Artificial Intelligence in Low-Dose Computed Tomography Lung Cancer Screening: Clinical Integration, Validation, and...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Lung cancer is a leading cause of mortality, often diagnosed late.
• Low-dose computed tomography (LDCT) screening reduces mortality but faces challenges like high false-positive rates and workflow burdens.
• Artificial intelligence (AI) offers potential solutions to improve LDCT screening's diagnostic consistency, efficiency, and risk stratification.

Purpose of the Study:

• To review current evidence on AI methodologies in LDCT lung cancer screening.
• To explore clinical applications, validation studies, and translation challenges of AI in this field.
• To synthesize findings on AI's role in nodule detection, characterization, risk prediction, and workflow optimization.

Main Methods:

• A structured literature search was performed in PubMed, Scopus, Embase, and Cochrane Library (January 2010 - September 2025).
• Keywords included AI, radiomics, and lung cancer screening.
• Studies focused on AI applications in LDCT for detection, characterization, risk prediction, and workflow optimization were selected.

Main Results:

• Deep learning and radiomics enable automated nodule detection and characterization, matching expert radiologist performance.
• Hybrid AI models integrating imaging and clinical data improve individualized risk prediction and personalized screening.
• AI-supported workflows enhance efficiency by reducing interpretation time while maintaining accuracy.

Conclusions:

• AI shows significant promise for improving LDCT lung cancer screening.
• Challenges remain in clinical translation, including validation, generalizability, interpretability, and workflow integration.
• Focusing on clinical integration and human-AI collaboration is crucial for realizing AI's full potential in lung cancer screening.