Search by:

Source

Usman Ali M, Miller J, Peirson L, et al. Screening for lung cancer: A systematic review and meta-analysis. Prev Med. 2016 Aug;89:301-314. PMID: 27130532.

Humphrey L, Deffebach M, Pappas M, et al. Screening for Lung Cancer: Systematic Review to Update the U.S. Preventive Services Task Force Recommendation. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Jul. PMID: 24027793.

Study Population: Individuals at high risk for lung cancer

Efficacy Endpoints

Death from any cause

Harm Endpoints

False positive results, incidental findings, complications of invasive procedures after a positive screening result

Narrative

Any means of reducing deaths due to lung cancer, the most frequently diagnosed cancer and leading cause of cancer death in the western world,1 would have major public health implications. We summarize two reviews of low dose computed tomography (LDCT) screening for lung cancer.2^, 3

Overall no mortality benefit was found in either review when comparing LDCT to usual care (no screening measures). Indeed two trials suggest LDCT worsens mortality.4^, 5 However there is nearly a mortality benefit (RR 0.94; .95CI 0.88-0.998) when pooling these trials together with an additional large trial, NLST, that compares LDCT to chest x-ray for screening.6

Notably, NLST, or the National Lung Screening Trial, published in 2011,6 randomized >53000 high-risk people to screening with low dose chest CT (LDCT) or annual chest x-ray. As the only lung screening trial ever to demonstrate a benefit the study has been highly cited and led to some guidelines and recommendations to use LDCT screening.

Unlike screening for most other cancers in which screening is done on a general population lung cancer screening trials target smokers at high risk since lung cancers overwhelmingly occur in previous or current smokers. NLST subjects were 55-74 with an average 36 pack-year smoking history.

In the 7 year follow-up report for NLST mortality was 0.46% lower with LDCT (NNS: 217),7 an encouraging and oft-cited finding.

Caveats

There are reasons to be wary of the NLST findings. Most obviously it is a single trial while meta-analyses, like the reviews for this summary, found no mortality benefit even including NLST.* This is because earlier trials find no benefit and in two studies mortality was increased, counterbalancing NLST.

Why the difference? As is often the case with multiple trials from multiple settings the answer is fuzzy. It may be the comparator: NLST compared LDCT to chest x-ray, which seems to be worse than nothing. Prior reviews including the two cited here find x-rays, when compared to nothing, cause false positives without saving lives and may lead to harms. LDCT could shine in comparison, even with a neutral or negative impact.

Or perhaps NLST was an anomaly. Almost 30% of the mortality advantage was made up of deaths unrelated to lung cancer. No explanation has been proffered for this, and LDCT did not seem to reduce deaths from other cancers or cardiovascular problems. Assuming this part of the difference is due to chance, removing it neutralizes the advantage (p=0.11).8

Conversely, the Canadian task force review, like the USPSTF, makes a “weak” recommendation for offering LDCT.4 These recommendations may or may not be premature. The Cochrane review points to short follow-up (7 years) and poorly understood harms. We are concerned there is a long history of irrational exuberance—often followed by reversal and regret—based on single studies not yet retested or reproduced.9

Challenges also exist for implementation of LDCT screening. For one, the cost will add billions to national health care expenditures annually,6 and while shared decision making has been mandated for LDCT this is largely being circumvented in practice.10

Although LDCT is the only screening strategy showing mortality benefit in a study it is associated with over-diagnosis and false positives, and the attendant physical and psychological consequences of both. Given the checkered history of LDCT, particularly when compared to usual care, the NLST results must be reproduced by other groups and in other settings before it can be considered a proven intervention.

Based on the (admittedly hopeful) possibility of benefit we have given this intervention a Yellow color designation, indicating more research is required.

*Editorial note: it is our policy to report, whenever possible, all-cause mortality in favor of disease-specific mortality, as the former is far more patient-centered. For those interested, in this case LDCT had a similar (no benefit) impact on ‘lung cancer mortality’.

Author

Saed Awadallah, MD; Michael Ritchie, MD
Supervising Editor: Shahriar Zehtabchi, MD

See theNNT.com's previous reviews of this topic:
CT Scanning for Lung Cancer Screening in High-Risk Smokers, August 7, 2011

Published/Updated

March 29, 2019