Pedersen, Jesper H. MD, DrMSci*; Ashraf, Haseem MD†; Dirksen, Asger DrMSci†; Bach, Karen MD‡; Hansen, Hanne MD‡; Toennesen, Phillip DrMSci†; Thorsen, Hanne MD§; Brodersen, John PhD§; Skov, Birgit Guldhammer DrMSci∥; Døssing, Martin DrMSci¶; Mortensen, Jann DrMSci#; Richter, Klaus PhD**; Clementsen, Paul DrMSci†; Seersholm, Niels DrMSci†
During the last decade the advent of low dose multislice computed tomography (CT) scanning has generated a widespread interest in lung cancer screening.1 Early observational reports stated that lung cancer could be detected in early stages in 85% of the cases2 and subsequently a 10-year survival rate of 88% for stage I lung cancer was estimated after CT screening.3 These reports have generated controversy,4–6 because data from other nonrandomized CT screening trials have predicted that no reduction in lung cancer mortality will be observed after screening and that harmful effects, overdiagnosis and overtreatment may be substantial.4–8 These conflicting results emphasize the need for randomized controlled trials to demonstrate the effects, benefits and harms of CT screening.4 Therefore, randomized trials are now being performed in the United States9,10 and Europe.11–13 In the United States CT screening is compared with plain chest radiography,9 whereas in the European studies CT screening is compared with no screening in the control arm.13,14 The aim of this article is to present the design and results from the prevalence round of the randomized Danish Lung Cancer Screening Trial (DLCST).
MATERIALS AND METHODS
Objective of the Study
Primary objective is to evaluate if annual low dose CT screening can reduce lung cancer mortality by more than 25%. Major secondary end points are; overall mortality in each study arm, numbers of lung cancer in each arm, 5-year survival after diagnosis, stage of lung cancer at diagnosis (stage distribution), surgical resection rate, effect on smoking behavior, frequency of false-positive diagnoses, and psycho-social consequences of these in addition to health economic evaluations of the cost of CT screening in Denmark.
The effect on lung cancer mortality will in the following years be evaluated in collaboration with the NELSON (Nederlands Leuvens Screening Onderzoek) CT screening trial in Belgium and the Netherlands which has included 16,000 individuals.12,14 The two trials combined are expected to have sufficient power to detect a reduction in lung cancer mortality of 25% 10 years after randomization.12,14
Individuals volunteered for the study in response to advertisements in local and regional free newspapers and weeklies that provided information regarding the general outline of the study, the eligibility requirements for the study, and the fact that the study was funded by a Governmental grant and performed in collaboration with the NELSON trial. Participants were men and women who were 50 to 70 years of age and without lung cancer related symptoms. The sample size was calculated on the assumption of a 1:1 randomization, a power of 80%, significance level of 5%, 95% compliance in the screen group, 5% contamination in the control group, and 10 years of follow-up after randomization.12
From October 1, 2004, to March 31, 2006, 4104 participants were enrolled and randomized at the initial visit after receiving both oral and written information about the trial, and signing the informed consent papers. A history of cigarette smoking of at least 20 pack-years was necessary for entrance into the study. Participants had to be current or former smokers, and former smokers had to have quit after the age of 50 years and less than 10 years ago. Participants had to be able to climb 2 flights of stairs (36 steps) without pausing. Spirometry is performed annually, and forced expiratory volume in 1 second had to be at least 30% of predicted normal at baseline to allow entry into the study.
Ineligible were those applicants with body weight above 130 kg or previous treatment for lung cancer, breast cancer, malignant melanoma, or hypernephroma. Individuals with a history of any other cancer within 5 years or tuberculosis within 2 years or any serious illness that would shorten life expectancy to less than 10 years were excluded, as was the case if a prior CT scan of the chest had been performed within the last year.
The trial is performed in one institution: Gentofte University Hospital in Copenhagen, Denmark. Participants were randomized by a computer program (random permuted blocks of 10 participants) to either annual screening by low-dose CT (the screening group) or the control group who were not offered CT screening. Prevalence CT scans were performed within 1 month after randomization. The screening is scheduled to last 5 years, i.e., an initial (prevalence) screening round followed by 4 annual (incidence) screening rounds (Figure 1). Follow-up is planned for 10 years after randomization.
At inclusion information on age, gender, educational background, and occupational status has been collected for each participant together with information on former and present smoking habits including the Danish version of the Fagerstroem scale.15 In addition the participants completed a questionnaire on their psychosocial status, the Consequences Of Screening (COS) questionnaire.16 This core-questionnaire, which has been developed and validated in Denmark contains questions relevant for participants in most cancer screening programs.17After focus group interviews conducted during the prevalence round questions especially relevant for participants in lung cancer screening have been added to the COS. This extended version of the COS, the COS-Lung Cancer will be used during the next years to investigate especially psychosocial consequences of having a false-positive screening result.
Lung Function Test
Spirometry was performed annually on all participants according to recommendations by the European Respiratory Society using a computerized system (Spirotrac IV; Vitalograph, Buckingham, UK), and expressed in absolute values and as percent of predicted values according to European reference equations.18
Imaging and Image Review
All CT scans of the study were performed on a MDCT scanner (16 rows Philips Mx 8000, Philips Medical Systems, Eindhoven, The Netherlands). Scans were performed supine after full inspiration with caudocranial scan direction including the entire ribcage and upper abdomen with a low dose technique, 120kV and 40 mAs. Scans were performed with spiral data acquisition with the following acquisition parameters: Section collimation 16 × 0.75 mm, pitch 1.5, rotation time 0.5 second. The obtained data were reconstructed in two ways and parameters were as follows: Section width 3 mm/1 mm, reconstruction increment 1.5 mm/1 mm, with a soft and hard algorithm, respectively. All image data were stored in DICOM format on MOD and PACS. All scans were read by two board certified radiologists (H.H. and K.S.B.) using cine/slab-viewing supplied with Maximum Intensity Projection, and in case of disagreement consensus was obtained. The location, size, demarcation and shape were registered, and the density classified as solid, semisolid or pure GGO lesions. The evaluation of size was based on linear measurement of the maximal diameter in axial slices. In selected cases volume of nodules were calculated by Philips nodule evaluation semiautomated software.
Classification of Nodules
All nodules were classified into four categories according to size and other characteristics: Nodules up to 15 mm in maximal diameter with benign characteristics (for calcified nodules up to 20 mm) (category 1) and nodules below 5 mm (category 2) were tabulated and no further action taken. Nodules with a diameter between 5 and 15 mm not classified as benign were considered indeterminate and were rescanned after 3 months (category 3). Nodules exceeding 15 mm (category 4) and all growing nodules (category 5) were referred for diagnostic investigation, in addition to nodules with suspicious morphology. After repeat CT scan, nodules were described as regressed, stable or growing by the radiologists. Growth was defined as an increase in volume of at least 25%.
Participants with nodules category 1 and 2 were regarded as screening test negative and those with nodules category 3, 4, or 5 regarded as screening test positive.
Diagnostic Work-Up of Positive Findings
Referral of participants for diagnostic evaluation was decided at weekly follow-up conferences between a pulmonologist (A.D.) and the radiologists. Indeterminate nodules were often evaluated using Fluorodeoxyglucose-positron emission tomography-CT, but will be reported separately. CT with contrast was performed before invasive procedures. Depending on the results of these initial procedures an individual diagnostic plan was made involving a variety of invasive procedures such as bronchoscopy, transthoracic needle aspiration biopsy, endoscopic ultrasound, endobronchial ultrasound and/or mediastinoscopy. In most cases video assisted thoracic surgery (VATS) had to be performed to reach a histologic diagnosis and staging of the disease. All diagnostic workup and treatment of participants with suspicious nodules was taken care of by units specialized in lung cancer and part of the public health service in Copenhagen.
Incidental findings on the CT scan outside the lungs or bronchi judged to be of clinical significance for the participant, was revealed to the participant and referred to relevant work up and treatment. The results are not reported here.
Surveillance of Vital Status and Occurrence of Lung Cancer
The vital status of all participants is checked annually in the Danish Civil Registration System which registers all national deaths within 2 weeks. In case of death, information was obtained from the Danish Causes of Death Register, which is updated with a lag time of 1 to 2 years, and from hospital and autopsy files when possible. Information on occurrence of lung cancer for the whole study group is obtained annually from the Danish Lung Cancer Register which registers over 90% of all lung cancers in Denmark. Both active participants and drop-outs will be followed for 10 years after randomization or until death. An international, independent death review board will be established.
Treatment was performed in two centers specialized in lung cancer treatment in Copenhagen. The indication for surgical or oncological intervention was in all cases decided at multi speciality conferences by board certified specialists in pulmonary medicine, thoracic surgery, pathology, oncology, and radiology. Pulmonary resection was in most cases performed in the same procedure as the diagnostic VATS, following frozen section pathology evaluation of the specimens obtained by wedge resection. Definitive surgical treatment was performed either by VATS or open thoracotomy depending on the preference of the surgeon. Danish national guidelines for lung cancer management were followed.
Resected tumors and specimens were analyzed according to the guidelines for screening-detected lung tumors from the EU-US pathology panel19 by one pathologist (B.G.S.) and all lung cancers were verified by the panel. All tumors were classified according to World Health Organization.20
Ethical and Legal Approval
The DLCST was approved by the Ethical Committee of Copenhagen County on January 31, 2003 and funded in full by the Danish Ministry of Interior and Health on June 23, 2004. Approval of data management in the trial was obtained from the Danish Data Protection Agency on February 11, 2005. The trial is registered in Clinical Trials.gov Protocol Registration System (identification no. NCT00496977).
From October 2004 to March 2006, 5861 subjects contacted the screening unit. Of these 1757 where excluded either because they did not meet the eligibility criteria or because they did not show up, and the remaining 4104 subjects (70%) were randomized to screening by CT (2052) or no screening (2052) after signing the informed consent form.
No statistically significant difference in age, sex, pulmonary function, smoking status, tobacco consumption, or duration of smoking cessation was seen between CT and control group (Table 1). In addition no difference in social status was found between the two groups (data not shown). In the CT arm all except five participants were scanned at base line (compliance 99.95%).
Radiologic and Clinical Findings
At baseline no nodules were found in 1458 participants (71%) and in the remaining 594 participants (29%) a total of 897 lung nodules were identified (Table 2). Seven hundred eight nodules did not require further diagnostic workup, either because the nodules were calcified (337 category 1 nodules) or because the size was under 5 mm (371 category 2 nodules) (Table 2). This leaves 189 nodules of which 151 noncalcified nodules in 142 participants were between 5 mm and 15 mm (category 3) and 38 nodules were larger than 15 mm (category 4) (Table 3). In 30 participants ground glass opacities (GGO lesions) were found, three of these having two GGO and three more than 2 GGO lesions (included in the above).
Based on the radiologic appearance, three participants with six category three nodules were referred for diagnostic evaluation, and all three participants had a malignant nodule. The remaining 145 category 3 nodules were rescanned after 3 months, where 29 (including 1 GGO) had regressed in size, 112 (including GGOs in 28 pts) were unchanged, and 4 showed growth and were removed. Of the growing nodules, 3 (including 1 GGO lesion) proved to be lung cancer, and one was a granuloma.
The 38 nodules larger than 15 mm (category 4) were found in 37 participants (Table 3). Five of these nodules were judged benign by morphology and presence of calcifications and no further action was taken except in one who had a 45 mm harmatoma removed by VATS due to the size. Seven nodules were judged malignant, and in six of these a malignant diagnosis was confirmed by further workup. The remaining 26 nodules (mean maximum diameter: 29.8 mm) (range, 16–93 mm) were rescanned after 3 months, 6 nodules regressed in size, 18 were unchanged in size but 5 showed changes in morphology and therefore sent for diagnostic evaluation. Two showed growth and proved to be malignant (Table 3).
Thus, at baseline 179 participants (8.7%) had a positive finding, and in 162 participants (7.9%) the finding proved to be false positive.
A total of 40 invasive diagnostic procedures were performed in 25 participants (Table 4). In eight participants no evidence of lung cancer was found. The most invasive diagnostic procedure was VATS wedge resection, which in nine cases was followed by definitive surgical treatment by lobectomy in the same anesthesia. In two cases VATS wedge resection was due to a false-positive diagnosis.
Lung Cancer Cases
A total of 17 cases of lung cancer were detected during the prevalence screening including 3 cases in whom biopsy from enlarged mediastinal lymph nodes showed lung cancer, but no obvious primary focus was found in the lung.
Treatment was surgical in 11 patients (65% resection rate), 5 had chemotherapy, and 1 patient refused treatment. Surgical treatment was minimal invasive (VATS) in 8 of 11 cases (72%). In 3 cases thoracotomy with lobectomy (n = 2) or pneumonectomy (n = 1) was performed (Table 5).
A 65-year-old man was discharged on day 10 after uneventful open lobectomy for a 12 mm adenocarcinoma in the right lung. He was readmitted on day 25 with pneumonia and died of myocardial infarction with congestive heart failure on day 34. No other treatment related deaths or complications were observed.
A morphologic diagnosis was obtained in all lung cancer cases, in 14 cases by histology and 3 cases by cytology: 2 squamous cell carcinoma, 12 adenocarcinoma, and 3 non-small cell carcinoma. No cases of pure bronchioloalveolar carcinoma were observed. In four patients who were treated with chemotherapy only, the final diagnosis was obtained by mediastinoscopy, fiberbroncoscopy or transthoracic fine needle biopsy.
The screening regimen used in this trial is based primarily on the reports by Claudia Henscke and the Early Lung Cancer Action Program (ELCAP) group2,3,21,22 and on the guidelines agreed by the EU-US Collaborative spiral CT working group under the IASLC. In addition, the regimen was harmonized with the currently running CT screening trial in The Netherlands and Belgium (the NELSON trial) 12 to allow pooling of results.
To our satisfaction the randomization process in this study did not result in any significant difference between the CT and the control group with regards to key parameters such as age, sex, pulmonary lung function, pack years, smoking status, and social class.
In screening trials the false-positive rate is often of concern. In the present study the false-positive rate at baseline was 7.9% only, which compares favorably with other trials3,8,23 where 12 to 19% false positives at baseline have been reported. In the Mayo Clinic trial almost 69% of participants had false-positive test results in 3 years of study,23,24 however, the criteria for a false-positive test differed from the one used here. The psychologic consequences of false-positive test results are currently evaluated and will be published separately.
The specificity of CT-detected nodules in the prevalence round of the present study is high (92.6%). However, high specificity usually implies low sensitivity, which is supported by a relatively small number of persons diagnosed with lung cancer. Therefore, some prevalent cancers will most likely appear at later screenings rounds, as was reported by Veronesi et al.25 The extent of this “delayed diagnosis” in the DLCST will be apparent in the following incidence screenings and will allow quantification of the number of false-negative test results.
We discovered 17 cases of lung cancer corresponding to a detection rate of 0.83% only. This is lower than previously reported by others: 2.7%,3 2.0%,19,23 1.9%,8 1.7%,26 1.3%27 but close to the detection rate in the NELSON trial (Rob van Klaveren, oral communication, November 2007). There are many possible explanations for this discrepancy; the risk profile of the Danish study population (smoking exposure, age etc) may differ; participation and self-selection for a screening trial in the United States and Europe may be different, and the radiologic expertise may differ. We used a modern multislice CT scanner, and both our radiologists have more than 10 years of experience in chest radiology.
The distribution of the stage of lung cancers detected in this trial showed that nine of total 17 LC were stage 1 (53%). This is lower than in the ELCAP reports3,26 but similar to other reports.19,27 We expect the proportion of stage 1 cancers to be higher in the following incidence rounds of screening. Furthermore, the use of growth and volume doubling time to detect malignant nodules differ from the ELCAP experience. Of the 17 patients with cancer only five were judged as growing by the radiologist, and nine nodules were removed because of malignant CT morphology at baseline without waiting for the results of a rescan after 3 months. Three were referred to oncological treatment because of metastasis to lymph nodes in the mediastinum. In our experience at baseline CT morphology is important for distinguishing benign and malignant nodules as also reported by Xu et al.28 Measurement of growth contributed to detection of malignancy in 5 of 17 (29%) of lung cancers, however, this rate may increase in the incidence screening rounds. The use of growth as a predictor for malignant disease has still to be evaluated in more detail. Although five of the six growing nodules were lung cancer, one did not show growth when evaluated in retrospect by Siemens software. Three of the five growing nodules had a VDT less than 400 days. Some studies have shown that a VDT less than 400 days is suggestive of a malignant nodule.29 VDT greater than 400 days may be slow growing tumors that do not develop into clinical cancers and thereby cause overdiagnosis.29,30 The extent of overdiagnosis is an important issue, but cannot be evaluated in this trial yet. Lung cancer cases in the incidence screening rounds and in the control group must be included in the evaluation, before any meaningful interpretation can be made. The detection and treatment of GGO lesions, that presumably represent very early lung cancer,31,32 may carry a risk of treating lesions that in some cases may resolve spontaneously and untreated. This is an exciting new area of research made possible by CT screening.
The procedure for evaluation and follow-up of nodules at baseline was apparently safe for the participants. The burden of diagnostic work up and invasive procedures was close to that reported by the ELCAP group3,19 and was lower than previously reported by investigators from the Mayo clinic.23,24 This may reflect differences in patient selection and criteria for use of invasive procedures. In the study by Swensen et al.23 of 1520 individuals 13 participants underwent 15 surgical procedures for benign disease. In this study (2052 screened) only 2 benign nodules were removed by local VATS resection. All other nodules that were removed turned out to be malignant. In a screening setting the number of invasive procedures should be reduced as much as possible, as is also reflected in the guidelines from the ACCP.33 However, in our study it remains to be seen to what extent the low false-positive rate and low frequency of invasive procedures has been achieved at the expense of a high rate of false-negative diagnoses. Truly, malignant nodules missed in the prevalence round will inevitably grow and present themselves in the incidence screening rounds.
In this study at baseline 11 of 17 patients (65%) with lung cancer could be offered surgical treatment, and a high proportion 8 of 11(72%) was performed as minimal invasive surgery (VATS lobectomy), illustrating that early detection of lung cancer implies more minimal invasive treatment options.
The randomized study design employed in this and other current trials will, hopefully, in the coming years, provide information on the efficacy, benefits and hazards of lung cancer screening. Such knowledge is essential for balanced and nonbiased recommendations for its eventual implementation in health practice.
1. Mulshine JL, Sullivan DC. Lung cancer screening. N Engl J Med 2005;352:2714–2720.
2. Henschke CI, McCauley DI, Yankelevitz DF, et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999;354:99–105.
3. Henschke CI, Yankelevitz DF, Libby DM, et al; The International Early Lung Cancer Action Program Investigators. Survival of patients with Stage I Lung Cancer Detected on CT screening. N Engl J Med 2006;355:1763–1771.
4. McMahon PM, Christiani DC. Computed tomography screening for lung cancer. Results of randomized trials are needed before recommending its adoption. BMJ 2007;334:271.
5. Black WC, Baron JA. CT screening for lung cancer. Spiraling into confusion? JAMA 2007;297:995–996.
6. Welch G, Woloshin S, Schwartz LM, et al. Overstating the evidence for lung cancer screening: the International Early Lung Cancer Action Program (I-ELCAP) Study. Arch Intern Med 2007;167:2289–2295.
7. Reich JM. A critical appraisal of overdiagnosis: estimates of its magnitude and implications for lung cancer screening. Thorax 2008;63:377–383.
8. Bach PB, Jett JR, Pastorino U, et al. Computed tomography screening and lung cancer outcomes. JAMA 2007;297:953–961.
9. Gohagan J, Marcus P, Fagerstrom R, et al. Baseline findings of a randomized feasibility trial of lung cancer screening with spiral CT scan vs chest radiograph. Chest 2004;126:114–121.
11. van Klaveren RJ, Habbema JDF, Pedersen JH, de Koning HJ, Oudkerk M, Hoogsteden HC. Lung cancer screening by low-dose spiral computed tomography. Eur Respir J 2001;18:857–866.
12. van Iersel CA, de Koning HJ, Draisma G, et al. Risk based selection from the general population in a screening trial: selection criteria, recruitment and power for the Dutch-Belgian randomised lung cancer multi-slice CT screening trial (NELSON). Int J Cancer 2006;120:868–874.
13. Mascalchi M, Belli G, Zappa M, et al. Risk-benefit analysis of X-ray exposure associated with lung cancer screening in the Italung-CT trial. AJR Am J Roentgenol 2006;187:421–429.
14. Xu DM, Gietema H, de Koning H, et al. Nodule management protocol of the NELSON randomised lung cancer screening trial. Lung Cancer 2006;54:177–184.
15. Fagerström KO. Towards better diagnoses and more individual treatment of tobacco dependence. Br J Addict 1991;86:543–547.
17. Brodersen J, Thorsen H. Consequences of Screening in Breast Cancer (COS-BC): development of a questionnaire. Scand J Prim Health Care 2008;26:4,251–256.
18. Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J 2005;26:948–968.
19. Thunnissen FB, Kerr KM, Brambilla E, et al. EU-US pathology panel for uniform diagnosis in randomised controlled trials for CT screening of lung cancer. Eur Resp J 2006;28:1186–1189.
20. Travis WD, Brambilla E, Muller-Hermelink HK, et al. (Eds.) World Health Organization Classification of Tumors. Pathology and Genetics of Tumors of the Lung, Pleura, Thymus and Heart. Lyon: IARC press; 2004.
21. Libby DM, Wu N, Lee IJ, et al. CT screening for lung cancer: the value of short term follow-up. Chest 2006;129:1039–1042.
22. Henschke CI, Yankelevitz DF, Naidich DP, et al. CT screening for lung cancer; suspiciousness of nodules according to size on base line scans. Radiology 2004;231:164–168.
23. Swensen SJ, Jett JR, Hartman TE, et al. CT screening for lung cancer: five-year prospective experience. Radiology 2005;235:259–265.
24. Lidell RM, Hartman TE, Swensen SJ, et al. Five year lung cancer screening experience. CT appearance, growth rate, location and histological features of 61 lung cancer cases. Radiology 2007;242:555–562.
25. Veronesi G, Bellomi M, Mulshine JL, et al. Lung cancer screening with low dose computed tomography: a non-invasive diagnostic protocol for baseline lung nodules. Lung Cancer 2008;61:340–349.
26. Shaham D, Breuer R, Copel L, et al. Computed tomography screening for lung cancer: applicability of an international protocol in a single-institution environment. Clin Lung Cancer 2006;7:262–267.
27. Diederich S, Wormanns D, Thomas M, et al. Screening for early lung cancer with low-dose spiral CT: prevalence in 817 asymptomatic smokers. Radiology 2002;222:773–781.
28. Xu DM, van Klaveren RJ, de Bock GH, et al. Role of baseline density and changes in density and nodule features in the discrimination between benign and malignant solid indeterminate pulmonary nodules. Eur J Radiol 2008, [Epub ahead of print]. doi: 10.1016/j.ejrad.2008.02.022.
29. Jennings SG, Winer-Muram HT, Tann M, Ying J, Dowdeswell I. Distribution of stage I lung cancer growth rates determined with serial volumetric CT measurements. Radiology 2006;241:554–562.
30. Thunnisen FB, Schuurbiers OC, den Bakker MA. A critical appraisal of prognostic and predictive factors for common lung cancers. Histopathology 2006;48:779–786.
31. Ikeda NI, Hayashi A, Miura Y, et al. Present strategy of lung cancer screening and surgical management. Ann Thor Cardiovasc Surg 2005;11:363–366.
32. Ohtsuka T, Watanabe K, Kaji M, et al. A clinicopathological study of resected pulmonary nodules with focal pure ground-glass opacity. Eur J Cardiothor Surg 2006;30:160–163.
33. Gould MK, Fletcher J, Iannettoni MD, et al. Evaluation of patients with pulmonary nodules: when is it lung cancer? ACCP evidence-based clinical practice guidelines (2nd edition). Chest 2007;132:108–130.