Standard treatment for prostate cancer often includes surgery, radiation and up to two years of hormonal therapy, which can cause fatigue, muscle loss, and raise the risk of heart problems. But what if not every patient needs the full course?

Researchers at Duke Cancer Institute, working with the health technology company ArteraAI, have developed a new artificial intelligence-powered biomarker that could help determine which patients need extended hormone therapy and which could safely avoid it.

Using digitized biopsy samples from a large clinical trial of 1,000 patients, the team trained the AI model to spot patterns linked to long-term outcomes, like the risk of cancer spreading to distant parts of the body, a major factor in prostate cancer survival.

“About two-thirds of the men had a positive biomarker, meaning that they benefited from the two years of hormonal therapy,” said Andrew Armstrong, MD, director of research for the DCI Center for Prostate and Urologic Cancers. “But one third of the men did not—they had no added risk of cancer spreading or coming back. That would save a third of all high-risk men that extra 18 months of hormonal therapy.”

The work published in the Journal of Clinical Oncology not only offers hope for more personalized prostate cancer care but also shows how AI tools could help reduce overtreatment and improve quality of life for cancer patients.

“We sought to develop a biomarker using tissue that could identify patients that really need those two years of hormonal therapy, or patients that could have excellent outcomes and not have to spend two years on hormonal therapy,” Armstrong said.

While the development and validation of the biomarker is complete, the team is now moving toward commercialization and creating a product that can be approved and used by doctors.

Armstrong acknowledged the importance of transparency in introducing AI tools into clinical care. He also noted that this may be just the beginning with more AI-driven innovations likely on the way.

“Digital pathology could be applied to many other contexts – kidney cancer, bladder, breast, colon, or lung,” Armstrong said. “There’s a lot of promise; as long as you have big data, long-term outcomes, and tissue samples, AI can uncover patterns that even expert pathologists might miss.”