Atherosclerotic cardiovascular disease (ASCVD) is a leading cause of morbidity and mortality, affecting approximately 4.6 million individuals in the United States (Arnett et al., 2019). Coronary artery disease (CAD), as the most common form of ASCVD, is characterized by the accumulation of atherosclerotic plaque in the coronary arteries, leading to myocardial infarction, angina, and other complications (Chan and Ramji, 2022).

While the exact etiology remains uncertain, several modifiable and non-modifiable risk factors have been identified, including high cholesterol, hypertension, smoking, insulin resistance, diabetes, obesity, sedentary lifestyle, age, and dietary habits. Although genetic predisposition and family history cannot be altered, a better understanding of and patient awareness of these risk factors can facilitate the adoption of preventive measures and promote healthier lifestyle choices (Said et al., 2019).

To better quantify risk, the American College of Cardiology introduced the Pooled Cohort Equations risk score, a widely used tool that estimates the 10-year probability of major cardiovascular events, including myocardial infarction and stroke (Goff et al., 2013). This tool incorporates variables such as age, sex, race, cholesterol levels, blood pressure, medication use, diabetes status, and smoking history. More recently, in 2023, the American Heart Association introduced the PREVENT risk tool, which incorporates additional clinical parameters, including body mass index (BMI), hemoglobin A1c, albuminuria, and estimated glomerular filtration rate (eGFR), offering a more comprehensive approach to cardiovascular risk assessment (Khan et al., 2023). However, while both models offer valuable insights, they do not account for genetic predisposition, a factor that may offer additional insights into coronary artery disease susceptibility and also refine individualized risk assessment (Hadley et al., 2021).

Polygenic risk scoring (PRS) is an innovative approach to estimating an individual’s genetic predisposition to traits or diseases based on genome-wide association study (GWAS) data (Choi et al., 2020; Lambert et al., 2019; Torkamani et al., 2018). By analyzing a patient’s genotype, PRS provides a probabilistic estimate of CAD risk, complementing conventional utilized clinical risk factors. The integration of PRS into coronary risk assessment frameworks has the potential to improve risk stratification and support personalized preventive strategies for physicians and their patients.

To address the need for a more comprehensive risk assessment framework, we developed an artificial intelligence (AI)-driven platform that integrates PRS with conventional CAD risk factors. Using machine learning techniques, we incorporated genetic, clinical, and lifestyle data to develop predictive models aimed at guiding treatment decisions, enhancing patient education, in order to and optimize coronary risk management. The software employs widely recognized algorithms, including PLINK (Purcell et al., 2007; Chang et al., 2015) and PRS-CS (Ge et al., 2019), to compute polygenic scores while accounting for linkage disequilibrium, allowing for a more comprehensive evaluation of genetic risk factors in complex diseases.

This study aims to evaluate whether an AI-generated CAD risk profile that integrates PRS with traditional coronary risk factors can realistically influence physician decision-making regarding both statin therapy prescription and dosage adjustments, while also enhancing patient education on coronary artery disease prevention.

This study evaluated the impact of integrating a polygenic risk score (PRS) for coronary artery disease (CAD) into clinical decision-making, revealing significant insights into how physicians utilized this tool in patient care. The findings indicate that PRS influenced physician decisions in 67% of cases (100 out of 150 participants), demonstrating its substantial role in shaping clinical approaches. While the majority of these decisions involved enhancing patient education and risk awareness, PRS also contributed to adjustments in treatment strategies, lifestyle recommendations, and targeted monitoring. These findings underscore PRS as a valuable adjunct to conventional risk assessment tools, offering additional layers of stratification beyond traditional clinical markers.

The results align with prior studies demonstrating the potential utility of PRS in cardiovascular risk prediction. Several studies have reviewed the theoretical benefits of PRS for CAD, emphasizing its potential to refine risk stratification and improve clinical decision-making (Harst et al., 2021; Klarin and Natarajan, 2022; Marston et al., 2023). Another study has explored how PRS information is incorporated into physician decision-making, contributing to the growing body of evidence on the influence of genetic risk scores in clinical practice (Kerman et al., 2023).

Findings from previous research also support the importance of physician education and structured implementation of PRS in healthcare (Adeyemo et al., 2021; Hao et al., 2022; Slunecka et al., 2022). Guideline-based studies have emphasized key challenges in PRS adoption, particularly the need for standardized recommendations and improved clinician training (Visseren et al., 2021). The 2021 ESC Guidelines on cardiovascular disease prevention have not yet incorporated PRS as a routine risk assessment tool, reflecting the ongoing need for structured implementation in clinical practice (Visseren et al., 2021). Additionally, prior research has demonstrated that physicians frequently use PRS as a tool for patient education and risk communication, reinforcing our finding that PRS serves an important role in guiding discussions about cardiovascular health (Lewis et al., 2022). Other studies further emphasized the significance of physician familiarity with PRS. Although clinicians generally recognize their potential, challenges such as the complexity of interpretation may still limit their widespread application (Slunecka et al., 2022).

The integration of PRS within an interactive dashboard streamlined the incorporation of genetic risk data with clinical parameters, enhancing its accessibility in routine practice. This structured approach enabled physicians to efficiently interpret PRS results and integrate them into patient discussions, contributing to increased engagement and usability. Our findings suggest that an intuitive system for PRS presentation can support its adoption in real-world settings, addressing some of the barriers previously identified in clinical genomics. Prior research has highlighted the importance of clear and actionable reporting formats for PRS to optimize their utility in patient care (Pacyna et al., 2023), and our study reinforces this need by demonstrating physician engagement with an automated PRS system.

Our findings suggest that PRS has the potential to refine risk assessment by providing additional stratification beyond traditional clinical markers. Physicians frequently used PRS to guide conversations about cardiovascular risk, motivate patients to adopt lifestyle changes, and reinforce existing treatment plans. This aligns with prior literature indicating that PRS can serve as a tool for shared decision-making rather than a sole criterion for prescribing interventions. In four cases, statins were prescribed despite LDL levels being below 130 mg/dL, based solely on high PRS scores, highlighting how genetic risk influenced clinical decisions beyond standard lipid thresholds.

Despite the promising implications, our study has limitations. The relatively small sample size (150 participants) may limit generalizability. Additionally, physician interpretations of PRS likely varied, as no standardized framework for its application currently exists. This lack of standardization raises ethical concerns that have been widely discussed in literature, including the risk of misinterpretation, potential psychological distress to patients, or even inappropriate clinical decisions (Andreoli et al., 2024; Chapman, 2023; Palk et al., 2019). Given that our cohort included physicians with an interest in genetics, results may not fully represent broader clinical populations, where familiarity with PRS may be lower. Moreover, the absence of a control group limits the ability to draw definitive conclusions regarding causality. This study, however, was not intended to assess causal effects, but rather to explore clinical uptake, interpretability, and integration of PRS in daily workflow. In addition, the present study did not evaluate the predictive accuracy of the PRS-integrated platform, as this was beyond its intended scope. Future studies should assess how this approach compares quantitatively to established risk estimation tools in terms of discrimination and calibration. Furthermore, the relatively young age distribution of the study population may have limited the extent to which pharmacologic decisions, such as statin initiation, were observed. This age range was intentionally selected to evaluate PRS use in younger individuals, a group where traditional risk tools may underestimate long-term cardiovascular risk. Finally, although the PRS used in this study was derived from a large-scale multiethnic GWAS, we did not perform ancestry-specific analyses, and genetic ancestry was not explicitly incorporated into model interpretation.Moving forward, future research should focus on the development of clearer and more standardized guidelines for PRS interpretation and reporting. While some recommendations may already exist, their applicability and clarity in clinical practice remain limited (O'Sullivan et al., 2022). Establishing more specific guidance would help ensure consistent and effective use of PRS across different settings. Additionally, greater emphasis should be placed on educating physicians in the proper application of PRS. Without sufficient training, PRS integration may remain inconsistent and underutilized. Developing structured training programs, including workshops and decision-support tools, could help bridge this gap and optimize PRS implementation in routine cardiovascular care. Furthermore, the importance of patient use of these dashboards should not be overlooked, as it contributes to improved adherence to and understanding of physician-suggested CAD interventions.

In conclusion, this study provides empirical evidence on the real-world utilization of PRS in clinical practice, demonstrating its role in both physician decision-making and patient education. While PRS was primarily used to enhance risk awareness rather than directly alter pharmacologic management, its integration into cardiovascular risk assessment offers a promising avenue for personalized medicine. As genetic risk stratification continues to evolve, well-defined implementation strategies will be critical in ensuring that PRS translates into cost effective and meaningful improvements in cardiovascular disease prevention and for patient outcomes.