MD Anderson presents 70 abstracts on precision radiation therapy at ASTRO 2026

Researchers at The University of Texas MD Anderson Cancer Center will bring nearly 70 abstracts to the 2026 American Society for Radiation Oncology Annual Meeting, with two studies offering particularly sharp insights into how radiation therapy is actually delivered and how to predict which patients will benefit most from treatment.

The first centers on a problem that sounds technical but carries real consequences: when radiation oncologists plan stereotactic radiosurgery—the highly focused beam work used to treat brain tumors and other targets—they assume that following the machine's specifications will produce consistent results. A team led by Lian Duan, a Ph.D. candidate in Radiation Physics, tested that assumption by examining more than 800 treatment plans created between 2013 and 2025 across more than 650 institutions. All the plans targeted the same spot inside a phantom—a standardized model designed to measure how radiation actually behaves in the head. What they found was striking: the amount of healthy brain tissue receiving high doses varied nearly fivefold depending on the plan, even though the target was identical.

The equipment mattered, but not as much as the planner's choices. On standard linear accelerators, the decisions made by the person designing the treatment—particularly how tightly they shaped the radiation field around the tumor—explained more of the variation than the machine itself did. "The same plan on different machines does not mean the same dose to healthy brain," Duan said. The implication is both practical and humbling: clinics need to understand their own equipment deeply and use that understanding as a baseline, then work to improve from there. Duan works within the Image and Radiation Oncology Core at MD Anderson, one of four centers nationwide funded by the National Institutes of Health to maintain quality standards for radiation therapy in cancer trials.

The second study addresses a different kind of precision—knowing which patients actually need aggressive treatment. The NRG/RTOG 9601 trial, conducted nearly two decades ago, showed that adding hormone therapy to radiation improved outcomes for men whose prostate cancer had returned after surgery. But hormone therapy carries significant side effects, and clinicians had no reliable way to identify which patients would actually benefit. Prostate-specific antigen testing helped somewhat, but predicting how aggressively a particular cancer would behave remained difficult.

Krishnan Patel, an associate professor of Radiation Oncology, led a team that took tumor samples from that landmark trial and analyzed them using a modern genomic test—one that examines the genetic material inside the cancer to forecast its behavior. The results were clear enough to reshape treatment decisions. Patients stratified into a low-risk group had a 75 percent survival rate 15 years after treatment, compared with 25 percent for those in the very high-risk group. The genomic test could work alongside standard blood tests, potentially identifying patients who looked low-risk by conventional measures but carried aggressive biological features—people who might benefit from hormone therapy despite normal PSA levels. Equally important, it could spare patients with high PSA but favorable genetics from unnecessary hormone treatment and its side effects.

"The ability to apply a modern genomic test to samples from a clinical trial conducted nearly two decades ago is incredibly valuable," Patel said. The finding points toward a future where treatment recommendations are increasingly tailored to individual tumor biology rather than broad categories. Some patients avoid toxicity they didn't need; others gain confidence that the harder treatment is worth it. That shift—from one-size-fits-most to precision matching—is what these two studies, presented among dozens of others at the conference, are quietly advancing.