not only to the immediate clinical applications of AI-driven technologies but also to the broader implications of enhancing cardiovascular health through ongoing research. The collaborative efforts between NHCS and national institutions showcase how interdisciplinary research can lead to innovative solutions tailored to meet the healthcare needs of diverse populations. The strategic partnerships aim to create a synergistic model of care that amplifies the impact of AI and reinforces the role of advanced technology in modern medicine.

As SENSE prepares for its rollout, healthcare professionals and patients alike stand to benefit from its implementation. The initiative underscores the relationship between technology and patient outcomes, illustrating how systematic enhancements in diagnostic processes can translate into more effective and swift treatment. As clinicians gain access to timely and highly detailed reports, the implications for preventative care and long-term health management possibilities are profound.

Furthermore, the operational efficiency introduced by SENSE allows healthcare providers to focus more on patient interaction and care rather than administrative burdens associated with traditional diagnostic metrics. It transforms the clinical environment, empowering clinicians to leverage high-quality data analytics in their decision-making processes. As this program evolves, it will undoubtedly set a precedent for future implementations of AI within this and other medical specialties.

In conclusion, the dawn of SENSE marks a transformative period in the realm of cardiac imaging and disease management. The collaboration between NHCS and A*STAR is pivotal in harnessing the capabilities of artificial intelligence, aiming for improved prognostic outcomes for patients dealing with coronary artery disease. As SENSE comes into effect, the impact on patient care delivery in Singapore will not only be immediate but will also serve as a model for implementing advanced AI systems in healthcare on a global scale.

Subject of Research: Artificial Intelligence in Cardiovascular Diagnostics
Article Title: National Heart Centre Singapore Launches SENSE: A Revolutionary AI Diagnostic Tool for Coronary Artery Disease
News Publication Date: May 20, 2025
Web References: National Heart Centre Singapore
References: NHCS Research Lab Publications
Image Credits: National Heart Centre Singapore

Keywords

AI, cardiovascular disease, coronary artery disease, diagnostic imaging, machine learning, healthcare technology, NHCS, SENSE, APOLLO, predictive analytics, patient outcomes.