23/05/2025

Each year, millions of women undergo mammography to screen for breast cancer, subtle signs such as microcalcifications are often difficult to detect. A research team has developed a deep-learning-based diagnostic system. This artificial intelligence (AI) approach improves the accuracy and reliability of detecting microcalcifications, even across varying imaging equipment and patient demographics.

Image: A set of annotated mammogram images comparing ground truth labels with AI-based predictions; Copyright: Q. Lin, W.-M. Tan, J.-Y. Ge et al.

AI-assisted mammography enables radiologists to detect microcalcifications — tiny early markers of breast cancer — more accurately and consistently across imaging devices.

Products and exhibitors related to this topic

Exhibitors and products related to this topic can be found in the database of MEDICA:

Key technical components

The AI system is based on two technical innovations:

Adaptive multi-scale detection:
By integrating a region-based convolutional neural network (R-CNN) with a feature-pyramid network (FPN), the model captures details at multiple scales. This enables the detection of both clusters and individual microcalcifications, without needing manual threshold adjustments.

Robust cross-center training:
The algorithm was trained on 4,810 biopsy-confirmed mammograms from three hospitals, covering 6,663 lesions (split between benign and malignant). The system standardizes input images, making it compatible with various scanners and clinical environments.

Diagnostic performance

In blind validation tests, the AI system achieved:

75% overall accuracy at the microcalcification-lesion level

76% sensitivity for malignant lesions

72% accuracy at the breast level

“This solution can be deployed directly on standard radiology workstations,” adds Dr. Yu. “By pre-marking suspicious regions on each mammogram, it enables radiologists to quickly focus on areas of concern, significantly reducing both missed diagnoses and unnecessary biopsies—thereby easing patient discomfort and lowering healthcare costs.”

MEDICA-tradefair.com; Source: Fundamental Research

More about Fundamental Research at: www.sciencedirect.com