Technological advances in radiation therapy enable radiation oncologists and physicists to aim the radiation beam precisely, while imaging provides visualization of the targets to hit and organs to avoid, according to presenter Dr. Atallah Baydoun of Case Western Reserve University in Cleveland.
"Imaging has been widely adopted in clinical care to avoid 'shooting blind,' " Baydoun said. "On the other hand, daily changes in shape, position, and treatment-induced volumetric changes result in moving targets."
The current practice -- treatment plans based on a single snapshot image acquired during CT simulation -- ignores changes in position and size. While care could be improved if treatment plans were updated daily, this approach faces hurdles related to the time, consistency, and labor-intensive aspects of manually contouring the tumors and organs at risk, he said.
As a result, the researchers sought to automate contour generation to account for daily, patient-specific anatomical and functional changes.
"Inspired by the human contouring processing model, we developed a novel methodology for autocontouring that is designed to support online [MRI-guided adaptive radiation therapy]," Baydoun said.
The AI method can be used for automatic generation of contours in daily MR-guided radiotherapy, and it has been tested in online MRI-guided adaptive radiation therapy for abdominal malignancies.
"Using a desktop computer, the contours were generated in approximately two minutes," Baydoun told AuntMinnie.com. "The next step will be to perform a virtual clinical trial."
The project being is performed by the Quantitative Imaging Laboratory at Case Western and the University Hospitals Cleveland Medical Center. You can learn all about it by attending this Wednesday afternoon talk.
