Ilana Gareen

Ph.D., M.P.H.

Brown University

Associate Professor

NLST (Learn more about this study)

NLST-505

Apr 29, 2019

Lung Screening: Efficacy versus Effectiveness

In 2011, the National Lung Screening Trial (NLST) reported a 20% reduction in lung cancer mortality for participants screened with low-dose CT (LDCT) as opposed to chest X-ray. Based largely on these findings, LDCT lung screening is now a covered service for 8.7 million high-risk current and former smokers in the United States. For high-risk individuals younger than 65, insurance coverage is mandated under the Affordable Care Act; for those 65 and over screening is covered by Medicare. Medicare covers lung screening only if patients are enrolled in a Centers for Medicare and Medicaid Services-approved lung screening registry. The American College of Radiology Lung Cancer Screening Registry (ACR-LCSR) is the only registry currently approved by CMS. Although most medical societies have endorsed CT for lung cancer screening, the American Academy of Family Physicians concluded that the evidence was insufficient to recommend either for or against lung screening, as the results of the NLST had not been replicated in a community setting. Concerns regarding the dissemination of lung screening focus on 1) the high rate of false positive (FP) screens, for which participants receive a positive screening result and require additional testing, but do not have lung cancer, 2) the relatively high rate of potentially clinically important significant incidental findings (SIFs) detected at lung screening that are unrelated to lung cancer, and 3) potential harms from the diagnostic evaluation of these FP and SIF abnormalities. It is unclear whether the rate of FPs and SIFs seen in the NLST, will be replicated in community practice, as opposed to a clinical trial setting. Higher rates of FPs and SIFs in the community or inefficient diagnostic evaluation may result in delayed cancer diagnosis, excessive testing, iatrogenic complications due to unnecessary testing, or decreased cost-effectiveness of lung screening in the community, as opposed to the NLST. For the proposed research, we plan to: 1) compare the rate and type of abnormalities suspicious for lung cancer and the rate and type of SIFs in the ACR-LCSR community registry data as opposed to the NLST, 2) compare the diagnostic pathways used to assess these abnormalities in the community as opposed to the NLST, and 3) use a decision-tree cost-effectiveness analysis to compare community lung screening with NLST cost-effectiveness assumptions with respect to rates of lung and SIF abnormalities and diagnostic pathways; and to identify most cost-effective diagnostic pathways for each type of abnormality. This information is of vital importance to ensure that the reduction in lung cancer mortality reported by the NLST is achieved in the community setting.

Aim 1: To compare the rate and type of lung abnormalities meeting the Lung RADs criteria as suspicious for lung cancer (Lung RADS 4A, 4B, 4X) and the rate and type of SIFs in community registry data as opposed to the NLST. We hypothesize that the rate of lung abnormalities and SIFs will be higher in the community as opposed to the NLST.

Aim 2: To compare the diagnostic pathways used to assess lung abnormalities suspicious for cancer and SIFs in the community as opposed to the NLST, and to characterize each pathway with respect to the duration of the time to final diagnosis. We hypothesize that the health care utilization for community detected lung abnormalities and SIFs will exceed that observed in the NLST, and that we will be able to identify specific efficient pathways for addressing lung and SIF findings on lung screening within strata of type of abnormality.

Aim 3: To use a decision-analysis cost-effectiveness framework to compare community lung screening with NLST cost-effectiveness assumptions with respect to rates of lung and SIF abnormalities and diagnostic pathways; and to identify the most cost-effective diagnostic pathways for each type of abnormality. We hypothesize that using this approach we can identify the most cost-effective approaches to evaluate lung nodules and SIFs detected at cancer screening.