MIcrosimulation SCreening Analysis (MISCAN) Lung Model
The Microsimulation Screening Analysis (MISCAN)-Lung Model (Erasmus University Medical Center) is a microsimulation model that simulates a population of individual life histories, development of preclinical and clinical lung cancer, survival of clinically detected lung cancer, death due to lung cancer, and death due to other causes in the presence and absence of screening. For each individual, a smoking history is generated using the Smoking History Generator (SHG), which was built by utilizing data on smoking habits in the US population (1).
Lung cancer is modeled through a multistep procedure. Once a person’s age at death from causes other than lung cancer is generated by the SHG, which is influenced by the person’s smoking exposure characteristics, the Two-Stage Clonal Expansion (TSCE) model is used to determine whether lung cancer develops in that individual (2,3). MISCAN-Lung distinguishes four histological types of lung cancers: squamous cell carcinoma, adenocarcinoma (which consists of the types adenocarcinoma, large-cell carcinoma, and bronchioloalveolar carcinoma), other (remaining non-small cell carcinoma), and small-cell carcinoma. Once lung cancer has developed, it will progress from less advanced to more advanced preclinical stages until it is clinically detected (Figure 1). Lung cancers can be detected either clinically (due to symptoms) or by screening. The incidence of clinically detected lung cancers depends on the onset, sojourn times and the probability of clinical detection, which both vary by cancer stage and histology (and by gender for the sojourn times). If the screening component of the model has not been activated, lung cancers can only be clinically detected. If a person dies of lung cancer before dying from other causes, his or her age of death is adjusted accordingly.
Figure 1 Lung cancer progression in the MISCAN-Lung model
Upon activating the screening component, properties of the screening test (such as the sensitivity and change in prognosis due to early detection) and screening policies (such as the starting age, stopping age, intervals between screening, requirements with regards to smoking, adherence to screening) can be specified. Preclinical lung cancers may be detected by screening, which may alter a person’s life history. Early detection by screening may cure a patient, allowing him/her to resume his/her normal (cancer free) life history. The probability of having such changes may differ by the stage of cancer at detection. MISCAN-Lung accounts for the effects of lead-time and over-diagnosis, i.e., detection of cancer due to screening which would have not been detected clinically during the person’s lifetime. Upon clinical or screen detection (without a history change) of lung cancer, the patient is assigned a histological type, stage- and gender-specific survival, which follows piecewise uniform distribution. The probability of curation of the disease (by stage) due to early detection by screening is based on data from the National Lung Screening Trial (NLST). The output of the MISCAN-Lung model consists of various simulated events in the presence and absence of screening, such as the number of diagnosed cases, the mortality due to the disease and other causes, and the number of life-years lost.
The MISCAN-Lung model has been used to evaluate the impact of tobacco control on US lung cancer mortality (4). Data from the NLST and the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) were used to calibrate MISCAN-Lung, from which information on the natural history and screen-detectability of lung cancer was derived (5-8). MISCAN-Lung contributed to the analyses of the CISNET Lung group in investigating the benefits and harms of 576 different lung cancer screening strategies (9). These analyses were used to inform the USPSTF for their recommendations for lung cancer screening (10). MISCAN-Lung was recently used to evaluate the cost-effectiveness of lung cancer screening in Ontario, Canada (11).
Tip: Hover your cursor over the dashed attribute links below for more information. View the details of this model in a grid with other lung models.