Nubeqa edges Xtandi on cognitive side effects | ASCO 2026

Prostate cancers treated with Nubeqa (darolutamide) had significantly less cognitive decline than those treated with Xtandi (enzalutamide), according to results of a head-to-head comparison between the two drugs that are being presented at the 2026 annual meeting of the American Society of Clinical Oncology next weekend in Chicago.

“The reason that I did this study is because I have had patients who have felt that when they started on these treatments, particularly enzalutamide, maybe things were a little bit slow for them — they were just cognitively a little slow,” lead author Alicia K. Morgans, M.D., M.P.H., a genitourinary medical oncologist and the director of the survivorship program at Dana-Farber Cancer Institute in Boston, said at a press conference on May 18 previewing some of the research findings at the annual meeting.

Although the results favored Nubeqa, Morgans said they shouldn’t be interpreted as a blanket recommendation to favor it over Xtandi.

“I think that this is informative of practice, but it doesn’t draw a black line here and say none of this drug and only that drug. I think that’s a very important take-home message,” she said.

Eric J. Small, co-director of the urological cancer service at University of California, San Francisco, and a commentator on the research presented at the press conference, noted that cognitive function decline is a “limiting feature of our common treatments for advanced prostate cancer” and that it shows up in a variety of ways: “memory loss, [loss of] executive function, falls, impaired functioning, lost jobs, getting lost in parking lot.” When it is severe, “it’s substantial,” he said. The results presented by Morgans, he said, fit with what has been observed by clinicians taking care of prostate cancer patients. “Potentially, I think this offers an option for preserving cognitive health in [patients with] with advanced prostate cancer. So, [this is] a very important study confirming what we've observed clinically.”

Nubeqa and Xtandi are among the second-generation antiandrogens that have become mainstays of the treatment of advanced prostate cancer treatment. Negative effects on cognition are well-recognized as a side effect of blocking the action of androgens, the most common of which is testosterone. Xtandi has the advantage of FDA approval for a broader set of indications that Nubeqa, but it crosses the blood-brain barrier while Nubeqa does not, raising the prospects of it having a greater negative effect duon cognition. Although there has been plenty of research into cognitive side effects of the second-generation antiandrogens, Morgans said there has been no prospective study in an U.S. population comparing the central nervous system effects of Nubeqa and Xtandi until this one.

Morgans and her colleagues enrolled 111 patients into the study, dubbed ARAGOG, between August 2021 and March 2025 and randomly assigned them to be treated with Nubeqa or Xtandi. The median age of the study participants was 71 and the vast majority (83%) were White individuals. They had been diagnosed with either patients with metastatic hormone-sensitive prostate cancer, metastatic castration-resistant prostate cancer or nonmetastatic castration-resistant prostate cancer. The researchers used CANTAB, a validated computer-based platform, to measure the study participants’ cognition. CANTAB’s assessments were organized into five domains of cognition: PALFAM-memory recall, SWM- spatial working memory, SSP-short-term visual memory, working memory, RVP-rapid visual processing and OTSMCC- executive function. Patients were tested with CANTAB atas they were enrolled into the study and then at 12, 24, and 48 weeks. They were allowed to crossover to the other treatment at 12 or 24 weeks for prespecified reasons, including but not limited to a fall. The primary end point was the percent change in the maximally changed cognitive domain (MCCD) from baseline to 24 weeks.

The results reported by Morgans and her colleague in the abstract showed MCCD between baseline and 24 weeks was -15.8% for patients randomly assigned to Nubeqa; the domain was the PALFAM-memory recall domain. For patients randomly assigned Xtandi, the MCCD was -36.1%; the domain was SWM- spatial working memory. Their ASCO abtractd says that individual cognitive domains suggest a possible learning effect (increased scores) for Nubeqa and stable or mild decline for Xtandi-treated patients.

The crossover patterns are also something of an endorsement for Nubeqa. Of the patients randomly assigned to Nubeqa, 34 were eligible to crossover to Xtandi and of those assigned to Xtandi, 36 were eligible to cross over to Nubeqa, according to data Morgans presented in her slide (which were slightly different than the data presented in the abstract. Thirty (83%) of the 36 of the study participants in Xtandi group crossed over to Nubeqa whereas none of the Nubeqa patients crossed over to Xtandi.