A real-world study of 7,319 adults undergoing low-dose computed tomography (LDCT) screening in Taiwan found that pathological lung malignancy detection was associated with female sex, older age, smoking history and chronic obstructive pulmonary disease. The retrospective cohort also identified a notable burden of lung cancer among people who did not smoke, particularly women, suggesting that smoking alone may not fully describe who is at risk of malignancy in this screening population.
Researchers analyzed adults who received LDCT between October 2018 and March 2023 through routine clinical practice in Taiwan. They collected information on demographics, smoking history, comorbidities and pathological outcomes, then used cause-specific Cox proportional hazards models to examine factors associated with lung malignancy detection. In these models, alternative mutually exclusive first pathological diagnoses were censored at their diagnosis dates, while participants without a pathological diagnosis were censored at their last available follow-up. During follow-up, 156 participants received pathological confirmation of lung malignancy.
The strongest association in the model using binary smoking status was chronic obstructive pulmonary disease (COPD), which was associated with an adjusted hazard ratio of 11.82 for lung malignancy detection, with a 95% confidence interval of 6.46 to 21.63. Age of 45 years or older was also associated with detection, with an adjusted hazard ratio of 3.81 and a 95% confidence interval of 2.19 to 6.63. Smoking history carried an adjusted hazard ratio of 2.48, with a 95% confidence interval of 1.47 to 4.18. Female sex was associated with a detection hazard ratio of 1.95, with a 95% confidence interval of 1.33 to 2.86. Because this was an observational cohort, these hazard ratios describe associations with the timing of detection and do not establish that the factors caused lung cancer.
Although smoking history remained significantly associated with lung malignancy, a substantial proportion of confirmed cases occurred among non-smokers, with the burden especially notable among women. Family history was not consistently associated with overall lung malignancy detection, although it was associated with specific pathological subtypes. The findings therefore point toward a more complex risk profile than smoking status alone: sex, age and COPD may help identify malignancy risk within an LDCT-screened population, while the results among non-smoking women indicate that the presence or absence of smoking should not be treated as a complete marker of susceptibility.
The study does not show that people outside existing smoking-based eligibility criteria would benefit from screening, nor does it establish the optimal age, sex-specific or clinical criteria for future programs. Its retrospective design, reliance on routine clinical practice and restriction to an LDCT-referred population limit how directly the results can be generalized to people who were never screened. The authors consequently call for further prospective studies to evaluate whether demographic and clinical factors can complement current smoking-based screening paradigms in Asian populations. For now, the cohort provides a real-world estimate of factors associated with lung malignancy detection in Taiwan and highlights the need to examine screening eligibility beyond smoking history alone.
Subject of Research: Clinical factors associated with lung malignancy detection in a Taiwanese LDCT screening cohort
Article Title: Risk stratification of lung malignancy in a real-world LDCT cohort: implications for screening beyond smoking criteria in Taiwan
Article References: Lu, C.-Y., Lin, S.-H., & Chang, Y.-J. (2026). Risk stratification of lung malignancy in a real-world LDCT cohort: implications for screening beyond smoking criteria in Taiwan. BMC Cancer. https://doi.org/10.1186/s12885-026-16830-5
Image Credits: AI Generated
DOI: 10.1186/s12885-026-16830-5
Keywords: lung cancer, LDCT screening, Taiwan, smoking history, never-smokers, women’s health, COPD, risk stratification, Risk, stratification, lung, malignancy
