Lung cancer is the most common cause of cancer-related death in Europe1 with an age-standardised 5-year survival of 13% for adult patients with cancer diagnosed in 2000–2007.2 In Spain, according to the Spanish Network of Cancer Registries (REDECAN), lung cancer will be in 2020 the most common cancer with an expected incidence rate (adjusted on the actual European population) of 64.1 per 100,000 individuals.3 For the period 2008–2013, the age-standardised 5-year relative survival was 13% and 18% for Spanish men and women respectively.4

Tobacco smoking is the main risk factor for developing lung cancer, with up to 90% of lung cancer attributed to smoking.5 Although smoking cessation has been identified as the most cost-effective strategy to prevent lung cancer,6 it may not be achieved in all current smokers; also, former smokers may still be at high risk of developing lung cancer. Thus, other strategies, alone or in combination, can be considered to better reduce lung cancer mortality.

Lung cancer screening with low-dose computed tomography (LDCT) has been proposed as an early-detection strategy to reduce lung cancer mortality.7 A meta-analysis, that combined the results of most randomized trials assessing the effects of lung cancer screening with LDCT, concluded that it might significantly reduce lung cancer mortality by 17%.8 These results support the European Union position statement published in 2017 that recommended the implementation of lung cancer screening with LDCT in Europe arguing that it can save lives.9 However, because lung cancer screening with LDCT is not exempt from side effects, this statement also recommended the adoption of specific actions, such as use of a risk stratification approach, to ensure a successful implementation lung cancer screening with LDCT. These side effects, that include physical and psychological harms due to overdiagnosis, surgery for benign lesions or radiation exposure,10 stress the need to carefully select the target population for lung cancer screening programmes and the importance of incorporating tobacco cessation practices in all settings.11

Recent studies have shown that a selection of candidates for lung cancer screening based on high-quality risk prediction models is superior to a selection based on criteria such as age and pack-years alone as it leads to fewer individuals being screened, more cancers being detected, and fewer false positives.12,13

In a previous paper, we compared different strategies to identify the proportion of the Spanish population at high risk of developing lung cancer, susceptible to be included in a lung cancer screening programme.14 Because in Spain health competences are transferred to the seventeen Autonomous Communities (ACs), and each autonomous region has competence to organize the prevision of preventive services, the present paper describes the proportions and absolute numbers of individuals at high risk of developing lung cancer in the different Spanish ACs.

Among the 6673 ENSE subjects, aged 50–74 years involved in this analysis, 3184 (2099 men and 1085 women) were ever-smokers (Table 1). Their median 6-year risk of developing lung cancer was 0.90 (IQR 0.35–1.96) for men and 0.53 (IQR 0.16–1.10) for women.

Overall, the proportion of 50–74 years old ENSE participants with a risk of lung cancer ≥2% significantly varied across sex. Among men, 511 ever-smokers showed a risk of developing lung cancer ≥2%; they represent almost a quarter (24.3%) of male ever-smokers and 15.9% of the male ENSE sample. In women, 122 ever-smokers had a risk of developing lung cancer ≥2%, which represents 11.2% of ever-smokers and 3.5% of the whole female ENSE sample.

In the ENSE sample, the ACs with the highest proportion of individuals at risk of lung cancer ≥2% were Ceuta (13.1%), Extremadura (13.0%), Andalucía (12.7%) and Melilla (11.5%), while the ACs with the lowest proportions of ever-smokers at risk ≥2% were Región de Murcia (6.6%), Galicia (7.0%), Aragón (7.3%), Comunidad de Madrid (7.5%) and País Vasco (7.5%). The proportions of individuals with a risk of lung cancer ≥2% observed in Andalucía and Extremadura were significantly above the proportion observed at national level (9.5%; p-values of 0.006 and 0.050, respectively).

We extrapolated these proportions into absolute figures for the Spanish population aged 50–74 years old and obtained that 1,341,483 individuals (1,087,327 men and 254,156 women) may have a risk of developing lung cancer ≥2% and might therefore benefit from lung cancer screening. The total number of individuals having a high-risk of lung cancer ranged from 8849 in La Rioja to 304,038 in Andalucía. While 17% of the Spanish population live in Andalucía, the number of individuals having a high-risk of lung cancer in this AC represented 23% of the total number of individuals having a high-risk of lung cancer.

The present study showed that 9.5% of the adult Spanish population had a 6-year risk of developing lung cancer, calculated using the model developed in the context of the PLCO trial, above the threshold of 2%. The proportion of subjects at high risk was significantly higher in men (15.9%) than in women (3.5%) and represented a total of more than 1,3 million people. However, the proportions observed in the different ACs were not very different from those observed at national level.

To our knowledge, only two papers have previously reflected on the use of LDCT for lung cancer screening in Spain.20,21 Ruano-Ravina and colleagues provided an estimation of the number of men and women who fulfilled lung cancer screening criteria in the different Spanish ACs.20 They estimated that a total of 1,714,683 individuals fulfilled lung cancer screening criteria. Although based on the same data (ENSE 2011–2012) this figure is above ours. This difference is explained by the age range selected in both studies (55–80 years old in their analysis and 50–74 years old in ours) and by the criteria used to define individuals at high-risk of developing lung cancer. While they used the NLST criteria (based on age and smoking history), we used a predictive model and selected individuals with a risk above the threshold of 2%; a selection strategy that previously showed to be more restrictive than the NLST criteria.14 The second paper, set up recommendations about the implementation of LDCT for lung cancer screening, but did not provide any estimation of the target population.21 The authors mentioned that, despite the accumulated scientific evidence on the benefits of lung cancer screening, some doubts still persist regarding the feasibility of its large-scale implementation mainly due to the uncertainty regarding the selection of candidates, the number and frequency of the scans and the management of findings. For this reason they recommend the design of pilot studies to analyze the benefits and risks of lung cancer screening in the Spanish environment.

In Spain, the Pamplona International Early Lung Cancer Detection Program (P-IELCAP) is the longest ongoing lung cancer screening programme.22 It has been ongoing in a private centre since 2000 as part of a single arm trial called I-ELCAP that started to screen with LDCT asymptomatic persons at risk for lung cancer in 1993.23 In 2015, Sanchez-Salcedo and colleagues published the first results of the P-IELCAP based on 2,989 men and women, aged 40 years and older, current or former smokers with a smoking history of ≥10 pack-years who had been screened between 2000 and 2014.22 While the authors concluded that the Spanish experience of lung cancer screening showed findings comparable with those observed in the rest of Europe and confirmed the feasibility and efficacy of lung cancer screening using LDCT, they also mentioned that the P-IELCAP performed in a private centre may not be entirely applicable to public institutions. This last point was also discussed in the reflection paper of Ruano-Ravina and colleagues, who wondered whether the Spanish public health system would be able to (i) identify the target population for lung cancer screening taking into account the information available in medical records (smoking history is not always registered); (ii) face the costs of a LDCT screening programme (tomographs and health professionals involved in screening); (iii) guarantee adequate participation of the target population and (iv) allocate extra medical resources to handle overdiagnosis and surgery for benign lesions.20 More recently, Ruano-Ravina et al. have indicated that the available evidence does not support the idea that the benefits of lung cancer screening with LDCT are greater than the harms, and recommended the funding of smoking cessation programmes in order to effectively reduce lung cancer burden.24 While funding smoking cessation programmes is necessary to reduce lung cancer incidence and improve cancer treatments,25 it is important to keep in mind that all current smokers may not be able to quit smoking and that some former smokers will still be at risk of developing lung cancer and may therefore benefit from lung cancer screening. Previous studies have evaluated the effect of introducing smoking cessation interventions in the context of lung cancer screening and showed that participation in a lung screening trial may promote cessation.26

Although some experts still actually think that the risk–benefit balance of lung cancer screening with LDCT is questionable, and do not recommend its implementation,24 expert panels have invited European countries to actively start a widespread implementation of lung cancer screening.9 Since screening of lung cancer with LDCT may produce physical and psychological harms, mainly due to overdiagnosis, surgery for benign lesions or radiation exposure,10 this prevention strategy needs to be offered to subjects at high risk of developing lung cancer in order to achieve the highest benefit-to-harm ratio. To achieve this goal, it is essential to use a risk assessment tool such as the PLCOm2012 that incorporates sociodemographic and health-related factors. This model previously showed to be superior to the NLST eligibility criteria (based on age and smoking history) and demonstrated excellent predictive performance in a large Australian cohort of 95,822 smokers, as well as higher positive predictive value and sensitivity, with minimal loss of specificity.19

To actively start the implementation of lung cancer screening, as recommended by European stakeholders, there is a crucial need to get precise information on the size of the target population, based on prediction models that include the most relevant risk factors for lung cancer, for which reliable data can be systematically obtained. Our analysis showed that the total number of high risk people does not only depend on the population size but on the prevalence of smoking. Indeed, while an AC such as Andalucia included 17% of the Spanish population, the number of individuals at high-risk of lung cancer in this AC represented 23% of the total number of individual at high-risk of lung cancer. Using the results described in the present paper, the health authorities of the different ACs willing to set-up a lung cancer screening programme in their region will be able to estimate the cost of such programme and/or design feasibility and pilot studies. Taking into account the actual economic constraints on the Spanish healthcare system, before implementing new prevention strategies, it is important to precisely evaluate the costs and the effectiveness of these strategies.

The present paper has limitations that need to be acknowledged. While the most recent ENSE data were published in 2017, we used data from 2011 to calculate to the proportion of individuals at high risk of lung cancer, as the latest survey did not include information on smoking intensity in ex-smokers and in current non-cigarette smokers (pipe or cigar). When we compared both sets of data, we observed similar proportions of current cigarette smokers (20% and 19% in 2011 and 2017, respectively) as well as comparable proportions of current cigarette smokers having a 6-year risk of developing lung cancer (7% and 8% in 2011 and 2017, respectively). For this reason, we think that if the analysis had been performed on the most recent ENSE data available, the results obtained would not significantly differ from those showed in the present paper. While the PLCOm2012 risk model includes family history of lung cancer and race/ethnicity as additional factors in the identification of ever-smokers at highest risk of developing lung cancer, we were not able to include these variables as this information was not gathered by the ENSE survey. Nevertheless, ethnicity is not such a relevant variable in Spain, as it can be in other countries, because the proportion of Caucasian is 93–95% and the remaining part of the population includes different ethnicities, Hispanic being the main one.27 For this reason, we think that assuming that our population was Caucasian may have produced a slight over estimation of the risk of lung cancer while assuming that no one had a family history of lung cancer may have resulted in a slight under estimation of this risk, taking into account the coefficients of the variables Hispanic ethnicity and history of lung cancer in Tammemägi's model. Introducing environmental risk factors such as radon or asbestos exposure in prediction models would also be interesting5; however, these factors are rarely systematically reported in medical files.

The main strengths of the present study relied on the use of a prediction model to assess the risk of developing lung cancer in 6-year time and the presentation of national and regional figures in a country in which autonomous regions have their own healthcare services responsible for the health centres, services and facilities of the regions (the Central Government retains healthcare management in the cities of Ceuta and Melilla).

The present study is the first one that evaluated the number of individuals at high risk of developing lung cancer in the different Spanish ACs using a predictive model to calculate this risk and selecting people with a 6-year risk above the threshold of 2%. This study showed that 9.5% of the overall population (more than 1.3 million of ever-smokers) had a high risk of developing lung cancer, with no major differences observed across ACs. Further studies should assess the cost and effectiveness associated to the implementation of a lung cancer screening programme to such population.

MG and CV conceived the study. NT performed the analysis and prepared the first draft. All authors contributed to the successive reviews of the manuscript and approved its final version.

This study has been funded by the Asociación Española Contra el Cáncer de Barcelona in the Call for proposals of Research in Oncology of 2016 (grant PROYBAR16909FU). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The Tobacco Control Research Group is partly supported by the Ministry of Business and Knowledge from the Government of Catalonia [2017SGR319] and by Instituto de Salud Carlos III, Government of Spain (CIBERES CB19/06/00004). EF was supported by the Instituto de Salud Carlos III, Government of Spain, co-funded by the European Regional Development Fund (FEDER) (INT16/00211 and INT17/00103). We thank CERCA Programme Generalitat de Catalunya for the institutional support.

The authors declare no competing interests.