Advertisement

Long-Term Active Surveillance of Screening Detected Subsolid Nodules is a Safe Strategy to Reduce Overtreatment

Open ArchivePublished:July 16, 2018DOI:https://doi.org/10.1016/j.jtho.2018.06.013

      Abstract

      Introduction

      Lung cancer presenting as subsolid nodule (SSN) can show slow growth, hence treating SSN is controversial. Our aim was to determine the long-term outcome of subjects with unresected SSNs in lung cancer screening.

      Methods

      Since 2005, the Multicenter Italian Lung Detection (MILD) screening trial implemented active surveillance for persistent SSN, as opposed to early resection. Presence of SSNs was related to diagnosis of cancer at the site of SSN, elsewhere in the lung, or in the body. The risk of overall mortality and lung cancer mortality was tested by Cox proportional hazards model.

      Results

      SSNs were found in 16.9% (389 of 2303) of screenees. During 9.3 ± 1.2 years of follow-up, the hazard ratio of lung cancer diagnosis in subjects with SSN was 6.77 (95% confidence interval: 3.39–13.54), with 73% (22 of 30) of cancers not arising from SSN (median time to diagnosis 52 months from SSN). Lung cancer–specific mortality in subjects with SSN was significantly increased (hazard ratio = 3.80; 95% confidence interval: 1.24–11.65) compared to subjects without lung nodules. Lung cancer arising from SSN did not lead to death within the follow-up period.

      Conclusions

      Subjects with SSN in the MILD cohort showed a high risk of developing lung cancer elsewhere in the lung, with only a minority of cases arising from SSN, and never representing the cause of death. These results show the safety of active surveillance for conservative management of SSN until signs of solid component growth and the need for prolonged follow-up because of high risk of other cancers.

      Keywords

      Introduction

      Lung cancer screening by low-dose computed tomography (LDCT) allows for early diagnosis of tumors that would otherwise present as symptomatic advanced disease.
      • Aberle D.R.
      • Adams A.M.
      • et al.
      National Lung Screening Trial Research Team
      Reduced lung-cancer mortality with low-dose computed tomographic screening.
      Early diagnosis and early treatment of lung cancer increase survival by reduction of cause-specific and overall mortality.
      • Aberle D.R.
      • Adams A.M.
      • et al.
      National Lung Screening Trial Research Team
      Reduced lung-cancer mortality with low-dose computed tomographic screening.
      However, screening carries the risk of diagnosing and treating malignancies that would have not affected the subjects’ life expectancy because of low tumor aggressiveness or concurrent comorbidities.
      • Esserman L.J.
      • Thompson Jr., I.M.
      • Reid B.
      Overdiagnosis and overtreatment in cancer: an opportunity for improvement.
      • Patz Jr., E.F.
      • Pinsky P.
      • Gatsonis C.
      • et al.
      Overdiagnosis in low-dose computed tomography screening for lung cancer.
      • Bach P.B.
      Overdiagnosis in lung cancer: different perspectives, definitions, implications.
      • Carter J.L.
      • Coletti R.J.
      • Harris R.P.
      Quantifying and monitoring overdiagnosis in cancer screening: a systematic review of methods.
      Such overdiagnosis and overtreatment are especially seen in slow-growing lung adenocarcinomas that are mostly represented by subsolid nodules (SSNs).
      • Kakinuma R.
      • Noguchi M.
      • Ashizawa K.
      • et al.
      Natural history of pulmonary subsolid nodules: a prospective multicenter study.
      • Travis W.D.
      • Asamura H.
      • Bankier A.A.
      • et al.
      The IASLC Lung Cancer Staging Project: proposals for coding T categories for subsolid nodules and assessment of tumor size in part-solid tumors in the forthcoming eighth edition of the TNM Classification of Lung Cancer.
      • Yip R.
      • Wolf A.
      • Tam K.
      • et al.
      Outcomes of lung cancers manifesting as nonsolid nodules.
      • Revel M.P.
      • Mannes I.
      • Benzakoun J.
      • et al.
      Subsolid lung nodule classification: a CT criterion for improving interobserver agreement.
      • Heidinger B.H.
      • Anderson K.R.
      • Nemec U.
      • et al.
      Lung adenocarcinoma manifesting as pure ground-glass nodules: correlating CT size, volume, density, and roundness with histopathologic invasion and size.
      Retrospective studies assessing precursor lesions of interval cancers and post-screen cancers showed that SSNs were visible at the tumor sites years before tumor diagnosis.
      • Scholten E.T.
      • Horeweg N.
      • de Koning H.J.
      • et al.
      Computed tomographic characteristics of interval and post screen carcinomas in lung cancer screening.
      Such retrospective evidence prompted more conservative algorithms in the most recent guidelines.

      Lung-RADS Version 1.0 Assessment Categories. April 28, 2014. http://www.acr.org/∼/media/ACR/Documents/PDF/QualitySafety/Resources/LungRADS/AssessmentCategories. Accessed January 1 2017.

      • Pinsky P.F.
      • Gierada D.S.
      • Black W.
      • et al.
      Performance of lung-RADS in the National Lung Screening Trial: a retrospective assessment.

      National Comprehensive Cancer Network Guidelines, Version 3.2018, Lung Cancer Screening. https://www.nccn.org/professionals/physician_gls/pdf/lung_screening.pdf. Accessed March 19, 2018.

      Smokers recruited in lung cancer screening show a heterogeneous range of concurrent causes of death, including multifocal thoracic malignancies, extrapulmonary cancer, and non-neoplastic disease.
      • Kotwal A.A.
      • Schonberg M.A.
      Cancer screening in the elderly: a review of breast, colorectal, lung, and prostate cancer screening.
      • Flores R.
      • Taioli E.
      • Yankelevitz D.F.
      • et al.
      Initiative for early lung cancer research on treatment: development of study design and pilot implementation.
      The resection of a slow-growing lung cancer might result in no clinical advantage or even detrimental effects for a subject with such comorbidities.
      • Bach P.B.
      • Mirkin J.N.
      • Oliver T.K.
      • et al.
      Benefits and harms of CT screening for lung cancer: a systematic review.
      • Leduc C.
      • Antoni D.
      • Charloux A.
      • et al.
      Comorbidities in the management of patients with lung cancer.
      Since 2005, the Multicenter Italian Lung Detection (MILD) study protocol implemented a prospective conservative approach to SSN, by long-term active surveillance. Such a conservative strategy for slow-growing tumors enabled us to evaluate the progression rate of SSNs and their long-term outcomes, as reported in a previous paper analyzing the first 5 years of follow-up.
      • Silva M.
      • Sverzellati N.
      • Manna C.
      • et al.
      Long-term surveillance of ground-glass nodules: evidence from the MILD trial.
      In the present study we describe risk of lung cancer and lung cancer–related death in subjects with unresected SSNs over a period of almost 10 years, and analyze whether the cancer actually arose from the SSN. In addition, we detail the cause of death in SSN subjects, including extrapulmonary cancer and non-neoplastic diseases.

      Materials and Methods

      The protocol of the MILD trial was approved by the Institutional Review Board and was registered in a publicly available database (www.clinicaltrials.gov identifier: NCT02837809), written consent was obtained from all recruited subjects.
      • Sverzellati N.
      • Silva M.
      • Calareso G.
      • et al.
      Low-dose computed tomography for lung cancer screening: comparison of performance between annual and biennial screen.
      The present study was based on the 2303 subjects randomized to the LDCT arm (mean age 58.1 ± 5.9 years, cumulative tobacco exposure 43.6 ± 21.5 pack years).
      For the present study, we selected both SSN prospectively detected by visual analysis and SSN retrospectively detected by computer-aided diagnosis (CAD).
      • Silva M.
      • Sverzellati N.
      • Manna C.
      • et al.
      Long-term surveillance of ground-glass nodules: evidence from the MILD trial.
      The integration of visual data with software detection was meant to provide the most complete and updated scenario of lung cancer screening with volumetric thin-slice LDCT. In particular, the CAD used for this study (CIRRUS Lung Screening, Nijmegen, the Netherlands) is the prototype version of Veolity (MeVis Medical Solutions AG, Bremen, Germany), which is a United States Food and Drug Administration– and Conformité Européene– approved reading workstation for lung cancer screening scans.
      • Jacobs C.
      • van Rikxoort E.M.
      • Twellmann T.
      • et al.
      Automatic detection of subsolid pulmonary nodules in thoracic computed tomography images.
      The baseline, 2-year, and 4-year LDCTs were used for detection of SSNs by the CAD. All CAD marks were jointly reviewed by two experienced thoracic radiologists (with 8 and 11 years of experience in screening LDCT, respectively) who selected CAD marks reflecting SSNs and discarded false-positive CAD marks. It was made sure that SSN identified by visual analysis during the original screening process were included in the final data set used for our analysis.
      All SSNs were classified into nonsolid nodules (NSNs) or part-solid nodules (PSNs), and were measured by volumetric semiautomatic segmentation.
      • Charbonnier J.P.
      • Chung K.
      • Scholten E.T.
      • et al.
      Automatic segmentation of the solid core and enclosed vessels in subsolid pulmonary nodules.
      The volume of SSN was measured including the whole nonsolid component (NSN and PSN) and the solid component (PSN).

      Diagnosis of Lung Cancer

      Lung cancer diagnosis during 9.3 ± 1.2 years of follow-up was established by histology from surgical resection or from biopsy (in patients not eligible for surgical treatment).
      The specific site of origin of each lung cancer and its relationship with the SSN were individually assessed on the basis of the pathology report (for resected case) or revision of LDCTs. Histology and pathologic stage of lung cancers not derived from SSNs were recorded over the same observation period of almost 10 years.

      Risk of Lung Cancer

      The risk of being diagnosed with any lung cancer throughout the follow-up period was calculated in subjects with at least one SSN, in subjects with only solid nodules, and in subjects with negative LDCT (no nodule or solid nodule less than 60 mm3 at baseline).
      • Sverzellati N.
      • Silva M.
      • Calareso G.
      • et al.
      Low-dose computed tomography for lung cancer screening: comparison of performance between annual and biennial screen.
      Hazard ratios (HRs) were calculated for subjects with SSNs and solid nodules using subjects with negative LDCT as a reference.

      Lung Cancer Derived From SSN

      According to the MILD protocol, NSNs and PSNs with a solid component less than 5 mm were managed conservatively by active surveillance (e.g., annual LDCT), regardless of the size of the nonsolid component. Management of a PSN with a new or growing solid component with a diameter 5 mm or greater was determined by a multidisciplinary team as follows: calculation of the volume doubling time of solid component (threshold less than 600 days), 18fluorodeoxiglucose positron-emission tomography (18FDG-PET), and computed tomographic–guided transthoracic needle aspiration.
      • Sverzellati N.
      • Silva M.
      • Calareso G.
      • et al.
      Low-dose computed tomography for lung cancer screening: comparison of performance between annual and biennial screen.
      • Horeweg N.
      • Scholten E.T.
      • de Jong P.A.
      • et al.
      Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers.
      The metabolic activity by 18FDG-PET was measured by means of the maximum standard uptake value (SUVmax). A definite cutoff value was not defined for SSN, whereas an SUVmax of 1.5 or more was deemed highly suspicious. Positive result of the work-up, either 18FDG-PET or transthoracic needle aspiration, or both, prompted surgical resection. Negative work-up granted continuous active surveillance by annual LDCT until evidence of solid component growth.
      The time between first LDCT detection of a SSN and surgical resection was recorded along with histology and pathologic stage.
      • Detterbeck F.C.
      • Chansky K.
      • Groome P.
      • et al.
      The IASLC Lung Cancer Staging Project: methodology and validation used in the development of proposals for revision of the stage classification of NSCLC in the forthcoming (eighth) edition of the TNM Classification of Lung Cancer.

      Survival of Subjects With SSN

      The survival of all screenees was retrieved April 1, 2016, by linking records with the vital status in the Italian National Registry Office database. The cause of death was categorized as: 1) lung cancer derived from SSN; 2) lung cancer not derived from SSN; 3) extrapulmonary cancer; and 4) non-neoplastic disease.

      Statistical Analysis

      Prevalence, incidence, and cumulative frequency of SSNs were calculated. Sociodemographic variables and pulmonary function data were compared between subjects with SSNs, subjects with solid nodules, and subjects without nodules.
      The Fisher’s exact test was used to test the method of detection (either visual or CAD) against any lung cancer and lung cancer mortality. Furthermore, subjects were divided into two groups according to the volume of SSN (either above or below the median) to test the association between volume of SSN and lung cancer not derived from SSN by the Fisher’s exact test.
      The risk of being diagnosed with lung cancer, lung cancer–specific mortality, and overall mortality were calculated by Cox proportional hazards model for subjects with SSNs and for subjects with solid nodules, using subjects without nodules as a reference. The corresponding 95% confidence interval (CI) was reported. The model was adjusted for age, sex, smoking history (years, cigarettes per day, and pack years), and smoking status (current versus former). A survival analysis was used to estimate the survival function of subjects with SSN under active surveillance, in particular assessing the following causes of mortality: 1) lung cancer derived from the SSN, 2) lung cancer not derived from the SSN, 3) extrapulmonary cancer, and 4) non-neoplastic disease. Survival curves were estimated using the Kaplan–Meier method and were compared by log-rank test, including stratification by age and sex.
      • Xie J.
      • Liu C.
      Adjusted Kaplan-Meier estimator and log-rank test with inverse probability of treatment weighting for survival data.
      The statistical analyses were performed using SAS 9.2 (SAS Institute, Cary, North Carolina). Figures were obtained using IBM SPSS Statistics for MacIntosh, Version 25.0 (IBM Corp., Armonk, New York).

      Results

      A total of 6541 nodules were detected in 55.5% (1277 of 2303) of the screenees. Using visual assessment, 2.9% (66 of 2303) of screenees were found to have an SSN. Adding CAD, the cumulative frequency of SSNs increased to 16.9% (389 of 2303) (Table 1). CAD detected 361 SSNs of which 323 were detected only by CAD. CAD detected 57.6% (38 of 66) of SSNs detected by visual reading. The median volume of SSN at first detection was 211 mm3. Visual reading showed a nonsignificant trend for detection of SSN with larger diameter and lower density. Subjects with SSNs detected by CAD showed a not significantly lower frequency of lung cancer (7.2%, 26 of 361) compared to those with visually detected SSNs (13.6%, 9 of 66; p = 0.089), with five subjects with lung cancer detected by both methods. Lung cancer mortality was not significantly lower in subjects with SSNs detected by CAD (1.4%, 5 of 361) compared to those with visually detected SSN (4.5%, 3 of 66; p = 0.111).
      Table 1Characteristics of Subjects With Subsolid Nodule, With Solid Nodule, or With Neither Type of Nodule
      Screenees (n)Visually Detected SSNs

      66
      CAD-Only Detected SSNs

      323
      All SSNs

      389
      Solid Nodules

      888
      SSN vs. Solid

      P
      No Nodule

      1,026
      SSN vs No Nodule

      p Value
      Age59 (54-64)58 (54-64)58 (54-64)58 (54-63)0.15257 (53-61)< 0.001
      Sex
       Female33 (50)124 (38.4)157 (40.4)262 (29.5)< 0.001331 (30.1%)< 0.001
       Male33 (50)199 (61.6)232 (59.6)626 (70.5)768 (69.9%)
      Body mass index, kg/m224 (21.6-27.5)25.4 (23.1-27.8)25.2 (22.8-27.7)25.5 (23.2-28.1)0.15925.9 (23.9-28.7)< 0.001
      Smoking history
       Years40 (35-43)39 (35-45)39 (35-44)38 (34-43)0.01137 (34-41)< 0.001
       Cigarettes per day20 (20-23)20 (20-30)20 (20-30)20 (20-26)0.04920 (20-28)0.405
       Pack-years41 (27-48)41 (32-59)41 (32-55)38 (30-50)0.00239 (31-50)0.012
       Current smoker50 (75.8%)237 (73.4)287 (73.8%)593 (66.8)0.013749 (68.2)0.040
      NLST eligible
      NLST eligibility was defined by age 55 years old or more and smoking history of 30 or more pack-years.
       Yes43 (65.2)186 (57.6)229 (58.9)485 (54.6)0.178580 (52.8)0.038
       No23 (34.8)137 (42.4)160 (41.1)403 (45.4)519 (47.2)
      Pulmonary function test
       FEV1 %predicted98 (89-108)97 (87-108)99 (88-110)99 (87-110)0.034100 (90-110)0.003
       Tiffeneau76 (71-80)74 (69-79)75 (70-80)75 (70-81)0.32576 (71-81)0.003
      SSNs are detailed according to method of detection, either visual detection or computer-aided diagnosis, including risk classification of SSN. The sociodemographic and pulmonary function data of all subjects with SSN is compared with subjects with solid nodule or subjects with neither type of nodule. Frequency is reported as absolute number and percentage or continuous variables are reported as median and 25th-75th percentile as appropriate.
      FEV1 %predicted: forced expiratory volume in 1 second, expressed as a percentage of the predicted normal for a person of the same sex, age and height; Tiffeneau: the ratio between forced expiratory volume in 1 second and forced vital capacity; SSN, subsolid nodule; NLST, National Lung Screening Trial.
      a NLST eligibility was defined by age 55 years old or more and smoking history of 30 or more pack-years.
      Subjects with SSNs were significantly more frequently female (40.4%) than subjects with solid nodules (29.5%) or subjects with negative LDCT (30.1%) (Table 1). Also, subjects with SSNs were significantly more frequently active smokers (73.8%) than subjects with solid nodules (66.8%), and were more frequently eligible by National Lung Screening Trial (NSLT) selection criteria (58.9%) than subjects with neither type of nodule (52.8%) (Table 1).
      • Aberle D.R.
      • Adams A.M.
      • et al.
      National Lung Screening Trial Research Team
      Reduced lung-cancer mortality with low-dose computed tomographic screening.
      Statistically significant differences were seen between subjects with SSNs and the subjects with solid nodules or no nodules also for continuous variables. However, the entity of such differences might be considered within the inherent clinical variability (Table 1; Supplementary data S1-S7).
      The prevalence of SSNs at baseline screening was 12.3% (284 of 2303). During follow-up, 105 SSNs were found, with an incidence rate of 2.3% (50 of 2172) at 2 years, and an additional 2.8% (55 of 1987) at 4 years. Thus, the weighted mean incidence rate was 2.5% per 2-year LDCT round.
      Thirty lung cancers were diagnosed in 389 subjects with SSNs, reflecting a 7.7% overall risk of being diagnosed lung cancer throughout the 9.3 ± 1.2 years of follow-up (8.6/1000 person-years).
      Compared to subjects without nodules, the HR for being diagnosed with lung cancer was 6.77 (95% CI: 3.39–13.54) in subjects with SSNs (389 of 2303) and 4.93 (95% CI: 2.55–9.51) in subjects with solid nodule (888 of 2303). In subjects with SSNs, lung cancer was more frequently diagnosed in areas different from the SSN (73.3%, 22 of 30) than in the SSN itself (26.7%, 8 of 30).

      Lung Cancer Derived From SSNs

      Surgical resection of SSNs was performed in 8 cases (0.37/1000 person-years), all showing a histology of adenocarcinoma. In no case of resected SSN was surgery performed for benign disease. At baseline, one PSN showed indication to work-up for solid component greater than 5 mm, with negative result (see note “a” in Table 2). Such PSN was therefore kept under active surveillance until interval growth (77 months later) prompting a further work-up that, instead, resulted positive and was followed by surgery.
      Table 2Detailed Description of the Evolution of the Seven Subsolid Nodules Undergoing Active Surveillance Before Eventual Resection
      Baseline Calipers, mmActive Surveillance, moPre-Surgical Calipers, mmStandard Uptake Value at 18FDG-PETLung Cancer Stage
      Prevalent NSNTotal 10.448Total 18.9, solid core 6.60.0Ia (T1N0M0)
      Prevalent PSNTotal 6.7, solid core 2.167Total 18.9, solid core 10.77.8Ia (T1N0M0)
      Prevalent PSN adjacent to thin-walled cystTotal 9.7, solid core 3.795Total 19.5, solid core 13.15.4IIIa (T3N1M0)
      Prevalent PSNTotal 14.7, solid core 9.5
      An 18FDG-PET was performed at baseline with SUVmax = 0.
      77Total 26.6, solid core 23.57.9Ib (T2N0M0)
      Prevalent PSN adjacent to thin-walled cystTotal 24.7, solid core 3.417Total 21.1, solid core 7.85.2Ib (T2N0M0)
      Prevalent NSNTotal 3.280Total 15.1, solid core 9.53.3Ib (T2N0M0)
      Prevalent PSN adjacent to thin-walled cystTotal 8.6, solid core 3.499Total 16.6, solid core 8.83.6Ia (T1N0M0)
      NSN: nonsolid nodule; PSN: part-solid nodule; 18FDG-PET, fluorodeoxyglucose positron-emission tomography; SUVmax, maximum standard uptake value.
      a An 18FDG-PET was performed at baseline with SUVmax = 0.
      Specifically, seven of eight screenees underwent surgery after a median active surveillance of 77 months (range, 17 to 99 months), as detailed in Table 2. The eighth screenee refused active surveillance and opted for surgical resection 8 months after first detection (protocol violation). Six cancers were resected in screenees with a single SSN (of which three were from subsolid bulla thickening), and two had multiple SSNs (of which one case showed two SSNs in the same lobe that were simultaneously resected by lobectomy, resulting in multifocal lung cancer). None of the SSNs presented as interval cancer. Seven of the eight resected lung cancers were stage Ia (n = 3, range of active surveillance 48 to 99 months; n = 1 protocol violation) or stage Ib (n = 3, range of active surveillance 17 to 80 months). One SSN was resected in stage IIIa after 95 months of active surveillance because of multifocal concurrent neoplastic foci in the same lobe and a metastatic interbronchial lymph node, about 2 cm from the primary site (T3N1M0) (Fig. 1). Twenty-four months after lung resection (119 months since the first detection of SSN), this patient developed an incident contralateral solid nodule that was followed for a further 27 months with evidence of growth and metabolic activity (SUVmax, 6.1); at this time the patient was 77 years old. The patient was assigned to stereotactic radiation therapy (detailed imaging findings are reported in the online supplement). The same patient also developed a gastrointestinal stromal tumor that was treated by wedge resection of the stomach.
      Figure thumbnail gr1
      Figure 1(A-M) Consecutive LDCT scans, 18FDG-PET, and histology show the history of multifocal lung cancer during active surveillance of a prevalent SSN. The prevalent SSN (A-F) showed slow evolution from November 2005 (A) through November 2008 (B) and January 2011 (C) until June 2013 (D) with progressive expansion of the solid component and positive 18FDG-PET (SUV = 5.4) in July 2013 (E), with ultimate resection with evidence of adenocarcinoma (F) (size of pathologic tissue, 13 mm; Ki-67 expression, 45%). The further nodule was not present at baseline in November 2005 (G), consequently appeared incidentally in November 2008 (H), grew through January 2011 (J) and June 2013 (K), until negative 18FDG-PET (L), with ultimate resection (lobectomy for the above principal site) and evidence of multifocal lung adenocarcinoma (M) (size of pathologic tissue, 8 mm; Ki-67 expression, 15%). Pathology showed also involvement of a lingular interbronchial lymph node approximately 2 cm from the principal site (size of pathologic tissue, 6 mm; Ki-67 expression, 35%), with final pT3N1M0 for multifocal lung adenocarcinoma of the left upper lobe. LDCT, low-dose computed tomography; 18FDG-PET, 18fluorodeoxiglucose positron-emission tomography; SSN, subsolid nodule; SUV, standard uptake value; Ki-67, marker of proliferation Ki-67.
      None of these subjects died from lung cancer within the follow-up period and no recurrences were observed during a median follow-up of 53 months after resection (range, 32 to 114 months). The integration by CAD analysis prompted earlier detection of six of eight (75%) prevalent SSNs developing into lung cancer.

      Lung Cancer Not Derived From SSNs

      Lung cancer not originating from the SSN was seen in 22 of 30 (73%) subjects with SSNs. Such lung cancers appeared after a median time of 52 months from detection of the SSN (range, 4 to 102 months), as detailed in Figure 2. Fifteen of 22 (68%) of these screenees were detected with SSNs only by CAD. Three of 22 (14%) cancers were diagnosed within 12 months of SSN detection (range, 4 to 11 months); they all grew from a dominant solid nodule detected at the same time as SSN. On the other hand, 19 of 22 (86%) cancers were diagnosed more than 24 months after SSN detection (range, 25 to 102 months), with only 3 of 19 (16%) cancers growing from the dominant solid nodule present at the time of SSN detection and the remaining 16 of 19 (84%) cancers growing either from nondominant nodule present at the time of SSN detection (4 of 16, 25%) or from nodules appeared after SSN detection (12 of 16, 75%).
      Figure thumbnail gr2
      Figure 2Timeline of events since the detection of SSN. The events are detailed as follows: diagnosis of lung cancer derived from SSN (n = 8), death from lung cancer derived from SSN (n = 0), diagnosis of lung cancer not derived from SSN (n = 22), death from lung cancer not derived from SSN (n = 8), diagnosis of extrapulmonary cancer (n = 43), death from extrapulmonary cancer (n = 7), and death from non-oncologic disease (n = 10). SSN, subsolid nodule.
      The stage at the time of diagnosis was Ia (n = 7, all adenocarcinomas), Ib (n = 3, two adenocarcinomas and one squamous cell carcinoma), IIIa (n = 4, two adenocarcinomas, one squamous cell carcinoma, and one large cell carcinoma), or IV (n = 7, five adenocarcinomas and two NSCLCs). The remaining one had a small cell lung cancer with limited disease. Resection was possible in 15 of 22 cases (11 lobectomies and 4 segmentectomies).
      The frequency of lung cancer in other areas of the lung was not significantly higher (p = 0.194) in subjects with volume of SSN above the median (14 lung cancers not derived from SSN in 195 subjects, 7.2%) compared to subjects below (eight lung cancers not derived from SSN in 194 subjects, 4.1%).

      Extrapulmonary Cancer

      Fifty-three subjects of 389 (13.6%) experienced an extrapulmonary cancer in their life, either before (10 of 389 [2.6%], median time 20 months before SSN detection, range 1 to 45 months) or after detection of SSN (43 of 389 [11%], median time 50 months after SSN detection, range 2 to 109 months).

      Survival of Subjects With SSN

      Twenty-five subjects (6.4%) died during the 9.3 ± 1.2 years of follow-up, the cause of death was as follows: 8 of 25 (32%) lung cancers not derived from SSN, 7 of 25 (28%) extrapulmonary cancers, and 10 of 25 (40%) non-neoplastic diseases (5 cardiovascular events, 2 respiratory failures, 2 acute abdominal diseases, and 1 Alzheimer’s disease). None of the screenees died because of lung cancer derived from SSN. Figure 2 provides the detailed timeline of events since the detection of SSN.
      During the follow-up period, lung cancer–specific survival among screenees with SSN-derived lung cancer was 100%, compared to 63.6% for lung cancers not derived from SSNs (stratified log-rank test p < 0.0001; Fig. 3).
      Figure thumbnail gr3
      Figure 3Kaplan-Meier plot shows the survival of screenees with SSN according to the diagnosed disease as grouped into four categories: 1) lung cancer derived from SSN (n = 8, including four cases with also extrapulmonary cancer; no death), 2) lung cancer not derived from SSN (n = 22, including 7 cases with also extrapulmonary cancer after detection of SSN; 8 deaths from lung cancer and 2 deaths from extrapulmonary cancer), 3) extrapulmonary cancer after detection of SSN (n = 42; 5 deaths from extrapulmonary cancer and 2 deaths from non-neoplastic disease), and 4) no malignancy (n = 320; 8 deaths from non-neoplastic disease). SSN, subsolid nodule.
      The HR for overall mortality was similar between subjects with SSN (HR = 1.34, 95% CI: 0.82–2.19) and subjects with solid nodules (HR = 1.21, 95% CI: 0.81–1.88), compared to subjects with no nodules. The HR for lung cancer–specific mortality was 3.80 (95% CI: 1.24–11.65) for subjects with SSN and 4.25 (95% CI: 1.59–11.34) for subjects with solid nodules.
      At the last LDCT round, 43 of 50 (86%) of PSNs were still under active surveillance (median overall size, 7.6 mm; range, 3.7 to 26.8 mm; median size of solid component if present, 3.4 mm; range, 1.1 to 10.6 mm). Fourteen PSNs with solid component greater than 5 mm were maintained with active surveillance based on the multidisciplinary team decision. The underlying reasons were lack of growth (n = 9) and minimal growth during long-term active surveillance (n = 5).

      Discussion

      This study shows that the majority of prognostically significant lung cancers did not derive from the SSN, but occurred in a different area of the lung, based on the longest active surveillance of SSN so far reported. Since the detection of SSN, the cause of death was never the SSN itself but always a different disease, namely: lung cancer not derived from SSN, extrapulmonary cancer, and non-neoplastic disease (Fig. 2). After a median active surveillance of 77 months (range, 17 to 99 months) before SSN resection, no screenee with lung cancer emerging from the SSN experienced disease relapse or death.
      Screening practice for any disease, either neoplastic or non-neoplastic, comes with the major issue of discerning clinically relevant findings from the large number of other abnormalities.
      • Esserman L.J.
      • Thompson Jr., I.M.
      • Reid B.
      Overdiagnosis and overtreatment in cancer: an opportunity for improvement.
      At cancer screening, the dichotomic classification of a finding into benign or malignant is everything but simple. It presents a continuous challenge to minimize psychologic, socioeconomic, and legal strain.
      • Walter L.C.
      • Covinsky K.E.
      Cancer screening in elderly patients: a framework for individualized decision making.
      This challenge is present for several tumor types, including lung cancer.
      • Patz Jr., E.F.
      • Pinsky P.
      • Gatsonis C.
      • et al.
      Overdiagnosis in low-dose computed tomography screening for lung cancer.
      In fact, lung cancer screening literature shows a large inter-individual variability of lung cancer aggressiveness.
      • Scholten E.T.
      • Horeweg N.
      • de Koning H.J.
      • et al.
      Computed tomographic characteristics of interval and post screen carcinomas in lung cancer screening.
      • Yousaf-Khan U.
      • van der Aalst C.
      • de Jong P.A.
      • et al.
      Final screening round of the NELSON lung cancer screening trial: the effect of a 2.5-year screening interval.
      • Silva M.
      • Galeone C.
      • Sverzellati N.
      • et al.
      Screening with low-dose computed tomography does not improve survival of small cell lung cancer.
      • Gierada D.S.
      • Pinsky P.F.
      • Duan F.
      • et al.
      Interval lung cancer after a negative CT screening examination: CT findings and outcomes in National Lung Screening Trial participants.
      SSNs, which are typically slow-growing, were found to be present in some 17% of subjects in MILD.
      • Yip R.
      • Wolf A.
      • Tam K.
      • et al.
      Outcomes of lung cancers manifesting as nonsolid nodules.
      • Silva M.
      • Sverzellati N.
      • Manna C.
      • et al.
      Long-term surveillance of ground-glass nodules: evidence from the MILD trial.
      • Yankelevitz D.F.
      • Yip R.
      • Smith J.P.
      • et al.
      CT screening for lung cancer: nonsolid nodules in baseline and annual repeat rounds.
      • Scholten E.T.
      • de Jong P.A.
      • de Hoop B.
      • et al.
      Towards a close computed tomography monitoring approach for screen detected subsolid pulmonary nodules?.
      • Yip R.
      • Henschke C.I.
      • Xu D.M.
      • et al.
      Lung cancers manifesting as part-solid nodules in the National Lung Screening Trial.
      We have shown that these SSNs are unlikely to develop into a deadly cancer within 10 years but that they are instead associated with concurrent deadly comorbidities, including more aggressive intra-individual lung cancer.
      Multifocal lung cancer was known as a relatively rare entity before the experience of lung cancer screening. In 2009, based on the International Early Lung Cancer Action Program (I-ELCAP) database, Vazquez et al.
      • Vazquez M.
      • Carter D.
      • Brambilla E.
      • et al.
      Solitary and multiple resected adenocarcinomas after CT screening for lung cancer: histopathologic features and their prognostic implications.
      reported individuals with multiple foci of lung cancer — notably adenocarcinoma — with extremely good prognosis at 10 years after resection. A recent survey from the International Association for the Study of Lung Cancer warned about the increasing incidence of multifocal lung cancer, which is expected to become more common in lung cancer screening.
      • Vazquez M.
      • Carter D.
      • Brambilla E.
      • et al.
      Solitary and multiple resected adenocarcinomas after CT screening for lung cancer: histopathologic features and their prognostic implications.
      • Leventakos K.
      • Peikert T.
      • Midthun D.E.
      • et al.
      Management of multifocal lung cancer: results of a survey.
      There is substantial heterogeneity of treatment choice among physicians (allegedly related to the lack of high-level evidence for therapeutic recommendation), thereby urging for prospective data to guide the management of multiple pulmonary lesions.
      • Leventakos K.
      • Peikert T.
      • Midthun D.E.
      • et al.
      Management of multifocal lung cancer: results of a survey.
      A systematic review of the literature on slow-growing adenocarcinoma derived from SSNs showed that the proportion of the solid component, but not the total size of the SSN, correlates with outcome measures, supporting a conservative management of slow growing tumors.
      • Yip R.
      • Li K.
      • Liu L.
      • et al.
      Controversies on lung cancers manifesting as part-solid nodules.
      The long-term outcome of subjects with SSNs from the MILD trial shows that it is a safe option to manage SSNs conservatively until signs of growth of the solid component. Although only a minority of SSNs were resected under the MILD protocol (8 of 389 including one protocol violation), none of the unresected persistent SSNs developed into a clinically perceivable cancer within the follow-up period of almost 10 years. All eight resected SSNs were adenocarcinomas, seven of which represented stage Ia or Ib. Our results are in concordance with the notion that persistent SSNs harbor adenocarcinomas with extremely low invasive potential.
      The MILD conservative management allowed a perceivable reduction of SSN resection compared to other screening trials that implemented a more aggressive approach to SSN.
      • Scholten E.T.
      • de Jong P.A.
      • de Hoop B.
      • et al.
      Towards a close computed tomography monitoring approach for screen detected subsolid pulmonary nodules?.
      • Yip R.
      • Henschke C.I.
      • Xu D.M.
      • et al.
      Lung cancers manifesting as part-solid nodules in the National Lung Screening Trial.
      • Henschke C.I.
      • Salvatore M.
      • Cham M.
      • et al.
      Baseline and annual repeat rounds of screening: implications for optimal regimens of screening.
      In the I-ELCAP study, the probability of SSN resection was 2.9/1000 person-years and 90% of SSNs (197 of 219) were resected at the baseline round, compared to only 0.37/1000 person-years and only one resected at baseline round (protocol violation) in the MILD trial. Fourteen NSNs were resected in I-ELCAP without histologic evidence of cancer.
      • Yankelevitz D.F.
      • Yip R.
      • Smith J.P.
      • et al.
      CT screening for lung cancer: nonsolid nodules in baseline and annual repeat rounds.
      • Henschke C.I.
      • Salvatore M.
      • Cham M.
      • et al.
      Baseline and annual repeat rounds of screening: implications for optimal regimens of screening.
      Within the MILD trial, no SSN was resected with ultimate diagnosis of benign disease, atypical adenomatous hyperplasia, or adenocarcinoma in situ. These two antipodal strategies of SSN management show no differences in mortality from SSN, hence early resection of SSN should indeed be classified as overdiagnosis and overtreatment.
      • Etzioni R.
      • Gulati R.
      Recognizing the limitations of cancer overdiagnosis studies: a first step towards overcoming them.
      We have also shown that subjects with SSN under active surveillance in a screening program will have a rate of growing SSNs requiring resection that is substantially lower than the rate of cancers developing elsewhere in their lungs, and also substantially lower than the rate of competing causes of death (extrapulmonary cancers and non-neoplastic disease). A SSN therefore represents a relevant biomarker on which personalized conservative management can be based.
      • Leventakos K.
      • Peikert T.
      • Midthun D.E.
      • et al.
      Management of multifocal lung cancer: results of a survey.
      • Yip R.
      • Li K.
      • Liu L.
      • et al.
      Controversies on lung cancers manifesting as part-solid nodules.
      • Thakur M.K.
      • Ruterbusch J.J.
      • Schwartz A.G.
      • et al.
      Risk of second lung cancer in patients with previously treated lung cancer: analysis of Surveillance, Epidemiology, and End Results (SEER) Data.
      In fact, delayed resection of SSN until signs of growth did not increase mortality in our series, validating the retrospective evidence from previous studies.
      • Yip R.
      • Wolf A.
      • Tam K.
      • et al.
      Outcomes of lung cancers manifesting as nonsolid nodules.
      • Yankelevitz D.F.
      • Yip R.
      • Smith J.P.
      • et al.
      CT screening for lung cancer: nonsolid nodules in baseline and annual repeat rounds.
      • Scholten E.T.
      • de Jong P.A.
      • de Hoop B.
      • et al.
      Towards a close computed tomography monitoring approach for screen detected subsolid pulmonary nodules?.
      • Yip R.
      • Li K.
      • Liu L.
      • et al.
      Controversies on lung cancers manifesting as part-solid nodules.
      • Henschke C.I.
      • Yip R.
      • Smith J.P.
      • et al.
      CT screening for lung cancer: part-solid nodules in baseline and annual repeat rounds.
      The conservative management of SSNs preserved the lung function in subjects that subsequently required resection of more aggressive cancer in other pulmonary sites. Early SSN resection could have caused deterioration of pulmonary function and quality of life, potentially hampering resectability of a second more aggressive lung cancer.
      • Thakur M.K.
      • Ruterbusch J.J.
      • Schwartz A.G.
      • et al.
      Risk of second lung cancer in patients with previously treated lung cancer: analysis of Surveillance, Epidemiology, and End Results (SEER) Data.
      • Sarna L.
      • Evangelista L.
      • Tashkin D.
      • et al.
      Impact of respiratory symptoms and pulmonary function on quality of life of long-term survivors of non-small cell lung cancer.
      • Falcoz P.E.
      • Conti M.
      • Brouchet L.
      • et al.
      The Thoracic Surgery Scoring System (Thoracoscore): risk model for in-hospital death in 15,183 patients requiring thoracic surgery.
      • Feinstein M.B.
      • Krebs P.
      • Coups E.J.
      • et al.
      Current dyspnea among long-term survivors of early-stage non-small cell lung cancer.
      Our results show for the first time that 73% (22 of 30) of clinically relevant lung cancers in subjects with SSN develop in the lung parenchyma away from the SSN, and that these lung cancers have a significantly poorer survival (mortality, 9 of 22) than those diagnosed in SSNs (mortality, none of eight) (Fig. 3). These data support the hypothesis that slow-growing lesions such as SSNs represent a phenotypical manifestation of individual lung cancer susceptibility, namely, a risk factor with more prognostic than diagnostic value. In this context, the retrospective application of CIRRUS software detected 68% of SSNs associated with lung cancer in other areas of the lung, which were not seen by visual reading. This suggests that CAD might allow the selection of subjects at higher risk of lung cancer, namely, those to be monitored by more intense surveillance.
      • Schreuder A.
      • Schaefer-Prokop C.M.
      • Scholten E.T.
      • et al.
      Lung cancer risk to personalise annual and biennial follow-up computed tomography screening.
      Hence, we believe the integration of CAD will provide remarkable information for risk stratification within lung cancer screening. Furthermore, we tested the potential association between the volume of SSN and the frequency of lung cancer in other sites of the lung, which showed a nonstatistically significant higher frequency in subjects with volume of SSN above median (7.2%) compared with those below (4.1%). The relatively small number of this population might have limited the statistical power of such association; we foster future investigation based on larger populations will undertake this question.
      Henschke et al.
      • Henschke C.I.
      • Yip R.
      • Smith J.P.
      • et al.
      CT screening for lung cancer: part-solid nodules in baseline and annual repeat rounds.
      reported an anecdotal death from myocardial infarction a few months after SSN resection in the I-ELCAP. Large-scale clinical series support the effect of cardiopulmonary stress of lung resection and show the impact of cardiovascular disease on survival of resected early stage lung cancer.
      • Falcoz P.E.
      • Conti M.
      • Brouchet L.
      • et al.
      The Thoracic Surgery Scoring System (Thoracoscore): risk model for in-hospital death in 15,183 patients requiring thoracic surgery.
      • Kravchenko J.
      • Berry M.
      • Arbeev K.
      • et al.
      Cardiovascular comorbidities and survival of lung cancer patients: Medicare data based analysis.
      Screening participants are selected among heavy smokers with relatively high risk of coronary artery disease and obstructive/restrictive pulmonary disease.
      • Sverzellati N.
      • Cademartiri F.
      • Bravi F.
      • et al.
      Relationship and prognostic value of modified coronary artery calcium score, FEV1, and emphysema in lung cancer screening population: the MILD trial.
      • Wille M.M.
      • Dirksen A.
      • Ashraf H.
      • et al.
      Results of the randomized Danish lung cancer screening trial with focus on high-risk profiling.
      Mortality due to extrapulmonary malignancies and non-oncologic comorbidities was particularly high in subjects with SSN. Five of eight patients with resected SSN lung cancer developed an extrapulmonary malignancy, suggesting a higher degree of individual susceptibility of subjects with SSNs. There is still much to investigate on the clinical significance of SSN as a biologic risk factor, rather than a marker of a slow-growing malignancy at the SSN site.
      Our study suffers from limitations. First, a proportion of the SSNs were detected retrospectively, using advanced detection software not yet available at the time of screening. However, such a retrospective approach provides a picture of the SSN frequency under future technical conditions. In fact, the performance of CAD software is increasing with the advent of deep learning approaches and, as a result, we believe that CAD will eventually be integrated into the reading process for lung cancer screening (e.g., reduction of reading time and increase in sensitivity and negative predictive value).
      • Zhao Y.
      • de Bock G.H.
      • Vliegenthart R.
      • et al.
      Performance of computer-aided detection of pulmonary nodules in low-dose CT: comparison with double reading by nodule volume.
      • Rubin G.D.
      • Roos J.E.
      • Tall M.
      • et al.
      Characterizing search, recognition, and decision in the detection of lung nodules on CT scans: elucidation with eye tracking.
      • Liang M.
      • Tang W.
      • Xu D.M.
      • et al.
      Low-dose CT screening for lung cancer: computer-aided detection of missed lung cancers.
      Second, the number of participants in this screening trial was relatively small compared to the NLST trial.
      • Aberle D.R.
      • Adams A.M.
      • et al.
      National Lung Screening Trial Research Team
      Reduced lung-cancer mortality with low-dose computed tomographic screening.
      Therefore, we wait for future studies to confirm our results in larger populations. Third, the conservative approach of the MILD protocol prevented histologic assessment of most SSNs. As a consequence, an extremely limited number of lung cancers emerging from these SSNs were reported in almost a decade (8 lung cancers in 2303 screenees) because only those with growing lesions were considered for resection. Finally, these results have been derived in the setting of lung cancer screening, where subjects are selected without a history of cancer in the previous 5 years or more. Caution must be exerted for generalization to other settings.
      In conclusion, the majority of subjects with SSNs who were diagnosed with lung cancer in the MILD cohort had developed a cancer elsewhere in the lungs. Lung cancers that arose from the SSN never represented the cause of death within the nearly 10-year follow-up period. Therefore, SSNs can be considered biomarkers of cancer risk, and should be managed by active surveillance until signs of growth of the solid component. This approach will reduce unnecessary surgery with cardiopulmonary damage in subjects with multiple comorbidities, including more aggressive lung cancers arising from lung sites other than the SSN. We suggest that subjects with SSNs might be the suitable target population for pharmacological smoking-cessation and chemoprevention trials.

      Acknowledgments

      The Multicenter Italian Lung Detection (MILD) trial was supported by grants from the Italian Ministry of Health ( RF 2004 ), the Italian Association for Cancer Research ( AIRC 2004 IG 1227 and AIRC 5xmille IG 12162 ), and Fondazione Cariplo ( 2004-1560 ). The authors thank Claudio Jacomelli and Claudio Citterio for data management; the MILD staff (Elena Bertocchi, Annamaria Calanca, Carolina Ninni, Stefano Sestini, and Paola Suatoni); and Drs. Antonello Cabras and Alessio Pellegrinelli for the iconographic contribution of pathology specimens.

      Supplementary Data

      References

        • Aberle D.R.
        • Adams A.M.
        • et al.
        • National Lung Screening Trial Research Team
        Reduced lung-cancer mortality with low-dose computed tomographic screening.
        N Engl J Med. 2011; 365: 395-409
        • Esserman L.J.
        • Thompson Jr., I.M.
        • Reid B.
        Overdiagnosis and overtreatment in cancer: an opportunity for improvement.
        JAMA. 2013; 310: 797-798
        • Patz Jr., E.F.
        • Pinsky P.
        • Gatsonis C.
        • et al.
        Overdiagnosis in low-dose computed tomography screening for lung cancer.
        JAMA Intern Med. 2014; 174: 269-274
        • Bach P.B.
        Overdiagnosis in lung cancer: different perspectives, definitions, implications.
        Thorax. 2008; 63: 298-300
        • Carter J.L.
        • Coletti R.J.
        • Harris R.P.
        Quantifying and monitoring overdiagnosis in cancer screening: a systematic review of methods.
        BMJ. 2015; 350: g7773
        • Kakinuma R.
        • Noguchi M.
        • Ashizawa K.
        • et al.
        Natural history of pulmonary subsolid nodules: a prospective multicenter study.
        J Thorac Oncol. 2016; 11: 1012-1028
        • Travis W.D.
        • Asamura H.
        • Bankier A.A.
        • et al.
        The IASLC Lung Cancer Staging Project: proposals for coding T categories for subsolid nodules and assessment of tumor size in part-solid tumors in the forthcoming eighth edition of the TNM Classification of Lung Cancer.
        J Thorac Oncol. 2016; 11: 1204-1223
        • Yip R.
        • Wolf A.
        • Tam K.
        • et al.
        Outcomes of lung cancers manifesting as nonsolid nodules.
        Lung Cancer. 2016; 97: 35-42
        • Revel M.P.
        • Mannes I.
        • Benzakoun J.
        • et al.
        Subsolid lung nodule classification: a CT criterion for improving interobserver agreement.
        Radiology. 2018; 286: 316-325
        • Heidinger B.H.
        • Anderson K.R.
        • Nemec U.
        • et al.
        Lung adenocarcinoma manifesting as pure ground-glass nodules: correlating CT size, volume, density, and roundness with histopathologic invasion and size.
        J Thorac Oncol. 2017; 12: 1288-1298
        • Scholten E.T.
        • Horeweg N.
        • de Koning H.J.
        • et al.
        Computed tomographic characteristics of interval and post screen carcinomas in lung cancer screening.
        Eur Radiol. 2015; 25: 81-88
      1. Lung-RADS Version 1.0 Assessment Categories. April 28, 2014. http://www.acr.org/∼/media/ACR/Documents/PDF/QualitySafety/Resources/LungRADS/AssessmentCategories. Accessed January 1 2017.

        • Pinsky P.F.
        • Gierada D.S.
        • Black W.
        • et al.
        Performance of lung-RADS in the National Lung Screening Trial: a retrospective assessment.
        Ann Intern Med. 2015; 162: 485-491
      2. National Comprehensive Cancer Network Guidelines, Version 3.2018, Lung Cancer Screening. https://www.nccn.org/professionals/physician_gls/pdf/lung_screening.pdf. Accessed March 19, 2018.

        • Kotwal A.A.
        • Schonberg M.A.
        Cancer screening in the elderly: a review of breast, colorectal, lung, and prostate cancer screening.
        Cancer J. 2017; 23: 246-253
        • Flores R.
        • Taioli E.
        • Yankelevitz D.F.
        • et al.
        Initiative for early lung cancer research on treatment: development of study design and pilot implementation.
        J Thorac Oncol. 2018; 13: 946-957
        • Bach P.B.
        • Mirkin J.N.
        • Oliver T.K.
        • et al.
        Benefits and harms of CT screening for lung cancer: a systematic review.
        JAMA. 2012; 307: 2418-2429
        • Leduc C.
        • Antoni D.
        • Charloux A.
        • et al.
        Comorbidities in the management of patients with lung cancer.
        Eur Respir J. 2017; 49: 1601721
        • Silva M.
        • Sverzellati N.
        • Manna C.
        • et al.
        Long-term surveillance of ground-glass nodules: evidence from the MILD trial.
        J Thorac Oncol. 2012; 7: 1541-1546
        • Sverzellati N.
        • Silva M.
        • Calareso G.
        • et al.
        Low-dose computed tomography for lung cancer screening: comparison of performance between annual and biennial screen.
        Eur Radiol. 2016; 26: 3821-3829
        • Jacobs C.
        • van Rikxoort E.M.
        • Twellmann T.
        • et al.
        Automatic detection of subsolid pulmonary nodules in thoracic computed tomography images.
        Med Image Anal. 2014; 18: 374-384
        • Charbonnier J.P.
        • Chung K.
        • Scholten E.T.
        • et al.
        Automatic segmentation of the solid core and enclosed vessels in subsolid pulmonary nodules.
        Sci Rep. 2018; 8: 646
        • Horeweg N.
        • Scholten E.T.
        • de Jong P.A.
        • et al.
        Detection of lung cancer through low-dose CT screening (NELSON): a prespecified analysis of screening test performance and interval cancers.
        Lancet Oncol. 2014; 15: 1342-1350
        • Detterbeck F.C.
        • Chansky K.
        • Groome P.
        • et al.
        The IASLC Lung Cancer Staging Project: methodology and validation used in the development of proposals for revision of the stage classification of NSCLC in the forthcoming (eighth) edition of the TNM Classification of Lung Cancer.
        J Thorac Oncol. 2016; 11: 1433-1446
        • Xie J.
        • Liu C.
        Adjusted Kaplan-Meier estimator and log-rank test with inverse probability of treatment weighting for survival data.
        Stat Med. 2005; 24: 3089-3110
        • Walter L.C.
        • Covinsky K.E.
        Cancer screening in elderly patients: a framework for individualized decision making.
        JAMA. 2001; 285: 2750-2756
        • Yousaf-Khan U.
        • van der Aalst C.
        • de Jong P.A.
        • et al.
        Final screening round of the NELSON lung cancer screening trial: the effect of a 2.5-year screening interval.
        Thorax. 2016; 72: 48-56
        • Silva M.
        • Galeone C.
        • Sverzellati N.
        • et al.
        Screening with low-dose computed tomography does not improve survival of small cell lung cancer.
        J Thorac Oncol. 2016; 11: 187-193
        • Gierada D.S.
        • Pinsky P.F.
        • Duan F.
        • et al.
        Interval lung cancer after a negative CT screening examination: CT findings and outcomes in National Lung Screening Trial participants.
        Eur Radiol. 2017; 27: 3249-3256
        • Yankelevitz D.F.
        • Yip R.
        • Smith J.P.
        • et al.
        CT screening for lung cancer: nonsolid nodules in baseline and annual repeat rounds.
        Radiology. 2015; 277: 555-564
        • Scholten E.T.
        • de Jong P.A.
        • de Hoop B.
        • et al.
        Towards a close computed tomography monitoring approach for screen detected subsolid pulmonary nodules?.
        Eur Respir J. 2015; 45: 765-773
        • Yip R.
        • Henschke C.I.
        • Xu D.M.
        • et al.
        Lung cancers manifesting as part-solid nodules in the National Lung Screening Trial.
        AJR Am J Roentgenol. 2017; : 1-11
        • Vazquez M.
        • Carter D.
        • Brambilla E.
        • et al.
        Solitary and multiple resected adenocarcinomas after CT screening for lung cancer: histopathologic features and their prognostic implications.
        Lung Cancer. 2009; 64: 148-154
        • Leventakos K.
        • Peikert T.
        • Midthun D.E.
        • et al.
        Management of multifocal lung cancer: results of a survey.
        J Thorac Oncol. 2017; 12: 1398-1402
        • Yip R.
        • Li K.
        • Liu L.
        • et al.
        Controversies on lung cancers manifesting as part-solid nodules.
        Eur Radiol. 2018; 28: 747-759
        • Henschke C.I.
        • Salvatore M.
        • Cham M.
        • et al.
        Baseline and annual repeat rounds of screening: implications for optimal regimens of screening.
        Eur Radiol. 2018; 28a: 1085-1094
        • Etzioni R.
        • Gulati R.
        Recognizing the limitations of cancer overdiagnosis studies: a first step towards overcoming them.
        J Natl Cancer Inst. 2015; 108 (piidjv345)
        • Thakur M.K.
        • Ruterbusch J.J.
        • Schwartz A.G.
        • et al.
        Risk of second lung cancer in patients with previously treated lung cancer: analysis of Surveillance, Epidemiology, and End Results (SEER) Data.
        J Thorac Oncol. 2018; 13: 46-53
        • Henschke C.I.
        • Yip R.
        • Smith J.P.
        • et al.
        CT screening for lung cancer: part-solid nodules in baseline and annual repeat rounds.
        AJR Am J Roentgenol. 2016; 207: 1176-1184
        • Sarna L.
        • Evangelista L.
        • Tashkin D.
        • et al.
        Impact of respiratory symptoms and pulmonary function on quality of life of long-term survivors of non-small cell lung cancer.
        Chest. 2004; 125: 439-445
        • Falcoz P.E.
        • Conti M.
        • Brouchet L.
        • et al.
        The Thoracic Surgery Scoring System (Thoracoscore): risk model for in-hospital death in 15,183 patients requiring thoracic surgery.
        J Thorac Cardiovasc Surg. 2007; 133: 325-332
        • Feinstein M.B.
        • Krebs P.
        • Coups E.J.
        • et al.
        Current dyspnea among long-term survivors of early-stage non-small cell lung cancer.
        J Thorac Oncol. 2010; 5: 1221-1226
        • Schreuder A.
        • Schaefer-Prokop C.M.
        • Scholten E.T.
        • et al.
        Lung cancer risk to personalise annual and biennial follow-up computed tomography screening.
        Thorax. 2018; 73: 626-633
        • Kravchenko J.
        • Berry M.
        • Arbeev K.
        • et al.
        Cardiovascular comorbidities and survival of lung cancer patients: Medicare data based analysis.
        Lung Cancer. 2015; 88: 85-93
        • Sverzellati N.
        • Cademartiri F.
        • Bravi F.
        • et al.
        Relationship and prognostic value of modified coronary artery calcium score, FEV1, and emphysema in lung cancer screening population: the MILD trial.
        Radiology. 2012; 262: 460-467
        • Wille M.M.
        • Dirksen A.
        • Ashraf H.
        • et al.
        Results of the randomized Danish lung cancer screening trial with focus on high-risk profiling.
        Am J Respir Crit Care Med. 2016; 193: 542-551
        • Zhao Y.
        • de Bock G.H.
        • Vliegenthart R.
        • et al.
        Performance of computer-aided detection of pulmonary nodules in low-dose CT: comparison with double reading by nodule volume.
        Eur Radiol. 2012; 22: 2076-2084
        • Rubin G.D.
        • Roos J.E.
        • Tall M.
        • et al.
        Characterizing search, recognition, and decision in the detection of lung nodules on CT scans: elucidation with eye tracking.
        Radiology. 2015; 274: 276-286
        • Liang M.
        • Tang W.
        • Xu D.M.
        • et al.
        Low-dose CT screening for lung cancer: computer-aided detection of missed lung cancers.
        Radiology. 2016; 281: 279-288