New results from two randomized clinical trials showed that analyzing a gene expression signature in tumors may help doctors customize radiation treatments for prostate cancer patients, improving outcomes while avoiding unnecessary side effects.
The gene expression signature, PrOstate Radiation Therapy Outcomes Score, or PORTOS, was first reported in 2016 by a research team from the University of Wisconsin School of Medicine and Public Health led by Dr. Shuang (George) Zhao, assistant professor of human oncology. Researchers used a machine learning model to analyze gene expression patterns in tumors, correlating the data with patients’ responses to radiation therapy. They identified 24 genes whose expression levels predicted radiation response, resulting in PORTOS.
Oncologists may treat prostate cancer by surgically removing the prostate or with radiation therapy, either as a primary treatment or to treat recurrences after surgery. Radiation therapy is a life-saving treatment, but it can cause toxic effects such as urinary and bowel tract issues as well as sexual dysfunction.
“The more radiation you give, the higher the likelihood of causing a negative side effect for the patient. You want to give as little as necessary, but just enough to treat the tumor,” Zhao said.
“Being able to make that decision based on an individual patient’s tumor, as opposed to a population average, means you can potentially improve outcomes or reduce the side effects for specific groups of patients.”
The results, which were published in Annals of Oncology in May, indicate PORTOS can help oncologists tailor radiotherapy strategies for prostate cancer patients, making it the first tool of its kind to be validated in randomized clinical trials.
Prostate cancer is the most commonly diagnosed cancer in men in two-thirds of all countries, and the fifth leading cause of cancer-related deaths worldwide.
In two large randomized phase III clinical trials, researchers examined tumor samples from patients to determine the outcomes of higher or lower doses of radiation to treat prostate cancer. One trial was run in Europe by the Swiss Group for Clinical Cancer Research and the other in the United States by NRG Oncology.
The European clinical trial focused on patients undergoing radiation therapy for recurrent cancer following surgical removal of the prostate gland — a treatment called salvage radiotherapy. The U.S. trial included patients being treated with radiation as a first option, called definitive radiotherapy. Tumor samples were tested from more than 200 patients in each trial to measure PORTOS.
Researchers correlated patients’ PORTOS with whether they benefited from higher doses of radiation. Then they compared disease progression probability in patients with high versus low PORTOS.
In both trials, the analysis showed that patients with high PORTOS benefited from higher doses of radiation and patients with low PORTOS did not.
In the American clinical trial, PORTOS levels were classified into three groups: low, average or high. Participants with low PORTOS showed no benefit from higher radiation doses, but those with average PORTOS fared better with higher doses. The beneficial effect was most pronounced for those with the highest PORTOS.
A key question for the researchers was which biological processes in prostate tumors are driven by the 24 genes measured using PORTOS. By assessing gene expression patterns in a large dataset from more than 73,500 prostate cancer patients, they found that high PORTOS is strongly associated with immune signatures in tumors and moderately associated with hypoxia, meaning tissue with low oxygen levels. Hypoxic tumors are known to be more radiation resistant.
The connection to the immune system is intriguing, Zhao noted, as scientists are already studying the relationship between radiation and the immune system. In a recent study, UW–Madison researchers found that mixed dose radiation enhances immune response to cancer.
Zhao and his team plan to continue studying how PORTOS can help predict response to radiation and personalize prostate cancer treatment by validating the gene signature in additional clinical trials.
– Author Mary Bosch is a 2025–26 science communications intern at the School of Medicine and Public Health
This research was supported in part by grants from the National Institutes of Health U24CA196067, 1DP2CA271832-01, 1R01CA271540, U54CA273956 and R01CA240991) and the Department of Defense (PC190039, PC210122, PC200334 and W81XWH-21-1-0296).
