WCLC 2025 - Posters & ePosters-EP.04.37 An Observational Study Evaluating the Benefits of AI-Assisted Lung Cancer Screening Using Chest X-Rays in Bangkok, Thailand

EP.04.37 An Observational Study Evaluating the Benefits of AI-Assisted Lung Cancer Screening Using Chest X-Rays in Bangkok, Thailand

Back to course

Pdf Summary

This document outlines a study involving the detection and verification process of lung nodules in a general population using artificial intelligence (AI) and radiologist assessments. The workflow begins with AI flagging potential nodules in chest images from a diverse group of individuals, which are then reviewed and verified by radiologists. Verified nodules are categorized as confirmed, no nodule, or inconclusive. Confirmed nodules are further classified into low-risk or high-risk groups requiring different follow-up strategies, such as CT scans, immediate follow-up, or tissue biopsy.<br /><br />The population sample consists of 7,784 individuals with demographic details: 37.3% male, 62.6% female, and 0.1% undisclosed gender. Exposure to environmental risk factors is highlighted, with 51.0% reporting dust or smoke exposure. Smoking status includes 5.9% active smokers with over 30 pack-years, 2.3% active smokers with fewer than 30 pack-years, 5.0% former smokers who quit over 15 years ago, 3.8% former smokers who quit more recently, 15.9% second-hand smokers, and 67.1% non-smokers.<br /><br />The study emphasizes agreement and disagreement rates between AI detection and radiologist confirmation, identifying the accuracy and reliability of the AI system for early lung nodule screening. The ultimate goal is refining the risk stratification of lung nodules to guide patient management decisions optimally.<br /><br />Acknowledgment is given to AstraZeneca for supporting the software implementation integral to the AI detection process. This combined approach leverages technology and expert analysis to improve lung nodule detection efficacy and potentially enhance early diagnosis of lung conditions, including cancer.

Asset Subtitle

Chuleekorn Soudorn

Meta Tag

Speaker Chuleekorn Soudorn

Topic Screening and Early Detection

Keywords

lung nodules

artificial intelligence

radiologist assessment

nodule verification

risk stratification

lung cancer screening

chest imaging

smoking status

environmental exposure

AstraZeneca collaboration