Researchers Identify Most Accurate Risk-Prediction Models for Lung Cancer Screening in Ever-Smokers

Key Points

  • The four risk-prediction models shown to be most accurate in selecting the highest-risk ever-smokers for lung cancer screening include: Bach; Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial Model 2012 (PLCOM2012); Lung Cancer Risk Assessment Tool (LCRAT); and the Lung Cancer Death Risk Assessment Tool (LCDRAT).
  • The data from this study can be used to inform future lung cancer screening guidelines.
  • The four lung cancer risk models should be further refined to improve their performance in certain subpopulations of ever-smokers.

Recent lung cancer screening guidelines from the National Comprehensive Cancer Network(®) permit the use of risk models to refer ever-smokers for screening. However, different models select different screening populations and the performance of each model in selecting ever-smokers is unknown. A study by Katki et al comparing nine risk models to refer ever-smokers for lung cancer screening has identified four models that most accurately predicted lung cancer risk and performed best in selecting ever-smokers for screening. The four models include Bach, PLCOM2012, LCRAT, and LCDRAT. The data from the study can be used to inform future lung cancer screening guidelines. The study is published in Annals of Internal Medicine.

Study Methodology

Researchers from the National Cancer Institute and the American Cancer Society compared the U.S. screening populations selected by nine risk models. The models included Bach; Spitz; Liverpool Lung Project (LLP); LLP Incidence Risk Model (LLPi); Hoggart; Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial Model 2012 (PLCOM2012); Pittsburgh Predictor; Lung Cancer Risk Assessment Tool (LCRAT); and the Lung Cancer Death Risk Assessment Tool (LCDRAT). 

The models selected screening populations in the United States using data from the National Health Interview Survey from 2010 to 2012. Model performances were evaluated using data from 337,388 ever-smokers in the National Institutes of Health–AARP Diet and Health Study and 72,338 ever-smokers in the Cancer Prevention Study II (CPS-II) Nutrition Survey cohort. 

The researchers used publicly available risk calculators or software provided by the model developers to confirm that all the models were correctly replicated. They then used publicly available R package “lcmodels” (ie, lung cancer risk models provided in the R programming language and software environment for statistical computing) to calculate risk from all nine models.

Study Results

At a 5-year risk threshold of 2.0%, the models chose U.S. screening populations ranging from 7.6 million to 26 million ever-smokers. These disagreements occurred because, in both validation cohorts, four models—Bach, PLCOM2012, LCRAT, and LCDRAT—were well-calibrated (expected-observed ratio range, 0.92–1.12) and had higher area under the curves (AUCs; range, 0.75–0.79) than five models that generally overestimated risk (expected–observed ratio range, 0.83–3.69) and had lower AUCs (range, 0.62–0.75).

The four best-performing models also had the highest sensitivity at a fixed specificity (and vice versa) and similar discrimination at a fixed risk threshold. These models showed better agreement on size of the screening population (7.6 million–10.9 million) and achieved consensus on 73% of persons chosen. 

“Ending the epidemic of smoking-related illness requires continued progress in smoking cessation and prevention. Effectively and efficiently targeting lung cancer screening to persons at highest risk can further reduce lung cancer mortality, the leading type of cancer death. Our findings suggest that four lung cancer risk models perform best in selecting U.S. ever-smokers for screening. The models should be further refined to improve their performance in certain subpopulations,” concluded the study authors.

Li C. Cheung, PhD, of the National Cancer Institute, is the corresponding author of this study.

Funding for this study was provided by the National Cancer Institute.

The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.


Advertisement

Advertisement



Advertisement

By continuing to browse this site you permit us and our partners to place identification cookies on your browser and agree to our use of cookies to identify you for marketing. Read our Privacy Policy to learn more.