Benchmarking MMAI Prostate Cancer AI Tool Using STAR-CAP Data

An artificial intelligence biomarker for prostate cancer generated risk scores that were tested against a nearly 20,000-patient database, according to data presented at the American Society for Radiation Oncology annual meeting. Researchers evaluated how the multimodal artificial intelligence tool, known as MMAI, compares to the STAR-CAP database to determine whether its risk predictions match what patients actually experience.

Key Clinical Takeaways:

• The MMAI biomarker incorporates clinical and digital pathology information to generate a score ranging from 0 to 100, grouping patients into low-, intermediate-, and high-risk categories alongside 10- or 15-year absolute risk estimates for metastasis or death.
• Data from the ASTuTE trial in Australia showed that three in four intermediate-risk prostate cancer patients recommended for short-term androgen deprivation therapy chose a different path after receiving their AI biomarker result.
• Clinical data presented at the 2026 American Society for Radiation Oncology meeting demonstrated consistent prognostic performance across global populations, showing algorithmic fairness without racial or age bias.

Benchmarking Risk Estimates Against STAR-CAP Data

To evaluate real-world reliability, researchers compared commercial MMAI forecasts against STAR-CAP, a well-established multi-center repository encompassing outcomes from nearly 20,000 prostate cancer patients who underwent surgery, radiation therapy, brachytherapy, or hormone treatment. The STAR-CAP resource provides a benchmark for testing whether a tool’s predicted risks align with actual patient outcomes. The MMAI tool evaluates clinical and digital pathology data to produce an absolute risk of developing metastatic disease or dying from prostate cancer at 10 or 15 years.

This calibration testing evaluates whether the numbers on a patient report are accurate, distinguishing calibration from standard validation studies. Validation studies typically test discrimination using hazard ratios, area under the curve, or C-index values. A tool can discriminate effectively while still misstating absolute risk, such as reporting a 20% risk of metastasis when the true clinical risk is 5%.

Artera Platform Provides Consistent Insights for Treatment Decisions

Clinical findings presented at the 2026 American Society for Radiation Oncology annual meeting demonstrated consistent prognostic and predictive results across five abstracts evaluating global patient cohorts and United States cooperative group trials. Artera CEO and co-founder Andre Esteva stated that the platform delivers consistent, data-backed insights across diverse populations to help clinicians address treatment decisions.

In the GenesisCare-led ASTuTE trial in Australia, investigators examined how the biomarker informs clinical decisions regarding short-term androgen deprivation therapy, or ST-ADT, for men with intermediate-risk prostate cancer. While ST-ADT improves radiotherapy efficacy by slowing testosterone-driven cancer growth, it provides an average survival benefit of around 5% across all patients, leaving many men to experience side effects without significant clinical gain. When MMAI risk profiles were taken into account, doctors and patients altered their use of ST-ADT in both directions—skipping hormone therapy when biological danger was low and opting for it when testing pointed to potential benefits. Investigators observed that three in four men recommended for hormone therapy in the trial chose a different course after receiving their AI biomarker result.

Whole-Pelvis Radiation Reduces Metastasis for High MMAI Risk Patients

Findings from the POP-RT trial, involving individuals with localized high-risk and very high-risk prostate cancer treated in India, were shared during a separate oral presentation at the conference.

The analysis revealed that whole-pelvis radiation significantly lowered rates of distant metastasis among participants classified as having high MMAI risk, whereas individuals with intermediate MMAI risk did not experience a comparable advantage.

Additional analyses of NRG/RTOG phase 3 post-radical prostatectomy trials demonstrated algorithmic fairness across racial and age subgroups. The analysis showed that elevated MMAI scores mapped to common gene expression profiles linked to tumor cell growth, aggressive clinical behavior, and distant metastasis across all demographics, with predictive performance holding steady regardless of a patient’s age or racial background. This consistency holds special importance for African American men, who experience disproportionately elevated prostate cancer death rates and have historically lacked adequate representation in the creation of diagnostic biomarkers.

Disclaimer: The information provided in this article is for educational and scientific communication purposes only and does not constitute medical advice. Always consult with a qualified healthcare provider regarding any medical condition, diagnosis, or treatment plan.