Research News

Researchers from Cleveland Clinic’s Lerner Research Institute and Heart, Vascular & Thoracic Institute have developed an artificial intelligence (AI) tool to help identify cancer patients at increased risk for adverse cardiac outcomes, according to findings published in PLOS Medicine.

Although advances in cancer care have helped patients live longer, many treatments are linked to cardiac dysfunction and cardiovascular disease, underscoring the need for clinical tools that can assess risk for cancer therapy-related cardiac dysfunction (CTRCD).

“With the help of our AI methodology, known as patient-patient similarity network-based risk assessment of cardiovascular disease, or psnCVD, cancer patients of unknown cardiac risk status can be classified based on their similarity to patients with known status,” said the study’s co-corresponding author Feixiong Cheng, PhD, Genomic Medicine.

The researchers, led by Dr. Cheng and Patrick Collier, MD, PhD, co-director of Cleveland Clinic’s Cardio-Oncology Center, used artificial intelligence technology and clinical data from more than 4,600 cancer patients who were referred to Cleveland Clinic’s cardio-oncology service between 1997 and 2019. For each patient, they verified their cardiac outcomes (i.e., atrial fibrillation, coronary artery disease, heart failure, myocardial infarction or stroke) and categorized them as either diagnosed before cancer therapy or after cancer therapy (called de novo CTRCD). They then used psnCVD to pinpoint clinically actionable patient subgroups that were significantly correlated with CTRCD risk and mortality.

“Notably, our analysis determined that mortality was more likely to occur within two to five years after the initiation of cancer therapy, indicating the vital role of early cardiac care in improving cancer patients’ survival,” noted Dr. Cheng.

They also found that levels of two well-established biomarkers for heart disease, NT-proBNP and Troponin-T, were higher among two patient groups: those with intermediate survival and the greatest de novo CTRCD risk and those with the worst survival and intermediate de novo CTRCD risk. The team’s findings suggest that these markers should be further studied for use in rapid cardiac risk assessment during cardio-oncology clinical practices.

“Altogether, our findings indicate that, compared to traditional risk models, psnCVD is clinically intuitive and excels at integrating large-scale, heterogeneous patient data to stratify cardiac dysfunction risk in cancer patients,” said Dr. Cheng. “If broadly applied, our AI methodology holds great promise for identifying novel cardiac risk subgroups and clinically actionable biomarkers that would advance precision medicine in the cardio-oncology field. As a next step, we are working to develop new risk calculators that integrate our AI models into Cleveland Clinic’s electronic health record system to help provide cardiovascular care for cancer patients.”

Yuan Hou, PhD, a postdoctoral fellow in Dr. Cheng’s lab; Yadi Zhou, PhD, a data scientist in Dr. Cheng’s lab; and Muzna Hussain, a cardiology research fellow at Cleveland Clinic, are co-first authors on the study, which was supported in part by the National Heart, Lung, and Blood Institute, the National Institute on Aging (both part of the National Institutes of Health) and Cleveland Clinic’s VeloSano Pilot Program.