The Evidence Behind Lung Cancer Screening: NLST, NELSON, and Number Needed to Screen

Two adequately powered randomized trials anchor the case for lung cancer screening. In the National Lung Screening Trial (NLST), three annual low-dose CT (LDCT) scans reduced lung-cancer mortality by 20.0 percent compared with chest radiography among older heavy smokers, and the European NELSON trial later reported a 24 percent reduction in men at ten years against a no-screening control. The benefit is genuine but modest in absolute terms: roughly 320 people had to be screened in NLST to prevent one lung-cancer death, and a share of screen-detected cancers may never have caused harm. Reading those numbers honestly is what separates a screening decision from a screening slogan.

Two trials that actually moved mortality

Most screening claims rest on studies that count how many cancers a test finds. That is the wrong currency. Finding more cancers can reflect earlier detection of lethal disease, or it can reflect discovery of slow-growing tumors that would never have mattered. Only a mortality endpoint in a randomized trial separates the two, and lung cancer screening has two such trials.

NLST: the American benchmark

NLST, published in the New England Journal of Medicine in 2011, enrolled 53,454 participants aged 55 to 74 with at least a 30 pack-year history who either still smoked or had quit within the previous 15 years. Participants were randomized to three annual rounds of LDCT or to single-view chest radiography. Notably, the comparator was not "no test" but an older imaging test, which makes the observed benefit a conservative estimate of screening versus nothing.

The headline result was a 20.0 percent relative reduction in lung-cancer mortality (95 percent confidence interval 6.8 to 26.7, P=0.004). All-cause mortality was 6.7 percent lower in the LDCT group, a secondary signal that the lung-cancer benefit was not simply offset by harms elsewhere. That combination, a disease-specific and an all-cause signal pointing the same direction, is why NLST changed practice.

NELSON: the European replication

A single trial, however strong, invites the question of whether the result repeats. NELSON, reported in the same journal in 2020, provided the answer. This Dutch-Belgian trial studied 13,195 men as its primary population, with 2,594 women analyzed separately, and used LDCT at baseline, year 1, year 3, and year 5.5, with lengthening intervals rather than annual scans. Its control group received no screening at all, a sharper contrast than NLST's chest radiography.

At ten years, the cumulative rate ratio for death from lung cancer among men was 0.76 (95 percent confidence interval 0.61 to 0.94), a 24 percent reduction that closely tracks NLST despite different methods, a different comparator, and a European population. Among women the point estimate was larger, a rate ratio of 0.67, but the confidence interval was wide (0.38 to 1.14) because far fewer women were enrolled, so that subgroup is best read as directionally consistent rather than definitive. Two trials, two continents, two protocols, one convergent conclusion: screening the right people lowers lung-cancer death.

Reading the number needed to screen

Relative percentages are persuasive and incomplete. A 20 percent reduction sounds identical whether the underlying risk is high or trivial, yet the absolute payoff differs enormously. The number needed to screen (NNS) restores that missing context by asking how many people must undergo the program for one to benefit.

In NLST the NNS to prevent one lung-cancer death was about 320 over roughly three rounds of screening and follow-up. That figure compares favorably with several accepted screening programs, but it is an average across an enrolled population already selected for high risk. Within that population the NNS is not a fixed constant: it falls sharply among the participants at highest baseline risk and climbs into the thousands among those at lowest risk. The practical lesson is that eligibility rules are doing quiet, heavy work, concentrating the benefit where risk is high enough that the arithmetic favors screening.

The costs: false positives and overdiagnosis

Every screening test detects things that are not cancer and things that are cancer but would never have hurt anyone. Both carry a price.

False positives dominated the day-to-day experience of NLST. About one in four LDCT screens was called positive, and 96.4 percent of those positive results were ultimately false alarms. Most were resolved with follow-up imaging, but a minority led to invasive procedures, with their own complication risk, plus radiation exposure and the anxiety of an ambiguous scan.

Overdiagnosis is subtler and harder to measure. It refers to detecting a real cancer that would never have become symptomatic within the person's lifetime. One analysis of NLST estimated that roughly 18.5 percent of LDCT-detected cancers were overdiagnosed, though extended follow-up of the trial suggests that early figure overstated the problem, since some apparently indolent tumors did eventually declare themselves. The direction is what matters for a reader: a nonzero fraction of screen-detected cancers lead to treatment that provides no benefit and can cause harm.

Who truly benefits

The trials define the beneficiaries by risk, not by curiosity. In 2021 the U.S. Preventive Services Task Force issued a Grade B recommendation for annual LDCT in adults aged 50 to 80 with a 20 pack-year history who currently smoke or quit within the past 15 years, lowering the age and pack-year thresholds it had set in 2013. The logic mirrors the NNS data: benefit is real where baseline risk is substantial and shrinks toward the harms as risk falls.

Two implications follow. First, screening a low-risk person, a light or never-smoker outside the criteria, imports the harms of false positives and overdiagnosis without the mortality upside that justified them. Second, screening is a complement to, never a substitute for, smoking cessation, which remains the single largest lever on lung-cancer death. The honest bottom line is that LDCT earns its place for a defined high-risk group, and that a good decision weighs the roughly 1-in-320 chance of preventing a lung-cancer death against a high likelihood of at least one false alarm along the way.

This article is educational and is not medical advice; screening decisions should be made with a qualified clinician who can weigh your individual risk.