Artificial intelligence, along with a $71-million expansion of radiation oncology services, is allowing University of Texas (UT) Southwestern Medical Center cancer physicians to pioneer a new PULSAR radiation-therapy strategy that improves tumor control compared with traditional daily therapy.

Personalized Ultra-Fractionated Stereotactic Adaptive Radiotherapy, or PULSAR—detailed in the International Journal of Radiation Oncology, Biology, Physics—achieved better tumor control by giving α-PD-L1 therapy during or after radiation, and spacing fractions 10 days apart rather than traditional daily fractions.

In the PULSAR paradigm, patients receive only a few large dose “pulses,” delivered with sophisticated, image-guided precision, at least a week, perhaps even months, apart. These “split treatments” are a radical break from the daily, long-course, conventional radiation treatments lasting six to nine weeks. They are less toxic and give oncologists time to fine-tune treatment after the new machines’ imaging shows the tumor’s changed shape, size, position, and its reaction to radiation.

“We unexpectedly found in these experiments that the time split between large, focused doses of radiotherapy will predict whether a certain class of immunotherapy drugs will work,” says Robert Timmerman, MD, professor of radiation oncology and neurological surgery, and member of the Harold C. Simmons Comprehensive Cancer Center at UT Southwestern.

“If they are 10 days apart, the drug therapy helps a lot for this model. If they are one to four days, it doesn’t help. Yet many of the clinical trials that are being done right now with radiotherapy and immunotherapy are using radiation schedules a day apart or every other day apart—perhaps the exact wrong timing.”

By allowing more time to assess changes for meaningful adaptation of an individual patient’s course, the PULSAR paradigm fulfills the promise of personalized cancer care. Instead of marching through a course of stand-alone, daily radiotherapy without interruption, PULSAR can be spaced out to integrate with surgery and drug therapy more thoughtfully.

“Using artificial intelligence, the team can replan cancer treatment in 30 minutes instead of the typical five to seven days,” says Timmerman. “The equipment and expertise under one roof should break new ground in fighting cancer.”

Collaborating on the PULSAR project are faculty from the departments of Radiation Oncology, Immunology, Pathology, and Neurological Surgery, and members of UT Southwestern’s Peter O’Donnell Jr. Brain Institute and Simmons Cancer Center—one of 51 designated comprehensive cancer centers in the U.S. by the National Cancer Institute, a member of the elite 30-member National Comprehensive Cancer Network, with its cancer program nationally ranked among the top 25 by U.S. News & World Report.

UTSW Radiation Oncology is pioneering PULSAR in a new, 71,000-square-foot expansion of Radiation Oncology services with seven new machines that image tumors and treat them with radiation.

The new adaptive radiotherapy machines in the Radiation Oncology expansion are:

  • Two Varian Ethos machines, which incorporate CT scans and artificial intelligence with radiotherapy
  • Two Elekta Unity machines, which incorporate MRI imaging with radiotherapy
  • Two Varian Halcyon machines, which incorporate cone-beam CT imaging with radiotherapy
  • One RefleXion, which incorporates PET imaging with radiotherapy

Artificial intelligence experts from radiation oncology and the Lyda Hill Department of Bioinformatics developed machine-learning algorithms, while radiation oncologists use the machines’ combined radiation and imaging abilities to make treatments as precise as possible, hitting the tumors and sparing healthy tissue.

Plans are underway to share expertise and data with other academic institutions, including Massachusetts General Hospital in Boston.