ECP 2026: Operational Gains with AI in Breast Pathology

Multi-center study reports benefits to diagnostic confidence, costs, and efficiencies after AI implementation

September 16, 2026
3 min

The Pathologist

Overview

The DeBORAH Study demonstrated high diagnostic accuracy and significant cost savings through the implementation of an AI algorithm in breast core biopsy reporting across five NHS centers. The study reported over 99% sensitivity and specificity for cancer diagnoses and substantial reductions in immunohistochemistry ordering and turnaround times.

Background

The integration of artificial intelligence (AI) in pathology is becoming increasingly relevant as it has the potential to enhance diagnostic accuracy and operational efficiency. The DeBORAH Study specifically addresses these aspects within breast pathology, a critical area given the rising incidence of breast cancer.

Data Highlights

Metric	Value
Sensitivity for overall cancer diagnosis	99%
Specificity for overall cancer diagnosis	99%
Sensitivity for invasive cancer diagnosis	99%
Sensitivity for DCIS diagnosis	96%
Reduction in immunohistochemistry ordering	Up to 72%
Cost reduction per patient	£1.80 to £17.70
Reduction in turnaround times	16 to 17 hours

Key Findings

The AI algorithm achieved over 99% sensitivity and specificity for overall cancer diagnosis.
Invasive cancer diagnosis sensitivity was also reported at 99%.
There was a 72% reduction in immunohistochemistry ordering in benign cases at one center.
Cost reductions ranged from £1.80 to £17.70 per patient across four of the five centers.
Turnaround times for biopsy reporting decreased by 16 to 17 hours.
Pathologists reported increased confidence and efficiency when using the AI algorithm.

Clinical Implications

The findings from the DeBORAH Study indicate that AI can enhance diagnostic processes in breast pathology.

Conclusion

The DeBORAH Study highlights the implementation of AI in breast core biopsy reporting, demonstrating improvements in diagnostic accuracy and operational costs.