Predictive AI Biomarker for ADT Benefit in Prostate Cancer Identified

Researchers from University Hospitals (UH) Seidman Cancer Center have validated the first-ever predictive artificial intelligence (AI) biomarker of androgen deprivation therapy (ADT) benefit in prostate cancer, according to a recent study published in NEJM Evidence.

The study indicates that ADT with radiotherapy has the potential to benefit patients with localized prostate cancer, but the treatment can have negative effects on quality of life. Further, the authors note that there is no validated predictive model to guide ADT’s use.

To address this, the researchers leveraged clinical data and digital pathology images from pretreatment prostate tissue to build a dataset to train an ADT benefit prediction model. These data were collected from 5,727 patients enrolled in five separate phase three randomized clinical trials, in which treatment was radiotherapy with or without ADT.

This information was used to develop and validate an AI model designed to predict which patients would develop distant metastasis, or cancer that spreads from a primary tumor to other areas of the body. The model used patient-specific baseline data to generate a binary output denoting whether an individual would likely benefit from ADT or not.

The tool was validated using data from a cohort of NRG Oncology/Radiation Therapy Oncology Group (RTOG) 9408 trial participants, all 1,594 of whom were randomly assigned to radiotherapy plus or minus four months of ADT during the research.

In the validation cohort, ADT significantly improved time to distant metastasis across 14.9 years of follow-up. Of these, 34 percent of individuals were model positive, indicating that ADT provided benefit compared to radiotherapy alone. The remaining 1,051 patients were model negative and saw no benefit from ADT.

These findings led the researchers to conclude that their model can accurately identify patients with intermediate risk for prostate cancer likely to benefit from short-term ADT.

“This is truly a milestone in the treatment for prostate cancer,” said first author Daniel Spratt, MD, the Vincent K. Smith chair in Radiation Oncology and professor in the Department of Radiation Oncology at UH Seidman Cancer Center and Case Western Reserve University, in a press release detailing the study. “With the first-ever predictive biomarker of ADT benefit in prostate cancer, created with AI, we are able to further realize the ability to create a personalized approach for the treatment of cancer.”

Other institutions are also working to leverage predictive analytics and AI to combat prostate cancer.

Last week, the United States Department of Veterans Affairs (VA) shared that research teams from five VA medical centers are working to develop an AI tool to predict aggressive prostate cancer.

The study will look at the metastatic progression of high-risk localized prostate cancer in US veterans, who are disproportionately impacted by the disease compared to the general population. The research will begin by analyzing data from over 5,000 veterans who have been diagnosed with high-risk prostate cancer and received initial treatment.

From there, the study is set to expand to 14 VA sites.