ASSESSING RISK PREDICTION MODELS FOR LDCT SCREENING: A SCOPING REVIEW OF CURRENT PRACTICES, LIMITATIONS, AND CHALLENGES
Author(s)
Monica Verma, MPH1, Anand Jha, MBA2, Vinita Devjani, MBA3.
1Associate Director, Ansea Consultants Pte Ltd, Singapore, Singapore, 2Ansea Consultants Pte Ltd, Singapore, Singapore, 3Ansea Consultants Pte Ltd, Ahmedabad, India.
1Associate Director, Ansea Consultants Pte Ltd, Singapore, Singapore, 2Ansea Consultants Pte Ltd, Singapore, Singapore, 3Ansea Consultants Pte Ltd, Ahmedabad, India.
OBJECTIVES: Large-scale randomized controlled studies, including 2013 National Lung Screening Trial (NLST) and the 2020 NELSON trial, have demonstrated the efficacy of low-dose computed tomography (LDCT) screening in enabling early detection of lung cancer. However, due to potential harms—such as false positives, overdiagnosis, cost, and radiation exposure, LDCT should be limited to high-risk populations. We conducted a scoping review to examine the current risk prediction models in practice to identify high-risk groups for screenings.
METHODS: We conducted a scoping review in accordance with PRISMA guidelines on electronic data bases including Pubmed, Web of Science and Scopus. Relevant keywords included “lung cancer,” “risk prediction models,” and “LDCT screening.” All relevant articles published in English between 2018 and 2025 were included, along with grey literature. Studies from three regions—North America, Europe, and the Asia-Pacific—were considered.
RESULTS: While smoking remains the primary risk factor for lung cancer, a substantial and growing proportion of cases occurs in individuals who have never smoked, particularly in the Asia-Pacific region. Despite this trend, most of the nine identified risk prediction models rely predominantly on age and cumulative smoking exposure. Key models include those developed in the United States (e.g., the Bach model and the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial 2012 model [PLCOm2012]) and Europe (e.g., the Liverpool Lung Project [LLP] model). To address the under-representation of never-smokers, newer models have been developed in Asia, including the TNSF-Q model (2020, Taiwan), the Yeo model (2021, South Korea), and the Guo model (2022, China). However, these models require further refinement and external validation to reliably identify high-risk never-smokers.
CONCLUSIONS: As the proportion of lung cancer cases among never-smokers continues to rise, developing more robust and widely applicable prediction tools is essential to effectively identify high-risk never-smokers and expand lung cancer screening strategies.
METHODS: We conducted a scoping review in accordance with PRISMA guidelines on electronic data bases including Pubmed, Web of Science and Scopus. Relevant keywords included “lung cancer,” “risk prediction models,” and “LDCT screening.” All relevant articles published in English between 2018 and 2025 were included, along with grey literature. Studies from three regions—North America, Europe, and the Asia-Pacific—were considered.
RESULTS: While smoking remains the primary risk factor for lung cancer, a substantial and growing proportion of cases occurs in individuals who have never smoked, particularly in the Asia-Pacific region. Despite this trend, most of the nine identified risk prediction models rely predominantly on age and cumulative smoking exposure. Key models include those developed in the United States (e.g., the Bach model and the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial 2012 model [PLCOm2012]) and Europe (e.g., the Liverpool Lung Project [LLP] model). To address the under-representation of never-smokers, newer models have been developed in Asia, including the TNSF-Q model (2020, Taiwan), the Yeo model (2021, South Korea), and the Guo model (2022, China). However, these models require further refinement and external validation to reliably identify high-risk never-smokers.
CONCLUSIONS: As the proportion of lung cancer cases among never-smokers continues to rise, developing more robust and widely applicable prediction tools is essential to effectively identify high-risk never-smokers and expand lung cancer screening strategies.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
HSD37
Topic
Epidemiology & Public Health, Health Policy & Regulatory, Health Service Delivery & Process of Care
Disease
Oncology