Benefits, harms, and cost-effectiveness of risk model-based and risk factor-based low-dose computed tomography screening strategies for lung cancer: a systematic review

Yin Liu; Qingchao Geng; Xin Lin; Chenxi Feng; Youlin Qiao; Shaokai Zhang

doi:10.1186/s12885-024-13356-6

. 2024 Dec 23;24:1567. doi: 10.1186/s12885-024-13356-6

Benefits, harms, and cost-effectiveness of risk model-based and risk factor-based low-dose computed tomography screening strategies for lung cancer: a systematic review

Yin Liu ^1,^#, Qingchao Geng ^1,^#, Xin Lin ¹, Chenxi Feng ¹, Youlin Qiao ^1,^2,^✉, Shaokai Zhang ^1,^✉

PMCID: PMC11665239 PMID: 39710662

Abstract

Background

It has been proposed that risk model-based strategies could serve as viable alternatives to traditional risk factor-based approaches in lung cancer screening; however, there has been no systematic discussion. In this review, we provide an overview of the benefits, harms, and cost-effectiveness of these two strategies in lung cancer screening application, as well as discussing possible future research directions.

Methods

Following the PRISMA guidelines, a comprehensive literature search was conducted across PubMed, Web of Science, Cochrane libraries, and EMBASE from January 1994 to April 2024. Studies comparing risk model-based and risk factor-based low-dose computed tomography(LDCT) screening strategies for lung cancer were included, with data extracted on study characteristics, screening criteria, and outcomes such as sensitivity, specificity, lung cancer deaths averted, false positive, biopsies, overdiagnosis, radiation-related cancer, and cost-effectiveness measures, et al.

Results

A total of 16 fulfilled articles were included, comprising 6 model simulation studies, 9 retrospective cohort studies, and 1 interim analysis of a prospective cohort study. Risk model-based strategies generally demonstrated higher sensitivity, comparable specificity and lower radiation-related harms compared to risk factor-based strategies. However, there were variations in life years gained, quality-adjusted life years gained, lung cancer deaths averted and overdiagnosis cases, highlighting the need for optimal risk threshold determination. Risk model-based strategies showed a potential for greater cost-effectiveness, particularly when tailored to individual risk profiles. Furthermore, subgroup analyses revealed a higher net benefit in women, emphasizing the importance of sex-specific eligibility criteria.

Conclusion

Risk model-based LDCT screening strategies present a more sensitive and potentially more efficient approach for lung cancer detection. Future research should explore optimal risk thresholds for broader applicability, with attention to sex-specific criteria and individual risk factor dynamics.

Keywords: Lung cancer, LDCT screening, Risk model, Risk factor, Systematic review

Background

In 2020, deaths from lung cancer reached 1.89 million globally, ranking first among malignant tumors [1]. Low-dose computed tomography (LDCT) screening can effectively reduce mortality; the National Lung Screening Trial (NLST) showed that LDCT screening reduced lung cancer-specific mortality (20%) and all-cause mortality (6.7%), after a 6.5 years follow- up [2]. Nevertheless, issues remain, including false positives, overdiagnosis, and radiation-related incidents [3, 4].

Accurate selection of individuals for LDCT screening is crucial to minimize harms and maximize benefits. Screening for individuals at high-risk of lung cancer generally employs one of two strategies: (i)Risk factor-based. This approach selects individuals based on established lung cancer risk factors and is often favored for its simplicity and ease of implementation. For example, in 2013, the US Preventive Services Task Force (USPSTF) recommended annual LDCT screening for lung cancer in asymptomatic adults aged 55 to 80 years, with a minimum of 30 pack-years of smoking within 15 years since cessation [5]. Similarly, the National Comprehensive Cancer Network Group 2 (NCCN-2) recommend screening for individuals over 50 with a smoking history of 20 pack-years coupled with at least one additional risk factor, such as a personal history of cancer or lung disease, family history of lung cancer, exposure to radon, or occupational carcinogens [6]. (ii) Risk model-based. This method selects individuals for screening based on their personalized lung cancer risk, as estimated by validated predictive models. A notable example is the UK Lung Screen trial, which utilized the Liverpool Lung Project (LLP) risk prediction model, setting a 5% risk threshold over five years as the criterion for screening eligibility [7].

The debate continues over the superiority of these strategies for LDCT screening eligibility eligibility. In this review, we examine the literatures to evaluate the comparative benefits, potential harms, and cost-effectiveness of risk model-based and risk factor-based strategies in identifying those at high risk of lung cancer. We also explore the prospects for future research that could inform and refine screening practices.

Materials and methods

This reporting was following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guideline [8]. Two reviewers independently performed literature screening, data extraction, and methodological quality evaluation of included studies, according to the inclusion and exclusion criteria. Disagreements were discussed and resolved, or a third-party reviewer requested.

Search strategy

The PubMed, Web of Science, Cochrane libraries, and EMBASE were searched using the key words: (i) [‘lung cancer’]; (ii) [screening]; (iii) [‘risk prediction model’ OR ‘risk model’ OR ‘risk assessment’ OR ‘risk-based’ OR ‘risk-targeted’]; AND (iv) [eligibility OR criterion]; (i), (ii), and (iii) were connected with AND, from the advent of screening with LDCT (January 1994) to April 1, 2024, and limited to English language. References in bibliographies of included literatures were manually searched.

Inclusion and exclusion criteria

Articles comparing the benefits, harms, and cost-effectiveness of risk model-based and risk factor-based strategies in selecting individuals for LDCT screening were included. Inclusion criteria were: (1) study population at high risk of lung cancer, including current or former smokers, individuals with a history of long-term occupational exposure to carcinogens, those with chronic lung disease(e.g., chronic obstructive pulmonary disease (COPD)), and people with a family history of lung cancer in a first-degree relative; (2) study design: published retrospective cohort, prospective cohort, cross-sectional, case–control, randomized controlled trials, non-randomized controlled trials, and modeling study; (3) at least one of the following outcomes reported: sensitivity, specificity, lung cancer deaths averted, false positive, biopsies, overdiagnosis, radiation-related harms, quality adjusted life years (QALYs) gained, life years (LYs) gained, incremental cost-effectiveness ratio (ICER), incremental net monetary benefit (INMB). Exclusion criteria were: informally published literature (conference abstracts and academic papers), review article, editorial, letter, commentary.

Data extraction

The data extraction was undertaken independently by two authors. An abstract table was formulated, including characteristics (first author, publication year, country/region, study design, study population, population source, duration of follow-up/time horizon, population age, sample size), LDCT lung cancer screening criteria used (risk model- and risk factor-based strategies), and study outcomes.

Results

Study characteristics

A total of sixteen studies were included following the search and review processes (Fig. 1), comprising six modeling studies [9–14], nine retrospective cohort studies [15–23], and one prospective cohort study [24] (Table 1).

Table 1.

Summary of study characteristics

No	First author	Year of publication	Country/region	Study design	Study population	Population source	Duration of follow-up/time horizon(years)	Population age (years)	Simulated results count / Sample size
1	Ten Haaf K [9]	2020	USA	Modeling study	Current or ever-smokers	1950 U.S. birth cohort	lifetime	45–90	100,000^a
2	Meza R [10]	2021	USA	Modeling study	Current or ever-smokers	1960 US birth cohort	lifetime	45–90	100,000^a
3	Toumazis I [11]	2023	USA	Modeling study	Current or ever-smokers	1960 U.S. birth cohort	lifetime	45–90	100,000^a
4	Liu Y [12]	2024	China	Modeling study	Current or ever-smokers	Henan Province CanSPUC	30	50–74	100,000^a
5	Cressman S [13]	2023	Canada, Australia, Hong Kong, and the UK	Modeling study	Current or ever-smokers	ILST	lifetime	55–80	100,000^a
6	Tomonaga Y [14]	2023	Switzerland	Modeling study	Current or ever-smokers	Birth cohorts 1940 to 1979 in Switzerland	lifetime	55–80	100,000^a
7	Katki HA [15]	2016	USA	Retrospective cohort study	Current or ever-smokers	NHIS 2010–2012	5	50 -80	43,413,257^b
8	Tammemägi MC [16]	2013	USA	Retrospective cohort study	Current or ever-smokers	PLCO intervention arm smokers	6	55–74	37,332^b
9	Tammemägi MC [17]	2014	USA	Retrospective cohort study	Current or ever-smokers	PLCO intervention arm smokers	11	55–74	37,327^b
10	Ten Haaf K [18]	2017	USA	Retrospective cohort study	Current or ever-smokers	PLCO intervention arm smokers	6	55–74	40,600^b
11	Tammemagi MC [19]	2021	Canadian	Retrospective cohort study	Current or ever-smokers	PLCO intervention arm smokers	6	55–74	NR
12	Cleven KL [20]	2020	USA	Retrospective cohort study	FDNY-WTC-exposed rescue/recovery workers (firefighters and EMS) who were current or ever-smokers	WTCHP	NCCN-2 = mean 3.8 USPSTF 2013 = mean 3.0 PLCO_M2012 = mean 3.0 Bach = mean 3.3	$>$ 50	3,953^b
13	Miranda-Filho A [21]	2021	Brazilian	Retrospective cohort study	Current or ever-smokers	Brazilian Surveillance System of Risk and Protective Factors for Chronic Diseases (Vigitel) Surve	5	55–79	2,300,000^b
14	Park B [22]	2021	Korea	Retrospective cohort study	Current or ever-smokers	Korean NHIS 2007–2008	mean 6.6	40–79	969,351^b
15	Landy B [23]	2019	USA	Retrospective cohort study	Current or ever-smokers	NHIS 2015	6	50–80	8,000,000^b
16	Tammemägi MC [24]	2022	Canada, Australia, Hong Kong, and the UK	Prospective cohort study	Current or ever-smokers	ILST	USPSTF 2013 = 2.1 PLCO_M2012 = 2.1	55–80	5,819^b

Open in a new tab

CanSPUC Cancer Screening Program Started in Urban China, ILST International Lung Screening Trial, NHIS National Health Insurance Service, PLCO Prostate, Lung, Colorectal and Ovarian, FDNY-WTC Fire Department of the City of New York World Trade Center, EMS Emergency Medical Service Providers, WTCHP The World Trade Center Health Program, PLCO_M2012 Prostate, Lung, Colorectal and Ovarian Model 2012, USPSTF US Preventative Services Task Force, NCCN-2 National Comprehensive Cancer Network Group 2

^aSimulated results count

^bSample size

The six modeling studies were predominantly conducted in the USA [9–11], China [12], and Switzerland [14]. Additionally, Cressman S et al [13] used data from the International Lung Screening Trial (ILST), encompassing diverse populations from Canada, Australia, Hong Kong, and the UK. These modeling studies uniformly adopted long-term or lifetime horizons to project lung cancer screening outcomes.

The nine retrospective cohort studies were predominantly U.S.-based, utilizing comprehensive datasets like National Health Insurance Service (NHIS) [15, 23], Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial [16–18], the World Trade Center Health Program (WTCHP) [20]. These studies had follow-up periods ranging from 5 to 11 years. One study described implementation of the PLCO_m2012 risk prediction model in the Canadian people [19]. Studies from Brazil [21] and Korea [22] used respective Brazilian Surveillance System of Risk and Protective Factors for Chronic Diseases (Vigitel) Surve and Korean NHIS, with follow-up durations of 5 to 6 years.

In addition, Tammemägi MC et al [24] conducted the ILST, a prospective cohort study with a follow-up of 2.1 years. This study compared the effectiveness of the USPSTF 2013 and PLCO_m2012 lung cancer screening criteria, providing valuable early insights into the performance of these models across diverse international cohorts.

LDCT screening strategies

The majority of studies, fifteen in total, targeted current or ever-smokers, detailed screening age and smoking pack-year criteria were shown in Table 2 [9–19, 21–24]. One distinct study concentrated on current/ever-smoker firefighters and emergency medical service providers [20].

Table 2.

LDCT screening strategy

No	Author	Risk prediction model(Risk threshold); Method for determining risk threshold	Risk factor-based strategies	Screening age	Frequency
1	Ten Haaf K [9]	Bach (2.8%) PLCO_M2012 (1.7%) LCDRAT (1.7%); Fixed-population size	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 80 years	Annual
2	Meza R [10]	PLCO_M2012 (0.4–2.3%), LCDRAT (0.24–2.3%) Bach (0.8–3.4%); A range of risk thresholds were considered, separated by 0.1%	Smoking criteria: (1) minimum pack-years: 20, 25, 30, 40 pack-years, (2) maximum years since quitting smoking:10, 15, 20, 25 years	Age to start screening: 45, 50, 55 years Age to stop screening: 75, 77, 80 years; Most of the efficient strategies started screening at age 50 or 55 and stopped at age 80	Annual and Biennial
3	Toumazis I [11]	PLCO_M2012 (1.2%); Cost-effectiveness	USPSTF 2021 USPSTF 2013	PLCO_M2012 (1.2%) and USPTSTF 2021: Age to start screening: 50 years Age to stop screening: 80 years USPSTF 2013: Age to start screening: 55 years Age to stop screening: 80 years	Annual
4	Liu Y [12]	Henan risk prediction model; Cost-effectiveness	China guideline 2021	Age to start screening: 50 years Age to stop screening: 74 years	Risk-stratified screening intervals: Screening intervals included annual, biennial, and triennial, with the high-risk group being screened more frequently than the low-risk group
5	Cressman S [13]	PLCO_M2012 (1.70%) Fixed-population size	USPSTF 2013	PLCO_M2012 (1.70%): Age to start screening: 55 years Age to stop screening: NR Risk factor-based strategy: Age to start screening: 55 years Age to stop screening: 80	Annual
6	Tomonaga Y [14]	PLCO_M2012 (1.60%); Cost-effectiveness	CSC1 to CSC4 report pack year-based strategies that follow suggestions in the recent CSC recommendations	Age to start screening: 55 years Age to stop screening: 80 years	Biennial
7(1)	Katki HA [15]	LCDRAT (1.9%); Fixed-population size	USPSTF 2013	LCDRAT (1.9%): Age to start screening: 50 years Age to stop screening: 80 years USPSTF 2013: Age to start screening: 55 years Age to stop screening: 80 years	Annual
7(2)	Katki HA [15]	LCDRAT (1.7%); Fixed-effectiveness (same NNS)	USPSTF 2013	LCDRAT (1.7%): Age to start screening: 50 years Age to stop screening: 80 years USPSTF 2013: Age to start screening: 55 years Age to stop screening: 80 year	Annual
8	Tammemagi MC [16]	PLCO_M2012 (1.3455%); Fixed-population size	NLST	Age to start screening: 55 years Age to stop screening: 74 years	Annual
9(1)	Tammemagi MC [17]	PLCO_M2012 (1.51%); The risk threshold determined by the risk cutoff point when the mortality rate in the NLST CT arm was consistently lower than CXR arm	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 80 years	Annual
9(2)	Tammemagi MC [17]	PLCO_M2012 (1.34%); Fixed-proportion size	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 80 years	Annual
10	Ten Haaf K [18]	PLCO_M2012 (1.35%) Bach (1.58%) TSCE (1.10%); Fixed-population size	NLST	Age to start screening: 55 years Age to stop screening: 74 years	Annual
11	Tammemagi MC [19]	PLCOM2012noRace (2%); Fixed-population size	Smoking criteria: (1)minimum pack-years: 40 pack-years, (2)maximum years since quitting smoking:10years	Age to start screening: 55 years Age to stop screening: 74 years	Annual
12	Cleven KL [20]	PLCO_M2012 (1.3455%) Bach (1.3455%); Fixed-population size	USPSTF 2013 NCCN-2	PLCO_M2012 (1.3455%) and Bach (1.3455%): Age to start screening: 50 years Age to stop screening: NR USPSTF 2013: Age to start screening: 55 years Age to stop screening: 80 years NCCN-2: Age to start screening: 50 years Age to stop screening: 80 years	Annual
13(1)	Miranda-Filho A [21]	LCDRAT (1.2%); Fixed-population size	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 79 years	Annual
13(2)	Miranda-Filho A [21]	LCDRAT; Age-specific risk-thresholds: select a similar number of ever-smokers in each age category as USPSTF 2013 Age 50–54 years: threshold = 0.36% Age 55–59 years: threshold = 0.64% Age 60–64 years:threshold = 0.99% Age 65–69 years: threshold = 1.55% Age 70–74 years:threshold = 2.57% Age 75–79 years:threshold = 3.46%	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 79 years	Annual
14	Park B [22]	Korean lung cancer risk model (2.1%); Fixed-population size	NLST	Korean lung cancer risk model (2.1%): Age to start screening: 40 years Age to stop screening: 79 years NLST: Age to start screening: 55 years Age to stop screening: 74 years	Biennial
15	Landy B [23]	PLCO_M2012(2.19%) LCDRAT (1.33%); Fixed-population size	USPSTF 2013	Age to start screening: 50 years Age to stop screening: 80 years	Annual
16	Tammemägi MC [24]	PLCO_M2012 (1.70%); Fixed-population size	USPSTF 2013	Age to start screening: 55 years Age to stop screening: 80 years	Annual

Open in a new tab

USPSTF 2021: 50–80 years-old, at least 20 pack-years smoked, and less than 15 years since quitting

USPSTF 2013: 55–80 years-old, at least 30 pack-years smoked, and less than 15 years since quitting

NLST: 55–74 years-old, at least 30 pack-years smoked, and less than 15 years since quitting

Preferred MISCAN: 55–75 years-old, at least 40 pack-years smoked, and less than 10 years since quitting

NCCN-2: 50–80 years-old, at least 20 pack-years smoked, no limit on duration of quitting smoking, and incorporating occupational exposure

China guideline 2021:50–74 years-old, at least 30 pack-years smoked, and less than 15 years since quitting

CSC1: 55–80 years-old, at least 20 pack-years smoked, and less than 10 years since quitting

CSC2: 55–80 years-old, at least 20 pack-years smoked, and less than 15years since quitting

CSC3: 55–80 years-old, at least 20 pack-years smoked, and less than 20 years since quitting

CSC4: 55–80 years-old, at least 20 pack-years smoked, and less than 25 years since quitting

NR not report, NNS number needed to screen, LCDRAT Lung Cancer Death Risk Assessment Tool, NLST National Lung Screening Trial, TSCE Two-Stage Clonal Expansion Model, USPSTF United States Preventive Services Task Force, NCCN-2 National Comprehensive Cancer Network Group 2, CSC Cancer Screening Committee

Most studies initiated screening at the age of 50 or 55. While a single modeling study pushed the starting age to 45 [10], the majority of efficient strategies opted for initiating screening at 50 or 55 years of age. Additionally, one retrospective cohort study started screening at the age of 40 [22]. The stop screening age was generally set between the ages of 74 to 80 years, with variations depending on the study.

In terms of screening frequency, twelve studies performed annual screening [9, 11, 13, 15–18, 20–24], two biennial [14, 22], and one both annual and biennial [10]. Liu Y et al [12] proposed a strategy to develop a personalized screening frequency based on individual’s risk profile, advocating for more frequent screenings for those at higher risk compared to those at lower risk.

Risk thresholds were used to determine eligibility criteria for risk-model based LDCT screening. Approaches to estimate risk thresholds in the included sixteen studies were summarized in Table 2. The most employed method was a fixed population size, in which risk-thresholds were set so participant number was equal to that achieved using risk factor-based strategies. Some studies determined risk-thresholds by matching efficiency (number needed to screen (NNS)) in reducing lung cancer death [15], selecting a similar proportion of eligible individuals [17], or selecting the same number of ever-smokers in each age group [21]. In addition, unlike above, Tammemägi MC et al [17] used PLCO_M2012, advocating for a 1.51% 6-year risk threshold as the risk cut-off point where mortality rate in the NLST CT arm was consistently lower than that in the chest X-ray arm. Two studies identified the most cost-effective thresholds for risk prediction models using ICER as the primary metric [11, 14]. Liu Y et al [12] recommended that annual screening for individuals with a 5-year risk threshold above 1.70%, biennial screening for individuals with a 5-year risk threshold of 1.03% to 1.69%, and triennial screening for individuals with a 5-year risk threshold less than 1.03% were most cost-effective.

Benefits

Sensitivity and specificity

Six studies reported on sensitivity and specificity [16–20, 22] (Table 3). In five of these studies, which focused on current or former smokers, risk prediction models demonstrated increased sensitivity, but the improvement in specificity was not obvious. For example, Tammemagi MC et al. [16] observed a sensitivity of 83.0% and a specificity of 62.9% when applying the PLCO_M2012-based risk threshold of 1.3455% for determining LDCT screening eligibility, in contrast, the NLST criteria yielded a sensitivity of 71.1% and a specificity of 62.7%. In a distinct study examining firefighters [20], risk model-based strategies achieved lower sensitivity (PLCO_M2012: 50.7%, Bach: 46.03%), but higher specificity (PLCO_M2012: 90.31%, Bach: 90.59%), than the NCCN-2 guideline (sensitivity: 79.37%, specificity: 16.19%), which incorporates occupational exposures, and reduced eligible screening age from 55 to 50 years and smoking history from 30 pack-years to 20, with no limit on duration of quitting smoking.

Table 3.

Outcomes of the included studies

No	Author	Sensitivity/Specificity	Lung cancer death averted; Number needed to screen (NNS) to prevent one lung cancer death	False positive	Biopsies
1	Ten Haaf K [9]	NR	Number of lung cancer death averted: Bach = 693 PLCO_M2012 = 698 LCDRAT = 696 USPSTF = 613 NNS: Bach = 466 PLCO_M2012 = 472 LCDRAT = 467 USPSTF 2013 = 533	NR	NR
2	Meza R [10]	NR	Number of lung cancer death averted: (Annual): PLCO_M2012, LCDRAT, and Bach = 388 to 662 Risk factor-based strategies = 327 to 578 NNS (Annual): PLCO_M2012, LCDRAT, and Bach = 37 to 50 Risk factor-based strategies = 29 to 46 Number of lung cancer death averted (Biennial): PLCO_M2012, LCDRAT, and Bach = 348 to 484 Risk factor-based strategies = 173 to 404 NNS (Biennial): PLCO_M2012, LCDRAT, and Bach = 55 to 70 Risk factor-based strategies = 42 to 63	Number of false-positive screens per individual screened (Annual): PLCO_M2012, LCDRAT, and Bach = 1.4 to 2.3 Risk factor-based strategies = 1.3 to 2.8 Number of false-positive screens per individual screened (Biennial): PLCO_M2012, LCDRAT, and Bach = 1 to 1.4 Risk factor-based strategies = 1.1 to 1.5	Number of biopsies (Annual): PLCO_M2012, LCDRAT, and Bach = 508 to 1015 Risk factor-based strategies = 422 to 933 Number of biopsies (Biennial): PLCO_M2012, LCDRAT, and Bach = 506 to 767 Risk factor-based strategies = 241 to 626
3	Toumazis I [11]	NR	Number of lung cancer death averted: PLCOm2012 = 493 LCDRAT = NR USPSTF2021 = 467 USPSTF2013 = 354	NR
4	Liu Y [12]	NR	Number of Lung cancer death averted: The most cost-effective risk model-based strategy = 606 China guideline 2021 = 770 NNS: The most cost-effective risk model-based strategy = 976 China guideline 2021 = 1392	Number of false-positive per individual screened: The most cost-effective risk model-based strategy = 0.46 China guideline 2021 = 0.83	NR
5	Cressman S [13]	NR	NR	NR	NR
6	Tomonaga Y [14]	NR	Number of Lung cancer death averted: PLCO_M2012=522 CSC1 = 475 CSC2 = 502 CSC3 = 526 CSC4 = 576 NNS: PLCO_M2012=33 CSC1 = 32 CSC2 = 33 CSC3 = 34 CSC4 = 34	Number of false-positive screens: PLCO_M2012=1,140 CSC1 = 1,130 CSC2 = 1,267 CSC3 = 1,396 CSC4 = 1,511	Number of biopsies resulting from false positive screenings PLCO_M2012=454 CSC1 = 445 CSC2 = 498 CSC3 = 548 CSC4 = 593
7(1)	Katki HA [15]	NR	Number of lung cancer death averted: LCDRAT model = 55,717 USPSTF 2013 = 46,488 NNS: LCDRAT model = 162 USPSTF 2013 = 194	Number of false positive screens per prevented lung cancer death: LCDRAT model = 116 USPSTF 2013 = 133	NR
7(2)	Katki HA [15]	NR	Number of lung cancer death averted: LCDRAT model = 62,382 USPSTF 2013 = 46,488 NNS: LCDRAT model = 194 USPSTF 2013 = 194	Number of false positives per lung cancer death prevented: LCDRAT model = 134 USPSTF 2013 = 133	NR
8	Tammemagi MC [16]	Sensitivity: PLCO_M2012 = 83.0% NLST = 71.1% Specificity: PLCO_M2012 = 62.9% NLST = 62.7%	NR	NR	NR
9(1)	Tammemagi MC [17]	Sensitivity: PLCO_M2012 = 80.1% USPSTF 2013 = 71.2% Specificity: PLCO_M2012 = 66.2% USPSTF 2013 = 62.7%	NR	NR	NR
9(2)	Tammemagi MC [17]	Sensitivity: PLCO_M2012 = 83.2% USPSTF2013 = 71.2% Specificity: PLCO_M2012 = 62.9% USPSTF 2013 = 62.7%	NR	NR	NR
10	Ten Haaf K [18]	Sensitivity: PLCO _M2012, Bach, and TSCE > 79.8% NLST = 71.4% Specificity: PLCO _M2012, Bach, and TSCE > 62.3% NLST = 62.2%	NR	NR	NR
11	Tammemagi MC [19]	Sensitivity: PLCOm2012noRace = 68.1% Preferred MISCAN = 59.6% Specificity: PLCOm2012noRace = 74.8% Preferred MISCAN = 74.8%	NR	NR	NR
12	Cleven KL [20]	Sensitivity: PLCO_M2012 = 50.7% Bach = 46.03% NCCN-2 = 79.37% USPSTF 2013 = 44.4% Specificity: PLCO_M2012 = 90.31% Bach = 90.59% NCCN-2 = 16.19% USPSTF 2013 = 20.3%	NR	Rate of False-positive findings: PLCO_M2012 = 259/309 Bach = 110/138 NCCN-2 = 129/161 USPSTF 2013 = 115/144	Proportions of over-biopsies resulting in non-lung cancer diagnosis: PLCO_M2012 = 2/34 Bach = 2/31 NCCN-2 = 4/54 USPSTF 2013 = 2/30
13(1)	Miranda-Filho A [21]	NR	Proportion of lung cancer death averted: LCDRAT = 61.9% USPSTF 2013 = 57.1%	NR	NR
13(2)	Miranda-Filho A [21]	NR	Proportion of lung cancer death averted: LCDRAT = 60.1% USPSTF 2013 = 57.1%	NR	NR
14	Park B [22]	Sensitivity: Korean lung cancer risk model = 60.1% NLST = 44.3% Specificity: Korean lung cancer risk model = 83% LST = 82.8%	NR	NR	NR
15	Landy B [23]	NR	Number of lung cancer death averted: PLCO _M2012 = 47,401 LCDRAT = 51,019 USPSTF 2013 = 41,298 NNS: PLCO _M2012 = 169 LCDRAT = 156 USPSTF 2013 = 194	Number of false positives per lung cancer death prevented: PLCO _M2012 = 119 LCDRAT = 112 USPSTF 2013 = 133	NR
16	Tammemägi MC [24]	NR	NR	NR	NR

Open in a new tab

^aOnly the outcomes of the efficient strategies were reported in the original study

Deaths averted

Eight studies reported on lung cancer deaths averted [9–12, 14, 15, 21, 23] (Table 3). Six studies indicated that, with the same screening ages and intervals, risk model-based strategies averted more deaths, with lower NNS to prevent one lung cancer death, than risk factor-based strategies [9–11, 15, 21, 23]. For example, Ten Haaf K et al [9] discovered that risk model-based strategies, which required similar screens among individuals aged 55–80 years as the USPSTF 2013 criteria (corresponding risk thresholds: Bach = 2.8%; PLCO_m2012 = 1.7%; LCDRAT = 1.7%), averted considerably more lung cancer deaths (Bach = 693; PLCO_m2012 = 698; LCDRAT = 696; USPSTF = 613). Liu Y et al [12] found that the most cost-effective risk model-based strategy, despite averting fewer lung cancer deaths because of the lower number of CT screens, had a lower NNS compared to China guideline 2021. The latter recommends annual screening for heavy smokers aged 50 to 74, with at least 30 pack-years of smoking, whether still smoking or having quit within the last 15 years [25]. Tomonaga Y et al [14] identified the most cost-effective strategy, targeting individuals aged 55 to 80 who met a 1.6% risk threshold as determined by the PLCO_M2012 model. This approach was more effective in preventing lung cancer deaths than the Switzerland Cancer Screening Committee's (CSC) recommendations 1 and 2, but not as many as the stricter CSC guidelines 3 and 4. All CSC recommendations suggest screening people aged 55 to 80 who have smoked for at least 20 years, with variations in the time since they quit smoking. Specifically, CSC 1 targets those who have quit within the last 10 years, CSC 2 extends this to 15 years, CSC 3 to 20 years, and CSC 4 to 25 years [26]. Notably, the new strategy had a similar NNS to prevent one death, just like all the CSC recommendations.

Additionally, Meza R et al [10] found that annual screening would avert more lung cancer deaths than biennial with a lower NNS; they also found that initiating screening at the age of 50 resulted in more lung cancer deaths averted compared to starting at age 55. No studies reported non-lung cancer deaths or all-cause deaths.

Life years and quality adjusted life years

Three studies reported on LYs [9, 10, 24], and three studies reported on both LYs and QALYs [11, 12, 14] (Table 4). Of them, three studies concluded that risk model-based strategies resulted in more LYs compared to risk factor-based approaches [9, 10, 24]. In a specific study highlighted by Toumazis I et al [11], the 1.2% threshold of the PLCO_M2012 model was the most cost-effective risk-based strategy for the 1960 U.S. birth cohort, yielding QALYs (1,857 per 100,000 individuals) equivalent to the USPSTF 2021 recommendation, surpassing those of the 2013 version (1,427 per 100,000 individuals). Similarly, the LYs gained (3,800 per 100,000 individuals) were less than the 2021 (4,171 per 100,000 individuals) but more than the 2013 USPSTF guidelines (2,950 per 100,000 individuals). Liu Y et al [12] discovered that the most cost-effective risk model-based strategy, while yielding fewer LYs and QALYs, required fewer screenings per LY and QALY gained, making it a more efficient approach compared to China guideline 2021. Tomonaga Y et al [14] compared the most cost-effective risk-based strategy (1.6% PLCO_M2012) with CSC 1 ~ CSC 4 strategies that varied in the duration since quitting smoking, and found that the 1.6% strategy resulted in fewer LYs and QALYs than the strategy allowing up to 10 years since quitting but outperformed those with longer cessation times of 15, 20, and 25 years.

Table 4.

Outcomes of the included studies

No	Author	Overdiagnosis	Radiation-related lung cancer	LYs/QALYs	Cost-effectiveness
1	Ten Haaf K [9]	Number of overdiagnosis lung cancer: Bach = 149 PLCO_M2012 = 147 LCDRAT = 150 USPSTF 2013 = 115 Overdiagnosis rate per screen-detected lung cancer: Bach = 7.8% PLCO_M2012 = 7.7% LCDRAT = 7.8% USPSTF = 7.3%	NR	LYs: Bach = 8,660 PLCO_M2012 = 8,862 LCDRAT = 8,631 USPSTF 2013 = 8,590	NR
2	Meza R [10]	Number of overdiagnosis lung cancer (Annual): PLCO_M2012, LCDRAT, and Bach = 80 to 120 Risk factor-based strategies = 61 to 95; Overdiagnosis rate per screen-detected lung cancer (Annual): PLCOM2012, LCDRAT, and Bach = 1.6 to 2.4% Risk factor-based strategies = 1.2 to 1.9%; Number of overdiagnosis lung cancer (Biennial): PLCO_M2012, LCDRAT, and Bach = 71 to 91 Risk factor-based strategies = 27 to 74; Overdiagnosis rate per screen-detected lung cancer (Biennial): PLCO_M2012, LCDRAT, and Bach = 1.4 to 1.8% Risk factor-based strategies = 0.5 to 1.5%	Number of radiation-related lung cancer deaths (Annual): PLCO_M2012, LCDRAT, and Bach = 10.9 to 43.3 Risk factor-based strategies = 13.8 to 55.0 Number of radiation-related lung cancer deaths (Biennial): PLCO_M2012, LCDRAT, and Bach = 11.0 to 25.7 Risk factor-based strategies = 6.8 to 23.6	LYs gained (Annual): PLCO_M2012, LCDRAT, and Bach = 4,087 to 8,387 Risk factor-based strategies = 4,058 to 8,186 LYs gained (Biennial): PLCO_M2012, LCDRAT, and Bach = 3,940 to 6,149 Risk factor-based strategies = 2,405 to 5,436 Women had higher LYs gained than men	NR
3	Toumazis I [11]	Overdiagnosis rate per screen-detected lung cancer PLCO_M2012 = 6.4% LCDRAT = NR USPSTF 2021 = 5.9% USPSTF 2013 = 6.1%	NR	LYs gained: PLCO_M2012 = 3800 LCDRAT = NR USPSTF 2021 = 4171 USPSTF 2013 = 2950 QALY gained: PLCO_M2012 = 1857 LCDRAT = NR USPSTF 2021 = 1857 USPSTF 2013 = 1427 Women had higher LYs and QALYs than men	ICER(compared with the strategy preceding it on the efficiency frontier): PLCO_M2012 = $94,659 per QALY LCDRAT = $97,284 per QALY USPSTF 2021 = $-28,783 per QALY USPSTF 2013 = $-1,425,040 per QALY Women had lower ICER values than men
4	Liu Y [12]	Number of overdiagnosed lung cancers: The most cost-effective risk model-based strategy = 38 China guideline 2021 = 22	Number of radiation-related lung cancer: The most cost-effective risk model-based strategy = 1.09 China guideline 2021 = 0.64	LYs gained The most cost-effective risk model-based strategy = 13,949 China guideline 2021 = 10,614 Average number of screens per LY gained: The most cost-effective risk model-based strategy = 77 China guideline 2021 = 56 QALYs: The most cost-effective risk model-based strategy = 1,153,425 China guideline 2021 = 1,153,601	INMB(compared with China guideline 2021): The most cost-effective risk model-based strategy = 1,032 CNY
5	Cressman S [13]	NR	NR	NR	ICER(compared with USPSTF 2013): PLCO_M2012=$355 per QALY INMB(compared with USPSTF 2013): PLCO_M2012(overall)₌$4294 PLCO_M2012(men) = $695 PLCO_M2012(women) = $6616
6	Tomonaga Y [14]	Overdiagnosis rate per screen-detected lung cancer: PLCO_M2012=4.9% CSC1 = 4.7% CSC2 = 4.7% CSC3 = 4.6% CSC4 = 4.6%	NR	LYs gained: PLCO_M2012=6678 CSC1 = 6441 CSC2 = 6810 CSC3 = 7133 CSC4 = 7394 QALYs gained: PLCO_M2012=5151 CSC1 = 4970 CSC2 = 5254 CSC3 = 5503 CSC4 = 5704	ICER(compared with the strategy preceding it on the efficiency frontier): PLCO_M2012=€29,852 per QALY CSC1 to CSC4 to be dominated by risk-based screening strategies
7(1)	Katki HA [15]	Proportion of overdiagnosed lung cancers to lung cancer deaths prevented: LCDRAT = 0.91 USPSTF 2013 = 0.93	NR	NR	NR
7(2)	Katki HA [15]	Proportion of overdiagnosed lung cancers to lung cancer deaths prevented: LCDRAT = 0.92 USPSTF 2013 = 0.93	NR	NR	NR
8	Tammemagi MC [16]	NR	NR	NR	NR
9(1)	Tammemagi MC [17]	NR	NR	NR	NR
9(2)	Tammemagi MC [17]	NR	NR	NR	NR
10	Ten Haaf K [18]	NR	NR	NR	NR
11	Tammemagi MC [19]	NR	NR	NR	NR
12	Cleven KL [20]	NR	NR	NR	NR
13(1)	Miranda-Filho A [21]	NR	NR	NR	NR
13(2)	Miranda-Filho A [21]	NR	NR	NR	NR
14	Park B [22]	NR	NR	NR	NR
15	Landy B [23]	NR	NR	NR	NR
16	Tammemägi MC [24]	NR	NR	LYs: PLCO_M2012 = 2,248.6 years USPSTF 2013 = 2,000.7 years Among patients diagnosed with lung cancer, women had a higher average life expectancy than men (16.8 years versus 11.7 years)	NR

Open in a new tab

LYs gained: increased Lys compared to no screening

QALYs gained: increased QALYs compared to no screening

NR not report, LCDRAT Lung Cancer Death Risk Assessment Tool, NLST National Lung Screening Trial, TSCE Two-Stage Clonal Expansion Model, USPSTF United States Preventive Services Task Force, NCCN-2 National Comprehensive Cancer Network Group 2, CSC Cancer Screening Committee, LYs life years, QALYs quality adjusted life years, ICER incremental cost-effectiveness ratio, INMB incremental net monetary benefit

Furthermore, Meza R et al [10] found that annual screening resulted in a greater gain of LYs compared to biennial screening, and initiating screening at an age of 50 led to a more significant increase in LYs saved than initiating at age 55.

Three studies conducted subgroup and sensitivity analyses, and revealed that more LYs or QALYs were gained in women than in men [10, 11, 24].

Harms

False positives and biopsies

Six studies reported on false-positive events (Table 3) [10, 12, 14, 15, 20, 23]. The majority of these studies consistently demonstrated that risk model-based strategies, when applied with identical screening ages and intervals, were more precise compared to the traditional risk factor-based strategies. This precision resulted in a lower incidence of false positives per lung cancer death prevented [15, 23], a reduced number of false-positive tests per screened individual [10, 12], and a decreased frequency of false-positive findings (i.e., suspicious nodules requiring follow-up that were not diagnosed as lung cancer by the end of the study) [20]. In the study conducted by Tomonaga Y et al [14], the most cost-effective risk-based strategy (1.6% PLCO_M2012) resulted in a higher number of false-positive screens compared to CSC1 (1,140 vs. 1,130 per 100,000 individuals), but it was lower than the numbers observed in CSC2 (1,267 per 100,000 individuals), CSC3 (1,396 per 100,000 individuals), and CSC4 (1,511 per 100,000 individuals).

Biopsies were reported in three studies [10, 14, 20]. Meza et al [10] indicated that risk model-based strategies were associated with more biopsies (annual: 508 to 1,015, biennial: 506 to 767, per 100,000 individuals) than the most efficient risk factor-based strategies (annual: 422, biennial: 241, per 100,000 individuals). Further, the proportions of over-biopsies resulting in non-lung cancer diagnosis were similar using NCCN-2 (4/54), USPTSF 2013 (2/30), PLCO_M2012 (2/34), and Bach (2/31) [20]. In the study conducted by Tomonaga Y et al [14], the number of biopsies resulting from false positive screenings was higher (454 per 100,000 individuals) when using the most cost-effective risk-based strategy compared to CSC 1 (445 per 100,000 individuals) and CSC 2 (498 per 100,000 individuals), yet lower than that observed with CSC3 (548 per 100,000 individuals) and CSC4 (593 per 100,000 individuals).

Additionally, Meza et al [10] further revealed that annual screening yielded a higher number of false positives and required more biopsies per person than biennial screening, and initiating screenings at age of 50 increased the incidence of false positives and the need for biopsies than initiating at age 55.

Overdiagnosis

One retrospective cohort study [15] and five modeling studies [9–12, 14] reported on the issue of overdiagnosis (Table 4). In this retrospective cohort study [15], the Lung Cancer Death Risk Assessment Tool (LCDRAT) model, when applied with thresholds of 1.9% and 1.7%, resulted in a slightly lower proportion of overdiagnosed lung cancers per lung cancer death prevented (1.9%: 0.91, 1.7%: 0.92) compared to USPSTF 2013 criteria (0.93). Four modeling studies applied natural history models to estimate long-term overdiagnosis rates across various screening strategies, and found that risk model-based strategies tended to identify a greater absolute number of overdiagnosed cases than those based on risk factors because of predominantly selecting older individuals [9–11, 14]. For example, Ten Haaf K et al. [9] found that, for a 1950 U.S. birth cohort, starting lung cancer screening between 55 and 80 years old with risk-model based strategies led to an average first screening age of 61.7 to 65.6 years. This was higher than the 55.6 years recommended by the USPSTF 2013 criteria. As a result, risk-model strategies had more overdiagnoses than the USPSTF 2013 criteria, with 18.5–45.9% more cases of overdiagnosed cancer. Nevertheless, the personalized screening intervals suggested by Liu Y et al. successfully decreased the overall number of overdiagnosed lung cancers by reducing the frequency of LDCT scans, compared to the standards set by the China guideline 2021 [12].

Additionally, Meza et al [10] found annual screening was associated with more overdiagnosis than biennial screening, and initiating screening at the age of 50 tended to result in more overdiagnosis than initiating at age 55.

Radiation-related cancer

Two studies separately modeled radiation-related lung cancer deaths and lung cancer cases associated with radiation exposure (Table 4) [10, 12]. Both of the two studies found that more LDCT scans correlated with higher rates of radiation-induced lung cancer. Meza R et al [10] observed that, for the 1960 U.S. birth cohort, the age of screening tended to shift to older ages for the risk model-based screening strategies compared with the risk factor-based strategies. This shift resulted in lower numbers of screens per person and radiation-related lung cancer deaths. Specifically, the average screening age was 65.1 years with the USPSTF 2013 recommendation, younger than the ages for screenings using the PLCO_M2012 (69.6 years), LCDRAT (70.1 years), and Bach-based recommendation (70.3 years). Consequently, the radiation-related lung cancer deaths was 20.6 per 100,000 individuals using the 2013 USPSTF recommendation, higher than that with the PLCO_M2012 (15.8 per 100,000 individuals), LCDRAT (15.6 per 100,000 individuals), and Bach-based s (15.5 per 100,000 individuals) recommendations. In another modelling study conducted by Liu Y et al [12], the most cost-effective risk-model based strategy also yield fewer LDCT screens compared to China guideline 2021(591,440 vs 1,071,810 per 100,000 individuals), as well as a lower number of radiation-related lung cancer (0.64 vs 1.09 per 100,000 individuals).

Furthermore, Meza R et al [10] found annual screening was linked to more radiation-related lung cancer deaths than biennial screening, and initiating screening at earlier age also resulted in more radiation-related lung cancer deaths than initiating at delayed age. No studies have compared the two screening strategies in terms of their potential to induce other types of cancers due to radiation exposure.

Cost-effectiveness

Four studies reported on the cost-effectiveness of LDCT screening employing various screening approaches (Table 4) [11–14]. These studies highlighted that the cost-effectiveness of risk model-based strategies hinged on choosing the correct risk threshold for determining LDCT screening eligibility. It is only with the right threshold that the cost-effectiveness of these risk model-based strategies can outperform those based on traditional risk factors. For example, Toumazis I et al [11] discovered that, for the 1960 U.S. birth cohort, only risk model-based screening strategies with a 6-year PLCO_M2012-based risk threshold of above 1.2% were deemed cost-effective. Additionally, two studies conducted subgroup analysis, revealing that LDCT screening was more cost-effective for women than for men [11, 13]. Sensitivity analyses found that excluding individuals with limited life expectancies (< 5 years) would further improve the cost-effectiveness of LDCT screening [11, 13].

Discussion

This is the first systematic review comparing the benefits, harms, and cost-effectiveness of risk model- and risk factor-based strategies in identifying individuals eligible for LDCT screening. While previous systematic reviews had concentrated on the development and potential uses of risk prediction models [27, 28], our comprehensive analysis of sixteen articles revealed that the risk model-based strategies could offer a balanced approach to the benefits and harms of LDCT screening. Identifying the optimal risk threshold is pivotal to enhancing the cost-effectiveness of risk-model based strategies, ensuring that screening initiatives are both beneficial and fiscally responsible.

For the heavy smokers, risk model-based strategies generally achieved higher sensitivity, without losing specificity. However, when heavy smokers were also firefighters, the risk-factor strategies, such as the NCCN-2 guidance, demonstrated higher sensitivity than risk model-based strategies like PLCO_M2012 and Bach, while increasing false positive scans [20]. This may be due to the NCCN-2's consideration of intense smoke exposure and its more lenient age-smoking criteria, which leads to high sensitivity but low specificity. Consequently, for individuals at an elevated risk of lung cancer, the application of risk-model based strategies should be approached with caution to avoid the possibility of missed diagnoses.

The risk-threshold is the cut-off risk level at which an individual is eligible for LDCT screening and serving as a crucial measure of the procedure's benefits and harms [29–31]. Risk model-based strategies have uniformly resulted in a lower incidence of radiation-induced lung cancers, a consequence of conducting fewer LDCT scans. However, their impact on enhancing LYs, QALYs, and on reducing lung cancer deaths and overdiagnosis has been inconsistent. Many studies have suggested that risk model-based strategies yielded only modest gains in LYs, but considerably more overdiagnosis, particularly when the applied risk thresholds aimed to match the efficiency of lung cancer death reduction, or when they select a similar number or proportion of eligible individuals as risk factor-based strategies [9, 10, 24, 32]. These approaches were more likely to identify older individuals and those with a history of comorbidities, who have a lower average benefit from screening, as eligible. Ideally, risk-thresholds should provide an optimal balance between long-term benefits and harms. Therefore, health economics evaluation is essential to provide a holistic framework for assessing these long-term outcomes, ensuring that screening strategies are not only effective but also economically viable and sustainable.

We found that the benefits and harms of LDCT screening were also influenced by screening interval and starting age. Annual screenings and those initiated at earlier ages generally prevented more lung cancer deaths and gained more life years compared to biennial screenings or those started later. However, they also generated more false positives, biopsies, overdiagnosis, and radiation-related lung cancer deaths. There has been discussion on the determination of optimal screening interval and age. Retrospective analyses of NLST and PLCO found that biennial screening maintain diagnostic sensitivity and significantly reduced the associated harms compared with annual screening [33]. Moreover, biennial screening was reported to offer equivalent or superior cost-effectiveness compared with annual screening [34, 35], suggesting that extending the screening interval appropriately could yield benefits; however, interval cancers were significantly increased and a higher proportion of late-stage lung cancer was detected with longer screening intervals [36]. Risk-stratified screening intervals, as proposed by Liu Y et al [12], presented a promising alternative. By customizing screening intervals for individuals at different risk levels, it could improve the cost-effectiveness of screening while simultaneously reduced the risk of overdiagnosis and radiation exposure. A previous Chinese study [37] set a risk-adapted starting age for LDCT screening by considering a comprehensive set of risk factors and using a 10-year cumulative risk of lung cancer for heavy smokers as the threshold. However, this study did not account for long-term benefits or potential risks, nor did it evaluate the cost-effectiveness of LDCT screening. Consequently, further research is essential to determine the optimal screening age and interval.

Women were found to gain more benefits, and LDCT screening for women was more cost-effective than that for men [10, 11, 13, 24]. This advantage is primarily attributed to the fact that women diagnosed with lung cancer are more likely to have early-stage diseases, longer life expectancy, greater adherence to treatment and smoking cessation, and fewer comorbidities than men [24, 38–40]. However, it is noteworthy that the risk of lung cancer and other significant cancers induced by the ionizing radiation in LDCT is elevated in women compared to men [41, 42]. The preclinical phase of lung cancer was also longer in women, potentially increasing the likelihood of overdiagnosis [43, 44]. Given their generally lower smoking rates and exposure to cigarette smoke, many female lung cancer patients often do not meet the current eligibility screening criteria, particularly within risk factor-based strategies [45, 46]. Therefore, to optimize the benefits and cost-effectiveness of LDCT screening, it is imperative to develop sex-specific eligibility criteria.

The studies included in this review had several limitations. First, the assessment of cancer risk was conducted at a single time-point across all cohort studies, thereby neglecting the dynamic trajectory of individual risk factors that evolve due to aging, behavioral changes such as smoking patterns, and other mutable risk factors [28]. The incorporation of dynamic risk assessments with timely recalibrations of LDCT screening protocols could potentially augment QALYs and curtail the rate of false positives, thereby surpassing the efficacy of a static, one-size-fits-all threshold strategy [47]. Second, the risk-models employed in the included studies considered only factors such as age and smoking behavior, neglecting the significant role of biomarkers in risk assessment and lung cancer screening. The inclusion of biomarkers is particularly crucial in Asian populations, given the high prevalence of lung cancer among individuals who have never smoked [48]. Third, the included cohort studies consisted of only one prospective study, while the others were retrospective, and most had relatively short follow-up periods. Such limitations may engender less reliable estimates of overdiagnosis. The overdiagnosis observed during shorter follow-up intervals may, in part, be an artifact of early cancer detection conferred by the lead time bias [49, 50]. Extended follow-up durations are capable of mitigating the impact of lead time, thereby reducing overdiagnosis [51, 52]. Modeling analyses indicate lead time associated with CT screening may be up to 12 years [53]. The most reliable method of estimating overdiagnosis remains the execution of well-designed, prospective randomized controlled trials with sufficient follow-up periods. Currently, the Yorkshire Lung Screening Trial is conducting a large prospective randomized controlled study to compare performance of PLCO_M2012 (threshold ≥ 1.51%), Liverpool Lung Project (V.2) (threshold ≥ 5%), and USPSTF 2013 criteria [54]. Fourth, most studies presumed LDCT screening adherence rates for risk model-based strategies to be equivalent to those for risk factor-based strategies (i.e., 100%). This assumption may introduce a bias in the comparative analysis, given that adherence rates to risk model-based strategies are often noted to exceed those of risk factor-based approaches [55]. Finally, research to date has focused mainly on developed countries; hence, the results may not be representative of developing countries, because of differences in genetics, smoking status, environmental exposures, and healthcare access [56].

This review has several limitations. First, it was restricted to English language and to published full-text articles; hence, the conclusions of the studies may be affected by publication bias. Second, a quantitative meta-analysis could not be perfomred as only published work was reviewed. Therefore, a qualitative analysis was conducted.

Conclusion

In conclusion, risk model-based strategies exhibited greater potential than risk factor-based strategies to enhance the benefit-to-harm ratio of screening and render it more cost-effective. Determination of an appropriate risk-threshold using a risk prediction model is a crucial consideration in the selection of individuals eligible for LDCT screening. Given that individuals at higher risk are often older and have a reduced life expectancy on average, the cost-effectiveness of LDCT screening must be carefully weighed against its potential health benefits when setting these thresholds. Risk-stratified screening intervals would be a promising alternative to enhance the benefits and cost-effectiveness of LDCT screening. There is also a compelling need for future research to develop sex-specific eligibility criteria for screening.

Authors’ contributions

QG and YL designed the study and write the manuscript. CF and XL responsible in data acquisition. YQ and SZ analyzed and interpreted the study results. All authors read and approved the final manuscript.

Funding

This work was supported by the Henan Province Key research and Development Project (grant number 221111310200), Henan Province Science and Technology Research Program Project (grant number 242102311158), and Henan Province Medical Science and Technology Public Relations Plan Province Department joint construction project (SBGJ202403020).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Yin Liu and Qingchao Geng first author.

Contributor Information

Youlin Qiao, Email: qiaoy@cicams.ac.cn.

Shaokai Zhang, Email: shaokaizhang@126.com.

References

1.Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49. 10.3322/caac.21660. [DOI] [PubMed] [Google Scholar]
2.Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395–409. 10.1056/NEJMoa1102873. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Brodersen J, Voss T, Martiny F, Siersma V, Barratt A, Heleno B. Overdiagnosis of lung cancer with low-dose computed tomography screening: meta-analysis of the randomised clinical trials. Breathe (Sheff). 2020;16(1):200013. 10.1183/20734735.0013-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Heleno B, Thomsen MF, Rodrigues DS, Jørgensen KJ, Brodersen J. Quantification of harms in cancer screening trials: literature review. BMJ. 2013;347:f5334. 10.1136/bmj.f5334. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Moyer VA. Screening for lung cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2014;160(5):330–8. 10.7326/m13-2771. [DOI] [PubMed] [Google Scholar]
6.Wood DE, Kazerooni EA, Aberle D, Berman A, Brown LM, Eapen GA, et al. NCCN guidelines® insights: lung cancer screening, version 1.2022. J Natl Compr Canc Netw. 2022;20(7):754–64. 10.6004/jnccn.2022.0036. [DOI] [PubMed] [Google Scholar]
7.Kavanagh J, Liu G, Menezes R, O’Kane GM, McGregor M, Tsao M, et al. Importance of long-term low-dose CT follow-up after negative findings at previous lung cancer screening. Radiology. 2018;289(1):218–24. 10.1148/radiol.2018180053. [DOI] [PubMed] [Google Scholar]
8.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535. 10.1136/bmj.b2535. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Ten Haaf K, Bastani M, Cao P, Jeon J, Toumazis I, Han SS, et al. A comparative modeling analysis of risk-based lung cancer screening strategies. J Natl Cancer Inst. 2020;112(5):466–79. 10.1093/jnci/djz164. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Meza R, Jeon J, Toumazis I, Ten Haaf K, Cao P, Bastani M, et al. Evaluation of the benefits and harms of lung cancer screening with low-dose computed tomography: modeling study for the US preventive services task force. JAMA. 2021;325(10):988–97. 10.1001/jama.2021.1077. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Toumazis I, Cao P, de Nijs K, Bastani M, Munshi V, Hemmati M, et al. Risk model-based lung cancer screening : a cost-effectiveness analysis. Ann Intern Med. 2023;176(3):320–32. 10.7326/m22-2216. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Liu Y, Xu H, Lv L, Wang X, Kang R, Guo X, et al. Risk-based lung cancer screening in heavy smokers: a benefit-harm and cost-effectiveness modeling study. BMC Med. 2024;22(1):73. 10.1186/s12916-024-03292-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Cressman S, Weber MF, Ngo PJ, Wade S, Behar Harpaz S, Caruana M, et al. Economic impact of using risk models for eligibility selection to the International lung screening Trial. Lung Cancer. 2023;176:38–45. 10.1016/j.lungcan.2022.12.011. [DOI] [PubMed] [Google Scholar]
14.Tomonaga Y, de Nijs K, Bucher HC, de Koning H, Ten Haaf K. Cost-effectiveness of risk-based low-dose computed tomography screening for lung cancer in Switzerland. Int J Cancer. 2024;154(4):636–47. 10.1002/ijc.34746. [DOI] [PubMed] [Google Scholar]
15.Katki HA, Kovalchik SA, Berg CD, Cheung LC, Chaturvedi AK. Development and validation of risk models to select ever-smokers for CT lung cancer screening. JAMA. 2016;315(21):2300–11. 10.1001/jama.2016.6255. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Tammemägi MC, Katki HA, Hocking WG, Church TR, Caporaso N, Kvale PA, et al. Selection criteria for lung-cancer screening. N Engl J Med. 2013;368(8):728–36. 10.1056/NEJMoa1211776. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Tammemägi MC, Church TR, Hocking WG, Silvestri GA, Kvale PA, Riley TL, et al. Evaluation of the lung cancer risks at which to screen ever- and never-smokers: screening rules applied to the PLCO and NLST cohorts. PLoS Med. 2014;11(12):e1001764. 10.1371/journal.pmed.1001764. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Ten Haaf K, Jeon J, Tammemägi MC, Han SS, Kong CY, Plevritis SK, et al. Risk prediction models for selection of lung cancer screening candidates: a retrospective validation study. PLoS Med. 2017;14(4):e1002277. 10.1371/journal.pmed.1002277. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Tammemagi MC, Darling GE, Schmidt H, Llovet D, Buchanan DN, Leung Y, et al. Selection of individuals for lung cancer screening based on risk prediction model performance and economic factors - the Ontario experience. Lung Cancer. 2021;156:31–40. 10.1016/j.lungcan.2021.04.005. [DOI] [PubMed] [Google Scholar]
20.Cleven KL, Vaeth B, Zeig-Owens R, Colbeth HL, Jaber N, Schwartz T, et al. Performance of risk factor-based guidelines and model-based chest CT lung cancer screening in world trade center-exposed fire department rescue/recovery workers. Chest. 2021;159(5):2060–71. 10.1016/j.chest.2020.11.028. [DOI] [PubMed] [Google Scholar]
21.Miranda-Filho A, Charvat H, Bray F, Migowski A, Cheung LC, Vaccarella S, et al. A modeling analysis to compare eligibility strategies for lung cancer screening in Brazil. EClinicalMedicine. 2021;42:101176. 10.1016/j.eclinm.2021.101176. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Park B, Kim Y, Lee J, Lee N, Jang SH. Risk-based prediction model for selecting eligible population for lung cancer screening among ever smokers in Korea. Transl Lung Cancer Res. 2021;10(12):4390–402. 10.21037/tlcr-21-566. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Landy R, Cheung LC, Berg CD, Chaturvedi AK, Robbins HA, Katki HA. Contemporary implications of U.S. preventive services task force and risk-based guidelines for lung cancer screening eligibility in the United States. Ann Intern Med. 2019;171(5):384–6. 10.7326/m18-3617. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Tammemägi MC, Ruparel M, Tremblay A, Myers R, Mayo J, Yee J, et al. USPSTF2013 versus PLCOm2012 lung cancer screening eligibility criteria (International Lung Screening Trial): interim analysis of a prospective cohort study. Lancet Oncol. 2022;23(1):138–48. 10.1016/s1470-2045(21)00590-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.He J, Li N, Chen WQ, Wu N, Shen HB, Jiang Y, et al. China guideline for the screening and early detection of lung cancer(2021, Beijing). Zhonghua Zhong Liu Za Zhi. 2021;43(3):243–68. 10.3760/cma.j.cn112152-20210119-00060. [DOI] [PubMed] [Google Scholar]
26.Cancer Screening Committee. Cancer screening committee recommendation on low-dose CT screening for lung cancer. 2022
27.Toumazis I, Bastani M, Han SS, Plevritis SK. Risk-based lung cancer screening: a systematic review. Lung Cancer. 2020;147:154–86. 10.1016/j.lungcan.2020.07.007. [DOI] [PubMed] [Google Scholar]
28.Ten Haaf K, van der Aalst CM, de Koning HJ, Kaaks R, Tammemägi MC. Personalising lung cancer screening: an overview of risk-stratification opportunities and challenges. Int J Cancer. 2021;149(2):250–63. 10.1002/ijc.33578. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making Nov-Dec. 2006;26(6):565–74. 10.1177/0272989x06295361. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Qu J, Li M, Wang Y, Duan X, Luo H, Zhao C, et al. Predicting the risk of pulmonary arterial hypertension in systemic lupus erythematosus: a Chinese systemic lupus erythematosus treatment and research group cohort study. Arthritis Rheumatol. 2021;73(10):1847–55. 10.1002/art.41740. [DOI] [PubMed] [Google Scholar]
31.Moons KG, Stijnen T, Michel BC, Büller HR, Van Es GA, Grobbee DE, et al. Application of treatment thresholds to diagnostic-test evaluation: an alternative to the comparison of areas under receiver operating characteristic curves. Med Decis Making Oct-Dec. 1997;17(4):447–54. 10.1177/0272989x9701700410. [DOI] [PubMed] [Google Scholar]
32.Hüsing A, Kaaks R. Risk prediction models versus simplified selection criteria to determine eligibility for lung cancer screening: an analysis of German federal-wide survey and incidence data. Eur J Epidemiol. 2020;35(10):899–912. 10.1007/s10654-020-00657-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.de Koning HJ, Meza R, Plevritis SK, ten Haaf K, Munshi VN, Jeon J, et al. Benefits and harms of computed tomography lung cancer screening strategies: a comparative modeling study for the U.S. Preventive Services Task Force. Ann Intern Med. 2014;160(5):311–20. 10.7326/m13-2316. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Field JK, Oudkerk M, Pedersen JH, Duffy SW. Prospects for population screening and diagnosis of lung cancer. Lancet. 2013;382(9893):732–41. 10.1016/s0140-6736(13)61614-1. [DOI] [PubMed] [Google Scholar]
35.Duffy SW, Field JK, Allgood PC, Seigneurin A. Translation of research results to simple estimates of the likely effect of a lung cancer screening programme in the United Kingdom. Br J Cancer. 2014;110(7):1834–40. 10.1038/bjc.2014.63. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Yousaf-Khan U, van der Aalst C, de Jong PA, Heuvelmans M, Scholten E, Lammers JW, et al. Final screening round of the NELSON lung cancer screening trial: the effect of a 2.5-year screening interval. Thorax. 2017;72(1):48–56. 10.1136/thoraxjnl-2016-208655. [DOI] [PubMed] [Google Scholar]
37.Wang C, Dong X, Tan F, Wu Z, Huang Y, Zheng Y, et al. Risk-adapted starting age of personalized lung cancer screening: a population-based, prospective cohort study in China. Chest. 2024;165(6):1538–54. 10.1016/j.chest.2024.01.031. [DOI] [PubMed] [Google Scholar]
38.Radkiewicz C, Dickman PW, Johansson ALV, Wagenius G, Edgren G, Lambe M. Sex and survival in non-small cell lung cancer: a nationwide cohort study. PLoS ONE. 2019;14(6):e0219206. 10.1371/journal.pone.0219206. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Becker N, Motsch E, Trotter A, Heussel CP, Dienemann H, Schnabel PA, et al. Lung cancer mortality reduction by LDCT screening-results from the randomized German LUSI trial. Int J Cancer. 2020;146(6):1503–13. 10.1002/ijc.32486. [DOI] [PubMed] [Google Scholar]
40.Pasquinelli MM, Tammemägi MC, Kovitz KL, Durham ML, Deliu Z, Guzman A, et al. Addressing sex disparities in lung cancer screening eligibility: USPSTF vs PLCOm2012 criteria. Chest. 2022;161(1):248–56. 10.1016/j.chest.2021.06.066. [DOI] [PubMed] [Google Scholar]
41.Larke FJ, Kruger RL, Cagnon CH, Flynn MJ, McNitt-Gray MM, Wu X, et al. Estimated radiation dose associated with low-dose chest CT of average-size participants in the National Lung screening trial. AJR Am J Roentgenol. 2011;197(5):1165–9. 10.2214/ajr.11.6533. [DOI] [PubMed] [Google Scholar]
42.Rampinelli C, De Marco P, Origgi D, Maisonneuve P, Casiraghi M, Veronesi G, et al. Exposure to low dose computed tomography for lung cancer screening and risk of cancer: secondary analysis of trial data and risk-benefit analysis. BMJ. 2017;356:j347. 10.1136/bmj.j347. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Han SS, Ten Haaf K, Hazelton WD, Munshi VN, Jeon J, Erdogan SA, et al. The impact of overdiagnosis on the selection of efficient lung cancer screening strategies. Int J Cancer. 2017;140(11):2436–43. 10.1002/ijc.30602. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Ten Haaf K, van Rosmalen J, de Koning HJ. Lung cancer detectability by test, histology, stage, and gender: estimates from the NLST and the PLCO trials. Cancer Epidemiol Biomarkers Prev. 2015;24(1):154–61. 10.1158/1055-9965.Epi-14-0745. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Zang EA, Wynder EL. Differences in lung cancer risk between men and women: examination of the evidence. J Natl Cancer Inst. 1996;88(3–4):183–92. 10.1093/jnci/88.3-4.183. [DOI] [PubMed] [Google Scholar]
46.Berlin I. Incidence of lung cancer among young women. N Engl J Med. 2018;379(10):988. 10.1056/NEJMc1808250. [DOI] [PubMed] [Google Scholar]
47.Toumazis I, Alagoz O, Leung A, Plevritis SK. A risk-based framework for assessing real-time lung cancer screening eligibility that incorporates life expectancy and past screening findings. Cancer. 2021;127(23):4432–46. 10.1002/cncr.33835. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Adams SJ, Stone E, Baldwin DR, Vliegenthart R, Lee P, Fintelmann FJ. Lung cancer screening. Lancet. 2023;401(10374):390–408. 10.1016/s0140-6736(22)01694-4. [DOI] [PubMed] [Google Scholar]
49.Welch HG, Black WC. Overdiagnosis in cancer. J Natl Cancer Inst. 2010;102(9):605–13. 10.1093/jnci/djq099. [DOI] [PubMed] [Google Scholar]
50.Brodersen J, Schwartz LM, Woloshin S. Overdiagnosis: how cancer screening can turn indolent pathology into illness. APMIS. 2014;122(8):683–9. 10.1111/apm.12278. [DOI] [PubMed] [Google Scholar]
51.de Koning HJ, van der Aalst CM, de Jong PA, Scholten ET, Nackaerts K, Heuvelmans MA, et al. Reduced lung-cancer mortality with volume CT screening in a randomized trial. N Engl J Med. 2020;382(6):503–13. 10.1056/NEJMoa1911793. [DOI] [PubMed] [Google Scholar]
52.Duffy SW, Parmar D. Overdiagnosis in breast cancer screening: the importance of length of observation period and lead time. Breast Cancer Res. 2013;15(3):R41. 10.1186/bcr3427. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Ten Haaf K, de Koning HJ. Overdiagnosis in lung cancer screening: why modelling is essential. J Epidemiol Community Health. 2015;69(11):1035–9. 10.1136/jech-2014-204079. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Crosbie PA, Gabe R, Simmonds I, Kennedy M, Rogerson S, Ahmed N, et al. Yorkshire Lung Screening Trial (YLST): protocol for a randomised controlled trial to evaluate invitation to community-based low-dose CT screening for lung cancer versus usual care in a targeted population at risk. BMJ Open. 2020;10(9):e037075. 10.1136/bmjopen-2020-037075. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Kim DD, Cohen JT, Wong JB, Mohit B, Fendrick AM, Kent DM, et al. Targeted incentive programs for lung cancer screening can improve population health and economic efficiency. Health Aff (Millwood). 2019;38(1):60–7. 10.1377/hlthaff.2018.05148. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Barta JA, Powell CA, Wisnivesky JP. Global epidemiology of lung cancer. Ann Glob Health. 2019;85(1):8. 10.5334/aogh.2419. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

No datasets were generated or analysed during the current study.

[CR1] 1.Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49. 10.3322/caac.21660. [DOI] [PubMed] [Google Scholar]

[CR2] 2.Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395–409. 10.1056/NEJMoa1102873. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Brodersen J, Voss T, Martiny F, Siersma V, Barratt A, Heleno B. Overdiagnosis of lung cancer with low-dose computed tomography screening: meta-analysis of the randomised clinical trials. Breathe (Sheff). 2020;16(1):200013. 10.1183/20734735.0013-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Heleno B, Thomsen MF, Rodrigues DS, Jørgensen KJ, Brodersen J. Quantification of harms in cancer screening trials: literature review. BMJ. 2013;347:f5334. 10.1136/bmj.f5334. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Moyer VA. Screening for lung cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2014;160(5):330–8. 10.7326/m13-2771. [DOI] [PubMed] [Google Scholar]

[CR6] 6.Wood DE, Kazerooni EA, Aberle D, Berman A, Brown LM, Eapen GA, et al. NCCN guidelines® insights: lung cancer screening, version 1.2022. J Natl Compr Canc Netw. 2022;20(7):754–64. 10.6004/jnccn.2022.0036. [DOI] [PubMed] [Google Scholar]

[CR7] 7.Kavanagh J, Liu G, Menezes R, O’Kane GM, McGregor M, Tsao M, et al. Importance of long-term low-dose CT follow-up after negative findings at previous lung cancer screening. Radiology. 2018;289(1):218–24. 10.1148/radiol.2018180053. [DOI] [PubMed] [Google Scholar]

[CR8] 8.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535. 10.1136/bmj.b2535. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Ten Haaf K, Bastani M, Cao P, Jeon J, Toumazis I, Han SS, et al. A comparative modeling analysis of risk-based lung cancer screening strategies. J Natl Cancer Inst. 2020;112(5):466–79. 10.1093/jnci/djz164. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Meza R, Jeon J, Toumazis I, Ten Haaf K, Cao P, Bastani M, et al. Evaluation of the benefits and harms of lung cancer screening with low-dose computed tomography: modeling study for the US preventive services task force. JAMA. 2021;325(10):988–97. 10.1001/jama.2021.1077. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Toumazis I, Cao P, de Nijs K, Bastani M, Munshi V, Hemmati M, et al. Risk model-based lung cancer screening : a cost-effectiveness analysis. Ann Intern Med. 2023;176(3):320–32. 10.7326/m22-2216. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Liu Y, Xu H, Lv L, Wang X, Kang R, Guo X, et al. Risk-based lung cancer screening in heavy smokers: a benefit-harm and cost-effectiveness modeling study. BMC Med. 2024;22(1):73. 10.1186/s12916-024-03292-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Cressman S, Weber MF, Ngo PJ, Wade S, Behar Harpaz S, Caruana M, et al. Economic impact of using risk models for eligibility selection to the International lung screening Trial. Lung Cancer. 2023;176:38–45. 10.1016/j.lungcan.2022.12.011. [DOI] [PubMed] [Google Scholar]

[CR14] 14.Tomonaga Y, de Nijs K, Bucher HC, de Koning H, Ten Haaf K. Cost-effectiveness of risk-based low-dose computed tomography screening for lung cancer in Switzerland. Int J Cancer. 2024;154(4):636–47. 10.1002/ijc.34746. [DOI] [PubMed] [Google Scholar]

[CR15] 15.Katki HA, Kovalchik SA, Berg CD, Cheung LC, Chaturvedi AK. Development and validation of risk models to select ever-smokers for CT lung cancer screening. JAMA. 2016;315(21):2300–11. 10.1001/jama.2016.6255. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Tammemägi MC, Katki HA, Hocking WG, Church TR, Caporaso N, Kvale PA, et al. Selection criteria for lung-cancer screening. N Engl J Med. 2013;368(8):728–36. 10.1056/NEJMoa1211776. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Tammemägi MC, Church TR, Hocking WG, Silvestri GA, Kvale PA, Riley TL, et al. Evaluation of the lung cancer risks at which to screen ever- and never-smokers: screening rules applied to the PLCO and NLST cohorts. PLoS Med. 2014;11(12):e1001764. 10.1371/journal.pmed.1001764. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Ten Haaf K, Jeon J, Tammemägi MC, Han SS, Kong CY, Plevritis SK, et al. Risk prediction models for selection of lung cancer screening candidates: a retrospective validation study. PLoS Med. 2017;14(4):e1002277. 10.1371/journal.pmed.1002277. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Tammemagi MC, Darling GE, Schmidt H, Llovet D, Buchanan DN, Leung Y, et al. Selection of individuals for lung cancer screening based on risk prediction model performance and economic factors - the Ontario experience. Lung Cancer. 2021;156:31–40. 10.1016/j.lungcan.2021.04.005. [DOI] [PubMed] [Google Scholar]

[CR20] 20.Cleven KL, Vaeth B, Zeig-Owens R, Colbeth HL, Jaber N, Schwartz T, et al. Performance of risk factor-based guidelines and model-based chest CT lung cancer screening in world trade center-exposed fire department rescue/recovery workers. Chest. 2021;159(5):2060–71. 10.1016/j.chest.2020.11.028. [DOI] [PubMed] [Google Scholar]

[CR21] 21.Miranda-Filho A, Charvat H, Bray F, Migowski A, Cheung LC, Vaccarella S, et al. A modeling analysis to compare eligibility strategies for lung cancer screening in Brazil. EClinicalMedicine. 2021;42:101176. 10.1016/j.eclinm.2021.101176. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Park B, Kim Y, Lee J, Lee N, Jang SH. Risk-based prediction model for selecting eligible population for lung cancer screening among ever smokers in Korea. Transl Lung Cancer Res. 2021;10(12):4390–402. 10.21037/tlcr-21-566. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Landy R, Cheung LC, Berg CD, Chaturvedi AK, Robbins HA, Katki HA. Contemporary implications of U.S. preventive services task force and risk-based guidelines for lung cancer screening eligibility in the United States. Ann Intern Med. 2019;171(5):384–6. 10.7326/m18-3617. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Tammemägi MC, Ruparel M, Tremblay A, Myers R, Mayo J, Yee J, et al. USPSTF2013 versus PLCOm2012 lung cancer screening eligibility criteria (International Lung Screening Trial): interim analysis of a prospective cohort study. Lancet Oncol. 2022;23(1):138–48. 10.1016/s1470-2045(21)00590-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.He J, Li N, Chen WQ, Wu N, Shen HB, Jiang Y, et al. China guideline for the screening and early detection of lung cancer(2021, Beijing). Zhonghua Zhong Liu Za Zhi. 2021;43(3):243–68. 10.3760/cma.j.cn112152-20210119-00060. [DOI] [PubMed] [Google Scholar]

[CR26] 26.Cancer Screening Committee. Cancer screening committee recommendation on low-dose CT screening for lung cancer. 2022

[CR27] 27.Toumazis I, Bastani M, Han SS, Plevritis SK. Risk-based lung cancer screening: a systematic review. Lung Cancer. 2020;147:154–86. 10.1016/j.lungcan.2020.07.007. [DOI] [PubMed] [Google Scholar]

[CR28] 28.Ten Haaf K, van der Aalst CM, de Koning HJ, Kaaks R, Tammemägi MC. Personalising lung cancer screening: an overview of risk-stratification opportunities and challenges. Int J Cancer. 2021;149(2):250–63. 10.1002/ijc.33578. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making Nov-Dec. 2006;26(6):565–74. 10.1177/0272989x06295361. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Qu J, Li M, Wang Y, Duan X, Luo H, Zhao C, et al. Predicting the risk of pulmonary arterial hypertension in systemic lupus erythematosus: a Chinese systemic lupus erythematosus treatment and research group cohort study. Arthritis Rheumatol. 2021;73(10):1847–55. 10.1002/art.41740. [DOI] [PubMed] [Google Scholar]

[CR31] 31.Moons KG, Stijnen T, Michel BC, Büller HR, Van Es GA, Grobbee DE, et al. Application of treatment thresholds to diagnostic-test evaluation: an alternative to the comparison of areas under receiver operating characteristic curves. Med Decis Making Oct-Dec. 1997;17(4):447–54. 10.1177/0272989x9701700410. [DOI] [PubMed] [Google Scholar]

[CR32] 32.Hüsing A, Kaaks R. Risk prediction models versus simplified selection criteria to determine eligibility for lung cancer screening: an analysis of German federal-wide survey and incidence data. Eur J Epidemiol. 2020;35(10):899–912. 10.1007/s10654-020-00657-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.de Koning HJ, Meza R, Plevritis SK, ten Haaf K, Munshi VN, Jeon J, et al. Benefits and harms of computed tomography lung cancer screening strategies: a comparative modeling study for the U.S. Preventive Services Task Force. Ann Intern Med. 2014;160(5):311–20. 10.7326/m13-2316. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Field JK, Oudkerk M, Pedersen JH, Duffy SW. Prospects for population screening and diagnosis of lung cancer. Lancet. 2013;382(9893):732–41. 10.1016/s0140-6736(13)61614-1. [DOI] [PubMed] [Google Scholar]

[CR35] 35.Duffy SW, Field JK, Allgood PC, Seigneurin A. Translation of research results to simple estimates of the likely effect of a lung cancer screening programme in the United Kingdom. Br J Cancer. 2014;110(7):1834–40. 10.1038/bjc.2014.63. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Yousaf-Khan U, van der Aalst C, de Jong PA, Heuvelmans M, Scholten E, Lammers JW, et al. Final screening round of the NELSON lung cancer screening trial: the effect of a 2.5-year screening interval. Thorax. 2017;72(1):48–56. 10.1136/thoraxjnl-2016-208655. [DOI] [PubMed] [Google Scholar]

[CR37] 37.Wang C, Dong X, Tan F, Wu Z, Huang Y, Zheng Y, et al. Risk-adapted starting age of personalized lung cancer screening: a population-based, prospective cohort study in China. Chest. 2024;165(6):1538–54. 10.1016/j.chest.2024.01.031. [DOI] [PubMed] [Google Scholar]

[CR38] 38.Radkiewicz C, Dickman PW, Johansson ALV, Wagenius G, Edgren G, Lambe M. Sex and survival in non-small cell lung cancer: a nationwide cohort study. PLoS ONE. 2019;14(6):e0219206. 10.1371/journal.pone.0219206. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Becker N, Motsch E, Trotter A, Heussel CP, Dienemann H, Schnabel PA, et al. Lung cancer mortality reduction by LDCT screening-results from the randomized German LUSI trial. Int J Cancer. 2020;146(6):1503–13. 10.1002/ijc.32486. [DOI] [PubMed] [Google Scholar]

[CR40] 40.Pasquinelli MM, Tammemägi MC, Kovitz KL, Durham ML, Deliu Z, Guzman A, et al. Addressing sex disparities in lung cancer screening eligibility: USPSTF vs PLCOm2012 criteria. Chest. 2022;161(1):248–56. 10.1016/j.chest.2021.06.066. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Larke FJ, Kruger RL, Cagnon CH, Flynn MJ, McNitt-Gray MM, Wu X, et al. Estimated radiation dose associated with low-dose chest CT of average-size participants in the National Lung screening trial. AJR Am J Roentgenol. 2011;197(5):1165–9. 10.2214/ajr.11.6533. [DOI] [PubMed] [Google Scholar]

[CR42] 42.Rampinelli C, De Marco P, Origgi D, Maisonneuve P, Casiraghi M, Veronesi G, et al. Exposure to low dose computed tomography for lung cancer screening and risk of cancer: secondary analysis of trial data and risk-benefit analysis. BMJ. 2017;356:j347. 10.1136/bmj.j347. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Han SS, Ten Haaf K, Hazelton WD, Munshi VN, Jeon J, Erdogan SA, et al. The impact of overdiagnosis on the selection of efficient lung cancer screening strategies. Int J Cancer. 2017;140(11):2436–43. 10.1002/ijc.30602. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Ten Haaf K, van Rosmalen J, de Koning HJ. Lung cancer detectability by test, histology, stage, and gender: estimates from the NLST and the PLCO trials. Cancer Epidemiol Biomarkers Prev. 2015;24(1):154–61. 10.1158/1055-9965.Epi-14-0745. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Zang EA, Wynder EL. Differences in lung cancer risk between men and women: examination of the evidence. J Natl Cancer Inst. 1996;88(3–4):183–92. 10.1093/jnci/88.3-4.183. [DOI] [PubMed] [Google Scholar]

[CR46] 46.Berlin I. Incidence of lung cancer among young women. N Engl J Med. 2018;379(10):988. 10.1056/NEJMc1808250. [DOI] [PubMed] [Google Scholar]

[CR47] 47.Toumazis I, Alagoz O, Leung A, Plevritis SK. A risk-based framework for assessing real-time lung cancer screening eligibility that incorporates life expectancy and past screening findings. Cancer. 2021;127(23):4432–46. 10.1002/cncr.33835. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Adams SJ, Stone E, Baldwin DR, Vliegenthart R, Lee P, Fintelmann FJ. Lung cancer screening. Lancet. 2023;401(10374):390–408. 10.1016/s0140-6736(22)01694-4. [DOI] [PubMed] [Google Scholar]

[CR49] 49.Welch HG, Black WC. Overdiagnosis in cancer. J Natl Cancer Inst. 2010;102(9):605–13. 10.1093/jnci/djq099. [DOI] [PubMed] [Google Scholar]

[CR50] 50.Brodersen J, Schwartz LM, Woloshin S. Overdiagnosis: how cancer screening can turn indolent pathology into illness. APMIS. 2014;122(8):683–9. 10.1111/apm.12278. [DOI] [PubMed] [Google Scholar]

[CR51] 51.de Koning HJ, van der Aalst CM, de Jong PA, Scholten ET, Nackaerts K, Heuvelmans MA, et al. Reduced lung-cancer mortality with volume CT screening in a randomized trial. N Engl J Med. 2020;382(6):503–13. 10.1056/NEJMoa1911793. [DOI] [PubMed] [Google Scholar]

[CR52] 52.Duffy SW, Parmar D. Overdiagnosis in breast cancer screening: the importance of length of observation period and lead time. Breast Cancer Res. 2013;15(3):R41. 10.1186/bcr3427. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Ten Haaf K, de Koning HJ. Overdiagnosis in lung cancer screening: why modelling is essential. J Epidemiol Community Health. 2015;69(11):1035–9. 10.1136/jech-2014-204079. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Crosbie PA, Gabe R, Simmonds I, Kennedy M, Rogerson S, Ahmed N, et al. Yorkshire Lung Screening Trial (YLST): protocol for a randomised controlled trial to evaluate invitation to community-based low-dose CT screening for lung cancer versus usual care in a targeted population at risk. BMJ Open. 2020;10(9):e037075. 10.1136/bmjopen-2020-037075. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Kim DD, Cohen JT, Wong JB, Mohit B, Fendrick AM, Kent DM, et al. Targeted incentive programs for lung cancer screening can improve population health and economic efficiency. Health Aff (Millwood). 2019;38(1):60–7. 10.1377/hlthaff.2018.05148. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Barta JA, Powell CA, Wisnivesky JP. Global epidemiology of lung cancer. Ann Glob Health. 2019;85(1):8. 10.5334/aogh.2419. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Benefits, harms, and cost-effectiveness of risk model-based and risk factor-based low-dose computed tomography screening strategies for lung cancer: a systematic review

Yin Liu

Qingchao Geng

Xin Lin

Chenxi Feng

Youlin Qiao

Shaokai Zhang

Abstract

Background

Methods

Results

Conclusion

Background

Materials and methods

Search strategy

Inclusion and exclusion criteria

Data extraction

Results

Study characteristics

Fig. 1.

Table 1.

LDCT screening strategies

Table 2.

Benefits

Sensitivity and specificity

Table 3.

Deaths averted

Life years and quality adjusted life years

Table 4.

Harms

False positives and biopsies

Overdiagnosis

Radiation-related cancer

Cost-effectiveness

Discussion

Conclusion

Authors’ contributions

Funding

Data availability

Declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Footnotes

Contributor Information

References

Associated Data

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases