Researchers Develop Personalized Model to Predict Risk of Subsequent of Cardiovascular Events

Despite advancements in treatment and prevention, patients with established atherosclerotic cardiovascular disease (ASCVD) still have particularly high risk of having another heart attack, stroke, or other cardiovascular event. Drawing on data from tens of thousands of patients from two independent sources, investigators from Mass General Brigham developed and validated new tools for quantifying this “residual risk” in patients with prior myocardial infarction, stroke, coronary artery disease, peripheral arterial disease, transient ischemic attack, or arterial revascularization. The novel scoring tools substantially improved risk discrimination compared with the guidelines-recommended model. Their findings are published in The Journal of the American College of Cardiology.

“By including a broader range of clinical variables and advanced modeling techniques, such as machine learning, our models demonstrated considerable improvement at predicting 10-year cardiovascular death than current guideline recommendations,” said co-senior author Olga Demler, PhD, of the Mass General Brigham Department of Medicine and the Division of Preventive Medicine at Brigham and Women’s Hospital and an assistant professor at Harvard Medical School.

“We demonstrated that existing model recommended by the guidelines can be improved,” said co-senior author Samia Mora, MD, a cardiologist with the Mass General Brigham Heart and Vascular Institute. “As a researcher and a practicing cardiologist, I view this as an important step, because once practitioners have more accurate and personalized risk estimates, we can formulate more appropriate treatment strategies to ultimately improve prognosis for our patients.”

ASCVD is a leading cause of death globally. In the U.S., approximately 26 million adults — or 10 percent of the population — have ASCVD, and are at higher risk of having additional cardiovascular events. ASCVD is thought to contribute to 400,000 deaths in the U.S. every year.

Many ASCVD risk models have been developed for primary prevention — that is, identifying patients at risk before they are diagnosed with ASCVD. Secondary prevention models that predict the likelihood a patient with ASCVD will experience a subsequent heart attack, stroke, or death, however, are uncommon. The American Heart Association and European Society of Cardiology have published the most widely used secondary prevention risk estimation models. However, these models fall short by either not providing individualized, quantitative risk estimates or by being applicable only to specific patient subgroups.

Mass General Brigham researchers, including lead author Olga Mineeva, PhD, developed two new models to address these shortcomings: RRS16, which considers 16 different factors that can contribute to secondary risk, and RRS24, which considers 24. They developed and validated the models with data from 32,994 United Kingdom Biobank participants and 54,969 Mass General Brigham participants, all with established ASCVD. Both models outperformed the American Heart Association guidelines but require further validation in more diverse populations.

Crucially, the model uses readily available clinical factors and blood assays to get an improved risk estimate for patients with ASCVD, enabling easier, faster clinical translation. Though the model is intended for research only, it is currently free and accessible online.

Authorship: Study was led by Dr. Olga Mineeva, forming part of her PhD thesis. In addition to Demler and Mora, Mass General Brigham authors include Franco Giulianini, Chunying Li and Vadim Bubes. Additional authors include Prof. Gunnar Raetsch and Simeon Häfliger from Swiss Federal Institute of Technology.

Disclosures: In work unrelated to the current study, Mora and Demler are listed as co-inventors on a patent application for a “Method for prediction of future cardiovascular disease risk via analysis of IgG glycome” assigned to GENOS d.o.o. and the Brigham and Women’s Hospital, Inc. Demler received funding from Kowa Research Institute for activities unrelated to current work. This research has been conducted using the UK Biobank Resource under Application Number 81959.

Funding: The study was funded by the National Heart, Lung, and Blood Institute (R21 HL156174) with additional funding from K24 HL136852, K01 HL135342, R21 HL167173, the American Heart Association (17IGMV33860009), the Swiss Federal Institute of Technology (ETH, Zurich, Switzerland), Dataspectrum4CVD from the Swiss Data Science Center #2022-812/Personalized Health & Related Technologies C22-15P, Zurich, Switzerland, and the National Human Genome Research Institute (U01HG008685). Mineeva was supported by the Max Planck ETH Center for Learning Systems and by ETH Core- funding (to Dr Raetsch).

Paper cited: Mineeva O et al. “Development and Validation of Novel Residual Risk Scores for Patients With ASCVD” JACC DOI: 10.1016/j.jacadv.2025.102162