Histology-Derived Signatures Predict Recurrence Risk and Chemotherapy Benefit in Randomized Trials of Early Breast...

Area of Science:

• Oncology
• Genomics
• Digital Pathology

Background:

• Routine hematoxylin and eosin (H&E) slides offer a rich source of information.
• Histology-derived gene-expression signatures can potentially enhance prognostic and predictive capabilities in cancer research.

Purpose of the Study:

• To determine if gene-expression signatures derived from H&E slides are prognostic for recurrence and predictive of chemotherapy benefit in early breast cancer patients.

Main Methods:

• A multi-cohort study analyzed whole-slide images from over 7,000 patients across four cohorts.
• 61 histology-derived signatures were generated per patient using a standardized pipeline.
• Prognostic performance was assessed using distant recurrence-free interval (DRFI) and distant recurrence-free survival (DRFS); predictive value was evaluated via signature-treatment interactions.

Main Results:

• The top histology signature demonstrated strong prognostic performance across validation cohorts (C-index 0.60-0.70) for 5-year DRFI.
• Significant predictive signals for chemotherapy benefit were observed, particularly for taxane and dose-dense chemotherapy.
• Combined clinical-histology models improved risk stratification, identifying low-risk groups with a 2%-10% risk of distant recurrence or breast cancer death.

Conclusions:

• Histology-derived signatures from H&E images are broadly prognostic for early breast cancer recurrence.
• These signatures show potential in predicting chemotherapy benefit, outperforming traditional clinical factors.
• Integrated clinical and histological data enhance risk stratification for personalized treatment strategies.