The Potential Role of Artificial Intelligence in Lung Cancer Screening Using Low-Dose Computed Tomography

Area of Science:

• Radiology
• Artificial Intelligence
• Public Health

Background:

• Low-dose CT (LDCT) screening significantly reduces lung cancer mortality in high-risk populations.
• Concerns regarding false positive rates, cost-effectiveness, and radiologist availability hinder widespread LCS implementation.
• Comorbid smoking-related diseases contribute significantly to mortality in smokers, often overlooked in current LCS programs.

Purpose of the Study:

• To review the performance of AI algorithms in interpreting LDCT scans for lung cancer screening.
• To assess the potential of AI as a second reader to improve LCS efficiency and accuracy.
• To explore AI's capability in identifying smoking-related comorbidities from CT scans for early intervention.

Main Methods:

• Systematic review of AI algorithms applied to various tasks in LCS CT scan interpretation.
• Evaluation of AI performance metrics for tasks such as nodule detection and characterization.
• Analysis of AI's potential for identifying comorbidities like cardiovascular disease, emphysema, and osteoporosis.

Main Results:

• AI algorithms show promise in enhancing the efficiency and accuracy of LCS interpretation.
• AI can serve as a valuable second reader, potentially reducing false positives and improving detection rates.
• AI demonstrates capability in identifying incidental findings of smoking-related comorbidities, offering opportunities for secondary prevention.

Conclusions:

• AI holds significant potential to optimize lung cancer screening programs by increasing efficiency and accuracy.
• Integrating AI into LCS can address current implementation barriers and improve patient outcomes.
• AI-driven analysis of CT scans can extend the value of LCS by enabling early detection and management of comorbid conditions in smokers.