Use of Synthetic MRI in Radiation Therapy

Treatment planning for radiation therapy for prostate cancer can be improved by the acquisition of both a CT and MRI scan during the treatment procedure. However, this leads to a longer treatment delivery workflow and may introduce uncertainty due to anatomical motion and rectal and bladder filling between scans. This project investigated the radiation dose and clinical workflow impact of synthesizing the MRI scan from a CT scan using artificial intelligence-based style and content transfer, instead of acquiring the real MRI. Our results showed no significant decrease in radiation treatment quality with a significant improvement in the workflow efficiency and throughput using synthetic MRI, relative to acquisition of the real MRI. This indicates the potential clinical utility of this method and opens the door to future prospective clinical trials. 

Alex Podgorsak, PhD, DABR, is a clinical Medical Physicist and Assistant Professor of Radiation Oncology at the Wilmont Cancer Institute in Rochester, NY. Having graduated from the University of Buffalo Department of Biomedical Engineering with a PhD in 2021 where he investigated the use of deep learning and artificial intelligence in endovascular intervention, Alex completed a clinical residency at Loyola University in Chicago in Therapy Medical Physics. At Loyola, Alex explored the use of artificial intelligence in brachytherapy and adaptive external beam radiotherapy. At the University of Rochester, Alex is a founding teaching faculty member in the Graduate Program of Medical Physics and is also the current Graduate Program Director. As a native Buffalonian, Alex has strong opinions about chicken wings and the great blue cheese vs ranch debate.