UBC Okanagan researchers, working with scientists at BC Cancer - Kelowna, have developed a groundbreaking AI model that can help identify breast cancer patients who may face a double-threat-that of cancer and cardiovascular disease.
Cardiovascular disease remains one of the leading causes of death worldwide, responsible for an estimated 17.9 million fatalities annually. According to the American Heart Association, breast cancer patients are more likely to experience cardiovascular complications and related issues than the general population.
"Breast cancer patients already face immense challenges. Cardiovascular disease is a critical and often overlooked health threat they endure," says Dr. Mohammad Shehata, a professor in the Irving K. Barber Faculty of Science's computer science department.
"The model we have created gives clinicians a new tool to proactively identify those patients who are at high-risk of developing cardiovascular disease, allowing them to intervene earlier and potentially save lives."
All breast cancer patients undergoing radiation therapy have chest CT scans to plan the treatment. This new research, published recently in Radiotherapy and Oncology , demonstrates how the CT scan can serve the added purpose of estimating cardiovascular risk.
"This research marks a significant step forward in how we assess cardiovascular risk in breast cancer patients," says Dr. Rasika Rajapakshe, a senior medical physicist at BC Cancer - Kelowna. "By combining routinely collected CT imaging with clinical health records, we can detect risk earlier and more accurately than ever before-without adding extra burden to patients or the health-care system."
Dr. Rajapakshe is a co-lead on this study, a clinical associate professor in UBC's Faculty of Medicine and an adjunct computer science professor at UBCO.
This study demonstrates how the AI model can compare a patient's CT scan with their electronic health records, including their general health, age, hypertension, diabetes and family history to forecast a more precise and personalized risk assessment.
Traditionally, risk models rely on clinical data, including age or any symptoms of hypertension or diabetes. This new research demonstrates how the AI models effectively identify subtle structural changes in the heart as seen on CT scans, alongside systemic risk factors.
"What makes this approach truly powerful is the use of multimodal artificial intelligence," says Dr. Shehata. "By integrating imaging and clinical data, the model can capture complex patterns that traditional tools simply miss, opening the door to more personalized and precise care."
The research team used powerful AI techniques to bring together different kinds of data to combine the information into a single system that helps predict health outcomes.
The model has demonstrated outstanding predictive accuracy-significantly outperforming existing methods.
"Our model achieved exceptionally high predictive performance, highlighting the value of combining advanced imaging analysis with cutting-edge clinical language models in a unified framework," says Dr. Rajapakshe.
Dr. Shehata says this research underscores the power of locally driven innovation to address global health challenges.
"This level of precision has the potential to identify high-risk patients early in their treatment while also tailoring interventions and care accordingly," he adds. "This approach may serve as a non-invasive, clinically valuable tool for early prediction of cardiovascular-related mortality, enabling timely identification of at-risk patients and improving their survival outcomes."
This research is supported by the BC Cancer Foundation .
