To evaluate the role and impact of artificial intelligence (AI) in breast imaging, particularly in mammography screening, with a focus on its effectiveness and safety.
Key Findings:
AI can improve lesion detection, specificity, and workflow efficiency in breast imaging, but current tools have not demonstrated improved long-term clinical outcomes.
AI systems have shown potential in enhancing mammography accuracy and reducing radiologist workload, yet their clinical effectiveness remains to be fully validated.
Interpretation:
While AI shows promise in improving breast imaging, its integration requires careful evaluation, ongoing research, and robust governance to ensure safety and efficacy.
Limitations:
Lack of comprehensive post-implementation evaluation studies for commercial AI tools limits understanding of their effectiveness.
Limited evidence on long-term outcomes associated with AI use in breast imaging raises concerns about its current application.
Variability in study designs and a limited number of studies, particularly for digital breast tomosynthesis, hinder conclusive findings.
Conclusion:
AI has the potential to significantly enhance breast cancer screening, but further research and validation are necessary before making strong recommendations for its widespread use, emphasizing the need for evidence-based practices.
