Abstract 3575: The effects of selection bias by lung cancer screening

Abstract

Abstract Background: Lung cancer is the most common cause of cancer mortality worldwide. The five-year survival rate of lung cancer is still extremely poor: only 16%. This is mainly due to the fact that most patients present with an advanced stage of the disease. Recent studies showed that screening of high risk persons is very effective in reducing the mortality from lung cancer. However, the inclusion criteria for these studies are rather strict. For instance the study by The National Lung Screening Trial Research Team Persons only included people with more than 30 pack-years (PY) and aged between 55 and 74 years at time. Aim: To evaluate the effects of selection bias in lung cancer screenings studies as the The National Lung Screening Trial. Methods: The inclusion criteria of The National Lung Screening Trial were extrapolated to our population based cohort: The Rotterdam Study. The Rotterdam Study is an ongoing prospective cohort study in Rotterdam, the Netherlands, in which all inhabitants of the Rotterdam suburb Ommoord aged ≤ 55 years were invited to participate in the study. A total of 7,983 participants (78%) were included between 1990-1993 in the first cohort. Up to 2009, 208 deaths due to incident lung cancer were found in this cohort. Results: In our cohort, only 62 cases (29.8%) complied with the criteria of the screening trial (persons with unknown smoking history were included in the >30 PY group). 12.5% of the lung cancer cases between 55-74 years of age were never smokers or had a history of <30 PY at baseline and the remaining 57.7% were above 74 years. Consequently, 70.2% of cases would not have been included. A relative reduction in mortality of 20%, as found in the screening trial, would lead to a reduction in mortality in the Rotterdam Study population of lung cancer of 6.0% (i.e. 0.2*62 = 12.4 persons out of the 208 would have been saved from mortality by lung cancer). Conclusion: Screening of lung cancer with the present inclusion criteria reduces lung cancer death rate compared to standard of care, but because only a minority of lung cancer patients comply with these criteria, we should be cautious in generalizing findings to the total population of such patients. In addition, mortality reduction may be different in countries with other demographics and health care systems. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3575. doi:1538-7445.AM2012-3575