Lung cancer models overview

General information

The Cancer Intervention and Surveillance Modeling Network (CISNET) lung group develops and applies population models for lung cancer, quantifying the impact of tobacco control and low-dose computed tomography (LDCT) on lung cancer and all-cause mortality. Nine lung cancer natural history models have been developed independently by investigators at seven institutions:*

  • Stanford University (Lung Cancer Outcomes Simulator; LCOS)
  • Massachusetts General Hospital and Harvard Medical School (Lung Cancer Policy Model; LCPM)
  • Fred Hutchinson Cancer Research Center (Longitudinal Multistage Observation Model; LMO)
  • Erasmus Medical Center (MIcrosimulation SCreening ANalysis Lung Model; MISCAN-Lung)
  • Georgetown University (SimSmoke Tobacco Control Policy Simulation Model; SimSmoke)
  • Georgetown University (Smoking-Lung Cancer Macro Model)
  • University of Michigan (Lung Cancer Screening Model [UM-LCSc])
  • University of Michigan (Lung Cancer Smoking Model [UM-LCSm]); and
  • Yale University (Yale Lung Cancer Model for population rates; YLCM).

Each CISNET-Lung model includes a dose-response module that relates individual detailed smoking histories to lung cancer risk and outcomes. A central component of the CISNET-Lung modeling is the smoking history generator (SHG), which simulates individual smoking histories based on the US population’s historical smoking patterns (1,2). These individual smoking histories, combined with simulations of the age at death from causes other than lung cancer, serve as inputs for the CISNET-Lung models. Each model then uses these smoking histories to simulate lung cancer incidence and mortality individually and at the population level under various tobacco control (3) or LDCT screening scenarios (4-7).

Data used by the models

The SHG was developed based on multiple data sources: National Health Interview Survey (NHIS), the Cancer Prevention Studies I & II (CPS-I & CPS-II), and the Human Mortality Database (HMD). The CISNET-Lung models’ dose-response modules were developed using data from various prospective cohort studies, such as the CPS II (8) and the Nurses’ Health Study (NHS) and the Health Professionals’ Follow-up Study (HPFS) (9,10), or a set of logistic regression models and tumor progression functions based on Surveillance, Epidemiology, and End Results (SEER) data (11).

For the simulations of LDCT screening impact, each CISNET-Lung model incorporated data from the two largest US lung cancer screening trials: the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) and the National Lung Screening Trial (NLST) (12,13). The data obtained from the PLCO and the NLST were used to derive information about the natural history of lung cancer by histology and the effectiveness of screening for lung cancer with chest radiography and LDCT.

Major contributions to public health policies

The CISNET-Lung group has made a number of major contributions to knowledge about lung cancer, and its work has informed the development of prevention and control strategies against lung cancer. The group’s work reconstructing smoking patterns in the US by birth cohort, age, and calendar year have provided new insights into the impact of tobacco control policies on smoking prevalence in the US (14,15). The group also showed the considerable impact of tobacco control on reducing lung cancer mortality and smoking-related mortality in the US since 1964 (3,16). Other projects include evaluating the health effects of raising the minimum age for purchasing tobacco products (17) and the development of a tobacco policy tool to project the impact of policies on smoking and health outcomes in the US.

The CISNET-Lung group has also investigated the long-term benefits and harms of hundreds of population-level lung cancer screening strategies that vary the screening frequency and eligibility criteria (4,6,7). These analyses were conducted in collaboration with the United States Preventive Services Task Force (USPSTF) and supported their revised lung cancer screening guidelines (4,5). Current work is extending the models to consider the impact of risk-based screening criteria (18) and to evaluate the cost-effectiveness of LDCT screening.

In-depth model information

Smoking dose-response models

The LCOS, MISCAN-Lung, SimSmoke, Smoking-Lung Cancer Macro Model,UM-LCSm (formerly known as FH-LC [Fred Hutchinson Cancer Research Center]), and YLCM models incorporate versions of the Two-Stage Clonal Expansion (TSCE) model as their central smoking dose-response model (8-10). The TSCE model is a biologically based mechanistic model, which models the effects of age and smoking on lung cancer development (19). Although six CISNET-Lung models used the TSCE model as a dose-response module, each group used it with a different parameterization. The LMO model includes a longitudinal multistage observation model (by histology), the UM-LCSc model includes a multistage clonal expansion model (by histology), and the LCPM model includes a probabilistic model (by histology) as a dose-response module.

Histologic types

Lung cancer can be classified into several different histologic types. The LCOS model includes adenocarcinoma, large cell, squamous, and small cell lung carcinoma (SCLC). The LCPM model includes six lung cancer cell types: adenocarcinoma, bronchioloalveolar carcinoma (BAC), large cell, squamous, SCLC, and other. The LMO model includes adenocarcinoma, large cell, squamous, BAC, other non-small cell lung carcinoma (O-NSCLC) and SCLC. The MISCAN-Lung model incorporates the following types: adenocarcinoma/BAC, large cell, squamous, SCLC, and O-NSCLC. The UM-LCSc model includes adenocarcinoma/BAC, SCLC, and O-NSCLC.

Stage progression

The LCOS and LCPM models incorporate stage progression based on tumor volume and metastatic burden. The LMO model incorporates stage progression based on tumor size and presence of metastasis. The MISCAN-Lung model incorporates a Markov state-transition by histology as the stage progression model. The UM-LCSc model incorporates a Markov state-transition by histology and sex (with rates proportional to tumor size).

Screening sensitivity

The LCOS model incorporates sensitivity by size/histologic features, and the LCPM model by size and location in the lung (correlates with histologic features). The MISCAN-Lung model incorporates stage/histology-specific sensitivities. The LMO and UM-LCSc models incorporate sensitivities by number of cells/histologic features/sex.

Screening effectiveness mechanism

The MISCAN-Lung model incorporates a cure model. The LMO model incorporates a combination of a cure model and a stage shift model. While the UM-LCSc model incorporates a stage shift model with adjustments for age of detection, the LCOS and LCPM models do not incorporate a stage shift model.

  • These models were denoted as follows in previous publications:

  • Model E (MISCAN-Lung, Erasmus Medical Center)

  • Model F (LMO, Fred Hutchinson Cancer Research Center)
  • Model S (LCOS, Stanford University)
  • Model M (LCPM, Massachusetts General Hospital)
  • Model U (UM-LCSc, University of Michigan)

References

  1. Jeon J, Meza R, Krapcho M, et al. Chapter 5: Actual and Counterfactual Smoking Prevalence Rates in the U.S. Population via Microsimulation. Risk Anal. 2012;32:S51-S68.
  2. Holford TR, Levy DT, McKay LA, et al. Patterns of Birth Cohort–Specific Smoking Histories, 1965–2009. Am J Prev Med. 2014;46(2):e31-e37.
  3. Moolgavkar SH, Holford TR, Levy DT, et al. Impact of Reduced Tobacco Smoking on Lung Cancer Mortality in the United States During 1975–2000. J Natl Cancer Inst. 2012;104(7):541-548.
  4. de Koning HJ, Meza R, Plevritis SK, et al. Benefits and harms of CT lung cancer screening programs for high risk populations. AHRQ Publication No. 13-05196-EF-2. Rockville, MD: Agency for Healthcare Research and Quality; 2013.
  5. de Koning HJ, Meza R, Plevritis SK, et al. Benefits and Harms of Computed Tomography Lung Cancer Screening Strategies: A Comparative Modeling Study for the U.S. Preventive Services Task Force. Ann Intern Med. 2014;160(5):311-320.
  6. Meza R, ten Haaf K, Kong CY, et al. Comparative analysis of 5 lung cancer natural history and screening models that reproduce outcomes of the NLST and PLCO trials. Cancer 2014;120(11):1713-1724.
  7. McMahon PM, Meza R, Plevritis SK, et al. Comparing Benefits from Many Possible Computed Tomography Lung Cancer Screening Programs: Extrapolating from the National Lung Screening Trial Using Comparative Modeling. PLoS ONE. 2014;9(6):e99978.
  8. Hazelton WD, Clements MS, Moolgavkar SH. Multistage carcinogenesis and lung cancer mortality in three cohorts. Cancer Epidemiol Biomarkers Prev. 2005;14(5):1171-1181.
  9. Meza R, Hazelton WD, Colditz GA, Moolgavkar SH. Analysis of lung cancer incidence in the nurses' health and the health professionals’ follow-up studies using a multistage carcinogenesis model. Cancer Causes Control. 2008;19(3):317-328.
  10. Hazelton WD, Jeon J, Meza R, Moolgavkar SH. Chapter 8: The FHCRC lung cancer model. Risk Anal. 2012;32 Suppl 1:S99-S116.
  11. McMahon PM, Kong CY, Johnson BE, et al. Chapter 9: The MGH-HMS lung cancer policy model: Tobacco control versus screening. Risk Anal. 2012;32 Suppl 1:S117-24.
  12. Aberle DR, Adams AM, Berg CD, et al. Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening. N Engl J Med. 2011;365(5):395-409.
  13. M.M. Oken, W.G Hocking, P.A. Kvale et al. Screening by chest radiograph and lung cancer mortality: The prostate, lung, colorectal, and ovarian (plco) randomized trial. JAMA. 2011;306(17):1865-1873.
  14. Holford TR, Levy DT, McKay LA, et al. Patterns of birth cohort-specific smoking histories, 1965-2009. Am J Prev Med. 2014;46(2):e31-7.
  15. U.S. Department of Health and Human Services. The health consequences of smoking: 50 years of progress. A report of the surgeon general. Chapter 13. 2014
  16. Holford TR, Meza R, Warner KE, et al. Tobacco control and the reduction in smoking-related premature deaths in the United States, 1964-2012. JAMA. 2014;311(2):164-171.
  17. IOM (Institute of Medicine). 2015. Public health implications of raising the minimum age of legal access to tobacco products. Washington, DC: The National Academies Press.
  18. Ten Haaf K, Jeon J, Tammemägi MC, et al. Risk prediction models for selection of lung cancer screening candidates: A retrospective validation study. PLoS Med. 2017;14(4):e1002277.
  19. Moolgavkar SH, Knudson AG,Jr. Mutation and cancer: A model for human carcinogenesis. J Natl Cancer Inst. 1981;66(6):1037-1052.