NELSON Results Guide CT Lung Cancer Screening

Computed tomography screening trial results demonstrate how best to improve lung cancer detection

medwireNews: The Lancet Oncology has published two reports from the NELSON trial detailing the significance of screening interval and node volume in low-dose computed tomography (CT) lung cancer screening.

“[T]hese two papers clarify important issues and add to the overall favourable outcomes emerging with this screening approach”, say James Mulshine, from Rush University in Chicago, Illinois, USA, and Claudia Henschke, from Icahn School of Medicine at Mount Sinai, New York, USA, in an accompanying comment.

“However, these analyses also underscore how much more work is needed to standardise methodology and optimise results.”

The NELSON trial includes 15,822 individuals aged 50 to 75 years at high risk for lung cancer after smoking at least 15 cigarettes per day for more than 25 years or at least 10 cigarettes per day for more than 30 years. The participants were randomly assigned to undergo CT screening at year 1, 2 and 4 of the study or to receive no screening.

The first of the two prespecified analyses included data for 7155 screened participants who were followed up for a median of 8.16 years, during which time 187 (3%) patients were diagnosed with 196 screen-detected cancers.

In addition, 34 (<1%) patients were diagnosed with a further 35 interval lung cancers, including 19 patients in the year after screening and 15 patients in the second year after CT, report Nanda Horeweg, from Erasmus University Medical Center in Rotterdam, the Netherlands, and co-authors.

When the results of the first three screening rounds were pooled with a 2-year follow-up, CT screening was 84.6% sensitive and 98.6% specific for lung cancer detection, with positive and negative predictive values of 40.4% and 99.8%, respectively.

Review of the last CT records revealed that interval cancers were not visible in 35% of affected patients and missed in the remainder, mostly due to radiological or interpretation errors.

Interval cancers were more likely to be diagnosed at stage III or IV than screen-detected tumours (83 vs 22%), were more likely to be small-cell carcinomas (20 vs 4%) and less likely to be adenocarcinomas (26 vs 52%), the researchers add.

“Lung cancer screening in the NELSON trial yielded high specificity and sensitivity, with only a small number of interval cancers”, the researchers conclude. “The results of this study could be used to improve screening algorithms, and reduce the number of missed cancers.”

The second study, also led by Nanda Horeweg, tested the ability of CT to differentiate between pulmonary nodules and lung cancer using data for 7155 screened patients with 9681 non-calcified nodules.

Analysis revealed that the probability of lung cancer within 2 years of CT was just 0.6% in patients with a nodule volume below 100 mm3 and 0.4% in patients with a maximum nodule diameter below 5 mm – these values did not significantly differ from the 0.4% risk of lung cancer in patients without nodules.

Patients with a nodule volume of 100 to 300 mm3 or a diameter of 5 to 10 mm had an intermediate risk of lung cancer, at 2.4% and 1.3%, respectively, and the probability of lung cancer was high in patients with a nodule volume above 300 mm3 or a diameter greater than 10 mm, at 16.9% and 15.2%, respectively.

Among patients whose largest nodule was between 50 and 500 mm3, the 2-year probability of lung cancer was no more than 1% for those with slow-growing nodules (volume doubling time ≥600 days) or nodules that were stable, shrunken or resolved at follow-up.

But the likelihood of lung cancer significantly increased to 4.0% for patients whose nodule doubling time was 400 to 600 days and to 9.9% for patients with a doubling time of up to 400 days.

Thus, immediate diagnostic follow-up is recommended for patients with nodule volume indicating a high risk, while those with intermediate-volume nodules should undergo volume doubling analysis with further follow-up for those with a doubling time of less than 600 days, the authors conclude.

“These results imply that use of lung cancer probability-based thresholds for nodule size and growth and volumetry in nodule management protocols can improve lung cancer detection, and reduce unnecessary follow-up CTs, invasive diagnostic procedures, and costs.”


Mulshine J, Henschke C. Lung cancer screening: achieving more by intervening less. Lancet Oncol 2014; Early Online Publication 2 October. doi:10.1016/S1470-2045(14)70418-8

Horeweg N, Scholten E, de Jong P, et al. Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers. Lancet Oncol 2014; Early Online Publication 2 October. doi:10.1016/S1470-2045(14)70387-0

Horeweg N, van Rosmalen , Heuvelmans M, et al. Lung cancer probability in patients with CT-detected pulmonary nodules: a prespecified analysis of data from the NELSON trial of low-dose CT screening. Lancet Oncol 2014; Early Online Publication 2 October. doi:10.1016/S1470-2045(14)70389-4

medwireNews ( is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2014