Multimodal artificial intelligence prediction of abiraterone efficacy in two STAMPEDE phase III trials of nonmetastatic very high-risk prostate cancer.
Ann OncolJun 6, 2026 (epub)
C T A Parker, H-C Huang, E Grist et al.
✦ AI-curated · Sources linked
In 30 seconds
This study evaluated a multimodal artificial intelligence (MMAI) model to predict abiraterone efficacy in nonmetastatic very high-risk prostate cancer patients from two STAMPEDE phase III trials. Among 1137 patients, those classified as MMAI very high-risk showed a significant improvement in metastasis-free survival (MFS) with abiraterone, with a hazard ratio of 0.47 and a 5-year MFS of 81% compared to 62% with standard treatment.
Key findings
- MMAI very high-risk patients had a 5-year MFS of 81% with abiraterone versus 62% with LT-ADT alone.
- The hazard ratio for MFS improvement in MMAI very high-risk patients was 0.47.
- Limited benefit from abiraterone was observed in MMAI standard high-risk patients, with a 5-year MFS of 82% versus 84%.
Why it matters
Identifying patients who will benefit most from abiraterone can optimize treatment strategies in high-risk prostate cancer, potentially reducing unnecessary exposure to toxicity and costs associated with ineffective therapies.