VelocityAI fills a gap in radiation therapy planning at the Oregon Health & Science University. Radiation oncologists use VelocityAI to speed adaptive planning midcourse and to assess tumor response during treatment follow-up.

"Velocity gives us a much more detailed
assessment of tumor change over time."

Martin Fuss, MD, OHSU

Sometimes patients' anatomy changes rapidly during the weeks that it takes to deliver the full course of a radiation treatment plan. The tumor may shrink, the surrounding normal tissues may shift, or the patients may lose weight, and these changes show up on the daily cone-beam CTs taken for image guidance. In these situations, doctors at Oregon Health & Science University (OHSU) use VelocityAI to assess the degree of change and decide if adaptive planning is warranted. "About five percent of our cases require adaptive replanning," says Martin Fuss, MD, director of the Program in Image Guided Radiation Therapy at OHSU. "Interestingly, many of these patients require more than one adaptive plan adjustment."

However, replanning is no small effort. It takes time and resources to acquire another simulation CT scan, delineate the target and normal structures again, and generate a new plan from scratch. "Velocity helps tremendously to reduce the overall efforts involved in adaptive replanning," explains Fuss. "We can assess if complete replanning is required. We can deform the original simulation CT scan plus all the delineated structures directly into a new treatment planning situation."

“When you are delineating structures two or three times, perfect consistency and reproducibility isn't humanly possible," adds Fuss. "Utilizing Velocity for adaptive recontouring helps in maintaining consistency."

Matching images from different sources
Deformable image fusion, another software gap that VelocityAI fills at the university's radiation oncology department, helps Fuss plan the liver, bile duct, and pancreatic cancer treatments that comprise approximately 30 percent of his practice. VelocityAI enables deformable image fusion from different image sources and patient positions, including diagnostic and simulation CTs and CBCT, MRI, PET, and SPECT imaging.

“In order to use the superior soft-tissue contrast of MRIs during the treatment planning process, I have to register the MRI onto the simulation CT scan, but the images will never perfectly match," says Fuss. "Distorting the second scan, pressing and pulling a little bit here and there so that the images match, is called deformable registration. Velocity provides an advanced deformable registration tool that I use frequently in my treatment planning process."