American Journal of Respiratory and Critical Care Medicine

Rationale: As of April 2015, participants in the Danish Lung Cancer Screening Trial had been followed for at least 5 years since their last screening.

Objectives: Mortality, causes of death, and lung cancer findings are reported to explore the effect of computed tomography (CT) screening.

Methods: A total of 4,104 participants aged 50–70 years at the time of inclusion and with a minimum 20 pack-years of smoking were randomized to have five annual low-dose CT scans (study group) or no screening (control group).

Measurements and Main Results: Follow-up information regarding date and cause of death, lung cancer diagnosis, cancer stage, and histology was obtained from national registries. No differences between the two groups in lung cancer mortality (hazard ratio, 1.03; 95% confidence interval, 0.66–1.6; P = 0.888) or all-cause mortality (hazard ratio, 1.02; 95% confidence interval, 0.82–1.27; P = 0.867) were observed. More cancers were found in the screening group than in the no-screening group (100 vs. 53, respectively; P < 0.001), particularly adenocarcinomas (58 vs. 18, respectively; P < 0.001). More early-stage cancers (stages I and II, 54 vs. 10, respectively; P < 0.001) and stage IIIa cancers (15 vs. 3, respectively; P = 0.009) were found in the screening group than in the control group. Stage IV cancers were nonsignificantly more frequent in the control group than in the screening group (32 vs. 23, respectively; P = 0.278). For the highest-stage cancers (T4N3M1, 21 vs. 8, respectively; P = 0.025), this difference was statistically significant, indicating an absolute stage shift. Older participants, those with chronic obstructive pulmonary disease, and those with more than 35 pack-years of smoking had a significantly increased risk of death due to lung cancer, with nonsignificantly fewer deaths in the screening group.

Conclusions: No statistically significant effects of CT screening on lung cancer mortality were found, but the results of post hoc high-risk subgroup analyses showed nonsignificant trends that seem to be in good agreement with the results of the National Lung Screening Trial.

Clinical trial registered with www.clinicaltrials.gov (NCT00496977).

Scientific Knowledge on the Subject

The U.S. NLST (National Lung Screening Trial) showed a 20% reduction in lung cancer mortality with three rounds of low-dose computed tomography screening versus plain chest radiography. Included participants were 55–74 years old and had a smoking history of 30+ pack-years. Thus, solid evidence for the effect of screening on lung cancer mortality for this high-risk group exists. The DLCST (Danish Lung Cancer Screening Trial) included participants of younger age (50–70 yr), 20+ pack-years, and >30% predicted FEV1, thus, a lower-risk population compared to the NLST.

What This Study Adds to the Field

The limited statistical power of this trial does not allow a conclusive statement about the efficacy of lung cancer screening. Expanding the screening to include subjects down to 50 years of age and a smoking history of 20 pack-years appears nonrecommendable, as these inclusion criteria resulted in no difference in mortality. However, post hoc analyses show interesting trends. Analysis of a high-risk subgroup (with chronic obstructive pulmonary disease and 35+ pack-years) within the DLCST shows significantly increased risk of death from lung cancer for this subgroup and a nonsignificant 20% lower hazard ratio for lung cancer mortality in the screening group compared to the control group. These trends seem in good agreement with the NLST results.

Lung cancer is not only an aggressive disease but also a rather common one worldwide, resulting in more deaths than any other cancer type (1). Thus, screening and intervening in curable, early-stage disease has been explored for decades with the aim of reducing mortality.

Previous lung cancer screening studies have found that significantly more early-stage cancers are detected with low-dose computed tomography (CT) screening (25). If this method is not accompanied by a reduction in advanced-stage cancers (i.e., an absolute stage shift), however, it is of no benefit. It may lead to overdiagnosis (i.e., detection of indolent cancers that would never have become symptomatic had they been left unfound, because of no or slow growth, or because death due to another cause interceded) (6, 7).

The NLST (National Lung Screening Trial) was a large, randomized controlled trial conducted in the United States. It included more than 50,000 participants between the ages of 55 and 74 years who had a smoking history of at least 30 pack-years. The results from the NLST showed an absolute stage shift, a 20% reduction in lung cancer mortality, and a 6.7% decrease in all-cause mortality with three rounds of low-dose CT screening versus plain chest radiography. However, the results also revealed a worrisomely large number of false-positive screens; the rate of positive screening tests was 24% in the CT screening group, and 96% of these turned out to be false-positive results (8).

Identifying high-risk individuals is crucial for the outcome of a screen because a high a priori risk of lung cancer in a particular cohort will increase the chances of finding early-stage malignancy and reducing lung cancer mortality with screening. Thus, the identification of risk factors and the selection of inclusion criteria are important for assessing outcomes in a lung cancer screening trial (911). Apart from smoking, age is a well-documented risk factor for lung cancer (12), and age at time of inclusion is probably a strong determinant of outcome in a screening study. Also, chronic obstructive pulmonary disease (COPD) and other lung diseases have been shown to be linked to an increased risk of lung cancer (13). Therefore, including lung function testing in a screening selection process may also affect the outcome of a screening. However, because COPD in itself reduces life expectancy and may prevent surgical treatment, the effect of lung cancer screening may depend in a more complex (nonlinear) way on COPD severity.

In this article, we report the follow-up of the DLCST (Danish Lung Cancer Screening Trial), with a focus on mortality and lung cancer findings. We hypothesize that careful selection of high-risk individuals is important for the outcome of lung cancer screening, and we consider that balancing benefits with harms, such as false-positive findings and overdiagnosis, can provide focus for high-risk profiling in participant screening. Participants in the DLCST were younger, had less smoking history, and possibly had better lung function than those in the NLST. Post hoc analyses of the effects of age, amount of smoking, and COPD at baseline on mortality among patients with lung cancer in the two randomization groups of the DLCST were explored to assess whether limiting lung cancer screening to a higher-risk group could have improved the effect of screening.

Some of the results of these studies have been reported previously in the form of an abstract (14).

Study Design

The overall design, baseline, and end-of-screening results of the DLCST have been published previously (4, 15, 16). The DLCST is a prospective, randomized screening trial comparing annual low-dose CT screening with no screening. The screening group had five annual low-dose chest CT scans divided into one baseline round and four incidence rounds. All participants, regardless of randomization group, had annual visits to the screening clinic in which spirometry and questionnaires concerning health, lifestyle, smoking habits, and psychosocial consequences of screening were completed.

The DLCST was approved by the ethics committee of Copenhagen County and fully funded by the Danish Ministry of Interior and Health. Approval of data management in the trial was obtained from the Danish Data Protection Agency. The trial is registered with ClinicalTrials.gov (NCT00496977). All participants provided written informed consent.

Participants

Enrollment of a total of 4,104 participants occurred from October 2004 to March 2006. Included were men and women aged 50–70 years who were current or former smokers with a minimum smoking history of 20 pack-years. If participants were former smokers, they had to have quit after the age of 50 years and within the previous 10 years. Lung function was measured by spirometry, and FEV1 had to be at least 30% of the predicted value. Participants had to be able to climb two flights of stairs (total of 36 steps) without pausing. Exclusion criteria were weight greater than 130 kg, history of cancer diagnosis and treatment, lung tuberculosis, shortened life expectancy less than 10 years, and chest CT screening during the past year for any reason.

After inclusion, participants were randomized to the screening group (n = 2,052) or the control group (no screening; n = 2,052). Participants were randomized by use of an in-house computer program developed by Asger Dirksen, M.D., D.Msc. (random permuted blocks of 10 participants) to either annual screening by low-dose computed tomography (the screening group) or the control group, which was not offered CT screening.

Procedures

All CT scans were performed at the same location on the same multislice CT scanner (16-slice Philips Mx 8000; Philips Healthcare, Best, the Netherlands). Scans were taken with the patient supine after full inspiration using a low-dose technique (120 kV and 40 mA) with the following specifications: section collimation, 16 × 0.75 mm, pitch 1.5, and rotation time 0.5 seconds, resulting in an effective dose of around 1 mSv. CT scans were read by two experienced chest radiologists, and the location, size, and morphology of nodules were registered.

Spirometry was performed according to the recommendations of the European Respiratory Society using a computerized system (Spirotrac IV; Vitalograph, Maids Moreton, UK) without the use of a bronchodilator before the test.

Follow-up

During the course of the screening study, participants were followed annually at study visits in which information regarding image review and possible diagnostic workup and treatment of nodules was collected. The results of these screenings have been published previously (4).

In April 2015 (5 years after the last screen), inquiries were made to the Danish Lung Cancer Registry, the Danish Cancer Registry, the Danish Causes of Death Registry, and the Danish Pathology Registry. In addition, mortality information was updated from the Danish Civil Registration System, where recent deaths are registered with a delay of only a few days. This data would not be available from the Danish Death Cause Registry, owing to lag time of up to 2 years. Therefore, for these recent deaths, the medical history, and cause of death were sought and obtained from medical records of hospitals, autopsy reports, and general practitioners. Final conclusions regarding cause of death were established by the local review board. To avoid bias, affiliations with randomization groups were not attached to these data until after causes of death were determined. Death cause was categorized according to the International Classification of Diseases, 10th Revision, classification and stage of lung cancer at time of diagnosis according to TNM classification (17).

Participants who emigrated from Denmark were lost to follow-up. In this study, the latest date of follow-up was April 7, 2015.

Outcomes

The primary outcome was assessment of lung cancer mortality and all-cause mortality in the two groups. Secondary outcomes were lung cancer diagnoses, survival, stages, and histology. Post hoc analyses of the effects of age, amount of smoking, and COPD at baseline on mortality of lung cancer were conducted.

Statistical Analysis

For frequency comparisons between the two randomization groups, the χ2 test and Fisher’s exact test were used. A two-sample t test was used to compare follow-up years in the two groups. Kaplan–Meier plots and log-rank tests were used to assess mortality and lung cancer diagnoses with 95% confidence intervals (CIs). P values less than 0.05 were considered statistically significant. Risk of lung cancer death was assessed using the Cox proportional hazards model. Hazard ratios (HRs) were calculated by log-rank test and adjusted only for covariates in the Cox proportional hazards model. Survival was calculated from the time of randomization (which was also the date of inclusion and baseline visit).

A power calculation showed that with a significance level of 5%, a sample size of 2 × 2,000 participants, an annual lung cancer mortality of 0.5%, 10 years of observation since randomization, 95% compliance in the screening group, 5% contamination in the control group, and a difference in mortality between the two groups of at least 20%, the power that the study would need to detect a difference was estimated to be 30%.

R version 3.1.3 software was used for statistical analysis.

Figure 1 shows the flowchart of the DLCST. A total of 4,104 participants were included, comprising 2,052 participants in each of the two randomization groups. The mean annual participation rates were 95.5 and 93.0% in the screening group and the control group, respectively, and the difference in participation rate was statistically significant in the second and third incidence rounds only (4).

In the control group, 153 participants had 199 chest CT scans outside the trial during the 4-year trial duration. Further details concerning the topic of contamination have been published previously (16).

By April 2015, 34 participants had been lost to follow-up due to emigration (20 from the CT group and 14 from the control group). Total person-years of follow-up were 19,439 person-years (mean, 9.47; median, 9.80) in the screening group and 19,547 person-years (mean, 9.53; median, 9.80) in the control group (P = 0.233).

Table 1 shows selected baseline characteristics of the study population. Lung function was missing for one participant in the screening group. No statistically significant differences regarding age, sex, current smoking status, or mean number of smoking pack-years were observed. However, there were significantly more participants with more than 35 pack-years of smoking in the screening group (P = 0.047). Regarding lung function, the lungs of participants in the screening group were significantly more obstructed than those in the control group (P = 0.029 for FEV1/FVC).

Table 1. Selected Baseline Characteristics of the DLCST Study Participants

CharacteristicsScreening Group (n = 2,052)Control Group (n = 2,052)P Value
Age at randomization, yr57.9 ± 4.857.8 ± 4.80.794
Male sex1,147 (56)1,120 (55)0.414
Current smokers1,545 (75.3)1,579 (76.9)0.227
Pack-years36.4 ± 13.435.9 ± 13.40.210
Participants with >30 pack-years1,278 (62.3)1,240 (60.4)0.236
Participants with >35 pack-years929 (45.3)865 (42.2)0.047
Lung function   
 FEV1, % predicted93.5 ± 16.994.3 ± 17.30.128
 FEV1/FVC0.70 ± 0.080.71 ± 0.080.029
 Participants with FEV1/FVC <0.7894 (43.6)856 (41.7)0.237

Definition of abbreviation: DLCST = Danish Lung Cancer Screening Trial.

Data presented are mean ± SD or n (%). Boldface type indicates statistically significant values.

By April 2015, 328 of the 4,104 participants had died. Of these, 165 were in the screening group and 163 were in the control group (all-cause mortality HR, 1.02; 95% CI, 0.82–1.27; P = 0.867) (Figure 2).

In the screening group, 39 died due to lung cancer (2.0 cases per 1,000 person-years), whereas 38 died as a result of lung cancer (1.9 cases per 1,000 person-years) in the control group (HR, 1.03; 95% CI, 0.66–1.6; P = 0.888 by log-rank test for lung cancer mortality) (Figure 2). Thus, no differences in mortality were observed. Among participants who died due to lung cancer, 49 had COPD and 28 did not, which was different from the rate of COPD in the overall trial (P < 0.001).

During five screening rounds, 1,029 (560 baseline and 469 incidence) noncalcified nodules were registered in 611 participants in the screening group. In the DLCST, a total of 153 cases of lung cancer were diagnosed in 149 participants; stages are summarized in Table 2. Significantly more cancers were detected in the screening group (n = 100; 5.1 cases per 1,000 person-years) than in the control group (n = 53; 2.7 cases per 1,000 person-years). Kaplan–Meier curves of lung cancer diagnoses are shown in Figure 3 (lung cancer diagnosis HR, 1.8; 95% CI, 1.3–2.6; P < 0.001 by log-rank test). In the screening group, the 100 cancers were detected in 96 persons. Of these, 68 cancers were screen detected and 32 were diagnosed after the end of screening. Of participants who developed lung cancer, 88 had COPD and 61 did not.

Table 2. Lung Cancer Disease Stages in the DLCST

TNM/StageAll (N = 4,104)Screening Group (n = 2,052)Control Group (n = 2,052)P Value
T1N0M047416<0.001
T4N3M1298210.025
Stage I58508<0.001
Stage II6420.687
Stage IIIa181530.009
Stage IIIb14860.789
Stage IV5523320.278
Unknown stage2020.500
Total15310053<0.001

Definition of abbreviations: DLCST = Danish Lung Cancer Screening Trial; TNM = tumor, node, metastasis.

Significantly more early-stage cancers (stages I and II) were detected in the screening group than in the control group (54 vs. 10, respectively; P < 0.001), and this was due mainly to a higher number of cancers in a very early stage (T1N0M0) in the screening group than in the control group (41 vs. 6, respectively; P < 0.001). Also, stage IIIa cancers were more frequent in the screening group than in the control group (15 vs. 3, respectively; P = 0.009). More stage IV cancers were found in the control group than in the screening group, although this was not statistically significant (32 vs. 23; P = 0.278), and the highest stage (T4N3M1) was significantly less frequent in the screening group than in the control group (8 vs. 21, respectively; P = 0.025).

Table 3 shows the histology of the lung cancers diagnosed in the DLCST. Significantly more lung adenocarcinomas (ACLs) were diagnosed in the screening group than in the control group (58 vs. 18, respectively; P < 0.001). There were no significant differences between the screening and control groups for other types of histology (squamous cell carcinoma, 15 vs. 9, respectively; P = 0.306; small cell lung cancer, 11 vs. 14, respectively; P = 0.688; non–small cell cancer, 14 vs. 13, respectively; P = 1).

Table 3. Histology of Lung Cancers in the DLCST

HistologyAll (N = 153)Screening Group (n = 100)Control Group (n = 53)
ACL58 (38)40 (40)18 (34)
ACL + BAC17 (11)17 (17)0 (0.0)
ACL + SQC1 (0.7)1 (1.0)0 (0.0)
BAC1 (0.7)1 (1.0)0 (0.0)
SQC23 (15)14 (14)9 (17)
NSCLC23 (15)14 (14)9 (17)
NSCLC + BAC1 (0.7)0 (0.0)1 (1.9)
SCLC + NSCLC3 (2.0)0 (0.0)3 (5.7)
SCLC22 (14)11 (11)11 (21)
Large-cell neuroendocrine carcinoma1 (0.7)1 (1.0)0 (0.0)
Carcinoid1 (0.7)0 (0.0)1 (1.9)
Unknown histology*2 (1.3)1 (1.0)1 (1.9)

Definition of abbreviations: ACL = adenocarcinoma; BAC = bronchioloalveolar carcinoma; DLCST =  Danish Lung Cancer Screening Trial; NSCLC = non–small cell lung cancer; SCLC = small cell lung cancer; SQC = squamous cell carcinoma.

Data presented are n (%).

*The two patients with unknown histology had widespread cancer and large lung tumors identified by computed tomography at the time of diagnosis. Because these patients had no prospect of survival, no histology was done.

Causes of death are shown in Table 4. Cancer was the most frequent cause with 181 deaths (55%), 77 (43%) of which were due to lung cancer.

Table 4. Causes of Death in the DLCST

Cause of DeathAll (N = 328)Screening Group (n = 165)Control Group (n = 163)
Cancer type   
 Lung77 (23)39 (24)38 (23)
 Pancreatic22 (6.7)9 (5.5)13 (8.0)
 Cerebral9 (2.7)5 (3.0)4 (2.5)
 Liver or biliary7 (2.1)3 (1.8)4 (2.5)
 Esophagus7 (2.1)4 (2.4)3 (1.8)
 Colon or rectal7 (2.1)5 (3.0)2 (1.2)
 Bladder7 (2.1)2 (1.2)5 (3.1)
 Prostate6 (1.8)3 (1.8)3 (1.8)
 Gastric5 (1.5)4 (2.4)1 (0.6)
 Other types*34 (10)18 (11)16 (9.8)
Ischemic heart disease22 (6.7)12 (7.3)10 (6.1)
Stroke16 (4.9)5 (3.0)11 (6.7)
COPD15 (4.6)7 (2.4)8 (4.9)
Alcohol addiction12 (3.7)3 (1.8)9 (5.5)
Alcoholic liver cirrhosis9 (2.7)5 (3.0)4 (2.5)
Aortic aneurism8 (2.4)4 (2.4)4 (2.5)
Sepsis5 (1.5)3 (1.8)2 (1.2)
Other50 (15)26 (16)24 (15)
Unknown10 (3.0)8 (4.8)2 (1.2)

Definition of abbreviations: COPD = chronic obstructive pulmonary disease; DLCST =  Danish Lung Cancer Screening Trial.

Data presented are n (%).

*Other types of cancer include fewer than five participants each and comprise breast cancer, sarcoma, malignant melanoma, leukemia, lymphoma, carcinoid cancer, tonsil cancer, oral cancer, and others.

Other causes of death include fewer than five participants each and comprise amyotrophic lateral sclerosis, heart failure, suicide, diabetes mellitus with complications, HIV, gastrointestinal hemorrhage, necrotic fasciitis, and others.

Table 5 and Figure 4 show the results of a Cox proportional hazards model with death due to lung cancer as the outcome and sex, age, COPD (FEV1/FVC <0.7), and more than 35 pack-years of smoking as explanatory risk factors. Age, but not sex, was associated with increased risk of death. With participants without COPD and less than 35 pack-years of smoking as references, participants with both COPD and more than 35 pack-years of smoking had a considerably increased risk of death, with HRs of 5.2 (P = 0.003) and 6.8 (P = 0.001) in the screening and control groups, respectively (Figure 4 and Table 5). The data in Table 5 make clear that only 13% (10 of 77) of the lung cancer deaths occurred in low-risk participants (without COPD and with <35 pack-years of smoking), even though they constituted the largest subgroup, and 48% (37 of 77) of lung cancer deaths occurred in the high-risk subgroup with both COPD and more than 35 pack-years of smoking.

Table 5. Results of Cox Proportional Hazards Model

Risk Factor/SubgroupNumber of SubjectsDeaths Due to Lung CancerHazard Ratio95% Confidence IntervalP Values*
Female sex1,837331.20.7–1.8P = 0.520
Age at baseline (per additional 10 yr)4,104772.41.5 –3.9P<0.001
No COPDPack-years <35Control70541.0ReferenceP = 0.420
Screening64061.70.5–6.0P = 0.420
Pack-years ≥35Control49182.60.8–8.6P = 0.125P = 0.625
Screening517103.21.0–10.4P=0.048
COPDPack-years <35Control39562.50.7–8.8P = 0.162P = 0.924
Screening40862.30.7–8.3P = 0.190
Pack-years ≥35Control461206.82.3–20.0P<0.001P = 0.425
Screening486175.21.7–15.6P=0.003

Definition of abbreviation: COPD = chronic obstructive pulmonary disease.

*In the P values columns, the left column relates directly to the hazard ratios (and 95% confidence intervals) and the right column gives P values for comparison of hazard ratios of the screening and control subgroups of the four strata defined by COPD and pack-years. Boldface type indicates statistically significant values.

There were no differences between the two groups in lung cancer mortality or all-cause mortality.

As of April 2015 (5 years after the last screen in the DLCST), almost twice as many lung cancers had been diagnosed in the screening group compared with the control group. These were mainly early-stage ACLs, and there was no significant difference in the number of high-stage cancers (stages III and IV) between the two groups, although the highest stage (T4N3M1) was significantly less common in the screening group. In other words, we observed a relative stage shift, and an absolute stage shift was observed for the highest stage (T4N3M1) only. The NLST investigators found a more profound stage shift that was statistically significant for stage IV as a whole, and not only for the worst subgroup (T4N3M1) of stage IV as in the DLCST.

Because the excess cancers in the screening group were primarily ACLs, we suggest that ACLs could more often have a slow volume doubling time, remain indolent for years, and thus be overdiagnosed. These considerations are in keeping with other studies (18, 19). Previous screening studies have shown that participants with normal lung function had a twofold higher prevalence of lung cancers with a longer volume doubling time compared with participants with COPD (20, 21). Hence, limiting lung cancer screening to high-risk individuals with obstructive lung function may reduce overdiagnosis.

DLCST Study Size and Participant Characteristics

The statistical power calculation and study size estimation of the DLCST was based on an annual incidence of lung cancer of 0.50% and a difference in lung cancer mortality between the two groups of at least 20%. The annual incidence of lung cancer in our control group was only half what we expected (i.e., 0.27% instead of 0.50%); thus, the study was underpowered, and the results of the DLCST do not allow us to make definitive conclusions about the efficacy of screening.

Before initiation of the DLCST in 2004, the possibility of future pooling of data with the larger NELSON (Dutch-Belgian Randomized Lung Cancer Screening Trial; ISRCTN63545820) was included in the protocol, with the aim of improving statistical power. However, the NELSON investigators subsequently prolonged their screening intervals from 1 year to 2 and then 2.5 years, which has protracted the screening activity of the NELSON trial and rendered the two studies less comparable. Therefore, when we finished screening in 2010 (4), we decided to report DLCST lung cancer and mortality data 5 years after the last CT screen, independent of the NELSON study.

On the basis of the NLST, we know that with the NLST inclusion criteria (55–74 years of age and a minimum smoking history of 30 pack-years), the difference in lung cancer mortality was in fact 20% (8). The reason that we did not see a difference in mortality may be small numbers; however, it may also be due to differences in inclusion criteria. The risk of lung cancer was probably lower in the DLCST cohort because participants were younger (50–70 years of age), had smoked less (lower limit was 20 pack-years), and thus possibly had better lung function (FEV1 >30% predicted) than the subjects in the NLST. In the Italian DANTE study, the median smoking exposure was 45.0 pack-years, which is 10 pack-years more than the median number of pack-years in the DLCST (22). The DLCST is a small screening study, and unfortunately by chance—and despite randomization—the screening group included more heavy smokers (>35 pack-years), as well as more participants with greater airway obstruction (FEV1/FVC), than the control group, implying shorter life expectancy for the screening group. This may explain the initial higher mortality in the screening group (Figure 2). Interestingly, this difference had decreased after about 8 years of follow-up, and after 10 years the curves crossed. Taking into account the significant results of the much larger NLST, the latter observation could be interpreted as the beneficial effect of screening seen in the NLST. In addition, the post hoc subgroup results of the DLCST (Figure 4) suggest limiting screening to high-risk individuals because the harms of screening seem to outweigh the benefit in subjects with a more moderate risk of lung cancer.

High-Risk Profiling

Lung cancer risk prediction studies have shown that older age, heavy smoking history, COPD (airway obstruction as well as emphysema), and interstitial lung disease are associated with an increased risk of lung cancer (9, 1113, 23). Indeed, in our analysis of higher-risk subgroups (obstructive lung function, more than 35 pack-years of smoking, and older age at baseline) within the DLCST, we consistently found a significantly increased risk of death due to lung cancer. Thus, for the subgroup with both COPD and more than 35 pack-years of smoking, the risk of death due to lung cancer was two to six times higher than in the other subgroups, and there was a nonsignificant lower HR for lung cancer mortality in the screening group compared with the control group (HRs, 5.2 vs. 6.8, respectively; P = 0.425). This finding may imply a beneficial effect of screening for this particular subgroup, which is in good agreement with the screening effect found in the NLST. However, the DLCST is underpowered, and, by limiting the analysis to higher-risk participants, the numbers become even smaller and, not surprisingly, differences in lung cancer mortality were not statistically significant. Nevertheless, the subgroup analyses seem to indicate that high-risk profiling is important for assessing outcomes. Therefore, we believe early risk stratification is of crucial importance when designing a screening protocol. The results of the DLCST are in keeping with a study within the NLST which showed that screening was most beneficial in the highest quintiles of risk (risk defined by age, body mass index, family history of lung cancer, pack-years of smoking, years since smoking cessation, and emphysema diagnosis) and that the cost-to-benefit ratio is best in the highest-risk group with a decreasing number of participants who would need to be screened to prevent one lung cancer death and with fewer false-positive screens (24).

False-Positive Screens

In the DLCST, diagnostic false-positive rates were 7.9% in the baseline round and 1.7, 2.0, 1.6, and 1.9% in four subsequent screening rounds, respectively (4). Thus, compared with the NLST, false-positive screens were substantially less common, and only one interval cancer was observed in the DLCST. This raises the question whether some of the high-stage cancers detected in the DLCST were a result of placing too few nodules in the “positive screen group” and waiting too long before resecting these nodules. Perhaps the price of a low false-positive rate is higher-stage cancers, and this could reduce or even eliminate the effect on mortality. Although many false-positive screens are neither ethically nor economically acceptable, there may be a trade-off between a high effect of screening on mortality and a low false-positive rate. Extra follow-up CT scans involve patient anxiety, additional radiation, and invasive procedures that are not without risk. We believe risk prediction tools for nodule evaluation could prove helpful in the diagnostic workup of nodules detected in screening, and the PanCan lung cancer risk prediction model performed very well in our cohort, in which areas under the receiver operating characteristic curves were 0.83–0.87 (25).

General Improvements in Cancer Survival in Denmark

In Denmark, several improvements in oncological treatments have been implemented during the last 15 years, and 5-year survival for lung cancer increased from 12 to 15% between 2000 and 2012. Availability of radiation therapy has increased, and faster diagnoses and treatments have been implemented (e.g., so-called cancer packages that reduce time from referral to treatment in the Danish health care system) (26). Palliative oncological treatment provided for patients with lung cancer has improved. Longer follow-up may be needed to see a potential effect of lung cancer screening on mortality.

Conclusions

The DLCST was statistically underpowered, and the results of the DLCST do not support a recommendation of lung cancer screening of current or former smokers as young as 50 years of age and with a smoking history of only 20 pack-years. Overdiagnosis could be a substantial problem associated with lung cancer screening in a population with these characteristics. There were both relatively few false-positive screens and few interval cancers in the DLCST. The randomization by chance in the DLCST was less favorable for the screening group, which came out with more airway obstruction and more pack-years of smoking. This may have reduced the effect of screening. Nevertheless, our results support risk stratification with a focus on age, smoking history, and obstructive lung disease when selecting candidates for lung cancer screening. In a high-risk subgroup, we observed trends in good agreement with the positive results of the NLST regarding stage shift and mortality.

1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:6990.
2. Henschke CI, McCauley DI, Yankelevitz DF, Naidich DP, McGuinness G, Miettinen OS, Libby DM, Pasmantier MW, Koizumi J, Altorki NK, et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999;354:99105.
3. Infante M, Cavuto S, Lutman FR, Brambilla G, Chiesa G, Ceresoli G, Passera E, Angeli E, Chiarenza M, Aranzulla G, et al.; DANTE Study Group. A randomized study of lung cancer screening with spiral computed tomography: three-year results from the DANTE trial. Am J Respir Crit Care Med 2009;180:445453.
4. Saghir Z, Dirksen A, Ashraf H, Bach KS, Brodersen J, Clementsen PF, Døssing M, Hansen H, Kofoed KF, Larsen KR, et al. CT screening for lung cancer brings forward early disease. The randomised Danish Lung Cancer Screening Trial: status after five annual screening rounds with low-dose CT. Thorax 2012;67:296301.
5. Swensen SJ, Jett JR, Hartman TE, Midthun DE, Mandrekar SJ, Hillman SL, Sykes AM, Aughenbaugh GL, Bungum AO, Allen KL. CT screening for lung cancer: five-year prospective experience. Radiology 2005;235:259265.
6. Humphrey LL, Deffebach M, Pappas M, Baumann C, Artis K, Mitchell JP, Zakher B, Fu R, Slatore CG. Screening for lung cancer with low-dose computed tomography: a systematic review to update the US Preventive Services Task Force recommendation. Ann Intern Med 2013;159:411420.
7. Marcus PM, Prorok PC, Miller AB, DeVoto EJ, Kramer BS. Conceptualizing overdiagnosis in cancer screening. J Natl Cancer Inst 2015;107:djv014.
8. Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM, Gareen IF, Gatsonis C, Marcus PM, Sicks JD; National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395409.
9. Cassidy A, Duffy SW, Myles JP, Liloglou T, Field JK. Lung cancer risk prediction: a tool for early detection. Int J Cancer 2007;120:16.
10. Spitz MR, Etzel CJ, Dong Q, Amos CI, Wei Q, Wu X, Hong WK. An expanded risk prediction model for lung cancer. Cancer Prev Res (Phila) 2008;1:250254.
11. Tammemägi MC, Katki HA, Hocking WG, Church TR, Caporaso N, Kvale PA, Chaturvedi AK, Silvestri GA, Riley TL, Commins J, et al. Selection criteria for lung-cancer screening. N Engl J Med 2013;368:728736.
12. Bach PB, Kattan MW, Thornquist MD, Kris MG, Tate RC, Barnett MJ, Hsieh LJ, Begg CB. Variations in lung cancer risk among smokers. J Natl Cancer Inst 2003;95:470478.
13. Mizuno S, Takiguchi Y, Fujikawa A, Motoori K, Tada Y, Kurosu K, Sekine Y, Yanagawa N, Hiroshima K, Muraoka K, et al. Chronic obstructive pulmonary disease and interstitial lung disease in patients with lung cancer. Respirology 2009;14:377383.
14. Pedersen JH, Winkler Wille MM, Dirksen A. The Danish Lung Cancer Screening Trial: results 5 years after last screening. Presented at the 16th World Conference on Lung Cancer (WCLC). September 6–9, 2015, Denver, CO. Abstract 2384.
15. Pedersen JH, Ashraf H, Dirksen A, Bach K, Hansen H, Toennesen P, Thorsen H, Brodersen J, Skov BG, Døssing M, et al. The Danish randomized lung cancer CT screening trial—overall design and results of the prevalence round. J Thorac Oncol 2009;4:608614.
16. Saghir Z, Ashraf H, Dirksen A, Brodersen J, Pedersen JH. Contamination during 4 years of annual CT screening in the Danish Lung Cancer Screening Trial (DLCST). Lung Cancer 2011;71:323327.
17. Goldstraw P, Crowley J, Chansky K, Giroux DJ, Groome PA, Rami-Porta R, Postmus PE, Rusch V, Sobin L; International Association for the Study of Lung Cancer International Staging Committee; Participating Institutions. The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh) edition of the TNM Classification of malignant tumours. J Thorac Oncol 2007;2:706714.
18. Patz EF Jr, Pinsky P, Gatsonis C, Sicks JD, Kramer BS, Tammemägi MC, Chiles C, Black WC, Aberle DR; NLST Overdiagnosis Manuscript Writing Team. Overdiagnosis in low-dose computed tomography screening for lung cancer. JAMA Intern Med 2014;174:269274.
19. Young RP, Hopkins RJ. Stage shift in computed tomography screening: possible role of indolent cancers, “histology shift,” and overdiagnosis. Am J Respir Crit Care Med 2013;188:10341035.
20. Young RP, Hopkins RJ. Estimating overdiagnosis of lung cancer. Ann Intern Med 2013;158:635.
21. Veronesi G, Maisonneuve P, Bellomi M, Rampinelli C, Durli I, Bertolotti R, Spaggiari L. Estimating overdiagnosis in low-dose computed tomography screening for lung cancer: a cohort study. Ann Intern Med 2012;157:776784.
22. Infante M, Cavuto S, Lutman FR, Passera E, Chiarenza M, Chiesa G, Brambilla G, Angeli E, Aranzulla G, Chiti A, et al.; DANTE Study Group. Long-term follow-up results of the DANTE Trial, a randomized study of lung cancer screening with spiral computed tomography. Am J Respir Crit Care Med 2015;191:11661175.
23. Wille MM, Thomsen LH, Petersen J. de Bruijne M, Dirksen A, Pedersen JH, Shaker SB. Visual assessment of early emphysema and interstitial abnormalities on CT is useful in lung cancer risk analysis. Eur Radiol 2016;26:487494.
24. Kovalchik SA, Tammemagi M, Berg CD, Caporaso NE, Riley TL, Korch M, Silvestri GA, Chaturvedi AK, Katki HA. Targeting of low-dose CT screening according to the risk of lung-cancer death. N Engl J Med 2013;369:245254.
25. Winkler Wille MM, van Riel SJ, Saghir Z, Dirksen A, Pedersen JH, Jacobs C, Thomsen LH, Scholten ET, Skovgaard LT, van Ginneken B. Predictive accuracy of the PanCan lung cancer risk prediction model - external validation based on CT from the Danish Lung Cancer Screening Trial. Eur Radiol 2015;25:30933099.
26. Danish Multidisciplinary Cancer Groups (DMCG); Danish Lung Cancer Group (DLCG). Benchmarking Consortium. Chapter 2: lung cancer 2000–2012 [in Danish; accessed 2015 Oct 29]. Available from: http://dmcg.dk/fileadmin/dmcg.dk/Benchmarking_Consortium/C_Lunge_06_11_2014.pdf
Correspondence and requests for reprints should be addressed to Mathilde M. W. Wille, M.D., Ph.D., Department of Imaging, Radiology Section, Nordsjællands Hospital, Dyrehavevej 29, DK-3400 Hillerød, Denmark. E-mail:

The Danish Lung Cancer Screening Trial was fully funded by the Danish Ministry of Interior and Health.

Author Contributions: M.M.W.W. completed the literature search; contributed to the study design; performed collection, analysis, and interpretation of data; made figures; and wrote the first manuscript draft. A.D. contributed to the literature search; contributed to study design; performed collection, analysis, and interpretation of data; completed figures; assisted in writing the first manuscript draft; and critically reviewed the manuscript. H.A. made a significant contribution to the implementation of the study and critically reviewed the manuscript. Z.S. made significant contributions to the implementation of the study and data collection, analysis, and interpretation and critically reviewed the manuscript. K.S.B. and H.H. made significant contributions to collection of data (assessed all computed tomographic scans) and critically reviewed the manuscript. J.B. made significant contributions to the implementation of the study, evaluated the analysis and interpretation of data, and critically reviewed the manuscript. P.F.C., J.M., and J.F.R. made significant contributions to the implementation of the study and data interpretation and critically reviewed the manuscript. K.R.L. made significant contributions to the study design, implementation of the study, and data interpretation and critically reviewed the manuscript. N.S. made significant contributions to the study design and implementation of the study and critically reviewed the manuscript. B.G.S. made significant contributions to the study design and collection of data and critically reviewed the manuscript. L.H.T. made significant contributions to the data analysis and interpretation and the drafting of the manuscript and critically reviewed the manuscript. P.T. made significant contributions to the implementation of the study and critically reviewed the manuscript. J.H.P. had primary responsibility for the study design; assisted in collection, analysis, and interpretation of data; assisted with figures; and critically reviewed the manuscript.

Originally Published in Press as DOI: 10.1164/rccm.201505-1040OC on October 20, 2015

Author disclosures are available with the text of this article at www.atsjournals.org.

Comments Post a Comment




New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine
193
5

Click to see any corrections or updates and to confirm this is the authentic version of record