|

Home / Research / Publications / A Multimodal-Multitask Deep Learning Model Trained in NSABP B-42 and Validated in TAILORx for Late Distant Recurrence Risk in HR+ Early Breast Cancer

Publications

A Multimodal-Multitask Deep Learning Model Trained in NSABP B-42 and Validated in TAILORx for Late Distant Recurrence Risk in HR+ Early Breast Cancer

Key Findings

  • Caris MI Clarity™ BCR showed strong, independent prognostic performance for both late and overall distant recurrence (DR) risk in the TAILORx cohort, with meaningful risk stratification up to 15 years.
  • The model remained prognostic after adjustment for clinical and genomic factors, demonstrating discrimination comparable to or better than existing tools across multiple analyses.
  • Together with prior NSABP B-42 results, these findings support the potential clinical utility of Caris MI Clarity BCR to inform long-term treatment decisions for patients with HR+ early breast cancer.
Download Publication
Learn More