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

Low-dose CT screening lowers the risk of dying from lung cancer, but only for a defined group of older people with a heavy smoking history, and only when the decision is made with the trade-offs in view. That is the whole story in one sentence, and the rest of this article is about why each qualifier belongs there.

Key points#

• Two randomized trials, NLST in the United States and NELSON in Europe, both found meaningful reductions in lung-cancer death from low-dose CT: about 20 percent and 24 percent, respectively.
• Roughly 320 high-risk people had to be screened in NLST to prevent one lung-cancer death, a number that gets much better in the highest-risk participants and much worse in the lowest-risk ones.
• The main harms are false alarms (very common) and overdiagnosis (uncommon but real): finding a cancer that would never have caused symptoms.
• U.S. guidelines now offer screening to adults aged 50 to 80 with a 20 pack-year history who still smoke or quit within the past 15 years.
• Screening supplements quitting; it does not replace it.

Why "finding more cancers" is the wrong scoreboard#

It is tempting to judge a screening test by how many cancers it catches. That instinct misleads. A test can find extra cancers because it caught aggressive disease early, when treatment still helps, or because it caught slow tumors that would never have shortened a life. Counting detections alone cannot tell those two situations apart.

The only way to separate them is a randomized trial that measures whether fewer people actually die of the disease, and lung cancer screening is fortunate to have two such trials on two continents, which is why its evidence base is unusually solid for a cancer screen.

The two trials that set the standard#

NLST: the American benchmark#

The National Lung Screening Trial, reported in 2011, enrolled 53,454 participants aged 55 to 74 who had smoked at least 30 pack-years and either still smoked or had stopped within the previous 15 years. Half received three yearly rounds of low-dose CT; the other half received a single-view chest X-ray. That comparison matters: the control arm still got an imaging test, so any advantage measured here understates the gap between CT and doing nothing.

The primary result was a 20.0 percent relative drop in lung-cancer deaths (95 percent confidence interval 6.8 to 26.7, P=0.004). Deaths from all causes were also 6.7 percent lower in the CT group, a reassuring sign that the lung-cancer gain was not being cancelled out by harms elsewhere. A disease-specific signal and an all-cause signal pointing the same way is what moved guidelines.

NELSON: the European check#

A single strong trial always raises the question of whether the finding will hold up under different conditions. NELSON, published in 2020, supplied that test. This Dutch and Belgian study followed 13,195 men as its main group, with 2,594 women analyzed separately, and scanned participants at baseline, year 1, year 3, and year 5.5, stretching the intervals rather than repeating annually. Its control group received no screening at all, a cleaner contrast than NLST's chest X-ray.

At ten years, the rate of lung-cancer death among men was 24 percent lower with CT (rate ratio 0.76, 95 percent confidence interval 0.61 to 0.94). Landing so close to NLST's figure, despite a different protocol, a different comparator, and a different population, is what makes the evidence convincing. Among women the estimated benefit looked even larger (rate ratio 0.67), but with far fewer women enrolled the range was wide (0.38 to 1.14), so that result is best treated as pointing in the right direction rather than settling the question.

From percentages to people#

A relative reduction sounds the same whether the underlying danger is large or tiny, which is exactly why it can mislead. The number needed to screen (NNS) fixes that by asking a concrete question: how many people have to enter the program for one of them to be spared a lung-cancer death?

In NLST that figure was about 320 over roughly three rounds of scanning and follow-up, which stacks up well against other accepted screening programs. But 320 is an average across a group that was already selected for high risk, and the average hides a wide spread. Among the participants at the very highest baseline risk, far fewer needed to be screened to save one life; among those at the lowest risk in the trial, the number climbed into the thousands. The eligibility rules are doing the heavy lifting here, steering the test toward the people for whom the arithmetic pays off.

What screening costs#

Every screening test turns up things that are not cancer, and occasionally things that are technically cancer but would never have mattered. Both carry a price you should see named plainly.

False alarms were the everyday reality of NLST. Roughly one in four CT scans was flagged as positive, and about 96 percent of those positives turned out to be nothing. Most were sorted out with a follow-up scan, but a smaller number led to invasive procedures, which carry their own risk of complications, along with additional radiation from repeat imaging and the stress of an uncertain result.

Overdiagnosis is subtler and harder to pin down. It means detecting a genuine cancer that would never have produced symptoms in the person's remaining lifetime. One analysis of NLST estimated that roughly 18 percent of CT-detected cancers were overdiagnosed, though longer follow-up suggested that early number was too high, because some tumors that looked harmless eventually declared themselves. What you should take from it is simply that the fraction is not zero: a share of screen-detected cancers lead to treatment that offers no benefit and can cause harm.

Who the evidence actually points to#

The trials define the people who benefit by their risk, not by their interest in a scan. In 2021 the U.S. Preventive Services Task Force gave a Grade B recommendation for yearly low-dose CT in adults aged 50 to 80 with a 20 pack-year history who currently smoke or quit within the past 15 years, widening the group it had defined in 2013 by lowering both the age and the pack-year thresholds.

Two practical points follow. If you are at low risk, a light or lifelong non-smoker outside these criteria, screening brings you the false alarms and the overdiagnosis without the survival advantage that justified them. And screening never substitutes for stopping smoking, which remains by far the largest single influence on lung-cancer death. The honest bottom line: low-dose CT earns its place for a specific high-risk group, and a good decision weighs your roughly 1-in-320 chance of preventing a lung-cancer death against a strong likelihood of at least one false alarm along the way.