Recommendations on screening for lung cancer

Benefits of screening

The systematic review performed for the task force included 33 studies on lung cancer screening; 13 RCTs studied the benefits of screening.12 Seven low-quality studies evaluated screening with chest radiography (with or without sputum cytology), compared with no screening or less intensive screening (e.g., screening with chest radiography at longer intervals or advice to have a chest radiograph) and found small benefits in terms of early disease detection. Screening with chest radiography detected more early-stage and fewer late-stage lung cancers compared with groups receiving usual care. However, such screening did not reduce lung cancer specific mortality (risk ratio [RR] 0.99, 95% confidence interval [CI] 0.92–1.07]) or all-cause mortality (RR 0.98, 95% CI 0.96–1.00) when compared with usual care.12

Three low-quality trials compared annual screening with low-dose CT to no screening or usual care and found no difference in lung cancer-specific mortality (RR 1.30, 95% CI 0.81–2.11) or all-cause mortality (RR 1.38, 95% CI 0.86–2.22) after five years or less of follow-up.12 A recent update of the literature search found one randomized study15 presenting interim results on early detection of lung cancer with low-dose multislice CT (compared with no screening) in Germany after five years of follow-up. The addition of these new results to the all-cause mortality analysis did not significantly alter the results (RR 1.20, 95% CI 0.83–1.73).

Two studies were found that compared low-dose CT to chest radiography. Although screening with chest radiography is not considered standard care in Canada, these studies were included in the evidence review because prior studies had shown no differences in mortality outcomes between chest radiography and usual care. One of the studies included mortality outcomes (National Lung Screening Trial).6 The National Lung Screening Trial (a high-quality RCT) reported a 15% reduction in lung cancer mortality (RR 0.85, 95% CI 0.75–0.96) and a 6% reduction in all-cause mortality (RR 0.94, 95% CI 0.88–1.00) associated with screening with low-dose CT compared with chest radiography after 6.5 years of follow-up.16 This means that screening 1000 people with low-dose CT three times at one-year intervals prevents three deaths from lung cancer compared with screening with chest radiography (number needed to screen = 322). Screening with low-dose CT reduced the absolute risk of lung cancer mortality by 0.31% and of all-cause mortality by 0.46%.12 Low-dose CT also detected significantly more cases of early-stage lung cancer (8 more per 1000 people screened) and significantly fewer cases of late-stage lung cancer (4 fewer per 1000 people screened) compared with chest radiography. Results from the second study17 comparing low-dose CT and chest radiography were not pooled with results of the National Lung Screening Trial in the systematic review because of an incompatible follow-up period (≤ 12 mo), the small number of reported events (lung cancers) and no reporting of mortality outcomes.12

Evidence from four studies showed no significant differences in rates of smoking cessation between the screened (low-dose CT or chest radiography) groups and the control groups.12 The risk of bias for these studies was unclear because of the self-reported nature of this outcome.

Harms of screening

The harms of screening and invasive follow-up tests were informed by 31 studies, many with observational designs.12 The main harms included false-positive results, death or major complications from invasive follow-up testing and overdiagnosis.

Data from the National Lung Screening Trial suggest that of every 1000 people screened three times with low-dose CT at one-year intervals, 391 would have at least one positive result, 40 would have lung cancer and 351 would have a false-positive result.6^,12 As such, follow-up tests, including minor invasive procedures (e.g., bronchoscopy, needle biopsy) or major invasive procedures (e.g., thoracotomy, thoracoscopy), are needed to determine whether a positive low-dose CT result is due to lung cancer. Although uncommon, there is a risk of major complications or death with these procedures. Based on data from the National Lung Screening Trial, 3 people per 1000 screened with low-dose CT experience major complications from invasive tests, and less than 1 person per 1000 screened die after an invasive test (within 60 d).6^,12 Some of these complications occur in people who receive a false-positive screening result. Data from 17 studies showed that 5 people per 1000 screened with low-dose CT received an unnecessary major invasive procedure for an ultimately benign condition (compared with 3 people per 1000 screened with chest radiography).12

Overdiagnosis occurs when people who are asymptomatic undergo screening for lung cancer and a slow-growing cancer that would have never caused them any harm during their lifetime is detected and diagnosed.18 Although current estimates of overdiagnosis for lung cancer vary by thresholds used and are based on limited follow-up, observational studies suggest that 2%–16% of lung cancers detected with chest radiography and 11%–26% of lung cancers detected with low-dose CT represent overdiagnosis.12 Overdiagnosis often leads to unnecessary treatment (overtreatment), which can cause harm.

Rationale

The findings of the National Lung Screening Trial6 suggest that screening with low-dose CT may reduce lung cancer mortality in patients at high risk. However, this benefit was shown in only one RCT, and there are possible harms related to screening and invasive follow-up testing. In addition, although there is no reason to suspect that screening with chest radiography would increase the risk of death from lung cancer, the National Lung Screening Trial compared screening with low-dose CT versus chest radiography rather than comparing low-dose CT with no screening (indirect evidence). For these reasons, the task force has made a weak recommendation to offer screening with low-dose CT for men and women at high risk of lung cancer (adults 55–74 years of age with at least a 30 pack-year smoking history who smoke or used to smoke). The overall body of evidence supporting this recommendation is rated as low quality. Whereas one high-quality study (National Lung Screening Trial) helped inform the recommendation,6 additional low-quality studies were used to address the critical outcomes. Following GRADE methods, the rating for the overall body of evidence must be based on the lowest-quality evidence used. Research currently underway may change our level of confidence in the benefit of low-dose CT screening for lung cancer, and we will reassess our recommendations once this new evidence is available.

The recommendation to screen the high-risk population places a higher value on a small benefit for reduced lung cancer mortality and the known poor prognosis of untreated lung cancer,19 but a lower value on the risk of adverse effects, overdiagnosis and the lack of data comparing low-dose CT with no screening.

There were no low-dose CT screening studies involving adults at low or moderate risk of lung cancer, or who had other risk factors unrelated to smoking (e.g., exposure to radon), that examined the outcomes the task force identified as critical or important. Therefore, the overall body of evidence was rated as very low quality. Men and women who are not at high risk of lung cancer (i.e., those who smoke with less than 30 pack-years of smoking history or those who quit smoking more than 15 years prior) would be expected to have lower absolute benefit of screening than the patients at high risk included in the National Lung Screening Trial but would still be susceptible to the potential harms (e.g., false-positive results, consequences from invasive follow-up tests and overdiagnosis) associated with screening. For these reasons, we strongly recommend against screening adults who are not at high risk or who have other risk factors for lung cancer unrelated to smoking.

We have made a strong recommendation against screening with chest radiography based on low-quality evidence because available evidence suggests that there is no benefit of screening with chest radiography on lung cancer–specific or all-cause mortality, but there are established harms (e.g., overdiagnosis, false-positive results and complications from follow-up testing).

Patient values and preferences

Patient preferences on screening for lung cancer were assessed in seven studies included in the systematic review,12 as well as de novo data collection conducted for the task force. The systematic review concluded that most participants in the high-risk group (i.e., adults 55–74 years of age who smoke or have quit smoking) had high willingness to participate in screening for lung cancer, motivated by their smoking history, belief that early detection improves health outcomes and a family history of lung cancer.12 Potential barriers to participation included inconvenience of screening and negative experiences with health care workers or settings. Patients who were screened for lung cancer with low-dose CT did not feel substantial distress or anxiety.12 In the judgment of the task force, the consistent evidence showing that patients at high risk are willing to be screened for lung cancer supports the recommendation for screening in this group.

The task force conducted a series of focus groups and a survey with 15 members of the public (mean age 63 yr [range 36 to 76 yr], 12 were female and 8 smoked or had quit smoking) to assess patient perceptions of these recommendations for screening for lung cancer.20 In general, participants agreed that these recommendations were appropriate, beneficial and feasible, although some had concerns about access to low-dose CT scans and limiting eligibility to adults between 55 and 74 years.

Other considerations

It is possible that longer or more intensive screening might yield additional benefits, but this is speculative, since there are no RCT data to support such a recommendation. Although ongoing screening (i.e., more than 3 screens) might further reduce mortality, it might also lead to more false-positive results and complicatons from invasive follow-up testing, potentially disrupting the balance between the benefits and harms as reported in the National Lung Screening Trial study.6

Smoking remains the primary risk factor for lung cancer; therefore, interventions to promote smoking cessation (which have benefits beyond decreasing the risk of lung cancer) should be incorporated into any screening program aimed at reducing lung cancer morbidity and mortality. Smoking cessation resources can be found in the Canadian Best Practices Portal of the Public Health Agency of Canada.21

Suggested performance measures

National-level quality indicators are currently being developed by the Pan-Canadian Lung Cancer Screening Network.22 Additional suggested performance measures, although difficult to measure, include rates of discussion of screening for lung cancer using low-dose CT among populations at high risk and the proportion of potentially eligible adults receiving accurate information about the risks and benefits of screening (ideally using an evidence-based decision aid). Reduction in use of chest radiography to screen for lung cancer is a potentially important implementation measure. Harms of screening and follow-up tests in different settings should also be continually assessed. Incidence and mortality data related to lung cancer should continue to be monitored at the provincial, territorial and national levels.