To evaluate the effectiveness of AI-supported mammography screening compared to standard double reading in reducing interval cancer rates and improving detection sensitivity, with a focus on the implications for clinical practice.
Key Findings:
Interval cancer rates were reduced by 12% in the AI-supported group (1.55 vs 1.76 per 1,000 patients).
The AI group had 16% fewer invasive interval cancers (75 vs 89).
Sensitivity was higher in the AI group (80.5%) compared to the control group (73.8%).
Specificity was consistent at 98.5% for both groups.
The AI-supported approach also reduced screen reading workload by 44%.
Interpretation:
The MASAI trial demonstrated that AI-supported mammography screening yields better outcomes in terms of interval cancer rates and detection sensitivity, suggesting potential for implementation in clinical practice, especially in light of radiologist shortages.
Limitations:
Further analyses are needed to assess long-term benefits and cost-effectiveness.
The study focused on a specific population (median age 54) and may not be generalizable to younger women or different demographics.
Conclusion:
AI-supported mammography screening shows promise for improving breast cancer detection and reducing workload, particularly in light of radiologist shortages, indicating a potential shift in clinical practice.
