2023 World Conference on Lung Cancer (Posters)-P1.14. Validation of Lung Cancer Risk Prediction Models in an Asbestos Exposed Population

Pdf Summary

This study aimed to validate lung cancer risk prediction models in an asbestos-exposed population. Participants from the Western Australian Asbestos Review Program underwent LDCT screening over a 10-year period. The study evaluated the performance of the PLCOm2012 and Liverpool Lung Project (LLPv2) risk models, as well as modified versions of these models that accounted for asbestos exposure. Model accuracy was compared against the USPSTF2021 and proposed Australian Lung Cancer Screening Program eligibility criteria. <br /><br />Of the 2,198 participants who underwent LDCT screening, 2,126 were available for risk evaluation. The majority of participants were men, former or never smokers with a median age of 70. A total of 24 prevalent and 22 incident lung cancers were identified. The LLPv2 model demonstrated the best discriminative performance, with an AUC of 0.72, compared to the PLCOm2012 and PLCOall014 models. The PLCO2012Occ model, which included a predictive term for asbestos, had identical performance to PLCOm2012. The USPSTF2021 criteria had high specificity but poor true positive rates, while the proposed Australian criteria had the lowest true positive rate. <br /><br />Overall, the LLPv2 risk model had the highest true positive rate and better sensitivity compared to the USPSTF2021 and Australian criteria. However, the study suggests the need for more bespoke models calibrated to high-risk populations, particularly those with occupational asbestos exposure. <br /><br />In summary, the study validated lung cancer risk prediction models in an asbestos-exposed population. The LLPv2 model demonstrated the best performance, highlighting the importance of accounting for asbestos exposure in risk prediction. Further research is needed to develop customized models for high-risk populations.

Asset Subtitle

Chellan Kumarasamy

Meta Tag

Speaker Chellan Kumarasamy

Topic Screening & Early Detection: Implementation

Keywords

lung cancer risk prediction models

asbestos-exposed population

LDCT screening

discriminative performance

AUC

prevalent lung cancers

incident lung cancers

specificity

sensitivity

occupational asbestos exposure