Dr Laura van’t Veer talks to ecancer at ESMO 2026 about the development and validation of an AI-based digital pathology assay designed to predict breast cancer recurrence risk using routine tissue samples.

The model was trained on 70,000 whole-slide images to reproduce the prognostic information provided by the 70-gene MammaPrint assay and validated in 372 patients from the RASTER study.

After a median follow-up of 10.3 years, the digital assay demonstrated strong concordance with MammaPrint (r=0.954) and successfully distinguished patients at high and low risk.

Ten-year breast cancer-specific survival was 98.1% in the low-risk group versus 86.7% in the high-risk group.

The digital assay also outperformed conventional clinical factors in predicting outcomes.

These findings suggest that AI-driven digital pathology could improve access to personalized recurrence risk assessment in early-stage breast cancer.