UAB One of the First in the Nation with New AI-Driven Linear Accelerator System

Dec 14, 2021 at 11:01 am by steve

Dennis Stanley, PhD | Samuel Marcrom, MD

Getting good results from radiation depends so much on delivering the right dose to the right place for the right amount of time. Up until now, the measuring and planning required to give each patient the optimum treatment to destroy malignant cells and keep healthy cells safe has been a formidable endeavor that can take one or two weeks.

In that time, patients can gain or lose weight. Maintaining the optimum placement is especially difficult in areas where the body is a moving target--the lungs, abdomen and pelvis where something as simple as what the patient had for lunch can throw calculations off by precious millimeters.

Now a new AI-driven linear accelerator system called Ethos® from Varian® is bringing personalized precision adaptive radiation therapy to UAB's oncology patients.

The Ethos is the first system that can quickly scan a patient when they are on the table prepping for their next treatment and adapt as needed. UAB Medicine and the Department of Radiation Oncology are also among the first in the nation to install an Ethos system. There are only 12 currently operating in the United States.

"We've been working to bring the platform to our patients since the benefits of the technology became clear a couple of years ago," oncologist Samuel Marcrom, MD of UAB's O'Neal Comprehensive Cancer Center said. " It should greatly enhance the efficacy of treatment and minimize radiation exposure to healthy tissue, especially in patients who have malignancies in areas that tend to be mobile."

After a thorough initial CAT scan, the oncologist and radiation physicist work together to plan the treatment strategy and meet again after each session.

"Together we evaluate each treatment and fine tune plans for the next treatment. Imaging while the patient is on the table allows us to adapt weekly or even daily," Marcrom said. " Since our first use treating patients in early August, five UAB oncologists have been using the system, and many more of our physicians will be using it to care for patients as they learn its capabilities and train in how to apply them."

The technical side of implementation was shepherded by UAB radiation physicist Dennis Stanley, PhD.

"Much of my time over the past year has been devoted to acquiring and installing the system, testing and evaluating it, and learning its capabilities and how to use them.," Stanley said. "We work closely with oncologists in planning and fine-tuning treatments. Staying still for radiation isn't easy. I can see if a patient moves or if their body has changed since the last treatment, and I can adapt the placement there and then. We don't lose the effect of that treatment on cancer cells, and we don't have healthy tissue affected by radiation.

"The Ethos system is made by Varian, a leader in this type of technology. It's very user friendly and works using a true neural net adaptive intelligence with the calculating power to provide solutions in near real time.

"The AI can calculate from the last position and quickly adapt to mark the correct position so treatment can continue on schedule. In the first month, we treated around 100 cases and the performance has been impressive. We should be seeing this technology used widely in the future."

Marcrom said, "Patients have been very excited about being treated using this system. They see it as giving them access to maximum benefits with minimal side effect."

It will take time to accumulate data to determine definitively what effect the greater precision of adaptive radiation will have on outcomes, but those using it in patient care are quite optimistic.

"This could well be a fundamental shift in radiation treatment," Marcrom said. "It gives us a much better tool to help our patients."

Sections: Clinical



Birmingham Medical News November 2024 Cover

November 2024

Nov 26, 2024 at 01:11 pm by kbarrettalley

Your November 2024 Issue of Birmingham Medical News is Here!