Jinzhong Yang, PhD, and Percy Lee, MD

Minimizing Risk of Cardiac Toxicity in Lung Cancer Radiotherapy
 

Jinzhong Yang, PhD, of the University of Texas MD Anderson Cancer Center and Percy Lee, MD, of the City of Hope National Medical Center are developing new artificial intelligence tools to manage cardiac toxicity in patients who are treated for lung cancer with radiotherapy. Heart disease is one of the most concerning side effects of radiotherapy for lung cancer, affecting about one in four patients who receive the treatment. Additionally, studies have shown that radiation-induced cardiac toxicity is related to overall survival in non-small cell lung cancer and that the radiation dose to specific cardiac substructures may be relevant.

Delivering radiotherapy using an MRI guided linear accelerator (MR-linac) shows great promise for minimizing cardiac toxicity because it allows for improved and real-time visualization of cardiac substructures during treatment and may reveal subtle changes to the heart due to radiation. To help realize the full potential of MRI guided radiotherapy (MRgRT) for non-small cell lung cancer, Dr. Yang and Dr. Lee are creating new AI tools to support contour automation, dose accumulation, and outcome prediction that will improve clinicians’ workflow for heart toxicity analysis and prediction. Their team is:

• Optimizing auto-segmentation for contouring the different chambers of the heart and comparing them to manually produced contours. Auto-segmentation uses AI to help identify the treatment target and nearby organs at risk, which could enhance accuracy and allow plans to be produced faster.
• Developing patch-based deep learning models that better account for respiratory motion and improve measurement of dose accumulation in specific substructures of the heart.
• Building and validating a model to predict cardiac diseases based on the dose delivered, imaging and clinical factors such as age, sex, disease stage and receipt of chemotherapy.

MRgRT holds great promise to offer better cardiac toxicity management for patients treated with radiation for lung cancer than traditional approaches, but more needs to be known about this emerging technology. Through their research, Dr. Yang and Dr. Lee are combining MRgRT with AI to identify heart toxicity from radiotherapy at an earlier treatment stage, which could enable real-time treatment plan adaptations. If successful, these new tools could reduce or eliminate cardiac toxicity from radiotherapy for lung cancer and help patients live longer, healthier lives.

Publications
 

• Deep Learning–Based Automatic Segmentation of Cardiac Substructures for Lung Cancers was published December 18, 2023, in Radiotherapy & Oncology.