Journal: Cancer discovery
An artificial intelligence model integrating diverse tumor and patient features was retrospectively compared with a widely used breast cancer genomic assay for predicting recurrence risk.
The multimodal AI approach—leveraging multiple data types rather than genomics alone—demonstrated significantly higher predictive accuracy for recurrence.
- Higher accuracy: The AI model outperformed the genomic assay in predicting recurrence risk.
- Comprehensive data: It integrated multiple tumor and patient features, not just genomic information.
- Clinical potential: If validated prospectively, it could improve how clinicians stratify recurrence risk.
- Treatment decisions: It may help refine decisions about adjuvant therapy and follow-up intensity in breast cancer care.