Abstract
Objective To review new evidence reported since the 2016 publication of the Canadian Task Force on Preventive Health Care recommendations and to summarize key facets of lung cancer screening to better equip primary care providers (PCPs) in anticipation of wider implementation of the recommendations.
Quality of evidence A new, large randomized controlled trial has been published since 2016, as have updates from 4 other trials. PubMed was searched for studies published between January 1, 2004, and December 31, 2020, using search words including lung cancer screening eligibility, lung cancer screening criteria, and lung cancer screening guidelines. All information from peer-reviewed articles, reference lists, books, and websites was considered.
Main message Lung cancers diagnosed at stage 4 have a 5-year survival rate of only 5% and have a disproportionate impact on those with lower socioeconomic status, rural populations, and Indigenous populations. By downstaging, or diagnosing lung cancers at an earlier and more treatable stage, lung cancer screening reduces mortality with a number needed to screen of 250 to prevent 1 death. Practical aspects of lung cancer screening are reviewed, including criteria to screen, appropriate low-dose computed tomography screening, and management of findings. Harms of screening, such as overdiagnosis and incidental findings, are discussed to allow PCPs to appropriately counsel their patients in the face of ongoing implementation of new lung cancer screening programs.
Conclusion Lung cancer screening, with its embedded emphasis on smoking cessation, is an excellent addition to PCPs’ preventive health care tools. The implementation of formal and pilot lung cancer screening programs across Canada means that PCPs will be increasingly required to counsel their patients around the uptake of lung cancer screening.
Lung cancer kills more people annually than colon, breast, and pancreatic cancers combined, accounting for 1 in 4 cancer deaths.1 The 5-year survival rate for lung cancer is 19%, primarily because approximately half of all lung cancers are diagnosed at stage 4, when survival is abysmal. Seventy-two percent of lung cancers are attributable to smoking, creating an opportunity for targeted screening for lung cancer. The primary benefit of lung cancer screening is downstaging—diagnosing lung cancers at an earlier stage when there are potential curative therapies with improved survival rates. Lung cancers diagnosed at stage 1 have a 3-year net survival rate of 71%, compared with 5% at stage 4.1
In 2016, the Canadian Task Force on Preventive Health Care (CTFPHC) recommended lung cancer screening (Box 1).2 Across Canada, there is currently a patchwork of lung cancer screening implementation, including an organized program in Ontario as of April 1, 2021, a program in British Columbia starting in 2022, and pilot programs in Quebec and Alberta. Business case proposals are being submitted by several other provinces. In the absence of organized programs, opportunistic screening is ongoing, a practice that potentially has increased harms, such as regular-dose computed tomography (CT) that has 4 to 5 times higher radiation, unnecessary repeat imaging studies, or biopsies, and there is no organized manner for abnormalities to be worked up.3
2016 CTFPHC lung cancer screening recommendation
The following is a weak recommendation with low-quality evidence.
In those 55 to 74 years of age who currently smoke or who quit within the past 15 years, and have ≥ 30 pack-years’ smoking history, screen with low-dose CT for up to 3 consecutive years, in centres with adequate expertise
CT—computed tomography, CTFPHC—Canadian Task Force on Preventive Health Care.
Data from the CTFPHC.2
The number needed to screen (NNS) to prevent 1 lung cancer death is 250,4 which compares favorably to the optimal NNS for mammography of 645.5 A simulation of lung cancer screening over the next 20 years projected 7000 to 17 000 fewer stage 4 diagnoses and 5000 to 11 000 fewer deaths should screening be implemented.1
Primary care provider (PCP) understanding of eligibility for screening and associated harms is suboptimal.6 This update on lung cancer screening will review new evidence reported since the 2016 publication of the CTFPHC recommendations and will summarize key facets of lung cancer screening to better equip PCPs in anticipation of wider implementation.
Quality of evidence
Since the meta-analysis that informed the 2016 CTFPHC recommendations,7 a new large randomized controlled trial, NELSON (Nederlands–Leuvens Longkanker Screenings Onderzoek),8 has been published, as have updates on several other trials (Table 1).8-12 We searched PubMed for studies published between January 1, 2004, and December 31, 2020, using search words including lung cancer screening eligibility, lung cancer screening criteria, and lung cancer screening guidelines. All information from peer-reviewed articles, reference lists, books, and websites was considered.
Lung cancer screening trials published since 2016
Main message
New evidence. A significant relative reduction in lung cancer mortality was shown in NELSON (24%) and the National Lung Screening Trial (NLST) (20%).8-10 Other smaller randomized controlled trials were not sufficiently powered to show a mortality difference due to screening; however, specific issues in low-dose CT (LDCT) screening were addressed.13 The MILD (Multicentric Italian Lung Detection) study highlighted that prolonged annual or biennial screening beyond 5 years is needed to show a lung cancer mortality reduction benefit.11 The German LUSI (Lung Cancer Screening Intervention) trial12 suggested women may benefit more from LDCT screening than men (Table 1).8-12 Recent meta-analyses have reported statistically and clinically significant relative reductions in lung cancer death of 17%14,15 and 19%,4 but no difference in all-cause mortality.
Cost effectiveness of LDCT. Low-dose CT is cost-effective, with an incremental cost of $20 724 per quality-adjusted life-year gained.16 Cost effectiveness is driven not only by lung cancer outcomes, but also by a reduction in tobacco-related diseases. The cost savings of lung cancer screening are amplified when the rising costs of immunotherapy and targeted therapy in stage 3 and 4 lung cancer treatment are considered. The average cost to screen an individual once is $500,16 while the cost of a single dose of pembrolizumab is $8800,17 which is given every 21 days for up to 2 years in metastatic lung cancer.
Who should be screened? Lung cancer screening is unique because it targets a specific high-risk group. More accurate identification of this high-risk cohort is critical to avoid harms of screening in those who have a lower incidence of lung cancer. The subgroup analysis of the NLST divided participants into risk quintiles and found an NNS of 61 in the highest-risk subgroup, compared with an NNS of 5276 in the lowest-risk subgroup. This translated into a decrease in false-positive results from 1648 in the low-risk quintile to only 65 in the high-risk quintile.18
The PLCOm2012 is a risk calculator that is available online (https://brocku.ca/lung-cancer-screening-and-risk-prediction/risk-calculators/) and has been validated as a superior method to determine eligibility for screening.19 This tool incorporates several predictors to generate a percentage risk of developing lung cancer in the next 6 years (Box 2).19 The prospective observational PanCan (Pan-Canadian Early Detection of Lung Cancer) study was not included in the CTFPHC guidelines or meta-analyses and assessed the efficacy of an earlier version of this risk-based model to select individuals for lung cancer screening with a 6-year lung cancer risk of 2% or greater. The PanCan study had an improved detection of lung cancer with statistically significant downstaging, with 77% of screen-detected lung cancers diagnosed at stage 1 or 2.20 When a PLCOm2012 score cutoff of 1.7% was compared with the criteria of age and pack-years alone, 18.5% more lung cancers were found when screening the same number of people.21 Primary care providers should be aware that Canadian screening programs are using risk stratification models like the PLCOm2012 to determine eligibility for screening. Individual health jurisdictions may adjust their screening thresholds based on health care resources. British Columbia screens individuals with a greater than 1.5% risk of developing lung cancer within a 6-year period, while Ontario uses a cutoff of 2% or greater.19,22
PLCOm2012 predictors
The following are predictors used by the PLCOm2012 to generate a percentage risk of developing lung cancer in the next 6 years:
Age
Education
Family history of lung cancer
Body mass index
Chronic obstructive pulmonary disease
Smoking duration
Smoking intensity
Smoking quit time
Personal history of cancer
Race or ethnic origin
Data from Tammemägi et al.19
What should screening consist of? Low-dose CT, in the form of multidetector helical CT with a slice thickness of 1.25 mm or less, is the only imaging modality recommended for lung cancer screening. There is no contrast required for LDCT and the radiation dose is 1.5 mSv or less, more than 4 times less than that of a traditional CT scan (8 mSv).23
The optimal frequency and length of screening are not yet known. Although the CTFPHC recommends 3 annual screens, prolonged annual or biennial screening beyond 5 years shows a lung cancer mortality reduction benefit11 (Table 1).8-12 The US Preventive Services Task Force recommends annual screening until the upper age limit of 80, unless the screenee is no longer eligible owing to comorbidities.24 A personalized approach to screening intervals is under investigation, with the aim of optimally balancing the costs and harms of screening with the risk of a missed cancer diagnosis.25
How should findings be managed? Nodules can be solid, semisolid, or nonsolid (eg, ground-glass nodules), and size can be expressed as diameter or volume. Calcified or “popcorn” and perifissural nodules are generally benign and do not require further follow-up.25 The greatest predictors of malignancy are size, rate of growth, and increase in density of semisolid or nonsolid nodules. An increase of 1.5 mm or greater in the diameter of an existing solid nodule is considered substantial and should be investigated. The appearance of new nodules after a baseline or subsequent LDCT is more suggestive of malignancy, triggering investigation at a smaller diameter.26 New nonsolid nodules have less malignant potential, and large nodules (> 20 mm) that appear suddenly are more likely infectious or inflammatory.27
In an organized screening program, standardized reports would be provided to the PCP and the screenee, indicating appropriate follow-up. For individuals with a low risk of lung cancer, this would include routine annual or biennial screening, while those individuals with nodules that should be monitored would be recalled for repeat LDCT in 1 to 6 months. Findings that arouse suspicion of cancer should trigger referral for diagnostic workup.25 In the absence of standardized reporting, Table 2 provides a guide to management.13,19,25 These recommendations are based on the Canadian Partnership Against Cancer Pan-Canadian Lung Cancer Screening Network evidence-based framework for management of screen-detected lung nodules.25
Management of screen-detected lung nodules
What are the harms of screening? The application of targeted screening selection criteria and standardized lung nodule workup protocols are paramount in mitigating harms of screening. False-positive findings are not uncommon and precipitate a workup that may include further radiation exposure, unnecessary procedures, and psychological distress. A systematic review of psychological distress associated with lung cancer screening showed that individuals with indeterminate scan results had an initial increase in psychological burden, but after 6 months, health-related quality of life and anxiety had returned to baseline.28 The NELSON trial had a 1.2% false-positive rate (264 of 22 600 participants), with 23% of these individuals undergoing invasive procedures.8 The NLST had a higher false-positive rate of 23%, which was related to the definition of positive screen findings.9 Protocols for lung nodule management have subsequently been refined.25,29 This is why organized screening instead of opportunistic or ad hoc screening is important.
Incidental findings in the thyroid, heart, lung, kidneys, adrenal glands, and liver are common in lung cancer screening. The identification of coronary calcification is also common and can trigger important lifestyle and risk management discussions for these patients.30 Although up to 20% of incidental findings will require investigation, their clinical significance is less than 1%.31,32 At about $12 per screen, the workup of incidental findings contributes substantially to the cost of screening programs.33 In the absence of specific recommendations in the radiology report, Table 3 provides a guide to management.34-36
Management of common incidental findings
Overdiagnosis refers to the diagnosis of cancers that would not have become clinically significant during the life of the individual. A comparison of the incidence of lung cancer in the screening and control groups over time generally allows for an estimation of this value. The 12.3-year follow-up of the NLST yielded an overdiagnosis estimate of 3%,9 the NELSON trial with 10 years of follow-up had an estimate of 8.9%,8 and the US Preventive Services Task Force has modeled a rate of 10% to 12%.24 A longer follow-up duration is required to determine the true rate of overdiagnosis.
Smoking cessation. A pillar of any lung cancer screening program must be an integrated smoking cessation program. Individuals presenting for screening may be uniquely receptive to smoking cessation given their increased awareness of the health impacts of smoking. The Alberta Lung Cancer Screening Study demonstrated the efficacy of embedded smoking cessation with a quit rate of 13%, more than double that of the general population (5%).37 Increased smoking cessation has been modeled to be cost-effective and, independent of screening, would prevent the development of lung cancers and improve mortality.38
Equity. A higher smoking incidence is strongly associated with lower socioeconomic status (SES), rurality, and Indigenous populations. The smoking rate in the highest income quintile is only 12% compared with 22% in the lowest income quintile. Smoking rates in First Nations people living off reserves are twice as high as in non-Indigenous populations, and 69% of Inuit living in Nunavut are daily smokers.39-41 As a result, these populations are disproportionately diagnosed with lung cancer. The lung cancers diagnosed in these individuals are typically later stage (stage 3 or 4), resulting in poorer survival rates compared with lung cancers diagnosed in urban populations, non-Indigenous individuals, and those of higher SES.1 Inequitable access to health care means that even when cancers are diagnosed at the same stage, those with lower incomes will have lower survival rates than individuals with higher incomes. Lung cancer screening presents a unique opportunity to address the disparities in health outcomes for those with low SES, Indigenous populations, and rural populations. For example, attaching lung screening to an in-community general health check (eg, detection of chronic obstructive pulmonary disease using spirometry and CT evidence of emphysema, detection of coronary artery disease from the screening LDCT, smoking cessation counseling), using telehealth, providing free transportation to screening centres, or using mobile screening may help to overcome the access barriers. Improved access to smoking cessation programs and coordinated diagnosis and management of lung cancers have considerable potential to have survival benefits greater than those reported in clinical trials.
Conclusion
This update has provided a review of the key evidence and the harms and benefits of screening, as well as the practicalities of results management, to better equip family doctors to support their patients. Lung cancer screening, with its embedded emphasis on smoking cessation, is an excellent addition to PCPs’ armamentarium of preventive health care tools. The innate ability of lung cancer screening to target people who have smoked or who are still smoking who are often found in underserved or underserviced populations with the poorest lung cancer outcomes provides a key means for PCPs to standardize access to care for their patient populations. The implementation of formal and pilot lung cancer screening programs across Canada means that PCPs will be increasingly required to counsel their patients on the uptake of lung cancer screening.
Notes
Editor’s key points
▸ Lung cancer screening is unique because it targets a specific high-risk group. More accurate identification of this high-risk cohort is critical to avoid harms of screening for those who have a lower incidence of lung cancer. Primary care providers should be aware that Canadian screening programs are using risk stratification models like the PLCOm2012 to determine eligibility for screening. Individual health jurisdictions may adjust their screening thresholds based on health care resources.
▸ Low-dose computed tomography (LDCT) is the only imaging modality recommended for lung cancer screening. There is no contrast required for LDCT and the radiation dose is 4 times less than a standard computed tomography scan. The optimal frequency and length of screening are not yet known. The greatest predictors of malignancy are size, rate of growth, and increase in density of semisolid or nonsolid nodules. An increase of 1.5 mm or greater in the diameter of an existing solid nodule is considered substantial and should be investigated.
▸ In an organized screening program, standardized reports would be provided to the primary care provider and the screenee, indicating appropriate follow-up. For individuals with a low risk of lung cancer, this would include routine annual or biennial screening, while those individuals with nodules that should be monitored would be recalled for repeat LDCT in 1 to 6 months.
Footnotes
Contributors
Both authors contributed to the literature review and interpretation, and to preparing the manuscript for submission.
Competing interests
None declared
This article is eligible for Mainpro+ certified Self-Learning credits. To earn credits, go to www.cfp.ca and click on the Mainpro+ link.
This article has been peer reviewed.
Cet article se trouve aussi en français à la page 823.
- Copyright © 2021 the College of Family Physicians of Canada