New York, NY (December 20, 2022) – Using machine learning and clinical data from electronic health records, researchers at the Icahn School of Medicine at Mount Sinai in New York constructed an in silico or derived marker of the coronary artery disease (CAD) computer to better measure clinically important characteristics of the disease.
The findings, published online Dec. 20 in The Lancet, may lead to more specific diagnosis and better disease management of CAD, the most common type of heart disease and one of the leading causes of death worldwide. world The study is the first known research to map CAD characteristics onto a spectrum. Previous studies have focused only on whether or not a patient has CAD.
CAD and other common conditions exist on a spectrum of diseases; each individual’s combination of risk factors and disease processes determines where they are on the spectrum. However, most of these studies divide this disease spectrum into rigid classes of cases (patient has disease) or control (patient does not have disease). This can lead to missed diagnoses, inappropriate management and poorer clinical outcomes, the researchers say.
“Information gained from this non-invasive disease staging could empower clinicians by more accurately assessing a patient’s condition and thereby inform the development of more specific treatment plans,” says Ron Do, PhD, lead author of the study and Charles Bronfman Professor in Personalized Medicine at the Icahn School of Medicine at Mount Sinai.
“Our model delineates populations of patients with coronary artery disease across a disease spectrum; this could provide more insight into disease progression and how those affected will respond to treatment. Having the ability to reveal different gradations of risk of disease, atherosclerosis and survival, for example, which might otherwise be missed with a conventional binary framework, is critical.”
In the retrospective study, researchers trained the machine learning model, called the In Silico Score for Coronary Artery Disease, or ISCAD, to accurately measure CAD across a spectrum using more than 80,000 electronic health records from two large biobanks based on the health system, the BioMe Biobank. at Mount Sinai Health System and UK Biobank.
The model, which the researchers called a “digital scorecard,” incorporated hundreds of different clinical features from the electronic medical record, including vital signs, lab test results, medications, symptoms and diagnoses, and compared it to a clinical score existing for CAD, which uses only a small number of predefined features and a genetic score for CAD.
The 95,935 participants included participants of African, Hispanic/Latino, Asian, and European ethnicities, as well as a large proportion of women. Most clinical and machine learning studies on CAD have focused on white Europeans.
The researchers found that the model probabilities accurately tracked the degree of narrowing of the coronary arteries (coronary stenosis), mortality and complications such as heart attack.
“Machine learning models like this could also benefit the broader healthcare industry by designing clinical trials based on appropriate patient stratification. It can also lead to more efficient data-driven individualized therapeutic strategies,” says lead author Iain S. Forrest , PhD, postdoctoral fellow in the laboratory of Dr. Don and PhD/MD student in the Medical Scientist Training Program at Icahn Mount. Sinai “Despite this progress, it is important to remember that physician- and procedural-based diagnosis and management of coronary artery disease are not being replaced by artificial intelligence, but can be supported by ISCAD as another tool powerful tool in the physician’s toolbox.”
Next, the researchers plan to conduct a large-scale prospective study to further validate the clinical utility and action of ISCAD, including in other populations. They also plan to evaluate a more portable version of the model that could be used universally across health systems.
The paper is titled “Machine Learning-Based Marker for Coronary Artery Disease: Derivation and Validation in Two Longitudinal Cohorts.” Additional co-authors are Ben O. Petrazzini, BS, Áine Duffy, MS, Joshua K. Park, BS, Carla Marquez-Luna, PhD, Daniel M. Jordan, PhD, Ghislain Rocheleau, PhD, Judy H. Cho, MD, Robert S. Rosenson, MD, and Jagat Narula, MD, and Girish N. Nadkarni, MD.
This work was supported by grants from the National Institute of General Medical Sciences of the National Institutes of Health (NIH) T32-GM007280 and R35-GM124836, and the National Heart, Lung, and Blood Institute of NIH grants R01-HL139865. and R01-HL155915.
/ Public communication. This material from the original organization/author(s) may be ad hoc in nature, edited for clarity, style and length. The views and opinions expressed are those of the author(s). See them in full here.