Location of AI risk markers and associated mammographic features in screening mammograms obtained years before screen-detected breast cancer - Scorecard - MDSpire
Advertisement
Location of AI risk markers and associated mammographic features in screening mammograms obtained years before screen-detected breast cancer
Clinical Scorecard: Identification of AI Risk Indicators and Corresponding Mammographic Characteristics in Screening Mammograms Prior to Breast Cancer Diagnosis
At a Glance
Category
Detail
Condition
Breast Cancer
Key Mechanisms
Mammographic screening and AI risk scoring
Target Population
Women aged 50–69 years
Care Setting
BreastScreen Norway program
Key Highlights
Breast cancer is the most frequently diagnosed cancer among females globally.
Mammographic screening reduces the incidence of advanced disease and breast cancer-specific mortality.
AI models can enhance early breast cancer detection by generating malignancy risk scores.
20–30% of screen-detected and interval cancers had visible signs on prior mammograms.
The study analyzed 130,031 screening examinations among 42,371 women.
Guideline-Based Recommendations
Diagnosis
Screen-detected cancer includes histologically verified ductal carcinoma in situ (DCIS) and invasive cancer.
Management
Mammograms assigned a score ≥ 2 by radiologists are reviewed for potential recall.
Monitoring & Follow-up
AI models generate continuous risk scores to assist in monitoring.
Risks
Mammographic screening is associated with potential harms, including overdiagnosis and false positives.
Patient & Prescribing Data
Women screened in the BreastScreen Norway program.
AI risk scores can identify cancers 2 to 4 years prior to diagnosis.
Clinical Best Practices
Utilize BI-RADS for consistent evaluation and communication of mammographic findings.
Incorporate AI risk scoring to enhance detection capabilities.
