DNA methylation patterns predict aggressive prostate cancer behavior

Prostate cancer is the most common malignancy in American men. Four million are living with it now, and another 330,000 will hear the diagnosis this year. Most of these men will be older, and most will face a slow-growing tumor that stays put, never threatening their lives. But several hundred thousand will not be so fortunate. Their cancers will move fast, spreading into lymph nodes and bone, becoming lethal. The central mystery of prostate cancer medicine has always been this: why does the same disease behave so differently from one man to the next?

For decades, researchers have known that mutations in cancer DNA shape how aggressive a tumor becomes. But mutations alone don't tell the whole story. A man can carry the same genetic changes as another and face a completely different disease trajectory. The question that has haunted oncologists is what happens between the mutation and the outcome—what translates the genetic code into clinical behavior.

Paul Boutros, a professor and director of the cancer center at Sanford Burnham Prebys Medical Discovery Institute, and his international team have now identified the answer: methylation. It's a process as old as cell biology itself. Methyl groups—chemical tags—attach to specific segments of DNA and switch genes on or off. In cancer, methylation patterns act like a Rosetta Stone, revealing how a tumor will behave and whether it will respond to treatment. Crucially, methylation happens after mutations occur, making it the bridge between what the cancer's genetics say and what actually happens to the patient.

To map this bridge, Boutros and colleagues assembled an unprecedented dataset: 3,001 prostate tissue samples spanning normal aging prostate through every stage of disease, from localized tumors to metastatic cancers that had spread throughout the body. The samples came from patients around the world, representing different ages and ancestries. A subset of 884 samples underwent deeper analysis, with researchers examining DNA and RNA alongside methylation patterns. What emerged from this massive undertaking was striking in its simplicity. Of the millions of possible methylation patterns in prostate cancer, just four recurred again and again across thousands of patients. These four "methylation subtypes" were not random noise. Each one corresponded to a distinct clinical reality. One subtype appeared almost exclusively in normal aging prostate and in cancers that grew so slowly they posed little threat. Two subtypes marked tumors that progressed at a moderate pace but rarely escaped the prostate itself. The fourth subtype appeared almost exclusively in cancers that metastasized and became lethal.

What made this discovery powerful was its consistency. The methylation patterns replicated across age groups, across different ancestries, and even across patients whose cancers carried entirely different mutations. Boutros described methylation as the way cancer "took the information from cancer genetics, patient lifestyle and other features and created a single signal that we could read." It was, in other words, the ideal biomarker—a test that could tell a clinician what kind of disease they were facing.

The research, published in Cancer Discovery in June 2026, required an unusual collaboration. Twenty-two institutions across the globe contributed expertise: cancer biologists working alongside statisticians, urologists consulting with radiation oncologists and pathologists. Takafumi Yamaguchi, a bioinformatician at Sanford Burnham Prebys and co-lead author, called it "a beautiful, complex dance," noting that many of the technologies and artificial intelligence techniques used in the analysis did not exist a decade ago.

The implications are immediate and practical. If methylation patterns can predict which men face aggressive disease, clinicians could spare many from unnecessary treatment while intensifying care for those who truly need it. But Boutros is already looking ahead. The team plans to bring methylation testing into prostate cancer clinics, translating laboratory discovery into patient care. They also noticed something else in the data: hints that oxygen flow in the prostate may play a role in cancer aggressiveness. They will soon test whether light exercise and similar interventions can reduce the incidence of aggressive disease in clinical trials. The question that has haunted prostate cancer medicine for decades now has an answer. What comes next is learning how to use it.