Predictive performance of risk prediction models for lung cancer incidence in Western and Asian countries: a...

Numerous prediction models have been developed to identify high-risk individuals for lung cancer screening, with the aim of improving early detection and survival rates. However, no comprehensive review or meta-analysis has assessed the performance of these models across different sociocultural contexts. Therefore, this review systematically examines the performance of lung cancer risk prediction models in Western and Asian populations. PubMed and EMBASE were searched from inception through January 2023. Studies published in English that proposed a validated model on human populations with well-defined predictive performances were included. Two reviewers independently screened the titles and abstracts, and the Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to assess study quality. A random-effects meta-analysis was performed, and a 95% confidence interval (CI) for model performance was reported. Between-study heterogeneity was adjusted for using the Hartung-Knapp-Sidik-Honkman test. A total of 54 studies were included, with 42 from Western countries and 12 from Asian countries. Most Western studies focused on ever-smokers (19/42; 45.2%) and the general population (17/42; 40.5%), and only two Asian studies developed models exclusively for never-smokers. Across both Western and Asian prediction models, the three most consistently included risk factors were age, sex, and family cancer history. In 45.2% (19/42) of Western and 50.0% (6/12) of Asian studies, models incorporated both traditional risk factors and biomarkers. In addition, 14.8% (8/54) of the studies directly compared biomarker-based models with those incorporating only traditional risk factors, demonstrating improved discrimination. Machine-learning algorithms were applied in eight Western models and two Asian models. External validation of PLCO_M2012 (AUC = 0.748; 95% CI: 0.719-0.777) outperformed other prediction models, such as Bach (AUC = 0.710; 95% CI: 0.674-0.745) and Spitz models (AUC = 0.698; 95% CI: 0.640-0.755). Despite showing promising results, the majority of Asian risk models in our study lack external validation. Our review also highlights a significant gap in prediction models for never-smokers. Future research should focus on externally validating existing Asian models or incorporating relevant Asian risk factors into widely used Western models (PLCO_M2012) to better account for unique risk profiles and lung cancer progression patterns in Asian populations.