University of Groningen. Optimization of nodule management in CT lung cancer screening Heuvelmans, Marjolein

Transcription

1 University of Groningen Optimization of nodule management in CT lung cancer screening Heuvelmans, Marjolein IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2015 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Heuvelmans, M. A. (2015). Optimization of nodule management in CT lung cancer screening [Groningen]: University of Groningen Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 2 Contributions of the European trials (European randomized screening Group) in CT lung cancer screening Marjolein A Heuvelmans Rozemarijn Vliegenthart Matthijs Oudkerk Published in Journal of Thoracic Imaging 2015, vol. 30, no. 2, pp

3 10 2. CONTRIBUTIONS OF THE EUROPEAN TRIALS Abstract Lung cancer is the leading cause of cancer related death worldwide. In 2011, the largest lung cancer screening trial worldwide, the U.S. National Lung Screening Trial published a 20% decrease in lung cancer specific mortality in the CT screened group, compared to the group screened by chest X-ray. Based on this trial, different U.S. guidelines recently have recommended CT lung cancer screening. However, several questions regarding the implementation of lung cancer screening need to be answered. In Europe, several lung cancer screening trials are ongoing. It is planned to pool the results of the lung cancer screening trials in European randomized lung cancer CT screening (EUCT). By pooling of the data, EUCT hopes to be able to provide additional information for the discussion of some important sub questions regarding the implementation of lung cancer screening by low-dose CT, for instance for the determination of the optimal screen population, for the comparison between a volume- and diameter-based nodule management protocol and for the determination of optimal screen intervals.

4 Introduction Lung cancer is a major health problem with no improvement in survival over the last decades. At time of diagnosis, lung cancer is often already in an advanced stage, with a 5-year survival of no more than 15%. Currently, several lung cancer screening trials investigating whether early detection of lung cancer in high-risk individuals will reduce lung cancer mortality are ongoing [1 4]. In 2011, the National Lung Screening Trial (NLST), was the first and to date only trial reporting a 20% decrease in lung cancer specific mortality when three rounds of annual low-dose computed tomography (CT) were compared with three annual rounds of chest X-ray screening [5]. Recently, several guidelines were published in the United States, including the U.S. Preventive Services Task Force (USPSTF) ([6 11], all recommending lung cancer screening in current or former heavy smokers. In Europe, active lung cancer screening in the general population has - up to now - not been implemented. First, some important remaining questions should be answered by the results of the European screening trials. In 2013, an overview of all European randomized lung cancer CT screening (EUCT) trials was published [12]. EUCT includes the Nederlands-Leuvens Longkanker Screenings Onderzoek (NELSON), the United Kingdom Lung Cancer Screening trial (UKLS), the Danish Lung Cancer Screening Trial (DLCST), the German Lung Cancer Screening Intervention Study (LUSI), the Multi-centre Italian Lung Detection Trial (MILD), the Italian Lung cancer Computed Tomography screening trial (ITALUNG), and the Detection and screening of early lung cancer by Novel imaging Technology and molecular assays (DANTE). By pooling of the data, EUCT hopes to be able to provide additional information for the discussion of some important issues regarding the implementation of lung cancer screening by low-dose CT, for instance for the determination of the optimal screen population, for the comparison between a volumeand diameter-based nodule management protocol and for the determination of optimal screen intervals [12]. European lung cancer screening trials Screen population in the EUCT trials The methodology of recruitment of study participants varied between the different EUCT trials. In the NELSON trial, eligible participants were selected after filling in questionnaires about general health, life style, and smoking habits [16]. On the basis of information from population registries, 335,441 individuals born between 1928 and 1953 received a questionnaire, 32% replied and 5% were selected for randomization based on their answers given in the questionnaire. Eligible participants were between 50 and 75 years of age and smoked at least 15 cigarettes per day for 25 years or 10 cigarettes per day for 30 years (Table 2.1). Ex-smokers quit less than 10 years before start of the trial could also be included. Three Italian trials with equal smoking criteria (at least 20 pack-years, and, in case of ex-smokers, quit smoking less than 10 years before the moment of inclusion) recruited their participants as follows: In MILD, participants were recruited by advertisements and arti-

5 12 2. CONTRIBUTIONS OF THE EUROPEAN TRIALS Table 2.1: Overview of the EUCT trials. NELSON [1] DLCST [13] MILD [14] LUSI [3] ITALUNG [4] DANTE [15] UKLS [2] Country The Netherlands, Belgium Denmark Italy Germany Italy Italy Great Britain Trial start Number of rounds or Screen interval (yrs) 1, 2, and or 2 (randomized) NA* Participants (N) 15,822 4,104 4,099 4,052 3,206 2,811 32,000 (planned) Males 84% 55% 68% 66% 65% 100% NA* Age criterion Mean age (yrs ±SD) 59 (6) 57 (5) 59 (6) 58 (5) 61 (4) 65 (5) NA* Pack years criterion > > NA* Mean pack years 42 (19) 36 (13) 43 (15) 36 (18) 43 (18) 47 (25) NA* (±SD) CT scanner Siemens and Philips Philips Siemens and Philips Toshiba and Siemens Siemens and General Electric Philips Siemens Number of rows 16 and and and 16 1 and Slice thickness 1 mm 1 and 3 mm 1 mm 1 mm 1 and 3 mm 5 mm 1 and 5 mm Reconstruction 0.7 mm 1 and 1.5 mm 1 mm 0.8 and 0.7 mm 1 and 1.5 mm 3 mm 1 and 2.5 mm interval Nodule evaluation 2D and 3D 2D and 3D 2D and 3D 2D and 3D 2D 2D 2D and 3D *NA = not available

6 cles published in the lay press and in television broadcasts [14]. Eligible participants were between 50 and 75 years old. The ITALUNG trial recruited participants via questionnaires sent to subjects of age years, registered with one of 269 general practitioners [4]. DANTE recruited subjects of age trough general practitioners, mass mailings, advertising leaflets and the local media [15]. In the DLCST, participants were recruited after responding to advertisements in local and regional free newspapers and magazines [13]. Eligible participants were between 50 and 70 years old, and smoked at least 20 pack-years. Ex-smokers who quit smoking less than 10 years before the moment of inclusion and were after age 50 could also be included. In LUSI, participants were recruited via questionnaires, sent on basis of information from population registries of a number of cities to persons between 50 and 69 years of age [3]. The same smoking criteria were used as in the NELSON trial. The UKLS sent questionnaires to subjects between 50 and 75 years old, randomly selected by local primary care trust records. Participants with at least 5% risk of developing lung cancer in the next five years, based on the Liverpool Lung Project model, were eligible for randomization [17]. The mean age of EUCT trial participants (Table 2.1) ranges from 57 years (DLCST) to 65 years (DANTE). The mean number of pack-years ranges from 36 years (DLCST and LUSI) to 47 years (DANTE). First results of the individual European Screening Trials The results of the first and second screening round of the seven EUCT trials are shown in Table 2.2. Up to date, three of the EUCT trials have published their mortality results, and, in contrast to the NLST results, none of these trials showed a significant reduction in lung cancer mortality [18 20]. The DANTE trial published their three-year results in 2009 [18]. In the CT screen arm, lung cancer was diagnosed in 60 of 1,276 (4.7%) participants, after a median follow-up period of 33.7 months, and in 34 of 1,196 (2.8%) participants of the control arm who received no screening examinations. The percentage of stage one lung cancers was significantly higher in the CT screen arm, compared to the control arm (33 versus 12 stage I cancers, P=0.004), however, no significant reduction in lung cancer mortality was found (1.6% versus 1.7%). The MILD trial published their five-year results in 2012 [19]. In this trial, three groups were compared; one group receiving no screenings and two groups receiving annual or biennial screenings, respectively. After a median follow-up period of 4.4 years, lung cancer was diagnosed in 34 of 1,190 (2.9%) participants in the annual screen arm, in 25 of 1,186 (2.1%) participants in the biennial screen arm and in 20 of 1,723 (1.2%) participants in the control arm. Again, the percentage of stage one lung cancers was significantly higher in the CT screen groups, but lung cancer mortality did not differ significantly between the control arm and the screen arms (12/1,190 (1.0%), 6/1,186 (0.5%), and 7/1,723 (0.4%), respectively). The third and final EUCT trial who has already published their individual results, in 2012, is the DLCST [20]. In the CT screen arm, lung cancer was diagnosed in 69 of 2,052 (3.4%) participants, after a median follow-up period of 4.8 years. In the control arm, receiving no screening examination, lung cancer was diagnosed in 24 of 2,052 (1.2%) participants. Also in this trial the number of low stage cancers was higher in the screen

7 14 2. CONTRIBUTIONS OF THE EUROPEAN TRIALS arm than in the control arm, but again, no difference in lung cancer specific mortality was found (15/2,052 (0.73%) versus 11/2,052 (0.54%), respectively. One of the reasons for the lack of significant difference in lung cancer specific mortality between the screen arm and the control arm in the above mentioned trials could be the relatively small number of study participants and the relatively short follow-up period. The results of the larger European lung cancer screening trials, and the pooling of the EUCT trials should be awaited, before definitive conclusions can be drawn. The results of the largest EUCT trial, the NELSON trial, will be published next year [21]. Table 2.2: Results of the first and second screening rounds. NELSON [1] DLCST [13] MILD [19] LUSI [3] ITALUNG [4] DANTE [18] Baseline test positive 2.6% 8.7% 2.2% 2.6% 4.2% 4.8% Baseline test indeterminate 19% 8% 13% 17% 26% 12% Second round test positive 1.8% 2.3% NA* 1.9% 1.0% 0.8% Second round test indeterminate 3.8% 1.5% 3% 28% 16% 3.4% Lung cancer detection rate baseline 0.9% 0.8% 0.7% 1.1% 1.5% 2.2% Lung cancer detection rate second round 0.8% 0.6% 0.8% NA* 0.2% 4.7% Stage I cancers baseline 64% 53% NA* 82% 48% 57% Stage I cancers second round 74% 66% NA* 57% 100% 70% *NA = not available. Results of UKLS currently not available Pooling By pooling of the data of the seven EUCT trials, a lung cancer screening population of over 37,000 participants can be created [12]. Although the seven EUCT trials all compared CT screening to no screening in heavy former and current smokers, they have some differences in their study designs. The number of screen rounds varied from one (UKLS [2]) to up to ten (MILD [19]) and the time-between-screens varied from one to 2.5 years (Table 2.1). There were also some differences in radiological aspects, wherein most striking the difference in nodule evaluation; where almost all EUCT trials applied threedimensional volume measurements for nodule evaluation, the DANTE and ITALUNG trials only used two-dimensional diameter measurements (Table 2.1). There are multiple subquestions that could be answered after pooling the EUCT data, besides the question on the effect of lung cancer screening on decrease in lung cancer specific mortality and all-cause mortality in a high risk population. These subquestions include the optimal screen population for lung cancer screening and the optimal nodule management protocol in terms of nodule measurement strategy, follow-up scans, number of screening rounds and time in-between screening rounds, the effect of overdiagnosis and the cost-effectiveness of lung cancer screening.

8 Optimal screen population Considering the millions of individuals that are potentially eligible for lung cancer screening [22, 23], it is of major importance to design a screening program with minimal harms and costs with preserved efficacy [24]. Screening programs should therefore focus on persons at higher risk for lung cancer, so that sensitivity, specificity and cost-effectiveness of lung-cancer screening programs will be improved [25, 26]. Ideally, individuals with the highest risk of developing lung cancer could be identified with biomarkers, and the same biomarkers could be used to determine the lung cancer risk of a detected lung nodule. Up till now, however, this way of determining the best target population for screening can not be used in practice [27]. Recently, a lot of attention was paid to the question what target population should be invited for LDCT screening, using data of a.o. the NLST and the Prostate, Lung Colorectal and Ovarian Cancer Screening Trial [28 33]. One of these studies [29] showed that a higher number of lung cancer deaths can be prevented in a quintile of participants with higher 5-year risk of lung-cancer death (risk of more than 2%), compared to group that was at lowest 5-year risk of lung-cancer death (risk of %). In the last group, only few lung-cancer deaths could be prevented. In this study, the risk quintiles were based on an absolute risk-prediction model for lung-cancer mortality. Factors that were included in this model were not only age, pack-years of smoking, and years since smoking cessation, as in the most lung cancer screening trials, but also body-mass index, family history of lung cancer, and emphysema diagnosis. The results of this study indicate that screen participant selection by using lung cancer risk models can be helpful to optimize lung cancer screening in terms of cost-effectiveness and test parameters. The results of former studies should be confirmed by using data of the EUCT trials. Optimal screen interval One of the issues in determining the optimal nodule management protocol is the length of the screen interval. If the screen interval is too short, participants have to undergo unnecessary CT examinations, leading to redundant radiation dose and costs. On the other hand, if the screen interval is too long, the number of interval cancers, mostly in a stage in which successful treatment is no longer an option, will increase dramatically. In most of the EUCT trials, as well as in the NLST, participants were screened by three to five annual screening rounds [3 5, 18, 20]. Two EUCT trials varied the scanning interval. In the NELSON study, different screen intervals (1 year, 2 years, and 2.5 years) were used to be able to determine the most optimal screen interval. It was found that the number of interval cancers increases by prolonged screen intervals [34]; five interval cancers were detected between baseline and the second round screening in year 2, seven and twelve in the first and second year after the second round screening, respectively, and seven and four in the first and second year after the third round screening, respectively. Overall, the number of interval cancers detected in the NELSON study is relatively low and the percentage of early stage lung cancers (stage I), ass well as the percentage of late stage lung cancers (stage IIIB and IV) did not differed significantly with the percentages in NLST, in which annual rounds of lung cancer screening were performed (62% stage I in NELSON versus 59% in NLST, and 17% late stage in NELSON versus 23% in NLST) [34]. In the MILD trial, participants were randomized to receive either annual or biennial screenings [19].

9 16 2. CONTRIBUTIONS OF THE EUROPEAN TRIALS In the five-year results, no difference in number of interval cancers was found between the two groups [19]. In future analyses, more results of determining the optimal screen interval are expected. A mathematical model for determining the optimal screen intervals to detect lung nodules with different growth rates showed that an interval of 7-10 months should be sufficient to detect even the most rapidly growing lung cancers (VDT <30 days) [35]. More common are lung cancers with a VDT of around days. Such a nodule would theoretically need months to grow from a volume of 5 mm 3 (not detectable) to 500 mm 3. Therefore, in this theoretical study, a screen interval of 18 months is suggested [35]. Nodule measurements in lung cancer screening Diameter measurements In two EUCT trials nodule management was solely based on diameter measurements [4, 15] (Table 2.1). In ITALUNG, nodules were measured manually by the radiologist with electronic calipers on workstations. At baseline, nodules with mean diameter 5 mm tested positive, otherwise negative. At repeat screenings, nodule management was based on growth (at least 1 mm in mean diameter tested positive) for existing nodules and size (mean diameter 3 mm tested positive) for new nodules [4]. In DANTE, participants with CT results suggestive of malignancy, such as nodules with a manually measured diameter of 10 mm, or 5-10 mm but showing spiculated margins, tested positive [15]. Two trials used diameter measurements only for newly detected nodules, for previously detected nodules growth was based on the volume-doubling time (VDT) [3, 13]. In the DLCST, nodule management of new nodules was based on maximal diameter, manually-measured in axial slices. Participants tested negative if maximal diameter was <5 mm, indeterminate in case of diameter 5-15 mm and positive if the maximal diameter exceeded 15 mm [13]. In the LUSI trial, participants tested negative if the largest nodule had a diameter 5 mm and positive if diameter 10 mm, otherwise they tested indeterminate [3]. Volume measurements The NELSON study was the first lung cancer screening trial in which nodule management is based on nodule volume, instead of transverse cross-sectional nodule diameter for new nodules, and nodule growth in terms of VDT for existing ones [36]. The UKLS, used a similar nodule management protocol as used in the NELSON trial [2]. The MILD trial also used a nodule management protocol based on nodule volume (Table 2.1), but they used different cutoff values (>250 mm 3 for a positive test result and <60 mm 3 for a negative test result [14]). In the DLCST and LUSI, nodule management of previously detected nodules was based on the VDT, using the same VDT cutoff values as NELSON [3, 13].

10 NELSON volume criteria for nodule stratification For solid nodules, and solid components of part-solid nodules, volume was calculated by three-dimensional volumetric computer assessment, using software for semi-automated volume measurements (LungCare, version Somaris/5: VA70C-W; Siemens Medical Solutions). The final screen result was based on the nodule with largest volume or fasted growth. In the NELSON management protocol, nodules were classified as negative if volume was <50 mm 3 (4.6 mm diameter if the nodule would have been perfectly spherical), leading to an invitation for the regular next-round CT, as positive if nodule volume was >500 mm 3 (>9.8 mm diameter), leading to referral to a pulmonologist for further workup, and as indeterminate in if nodule volume was mm 3. Indeterminate nodules underwent a short-term follow-up low-dose CT for growth assessment (Figure 2.1) [36]. Figure 2.1: Decision tree of baseline CT examination in the NELSON study. Volumetric growth assessment of pulmonary nodules After a nodule has been selected by a radiologist, nodule volume was semi-automatically calculated by the LungCare software. Information was saved in the NELSON Management System (NMS), which calculated nodule growth in case of a previously detected nodule. Growth was defined as a change in volume of 25% between two subsequent scans according to the formula [36]: Percentage Volume Change = V 2 V 1 V 1 100% (2.1) In which V 2 = volume at last CT, and V 1 = volume at previous examination.

11 18 2. CONTRIBUTIONS OF THE EUROPEAN TRIALS Determination of the volume-doubling time For solid nodules, or solid components of partial-solid nodules with percentage volume change of more than 25%, the VDT was calculated by NMS according to the formula [36]: VDT = ln(2) t ln(v2 /V1 ) (2.2) VDT is used to classify a screen as positive (VDT <400 days), requiring additional diagnostic procedures, indeterminate (VDT days), requiring an extra follow-up CT 12 months after the regular round CT, or negative (VDT >600 days) (see e.g. Figure 2.2) [36]. Using this two-step approach of volume and growth assessment, less participants tested negative compared to diameter based protocols, and compared to other screening trials, a higher proportion tested true-positive. The negative predictive value was comparable to diameter-based protocols [1]. Figure 2.2: Example of a growing malignant lesion in a 51 year-old current male smoker (46 pack-years). Axial computed tomography (CT) scan (B) shows a nodule (arrow) with a volume of 277 mm3 in the right lower lobe, which was not present at the previous examination (A), resulting in an indeterminate screen result. At the follow-up CT seven weeks later (C), the nodule volume was 569 mm3. The VDT at that time was 47 days. Work-up revealed a stage IA adenocarcinoma. M.a.b. = months after baseline, FU = follow-up. Comparison of volumetric and diameter assessment of pulmonary nodules Lung nodules often do not have a perfectly spherical shape, so a management protocol based on the maximal transversal diameter lead to an overestimation of nodule size in the majority of cases. In lung cancer screening trials in which nodule management is based on maximal diameter, the number of false-positive screens is higher as in semi-automated volume-based lung cancer screening [1, 5].

12 Furthermore, volume measurements have been found to be more reliable than diameter measurements for determining nodule size [37, 38]. In a lung phantom study [39], measurements of the manually-measured maximal transverse diameter and semi-automated measurements of diameter and nodule volume were compared with the actual properties. In both the manual and the semi-automated measurement method, diameter and volume of the spherical solid nodules were significantly underestimated, although semi-automated measurements were found to be more reliable than manual measurements. For measurements of nodule diameter, the overall underestimation was about 10% using the manual method, and less than 4% using the semi-automated method. For measurements of nodule volume, the overall underestimation was about 25% (corresponding to 8% diameter underestimation) using the manual method, compared to less than 8% (corresponding to 2.5% diameter underestimation) using the semi-automated method. Nodule growth can be detected after a shorter time period when using volume measurements compared to using diameter measurements. For example, regarding a spherical nodule, a 25% increase in diameter from 6 mm to 7.5 mm represents nearly a doubling in volume (from 113 mm 3 to 221 mm 3 ). In other words, it is impossible to detect a diameter growth of 25% before the lung nodule doubles in volume. On contrary, a 25% increase in volume from 60 mm 3 to 75 mm 3 reflects an increase in diameter from only 4.9 mm to 5.2 mm, with a corresponding percentage diameter change of less than 8%. This increase in diameter lies within the standard deviation of CT diameter measurements for small nodules, and could not be measured manually [39]. Recently a study on lung cancer probability in the NELSON study, in which different nodule management strategies based on nodule volume, VDT, and diameter were simulated on the data of the baseline screening round of the NELSON study was published [40]. A protocol based on nodule volume (cut-offs for an indeterminate and positive test result 100 mm 3 and 300 mm 3, respectively) and VDT was compared to a protocol based on nodule diameter (cut-offs for an indeterminate and positive test result 5 mm and 10 mm, respectively). It was shown that the volume-based protocol had comparable sensitivity and negative predictive value to the diameter-based protocol (90.9% (95%-confidence interval (CI): ) versus 92.4% (95%-CI: ), and 99.9% (95%-CI: ) for both studies, respectively), while the specificity and positive predictive value were favourable for the volume-based protocol (94.9% (95%-CI: ) versus 90.0% (95%-CI: ), and 14.4% (95%-CI: ) versus 7.9% (95%-CI: ), respectively) [40]. The percentage of positive test results in the first two screening rounds of the NELSON trail, as published in 2009 [1], was considerably lower than that of the NLST (2.6% at baseline and 1.8% at first repeat round versus 27.3% at baseline and 27.9% at first repeat round, respectively [1, 5]), with comparable negative predictive value ( % versus 99.9%, respectively). This suggests that the relatively stringent volume and VDT based nodule management protocol of the NELSON study, which is also used in the LUSI, UKLS and DLCST trials, is more efficient with less co morbidity and lower costs. However, differences between volume and diameter based nodule management protocols in terms of early lung cancer detection, morbidity, mortality, radiation exposure and costs remain to be demonstrated, and results of the NELSON trial need to be awaited, before positive recommendations on lung cancer screening in Europe can be given.

13 20 References [1] van Klaveren RJ, Oudkerk M, Prokop M, Scholten ET, Nackaerts K, Vernhout R, et al. (2009). Management of lung nodules detected by volume CT scanning. The New England journal of medicine, 361(23): [2] Baldwin DR, Duffy SW, Wald NJ, Page R, Hansell DM, and Field JK (2011). UK Lung Screen (UKLS) nodule management protocol: modelling of a single screen randomised controlled trial of low-dose CT screening for lung cancer. Thorax, 66(4): [3] Becker N, Motsch E, Gross ML, Eigentopf A, Heussel CP, Dienemann H, et al. (2012). Randomized study on early detection of lung cancer with MSCT in Germany: study design and results of the first screening round. Journal of cancer research and clinical oncology, 138(9): [4] Lopes Pegna A, Picozzi G, Mascalchi M, Maria Carozzi F, Carrozzi L, Comin C, et al. (2009). Design, recruitment and baseline results of the ITALUNG trial for lung cancer screening with low-dose CT. Lung cancer, 64(1): [5] National Lung Screening Trial Research Team, Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, et al. (2011). Reduced lung-cancer mortality with low-dose computed tomographic screening. The New England journal of medicine, 365(5): [6] US Preventive Services Task Force (2013). Screening for Lung Cancer: U.S. Preventive Services Task Force Recommendation Statement. URL uspreventiveservicestaskforce.org/draftrec.htm. [7] Detterbeck FC, Mazzone PJ, Naidich DP, and Bach PB (2013). Screening for lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest, 143(5 Suppl):e78S 92S. [8] Wender R, Fontham ET, Barrera Jr E, Colditz GA, Church TR, Ettinger DS, et al. (2013). American Cancer Society lung cancer screening guidelines. CA: a cancer journal for clinicians, 63(2): [9] Jaklitsch MT, Jacobson FL, Austin JH, Field JK, Jett JR, Keshavjee S, et al. (2012). The American Association for Thoracic Surgery guidelines for lung cancer screening using low-dose computed tomography scans for lung cancer survivors and other highrisk groups. The Journal of thoracic and cardiovascular surgery, 144(1): [10] National Comprehensive Cancer Network (2013). NCCN Clinical Practice Guidelines in Oncology. Lung Cancer Screening. [11] Jacobson FL, Austin JH, Field JK, Jett JR, Keshavjee S, MacMahon H, et al. (2012). Development of The American Association for Thoracic Surgery guidelines for lowdose computed tomography scans to screen for lung cancer in North America: recommendations of The American Association for Thoracic Surgery Task Force for Lung Cancer Screenin. The Journal of thoracic and cardiovascular surgery, 144(1):25 32.

14 [12] Field JK, van Klaveren R, Pedersen JH, Pastorino U, Paci E, Becker N, et al. (2013). European randomized lung cancer screening trials: Post NLST. Journal of surgical oncology, 108(5): [13] Pedersen JH, Ashraf H, Dirksen A, Bach K, Hansen H, Toennesen P, et al. (2009). The Danish randomized lung cancer CT screening trial overall design and results of the prevalence round. Journal of thoracic oncology, 4(5): [14] Pastorino U (2006). Early detection of lung cancer. Respiration, 73(1):5 13. [15] Infante M, Lutman FR, Cavuto S, Brambilla G, Chiesa G, Passera E, et al. (2008). Lung cancer screening with spiral CT: baseline results of the randomized DANTE trial. Lung cancer, 59(3): [16] van Iersel CA, de Koning HJ, Draisma G, Mali WP, Scholten ET, Nackaerts K, et al. (2007). 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). International journal of cancer.journal international du cancer, 120(4): [17] McRonald FE, Yadegarfar G, Baldwin DR, Devaraj A, Brain KE, Eisen T, et al. (2014). The UK Lung Screen (UKLS): demographic profile of first 88,897 approaches provides recommendations for population screening. Cancer prevention research, 7(3): [18] Infante M, Cavuto S, Lutman FR, Brambilla G, Chiesa G, Ceresoli G, et al. (2009). A randomized study of lung cancer screening with spiral computed tomography: threeyear results from the DANTE trial. American journal of respiratory and critical care medicine, 180(5): [19] Pastorino U, Rossi M, Rosato V, Marchiano A, Sverzellati N, Morosi C, et al. (2012). Annual or biennial CT screening versus observation in heavy smokers: 5-year results of the MILD trial. European journal of cancer prevention, 21(3): [20] Saghir Z, Dirksen A, Ashraf H, Bach KS, Brodersen J, Clementsen PF, et al. (2012). 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, 67(4): [21] Ru Zhao Y, Xie X, de Koning HJ, Mali WP, Vliegenthart R, and Oudkerk M (2011). NELSON lung cancer screening study. Cancer imaging, 11 Spec No:S [22] Ma J, Ward EM, Smith R, and Jemal A (2013). Annual number of lung cancer deaths potentially avertable by screening in the United States. Cancer, 119(7): [23] Pinsky PF and Berg CD (2012). Applying the National Lung Screening Trial eligibility criteria to the US population: what percent of the population and of incident lung cancers would be covered? Journal of medical screening, 19(3): [24] Bach PB, Mirkin JN, Oliver TK, Azzoli CG, Berry DA, Brawley OW, et al. (2012). Benefits and harms of CT screening for lung cancer: a systematic review. Journal of the American Medical Association, 307(22):

15 22 [25] Field JK, Aberle DR, Altorki N, Baldwin DR, Dresler C, Duffy SW, et al. (2014). The International Association Study Lung Cancer (IASLC) Strategic Screening Advisory Committee (SSAC) response to the USPSTF recommendations. Journal of thoracic oncology, 9(2): [26] Field JK, Chen Y, Marcus MW, Mcronald FE, Raji OY, and Duffy SW (2013). The contribution of risk prediction models to early detection of lung cancer. Journal of surgical oncology, 108(5): [27] Field JK, Hansell DM, Duffy SW, and Baldwin DR (2013). CT screening for lung cancer: countdown to implementation. The lancet oncology, 14(13):e [28] Kovalchik SA, Tammemagi M, Berg CD, Caporaso NE, Riley TL, Korch M, et al. (2013). Targeting of low-dose CT screening according to the risk of lung-cancer death. The New England journal of medicine, 369(3): [29] Tammemagi MC, Katki HA, Hocking WG, Church TR, Caporaso N, Kvale PA, et al. (2013). Selection criteria for lung-cancer screening. The New England journal of medicine, 368(8): [30] Tammemagi CM, Pinsky PF, Caporaso NE, Kvale PA, Hocking WG, Church TR, et al. (2011). Lung cancer risk prediction: Prostate, Lung, Colorectal And Ovarian Cancer Screening Trial models and validation. Journal of the National Cancer Institute, 103(13): [31] Spitz MR, Hong WK, Amos CI, Wu X, Schabath MB, Dong Q, et al. (2007). A risk model for prediction of lung cancer. Journal of the National Cancer Institute, 99(9): [32] Bach PB, Kattan MW, Thornquist MD, Kris MG, Tate RC, Barnett MJ, et al. (2003). Variations in lung cancer risk among smokers. Journal of the National Cancer Institute, 95(6): [33] Raji OY, Duffy SW, Agbaje OF, Baker SG, Christiani DC, Cassidy A, et al. (2012). Predictive accuracy of the Liverpool Lung Project risk model for stratifying patients for computed tomography screening for lung cancer: a case-control and cohort validation study. Annals of Internal Medicine, 157(4): [34] Horeweg N, Th Scholten E, de Jong PA, van der Aalst CM, Weenink C, Lammers JW, et al. (2014). Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers. The Lancet Oncology, 15(12): [35] Poullis M (2014). Computed tomography surveillance scanning after lung cancer surgery: mathematical optimization of scanning interval based on tumour biology. Interactive cardiovascular and thoracic surgery, 18(2): [36] Xu DM, Gietema H, de Koning H, Vernhout R, Nackaerts K, Prokop M, et al. (2006). Nodule management protocol of the NELSON randomised lung cancer screening trial. Lung cancer, 54(2): [37] Yankelevitz DF, Reeves AP, Kostis WJ, Zhao B, and Henschke CI (2000). Small pulmonary nodules: volumetrically determined growth rates based on CT evaluation. Radiology, 217(1):

16 [38] Revel MP, Bissery A, Bienvenu M, Aycard L, Lefort C, and Frija G (2004). Are two-dimensional CT measurements of small noncalcified pulmonary nodules reliable? Radiology, 231(2): [39] Xie X, Zhao Y, Snijder RA, van Ooijen PM, de Jong PA, Oudkerk M, et al. (2013). Sensitivity and accuracy of volumetry of pulmonary nodules on low-dose 16- and 64-row multi-detector CT: an anthropomorphic phantom study. European radiology, 23(1): [40] Horeweg N, van Rosmalen J, Heuvelmans MA, van der Aalst CM, Vliegenthart R, Th Scholten E, et al. (2014). Lung cancer probability in patients with CT-detected pulmonary nodules: a prespecified analysis of data from the NELSON trial of lowdose CT screening. The Lancet Oncology, 15(12):

17