AI for triple-negative breast cancer: bridging imaging to multi-omics

AI applied to MRI, ultrasound, mammography, and multi-omics shows promise for TNBC tasks like response prediction and risk stratification, but most studies lack external validation and adequate cohort sizes.

• MRI, ultrasound, mammography, whole-slide histopathology, and transcriptomics provide complementary data for AI models.
• Multimodal fusion and radiogenomic frameworks appear most promising for capturing TNBC heterogeneity.
• Current evidence is limited by small cohorts, inconsistent endpoints, non-patient-level splitting, and domain shift.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.