Literature review demonstrates improved diagnostic accuracy with artificial intelligence in breast cancer screening, highlighting dataset diversity and clinical integration barriers.
Key Points
To evaluate the current clinical applications, diagnostic performance benefits, and implementation barriers of artificial intelligence across various breast cancer screening modalities.
Reviewed literature evaluating machine learning and deep learning algorithms integrated into mammography, digital breast tomosynthesis, ultrasound, and magnetic resonance imaging.
Assessed algorithmic contributions to lesion detection, image classification, workflow efficiency, risk prediction, and clinical translation hurdles.
Artificial intelligence algorithms improve lesion detection, image classification accuracy, and diagnostic workflow efficiency while mitigating radiologist interpretation variability and sensitivity deficits in dense tissue.
Implementation barriers persist across clinical environments, notably driven by restricted dataset diversity, algorithmic bias, uninterpretable model outputs, data privacy challenges, and unstandardized regulatory pathways.