Characteristics of lung cancer screening eligible population in the US and prediction of the eligibility with simplified criteria

Introduction

Lung cancer stands as the primary cause of cancer-related mortality in the United States (US) with about 155,870 deaths each year (1). Lung cancer screening (LCS) with low-dose computed tomography (LDCT) can reduce lung cancer mortality by 20% among the high-risk population (2,3). In 2013, the US Preventive Services Task Force (USPSTF) recommended annual LCS with LDCT to individuals aged 55–80 years, current or former smokers who quit smoking within 15 years, and with ≥30 pack-years smoking history (2). In 2021, the USPSTF updated these guidelines, now recommending annual LCS for those aged 50–80 years, current or former smokers who quit within 15 years, and with a history of ≥20 pack-years of smoking (4). Little is known about the characteristics of LCS eligible population and its frequency by social and clinical factors under these updated guidelines.

The successful implementation of LCS is a critical step in harnessing its potential survival benefits. Yet, the LCS utilization at the population level remains low, with less than 20% of the eligible individuals undergoing screening (5,6). One of the significant challenges in LCS implementation pertains to the identification of the target population (7-9). This challenge arises because data on lifetime smoking intensity (smoking pack-years) required to determine LCS eligibility are often unavailable or incomplete in population-based surveillance data or electronic health records (EHRs), and difficult to accurately assess. Previous studies have proposed the use of simplified criteria, based on age and current smoking status, as prescreening tools, which can be combined with subsequent formal assessment to identify the LCS eligible population (10-12). These simplified criteria have demonstrated good performance in estimating the LCS eligibility as per previous USPSTF guidelines (13). However, it is unknown whether these simplified criteria can effectively discriminate LCS eligible populations defined by the current guidelines. In this study we aimed to (I) estimate the proportion and examine characteristics of individuals eligible for LCS among people aged 50–80 years in the US based on current guidelines; (II) evaluate the performance of five simplified criteria in predicting LCS eligibility among population aged 50–80 years and subpopulations defined by sex, race/ethnicity, and education. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-1942/rc).

Methods

Data source and study population

We used the 2013–2018 National Health and Nutrition Examination Survey (NHANES) data (14). NHANES is a national survey designed to study the health and nutrition of adults and children in the US. A complex, multistage, probability sampling design was used to select participants representative of the civilian, non-institutionalized US population (15). In this study, we included participants aged 50–80 years with complete information on the smoking questionnaire. Since NHANES data is a fully publicly available and de-identified dataset, this study is exempted from the Institutional Review Boards review. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Variables

LCS eligibility was determined based on 2021 USPSTF criteria, i.e., individuals aged 50–80 years, currently smoking or quitting smoking within 15 years, and with ≥20 pack-years smoking history (4). Participants who smoked at least 100 cigarettes in life were defined as ever smokers; current smokers were those who smoke every day or some days currently; and former smokers were those who do not smoke currently. Lifetime smoking intensity was calculated as the product of smoking duration (years between starting and quitting smoking for former smokers; between starting smoking to the year of survey for current smokers) and number of packs of cigarettes smoked every day.

The five simplified criteria based on age and smoking status were: (I) ever smoker, defined as an individual with any positive smoking history; (II) current or former smoker, an individual with any positive smoking history or who quit smoking within 15 years; (III) current smoker, an individual currently smoking; (IV) current smoker, an individual currently smoking >0.5 packs per day (ppd); (V) Current smoker, a person currently smoking >1 ppd. Other sociodemographic and clinical variables used in the analyses included: age, sex, race/ethnicity, education, marital status, insurance, food security, housing characteristics, family poverty level index, general health condition, availability of primary care provider, smoking status, history of respiratory illnesses, history of cancer, and history of other comorbidities (coronary heart disease, liver conditions, arthritis, or gout).

Statistical analysis

Chi-squared tests were employed to compare categorical variables. With the USPSTF LCS guideline as the gold standard criteria, we calculated the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of each of the five simplified criteria. Participants eligible for LCS and meeting the simplified criterion were considered as true positive (Table 1); those eligible for LCS but not meeting the criterion were false negative; those not eligible for LCS but meeting the criterion were false positive; and those not eligible for LCS and not meeting the criterion were true negative. Sensitivity was calculated as true positive/(true positive + false negative); specificity as true negative/(false negative + true negative); PPV as true positive/(true positive + false positive); NPV as true negative/(false negative + true negative); and accuracy as (true positive + true negative)/(true positive + false positive + false negative + true negative). Levels of significance was set at P<0.05. Statistical analysis was conducted using SAS software version 9.4, accounting for survey weights and clustering by sampling units to derive nationally representative estimates.

Results

A total of 8,309 participants from 2013–2018 NHANES were identified, representing 104.36 million US residents aged 50–80 years, among whom 16.01% (representing 16.70 million US residents) were eligible for LCS (Table 2). The proportion of participants eligible for LCS was higher in those aged 50–64 years (18.35%), male (20.42%), non-Hispanic White (17.63%), with high school or lower education (21.52%), single or never married (21.14%), without insurance (24.70%), with low family poverty level index (23.76%), with very low food security (31.83%), renting houses or with other housing arrangements (21.34%), with fair or poor general health condition (24.47%), without primary care provider (22.57%), with a history of respiratory illnesses (29.35%), and other comorbidities (18.26%) (P<0.001 for each). In addition, 56.49% of current smokers and 23.40% of former smokers were eligible for LCS.

Table 3 describes the characteristics of the population eligible for LCS (weighted N=16.70 million). Among all LCS eligible individuals, the majority were aged 50–64 years (66.99%), male (59.03%), non-Hispanic White (79.14%), married or living with partner (57.17%), with full household food security (66.12%), owning a house (72.13%), with good to excellent general health (65.99%), having primary care provider (81.80%), without respiratory illnesses (65.01%) or cancer (81.12%), and with other comorbidities (62.04%). About half were current smokers (55.76%), having high school or lower education (52.32%), private insurance (51.56%), or with family poverty level index >1.85 (55.80%).

Simplified criteria based on current smoking status and intensity performed differently in identifying individuals eligible for LCS among people aged 50–80 years (Table 4). Ever smoker (current or former smoker) as a simplified criterion had a sensitivity of 1.00, specificity of 0.64, PPV of 0.35, NPV of 1.00, and accuracy of 0.70. Compared to ever smoker, adding years since quitting smoking, i.e., current smoker or former smoker quitting smoking <15 years, can differentiate LCS eligible population better with sensitivity as 1.00, specificity as 0.87, PPV as 0.59, and accuracy as 0.89. The sensitivity decreased (0.56) but specificity increased (0.92) when using only current smoker as a simplified criterion. For the criteria including both current smoking status and smoking intensity, current smoker with >0.5 ppd can identify LCS eligibility more accurately than current smoker with >1 ppd. Both criteria had high specificity (>0.99 and 1.00, respectively), PPV (0.97 and 1.00, respectively), NPV (0.91 and 0.85, respectively), and accuracy (0.91 and 0.85, respectively), but current smoker >1 ppd had very low sensitivity (0.08).

There were no substantial differences in the performance of the simplified criteria in identifying LCS eligible individuals among subpopulations stratified by sex, race/ethnicity, and education (Tables S1-S3). Current smoker with >1 ppd had lower sensitivity in women (0.05) than in men (0.10), in non-Hispanic Black people (0.03) than in non-Hispanic White people (0.09), and in individuals with college or above education (0.07) than in those with high school or lower education (0.09).

Discussion

The USPSTF recently updated LCS guidelines which expanded the LCS eligible population to include those at younger age and with lower lifetime smoking intensity. With a nationally representative sample, about 16.70 million individuals, accounting for 16.01% of people aged 50–80 years, were eligible for LCS in the US. The proportion of LCS eligibility is higher in those with lower socioeconomic status and worse clinical status. To identify LCS eligibility, simplified criterion as current smoker with >1 ppd had low sensitivity. Other simplified criteria based on age, current smoking status, and current smoking intensity, had sensitivity ranging from 0.45 to 1.00, which can be used as prescreening tools to identify the high-risk populations for targeted interventions.

As the 2013 USPSTF guidelines identified about 8 to 9 million individuals eligible for LCS in the US (9.9–12.7% of the population aged 55–80 years) (5,16), the 2021 USPSTF guidelines almost doubled the size (16.70 million, representing 16.01% of the population aged 50–80 years). The updates in the guidelines increased the proportion of people younger than 65 years (from 56.6% to 66.7%), males (from 49.6% to 59.0%), and current smokers (from 45.4% to 55.8%). The frequency of LCS eligibility remained higher in people with worse general health status or with respiratory or other health conditions. In addition, two changes in the characteristics of LCS eligible population were worthy of note. The first was the decrease in the proportion of non-Hispanic White people (from 86.2% to 79.14%) among all LCS eligible individuals. Non-Hispanic Black people tend to initiate smoking at an older age and have lower lifetime smoking intensity, thus less likely to be eligible for LCS but still at high risk of lung cancer than non-Hispanic White smokers (17-23). The 2013 guidelines had lower accuracy in identifying the non-Hispanic Black high-risk population (18). Our findings and studies using other national survey data confirmed that the update of LCS guidelines decreased the racial differences in LCS eligibility (24). The second was the increase in the proportion of uninsured individuals (from 7.9% to 13.3%) and those without a primary care provider (from 11.7% to 18.2%) among all LCS eligible individuals. For the first time, we found that a substantial 31.8% of individuals experiencing very low food security were eligible for LCS, in contrast to 13.5% of those with full food security. Different from other cancer screenings recommended to an entire age group, LCS is recommended based on an individual’s risk. People of lower socioeconomic status are less likely to undergo LCS but more likely to be eligible for LCS (5,25). The updated guidelines resulted in a disproportionate increase in the number of individuals eligible for LCS with lower socioeconomic status. This shift has the potential to exacerbate disparities in LCS utilization. Hence, there is an immediate and pressing need to enhance LCS implementation within these underserved subpopulations.

A significant barrier to the effective implementation of LCS is the challenge of identifying patients eligible for LCS, ultimately impeding physician referrals. Previous research demonstrated that up to 54.7% of barriers in LCS utilization were accounted for by the failure of EHR notification regarding patients’ LCS eligibility (26). LCS full eligibility was determined based on age, smoking status, and lifetime smoking intensity. Lifetime smoking intensity, measured by smoking pack-years, usually is unavailable or incomplete in EHRs. There is also significant discordance between EHR recorded and patient-self-reported smoking intensity (27). Thus, researchers proposed to use simplified criteria based on easily accessible information as a prescreening tool to identify the high-risk population for further evaluation and interventions of LCS. A study in 2019 evaluated five simplified criteria based on age, current smoking status, and current smoking intensity, and found that all the five criteria demonstrated good performance in discriminating LCS full eligibility (10). With the significant changes in USPSTF LCS guidelines, we revisited this topic and evaluated whether the five simplified guidelines can still work well as LCS prescreening tools. We found that the criterion of current smoker with >1 ppd had much lower sensitivity than it did in the previous report. Only 8% of individuals meeting current LCS guidelines are current smokers smoking >1 pack of cigarettes per day. The reason for the difference is the substantial decrease in the required lifetime smoking intensity in the current LCS guidelines (from 30 to 20 pack-years), thus smokers with very high smoking intensity represent only a small proportion of all LCS eligible population. However, the PPV of this criterion was 1.00 indicating that 100% of all current smokers with >1 ppd meet LCS eligibility. This indicator suggests that when resources are limited and no additional efforts are available for further evaluation, current smokers with >1 ppd can still be used to identify the very high-risk patient population for LCS implementation.

The performance of the other four simplified criteria in determining LCS eligibility remained relatively similar based on current and previous guidelines. A noteworthy difference was the higher PPV for all criteria according to current LCS guidelines. For example, about 35% of ever smokers, 59% of current smokers or former smokers who quit smoking within 15 years, 57% of current smokers, and 97% of current smokers with >0.5 ppd met current LCS eligibility (compared to 25%, 48%, 44%, and 66%, respectively, based on previous guidelines). This evidence indicates that the simplified criteria can identify the target population more efficiently according to current guidelines. The PPV can even increase when the screening tool is used in a population with a higher prevalence of LCS eligibility, such as people with lower socioeconomic status or more respiratory comorbidities. Our stratified analyses did not show significant differences in the performance of simplified criteria among subpopulations. The lower PPVs in women than in men, in non-Hispanic Black than in non-Hispanic White people, and in people with college or above education than in people with high school or lower education were likely due to the lower prevalence of LCS eligibility in these subpopulations.

The USPSTF guidelines on screening individuals at high risk for lung cancer creates a demand to implement efficient and effective LCS programs. LCS implementation requires an intensive interaction between smokers and the healthcare system. Simplified criteria based on easily accessible information can be used as a prescreening tool to initiate the communication on LCS. Depending on the patient volume and resource availability, LCS programs can opt for specific simplified criterion to establish routine alerts within EHRs. These alerts can aid healthcare providers in recognizing LCS eligible populations. Our findings validate the accuracy and effectiveness of each simplified criterion. With the fast development of artificial intelligence (AI), the prediction of LCS eligibility can be enhanced by AI tools (28,29). The simplified criteria we proposed can serve as foundational information for further AI tools or can be used synergistically with other AI tools. which could empower healthcare professionals with comprehensive and timely information crucial for informed decision-making in LCS protocols.

The major strength of this study is the large, contemporary, and nationally representative sample, which allows generalizability of our findings to the US population. Another strength is the inclusion of specific measures of socioeconomic and clinical status, such as the availability of primary care provider, household food security, housing, general health status, and respiratory and other comorbidities. These data can provide detailed descriptions of LCS eligibility among US subpopulations. Despite its strengths, the study is also subjected to limitations including being entirely comprised of survey data and lack of information on state variation. In addition, self-reported smoking history is a potential limitation of the dataset. However, previous research validated NHANES smoking data and showed substantial consistency between self-reported smoking and biomarkers (30).

Conclusions

Following the 2021 updates to the USPSTF LCS guidelines, 16.70 million smokers are currently eligible for LCS in the US. As populations with lower socioeconomic status experienced greater increases in the proportion of LCS eligibility, the changes in the guidelines could exacerbate disparities in LCS utilization. Given the current low utilization of LCS, there is a pressing need for interventions to enhance LCS implementation. Simplified criteria based on easily accessible information such as age, current smoking status, and current smoking intensity, can be used as prescreening tools to identify target populations.

Acknowledgments

Funding: This study was funded by American Cancer Society Discovery Boost Grant (DBG-23-1036396-01-HOPS), Prisma Health Seed Grant, and Dabo’s All in Team Foundation. Funders did not play any role in study design, data analysis, and results interpretation.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-1942/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-1942/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-1942/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017. CA Cancer J Clin 2017;67:7-30. [Crossref] [PubMed]
2. Ostrowski M, Marjański T, Rzyman W. Low-dose computed tomography screening reduces lung cancer mortality. Adv Med Sci 2018;63:230-6. [Crossref] [PubMed]
3. National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395-409. [Crossref] [PubMed]
4. US Preventive Services Task Force. Screening for Lung Cancer: US Preventive Services Task Force Recommendation Statement. JAMA 2021;325:962-70. [Crossref] [PubMed]
5. Narayan AK, Gupta Y, Little BP, et al. Lung cancer screening eligibility and use with low-dose computed tomography: Results from the 2018 Behavioral Risk Factor Surveillance System cross-sectional survey. Cancer 2021;127:748-56. [Crossref] [PubMed]
6. Kee D, Wisnivesky J, Kale MS. Lung Cancer Screening Uptake: Analysis of BRFSS 2018. J Gen Intern Med 2021;36:2897-9. [Crossref] [PubMed]
7. Mulshine JL, D'Amico TA. Issues with implementing a high-quality lung cancer screening program. CA Cancer J Clin 2014;64:352-63. [Crossref] [PubMed]
8. Eberth JM, Sercy E. Implementation of Lung Cancer Screening in the United States: Changing Trends Based on a Survey of Society of Thoracic Radiology Members. J Thorac Imaging 2015;30:W60-2. [Crossref] [PubMed]
9. Zeliadt SB, Hoffman RM, Birkby G, et al. Challenges Implementing Lung Cancer Screening in Federally Qualified Health Centers. Am J Prev Med 2018;54:568-75. [Crossref] [PubMed]
10. Triplette M, Donovan LM, Crothers K, et al. Prediction of Lung Cancer Screening Eligibility Using Simplified Criteria. Ann Am Thorac Soc 2019;16:1280-5. [Crossref] [PubMed]
11. Rendle KA, Steltz JP, Cohen S, et al. Estimating Pack-Year Eligibility for Lung Cancer Screening Using 2 Yes or No Questions. JAMA Netw Open 2023;6:e2327363. [Crossref] [PubMed]
12. Burnett-Hartman AN, Powers JD, Hixon BP, et al. Development of an Electronic Health Record-Based Algorithm for Predicting Lung Cancer Screening Eligibility in the Population-Based Research to Optimize the Screening Process Lung Research Consortium. JCO Clin Cancer Inform 2023;7:e2300063. [Crossref] [PubMed]
13. Moyer VA. U.S. Screening for lung cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2014;160:330-8. [Crossref] [PubMed]
14. Centers For Disease Control And Prevention. NHANES Questionnaires, Datasets, and Related Documentation: National Center for Health Statistics; 2017. Available online: https://wwwn.cdc.gov/nchs/nhanes/Default.aspx
15. Centers For Disease Control And Prevention. About the National Health and Nutrition Examination Survey: National Center for Health Statistics; 2017. Available online: https://www.cdc.gov/nchs/nhanes/about_nhanes.htm
16. Richards TB, Soman A, Thomas CC, et al. Screening for Lung Cancer - 10 States, 2017. MMWR Morb Mortal Wkly Rep 2020;69:201-6. [Crossref] [PubMed]
17. Ho MK, Faseru B, Choi WS, et al. Utility and relationships of biomarkers of smoking in African-American light smokers. Cancer Epidemiol Biomarkers Prev 2009;18:3426-34. [Crossref] [PubMed]
18. Aldrich MC, Mercaldo SF, Sandler KL, et al. Evaluation of USPSTF Lung Cancer Screening Guidelines Among African American Adult Smokers. JAMA Oncol 2019;5:1318-24. [Crossref] [PubMed]
19. Ho MK, Mwenifumbo JC, Al Koudsi N, et al. Association of nicotine metabolite ratio and CYP2A6 genotype with smoking cessation treatment in African-American light smokers. Clin Pharmacol Ther 2009;85:635-43. [Crossref] [PubMed]
20. Ross KC, Dempsey DA, St Helen G, et al. The Influence of Puff Characteristics, Nicotine Dependence, and Rate of Nicotine Metabolism on Daily Nicotine Exposure in African American Smokers. Cancer Epidemiol Biomarkers Prev 2016;25:936-43. [Crossref] [PubMed]
21. Trinidad DR, Pérez-Stable EJ, White MM, et al. A nationwide analysis of US racial/ethnic disparities in smoking behaviors, smoking cessation, and cessation-related factors. Am J Public Health 2011;101:699-706. [Crossref] [PubMed]
22. Hecht SS. Cigarette smoking and lung cancer: chemical mechanisms and approaches to prevention. Lancet Oncol 2002;3:461-9. [Crossref] [PubMed]
23. Fiscella K, Winters P, Farah S, et al. Do Lung Cancer Eligibility Criteria Align with Risk among Blacks and Hispanics? PLoS One 2015;10:e0143789. [Crossref] [PubMed]
24. Reese TJ, Schlechter CR, Potter LN, et al. Evaluation of Revised US Preventive Services Task Force Lung Cancer Screening Guideline Among Women and Racial/Ethnic Minority Populations. JAMA Netw Open 2021;4:e2033769. [Crossref] [PubMed]
25. Zahnd WE, Murphy C, Knoll M, et al. The Intersection of Rural Residence and Minority Race/Ethnicity in Cancer Disparities in the United States. Int J Environ Res Public Health 2021;18:1384. [Crossref] [PubMed]
26. Coughlin JM, Zang Y, Terranella S, et al. Understanding barriers to lung cancer screening in primary care. J Thorac Dis 2020;12:2536-44. [Crossref] [PubMed]
27. Modin HE, Fathi JT, Gilbert CR, et al. Pack-Year Cigarette Smoking History for Determination of Lung Cancer Screening Eligibility. Comparison of the Electronic Medical Record versus a Shared Decision-making Conversation. Ann Am Thorac Soc 2017;14:1320-5. [Crossref] [PubMed]
28. Liu S, McCoy AB, Aldrich MC, et al. Leveraging natural language processing to identify eligible lung cancer screening patients with the electronic health record. Int J Med Inform 2023;177:105136. [Crossref] [PubMed]
29. Ruckdeschel JC, Riley M, Parsatharathy S, et al. Unstructured Data Are Superior to Structured Data for Eliciting Quantitative Smoking History From the Electronic Health Record. JCO Clin Cancer Inform 2023;7:e2200155. [Crossref] [PubMed]
30. Li W, Liu B. Comparing cotinine and NNAL verification of self-reported smoking status among lung cancer screening eligible population from the 2007-2014 National Health and Nutrition Examination Survey (NHANES). Biomarkers 2021;26:45-54. [Crossref] [PubMed]