If you’re stymied by cancer’s unpredictable nature, you’ll be interested in this international study investigating how prostate cancer progresses over time.

The researchers, funded in part by NCI’s Cancer Systems Biology Consortium, used artificial intelligence (AI) to help grade biopsies. They coupled those findings with genomic data to predict prostate cancer’s trajectory, even decades after the initial diagnosis.

The researchers identified evolutionary biomarkers linked to certain genetic and morphological characteristics. People carrying those biomarkers were significantly more likely to experience a reoccurrence of their cancer.

Corresponding author and Emeritus Professor David Dearnaley, of The Institute of Cancer Research in London, noted some key advantages to their approach. His team used routine diagnostic biopsies (i.e., formalin-fixed, paraffin-embedded). And their genetic analysis needed only low-coverage whole genome sequencing. This makes their approach less expensive and more likely to work in a clinical practice setting.

The researchers will be further validating these findings, but they’re encouraged by their results. Co-corresponding author, Dr. Andrea Sottoriva, from Human Technopole, a new life science institute in Milan, said, “Someday, we may be able to combine these ‘evolvability’ metrics with traditional clinical data and commercially available transcriptomic tests to better predict when (or if) prostate cancer will re-occur, especially in individuals who are at particularly high-risk or who have locally advanced disease.”