Researchers have utilized artificial intelligence (AI) tools to enhance our understanding of the risks associated with specific cardiac arrhythmias when different areas of the heart are subjected to varying levels of radiation as part of lung cancer treatment.
A team of researchers from Brigham and Women’s Hospital, which is a founding member of the Mass General Brigham healthcare system, has employed AI tools to improve understanding of the risks of particular cardiac arrhythmias when different sections of the heart are exposed to various radiation thresholds during lung cancer therapies. Their findings have been published in JACC: CardioOncology.
“Exposure to radiation targeting the heart during lung cancer treatment can lead to significant and immediate impacts on a patient’s heart health,” remarked Raymond Mak, MD, the lead author from the Department of Radiation Oncology at Brigham and Women’s Hospital. “We aim to educate not only oncologists and cardiologists but also the patients undergoing radiation therapy about the cardiac risks when treating lung cancer.”
The introduction of AI tools in healthcare has been revolutionary, with the potential to profoundly improve the overall care process and shape treatment strategies for cancer patients. Mass General Brigham, recognized as one of the top integrated academic health systems and the largest innovation enterprise in the country, is at the forefront of conducting thorough research on new technologies to ensure the accurate integration of AI into patient care.
Patients undergoing radiation treatment for non-small cell lung cancer (NSCLC) frequently experience arrhythmias, or irregular heart rhythms. This is due to the heart’s proximity to the lungs, as NSCLC tumors can be situated close to the heart, resulting in unintended radiation exposure to cardiac tissues while targeting cancerous tumors. Previous research has shown that such radiation exposure relates to general cardiac health issues. However, this recent investigation revealed that the risk for various arrhythmia types can significantly differ, depending on the cardiac structures and their physiological responses to different radiation levels.
The researchers performed a retrospective analysis on 748 patients from Massachusetts treated with radiation for locally advanced NSCLC to categorize the types of arrhythmias linked to radiation exposure in specific cardiac areas. The arrhythmia subtypes identified included atrial fibrillation, atrial flutter, other supraventricular tachycardia, bradyarrhythmia, and ventricular tachyarrhythmia or asystole.
The statistical results indicated that approximately one out of six patients experienced at least one grade 3 arrhythmia, with a median time of 2.0 years until the onset of the first arrhythmia. Grade 3 events are classified as serious and typically necessitate intervention or hospitalization. Furthermore, nearly one-third of patients facing arrhythmias also encountered significant adverse cardiac events.
While the arrhythmia categories considered in the study do not cover every possible heart rhythm disorder, the authors suggest that these findings enhance our comprehension of the underlying physiological mechanisms and introduce possible strategies to reduce cardiac toxicity after radiation therapy. Their research also presents a predictive model for radiation dose exposure and the anticipated types of arrhythmias.
Looking ahead, the researchers advocate for enhanced collaboration between radiation oncologists and cardiology specialists to further understand the mechanisms of heart injury related to radiation treatment. They also emphasize using advanced radiation techniques to specifically minimize exposure to high-risk cardiac areas known for causing arrhythmias. According to Mak, this study and previous research will assist in monitoring, screening, and guiding radiation oncologists on which heart regions should limit radiation exposure, consequently reducing potential complications.
“A notable aspect of our work involved applying AI algorithms to define structures such as the pulmonary vein and parts of the conduction system, allowing us to measure radiation dose exposure across over 700 patients. This saved us a considerable amount of manual effort,” stated Mak. “Thus, this work not only has significant clinical implications but also paves the way for utilizing AI in radiation oncology research to facilitate discovery and expand datasets.”
