Applying artificial intelligence to ensure high quality and equitable lung cancer screening

Area of Science:

• Radiology
• Medical Informatics
• Public Health

Background:

• Lung cancer screening (LCS) using low-dose computed tomography (LDCT) aims for early detection but current criteria may exclude high-risk, underrepresented groups.
• Racial minorities and rural populations often face barriers to LCS access and quality interpretation, potentially exacerbating health disparities.

Purpose of the Study:

• To explore the potential of artificial intelligence (AI) to enhance the effectiveness and equity of lung cancer screening.
• To identify how AI can refine risk stratification, improve access, support radiologists, and manage findings in LCS.

Main Methods:

• Review of AI applications in refining LCS eligibility criteria beyond age and smoking history.
• Analysis of AI's role in improving image acquisition, remote interpretation, nodule detection, and malignancy risk assessment.
• Examination of challenges and considerations for AI implementation in LCS, including regulatory, infrastructural, and bias mitigation.

Main Results:

• AI can enhance LCS by incorporating diverse data for risk stratification, identifying individuals missed by current criteria (e.g., Black Americans, women).
• AI may improve screening access through enhanced imaging and remote interpretation, while supporting radiologists in nodule detection and risk assessment.
• Barriers to AI implementation include regulatory issues, cost, infrastructure needs, training, and the critical need for fairness-aware frameworks to prevent bias.

Conclusions:

• Integrating AI into LCS workflows holds promise for addressing disparities and improving screening quality and reach.
• Careful evaluation is essential to ensure AI implementation achieves its goals of enhanced effectiveness and equity in lung cancer screening.
• Addressing potential racial bias in AI algorithms and ensuring equitable access are critical for successful LCS advancement.