Impact of Low-Dose CT Screening on Lung Cancer Incidence and Outcomes

Abstract

Purpose of review:

To review findings from clinical trials of lung cancer screening (LCS), assess contemporary issues with implementation in clinical practice, and review emerging strategies to increase the uptake and efficiency of LCS.

Recent findings:

In 2013, the USPSTF recommended annual screening for individuals aged 55–80 years and currently smoke or quit within the past 15 years based on reduced mortality from lung cancer with annual low-dose computed tomography (LDCT) screening in the National Lung Screening Trial. Subsequent trials have demonstrated similar mortality outcomes in individuals with lower pack-year smoking histories. These findings, coupled with evidence for disparities in screening eligibility by race, resulted in updated guidelines by USPSTF to broaden eligibility criteria for screening. Despite this body of evidence, implementation in the US has been suboptimal with fewer than 20% of eligible individuals receiving a screen. Barriers to efficient implementation are multifactorial and include patient, clinician, and system level factors.

Summary:

Multiple randomized trials have established that annual LCS reduces mortality from lung cancer, however several areas of uncertainty exist on the effectiveness of annual LDCT. Ongoing research is examining approaches to improve the uptake and efficiency of LCS such as the use of risk-prediction models and biomarkers for identification of high-risk individuals.

Introduction

In 2023, an estimated 238,430 lung cancers will be diagnosed in the United States resulting in approximately 127,070 deaths.(1) Since tobacco use represents the most important risk factor for the disease, lung cancer incidence closely mirrors tobacco use patterns in the population with a lag of several decades.(2) Although lung cancer survival has improved over the last two decades, lung cancer remains the leading cause of cancer-related death.

In 2013, the United States Preventive Services Task Force (USPSTF) published initial guidelines for the use of low-dose CT (LDCT) for lung cancer screening (LCS), recommending annual screening for individuals aged 55 to 80 who have a smoking history of at least 30 pack-years without cessation in the preceding 15 years.(3) This recommendation was primarily based on the National Lung Screening Trial (NLST), a US-based randomized clinical trial (RCT) that allocated 53,454 high-risk individuals to three rounds of screening using either annual LDCT or chest x-ray. The LDCT arm demonstrated a 20% relative reduction in lung cancer mortality and a 6.7% relative reduction in overall mortality at a median follow-up of 6.5 years.(4) Long-term follow-up of the NLST population resulted in two important observations. First, based on the number of lung cancer deaths across arms, the number needed to screen (NNS) was 303, similar to the original NNS estimate of 320.(5) Second, the numbers of total cancers across the two arms were nearly identical, resulting in a decrease in the estimate of overdiagnosis rate from 18% in the original analysis to 3% in the extended analysis.(5)

The mortality benefit observed in the NLST was coupled with a significant shift in the distribution of cancer stage between the LDCT and chest x-ray arms, with more early-stage and fewer advanced-stage cancers identified in the LDCT arm. The presence of this ‘stage shift’ in conjunction with subsequent timely diagnosis and treatment likely represents the primary mechanism through which annual screening resulted in improved lung cancer-specific mortality. Despite the evidence from the NLST on the presence of stage shift and improvements in lung cancer-specific mortality, uptake of LCS in the US has been modest. Data on screening rates since the release of the USPSTF guidelines demonstrate persistently low rates of screening. With fewer than 20% of eligible individuals screened, the LCS rate lags behind similar guideline-recommended cancer screening rates such as for breast cancer or colon cancer.(6) Despite the low uptake of LCS, population level data demonstrates a steady increase in the incidence of early-stage disease from 2014 through 2018,(7,8) which is at least partially due to LCS availability. Nevertheless, more than half of current lung cancer diagnoses occur during late-stage disease.(1) Although treatment paradigms have improved for individuals with advanced lung cancer, increasing the 5-year survival rate to 23%,(9) this remains considerably lower than survival rates observed in other common cancers such as breast, colon, and prostate.

A Change in Screening Guidelines

Critics noted that the initial guidelines, based entirely on NLST, had limited information on the performance of screening among Black individuals who develop lung cancer at younger ages and with fewer pack-years than white individuals. In 2021, the USPSTF broadened their LCS recommendations to individuals aged 50 through 80 years old who have a smoking history of at least 20 pack-years and who currently smoke or have quit smoking within the past 15 years.(10) Reductions in age at screening initiation and number of pack-years required to meet eligibility came about for two main reasons. First, two European RCTs reported positive results from LDCT screening studies. The Dutch-Belgian lung cancer screening trial (NELSON), which enrolled individuals with fewer pack-years (minimum of approximately 15 pack-years) than NLST, confirmed the mortality benefit of LDCT screening by demonstrating a 24% reduction in lung cancer mortality in men and a 33% reduction among women.(11) Similarly, the Multicentric Italian Lung Detection (MILD) trial, a study of either annual or biennial LDCT compared to no screening, enrolled participants with a greater than 20 pack-year history and demonstrated a 39% reduction in lung cancer mortality at 10 years.(12) Second, data on disparities by race in lung cancer screening eligibility criteria were reported from the Southern Community Cohort Study, which showed that only 17% of Black adults who smoked were eligible for 2013 USPSTF guidelines compared to 31% of white individuals who smoked, and Black adults diagnosed with lung cancer were less likely to be eligible for screening.(13) Expanding the age and cigarette smoking criteria nearly doubled the number of individuals eligible for LCS across the US, with less disparity in screening eligibility by sex and race/ethnicity. Microsimulation modeling estimates suggest that the expanded criteria will result in further improvements in lung cancer mortality reduction, improving to 13.0% with USPSTF-2021 compared to 9.8% with USPSTF-2013.(14)

Impact of Screening in Real-World Settings

Although LCS trials have now shown improvements in lung cancer mortality, the impact of LDCT screening in clinical practice has not been clearly demonstrated. There are several reasons why trial results may not translate to the outcomes and harms observed in screening practices in real-world settings. First, both the NLST and NELSON trials were well-conducted, achieving rates of adherence >95%, and were performed largely in centers of excellence that may have led to optimal outcomes from diagnostic evaluation and subsequent cancer treatment. Enrolled participants were also younger, more educated, and had fewer comorbidities than the population of individuals that are potentially eligible for screening in real-world settings.(15,16) Additionally, concerns have been raised that harms from screening may be greater in clinical practice; these concerns have persisted and may be contributing to the slow uptake of LCS in the US.(17,18)

Population-based analyses demonstrate that, although screening rates have increased since 2014, only 5–20% of eligible US adults have completed screening.(19) The American College of Radiology’s Lung Cancer Screening Registry (LCSR) showed that, among the first million people who received LCS in the US, an estimated 9% of screening was performed among individuals who did not meet USPSTF criteria. Additionally, there were particularly low rates of screening among men, people who formerly smoked, and younger individuals. Other studies have identified lower rates of LDCT completion among Black adults and those without a higher education.(19–21) These inequities are particularly concerning given persistent disparities in lung cancer and highlight the need for new approaches.(22,23) Emerging evidence also suggests that rate of adherence to annual screening is considerably lower, particularly among marginalized populations, than the rates observed in the NLST and NELSON trials, which may diminish the mortality benefit observed in trial settings.(24–26)

Multiple challenges exist in LCS program implementation in practice. Considerable resources are needed to establish clinical teams that coordinate multidisciplinary care. These teams can effectively and equitably expand screening uptake among eligible individuals as well as increase adherence to annual screening and follow-up of abnormal findings. The Centers for Medicare and Medicaid Services’ national coverage of annual LCS includes the unique mandate that a high-quality shared decision-making (SDM) visit occur prior to ordering LDCT.(27) The SDM visit must include both a collaborative discussion of the potential benefits and risks of LDCT in the context of patient values and tobacco cessation counseling.(28) Despite the recommendation for and reimbursement of SDM visits, completion of SDM visits is remarkably low, with known disparities by race, sex, and socioeconomic status.(29,30) Even when completed, the quality of SDM is suboptimal. As such, current implementation of SDM may also represent a barrier to screening.(31–33) Other barriers to uptake include limited knowledge among patients, inadequate knowledge among primary care and other providers, concerns regarding the risk of false positive results and overdiagnosis, high rates of incidental findings, and more.(31,34) (Figure 1)

Figure 1.

Barriers listed summarized from Carter-Harris, et al.(31) and Cavers, et al.(34)

Despite the many challenges associated with LCS implementation, evidence is emerging on the balance of harms and benefits in practice. Although screening outcomes, such as measurement of lung cancer mortality, are difficult to assess outside of a clinical trial and may only be observed a decade after LCS uptake in clinical practice, the distribution of cancer stage among screened individuals may allow for an early indication of the impact of screening on lung cancer outcomes. Analyses of cancer registries and individual LCS programs have started to report initial outcomes and many, though not all, have reported favorable distributions of cancer stage among their population of patients diagnosed after screening initiation.(35–38) Although promising, most of these studies have been limited by their inability to identify the source population, thereby limiting meaningful comparisons between individuals with lung cancer diagnosed among screened and unscreened groups.

Identification of High-Risk Individuals

While current guidelines recommend eligibility based on categorization of age, pack-years, and years since quitting, an alternative approach uses prediction models that include multiple additional variables and estimate lung cancer risk over a specified period.(39,40) Retrospective analyses demonstrate that selecting individuals with the use of models results in higher sensitivity and in more lung cancer deaths averted. Models with the greatest promise to date include Lung Cancer Death Risk Assessment Tool (LCDRAT),(41) the Liverpool Lung Project (LLP) model, and PLCO_m2012 (model based on data from the Prostate, Lung, Colorectal and Ovarian trial).(42)

PLCO_m2012 considers age, race, smoking history, personal history of cancer, family history of lung cancer, presence of chronic obstructive pulmonary disease (COPD), education level, and body-mass index (BMI).(42) An analysis from the International Lung Screening Trial, a prospective cohort study designed to compare the accuracy of the PLCO_m2012 model to the USPSTF criteria, provided evidence that PLCO_m2012 was more sensitive at identifying those who will be diagnosed with lung cancer than the USPSTF-2013 criteria.(43) Furthermore, the cumulative potential life-years gained in individuals diagnosed with lung cancer was significantly higher with a PLCO_m2012 threshold of > 1.70% at 6 years than in patients selected based on USPSTF-2013 criteria. A recent microsimulation analysis also suggests that risk model-based screening strategies are more cost-effective than the USPSTF-2021 recommendation.(44) In particular, strategies using a 6-year risk threshold of 1.2% or greater for PLCO_m2012 (or 1.1% or greater for LCDRAT) had incremental cost-effectiveness ratios (ICERs) less than $100,000 per quality-adjusted life-year (QALY).

Another emerging strategy to identify high-risk individuals is the use of blood-based biomarkers. The ability to non-invasively determine protein, auto-antibody, genomic, or epigenomic changes in blood has the ability to transform how selection criteria are assessed for LDCT screening.(45) Despite this potential, blood-based strategies are still at relatively early stages of development, most often limited by diagnostic sensitivity for early-stage tumors. It remains to be determined if a molecular assay will achieve high enough sensitivity and acceptable specificity to allow it to be used prior to LDCT or if biomarker approaches are better utilized in conjunction with LDCT to more accurately differentiate between benign and malignant nodules.(46)

Screening of Individuals Who Have Never Smoked

Despite the recent broadening of USPSTF criteria and the emergence of risk prediction models as tools for improved eligibility assessment, current approaches to selection exclude a substantial proportion of individuals who will go on to be diagnosed with lung cancer. The proportion of lung cancers diagnosed in individuals that have never smoked is increasing over time, accounting for 25% of all lung cancers. If considered as a distinct disease entity, non-smoking related lung cancer would rank as the seventh most common cause of cancer-related death worldwide.(47) In Asia, 30–40% of all lung cancers and 60–80% of lung cancers in women occur in never-smokers,(48,49) likely reflecting exposures to environmental factors including outdoor air pollution and household burning of solid fuels for heating and cooking.(50) Of note, outdoor air pollution is estimated to cause 20.5% of lung cancer deaths in China, compared to 4.7% in the United States.(51)

Given these observations, further evaluation of screening approaches and development of risk prediction tools to estimate lung cancer risk among never-smokers may suggest a benefit from LDCT screening in this population. One such model, developed from the China National Lung Cancer Screening (NLCS) program, included five variables (age, female sex, BMI, family history of lung cancer, and chronic respiratory disease) and performed slightly better than PLCO_all2014 (a prior model developed by Tammemagi and colleagues that is analogous to PLCO_m2012 but configured to include never smokers).(50,52)

Preliminary findings from the Taiwan Lung Cancer Screening for Never Smoker Trial (TALENT), a multicenter single arm cohort of LDCT screening among 12,011 never-smokers aged 55–75 with at least one risk factor (family history of lung cancer within a third-degree relative, passive smoking exposure, history of tuberculosis or COPD, and exposure to cooking fumes), demonstrate an overall lung cancer prevalence of 3.2% at baseline screening.(53) The cancer stage data (78% diagnosed at Stage 1A/1B) is encouraging, however the results from TALENT also raise some concern regarding the potential for lung cancer overdiagnosis. Although the cancer detection rate of 3.2% at baseline was higher than baseline cancer rates in NLST or NELSON, 18% (57 of 311) of cancers were diagnosed at Stage 0 (i.e., carcinoma in situ).

Additional evidence for overdiagnosis from screening of low smoking prevalence populations was recently provided in an analysis by Gao and colleagues that assessed stage-specific lung cancer incidence using data from the Taiwan National Cancer Registry.(54) Following the introduction and marketing of LDCT screening in the mid-2000s, the incidence of stages 0-I lung cancer in women increased more than 6-fold (from 2.3 to 14.4 per 100,000) between 2004 and 2018; however, there was no change in the incidence of Stage II-IV lung cancer. The pattern of increasing early-stage disease not accompanied by a decrease in late-stage disease suggests a substantial contribution of overdiagnosis. The likelihood of considerable overdiagnosis was further supported by a minimal decrease in lung cancer mortality (17 to 16 per 100,000) but a substantial change in 5-year survival from 18% to 40% during the same time period. (54)

Thus, the balance of benefits and harms of expanding annual screening to individuals who have never smoked remains uncertain. Further research will be required to elucidate the impact of risk factor or model-driven approaches to patient selection for screening and the impact on lung cancer mortality, overdiagnosis, biopsy and complication rates, and cost.

Conclusion and Future Directions

Despite the evidence on reduction in lung cancer mortality from multiple RCTs, uptake of lung cancer screening in the US and worldwide remains suboptimal and multiple areas of uncertainty exist regarding optimization of clinical effectiveness and care delivery. A number of ongoing trials will help address some of these unanswered questions (Table 1). Existing areas of investigation span the spectrum of LCS implementation, including investigations of blood-based screening tests, risk prediction model-based selection, and assessment of personalized screening intervals. Findings from these studies, and other emerging evidence from observational and modeling approaches, will help refine evidence-based strategies for successful LCS implementation across diverse clinical settings and populations.

Table 1.

Ongoing Clinical Trials of Lung Cancer Screening

Trial	Study Start Year	Sample Size	Study Population	Study Arms	Primary Outcomes
4-IN-THE-LUNG-RUN	2020	24,000	Age 60–79 years; ≥35 pack-year smoking history; currently smoke or quit less than 10 years ago or age ≤79 years and PLCOm2012noRace 6-year lung cancer risk >2·60%	Annual vs biennial screening for participants with normal baseline low-dose CT	Effectiveness of risk-based screening interval
SUMMIT	2019	25,000	Age 55–77 years; ≥30 pack-year smoking history (or at least 20 years duration) and currently smoke or quit <15 years ago or PLCOm2012 6-year lung cancer risk of ≥1·3%	Annual vs biennial screening for participants with normal baseline low-dose CT; MCED at baseline for all participants	Performance of a multi-cancer early detection blood test; multiple measures of LDCT efficacy
TALENT	2015	12,000	Age 55–75 years, never smoked, and with one of the following risk factors: family history of lung cancer within third-degree, passive smoking exposure, tuberculosis/COPD, cooking index ≥110, and not using ventilator during cooking	Baseline low-dose CT, annual low-dose CT for 2 years and biennial low-dose CT for 6 years if no lung cancer is detected	Validity of LDCT for lung cancer among non-smokers
ILST	2017	5819	Age 55–80 years, currently smoke or previously smoked and met USPSTF (2013) criteria or 6-year lung cancer risk ≥1·51% (PLCOm2012)	Baseline low-dose CT and one repeat in 1–2 years (n=5819)	Number of lung cancers, nodules, and change in QOL.
Improving Utilization of LCS in Underserved PA Populations	2023	2500	Age 50–77 years, history of smoking in EMR	Centralized outreach and education vs usual care	LCS rates
Watch the Spot	2017	35200	Age ≥ 35, <15mm pulmonary nodule detected incidentally or by screening	More intensive surveillance with Lungs-RADS or Fleischner guidelines follow up vs less intensive surveillance with modified Lung-RADS or Fleischner Guidelines follow up	Tumor progression beyond AJCC 7 stage T1a (tumor size <20mm)
TRIO	2022	6618	Age 50–80; minimum 20-pack year history (group 1); First degree relative of lung cancer patient (group 2); Age 50–80 with high-risk occupational or environmental exposure (group 3)	Lung cancer screening strategy using Lung-RADS	LDCT completion; lung cancer detection rate

Key Points:

1. Clinical trials of lung cancer screening demonstrate favorable shifts in lung cancer stage and improvements in lung cancer specific mortality.

2. Uncertainty exists regarding the balance of benefits and harms with implementation of lung cancer screening in clinical practice.

3. Future research on outcomes of lung cancer screening should focus on strategies to reduce barriers to screening and evaluate novel approaches for patient selection.