KEY POINTS
- ASTuTE is a prospective Australian multicenter implementation trial evaluating an AI-based multimodal biomarker in men with intermediate-risk prostate cancer receiving curative radiotherapy. This prespecified interim analysis included the first 200 of a planned 800 patients across 28 centers, including 150 unfavorable- and 50 favorable-intermediate-risk patients.
- The commercially available ArteraAI platform analyzes digitized H&E-stained biopsy slides together with clinicopathologic data and provides both a 10-year prognostic risk estimate and a prediction of benefit from short-term androgen deprivation therapy.
- The primary endpoint was met decisively: biomarker testing changed the final shared ADT decision in 55 of 200 patients, or 27.5% (95% CI 21.4–34.2%).
- Of 74 patients initially planning short-term ADT, 52 (70.3%, 95% CI 58.5–80.3%) changed to no ADT after testing. By contrast, only 3 of 126 (2.4%, 95% CI 0.5–6.8%) initially planning no ADT changed in the opposite direction (p<0.001). Final planned ADT use therefore fell from 37.0% to 12.5%.
- Only 30 of 200 patients (15%) had a predictive biomarker suggesting benefit from short-term ADT, while 85% were biomarker-negative. Overall, 85.5% of final shared decisions aligned with predictive biomarker status.
- Decision changes were concentrated in unfavorable-intermediate-risk disease: 35% of these patients changed treatment decisions compared with only 4% of favorable-intermediate-risk patients (p<0.001). The biomarker also demonstrated substantial risk overlap between conventional clinical groups, with 28% of each group showing prognostic estimates resembling the other group.
- Before testing, clinicians favored ADT more often than patients (46.0% vs 29.5%, p<0.001). After biomarker disclosure, this gap largely disappeared (15% vs 12%, p=0.38), suggesting the test also influenced clinician–patient agreement.
- These are decision-making outcomes, not cancer-control outcomes. The trial is open-label and single-arm, 85% of biomarker results pointed toward no ADT benefit, and the study is conducted in collaboration with the test manufacturer; mature oncologic, quality-of-life and cost-effectiveness results remain pending.
CLINICAL TAKEAWAY
The magnitude of de-escalation is substantial: planned short-term ADT use fell from 37% to 12.5% after AI biomarker testing. That makes ASTuTE highly relevant, but it does not yet prove that omitting ADT on this basis preserves cancer outcomes—the long-term follow-up will determine whether this is genuine personalization or simply less treatment.