Risk stratification of lung malignancy in a real-world LDCT cohort: implications for screening beyond smoking criteria in Taiwan

Abstract

Low-dose computed tomography (LDCT) screening has been widely implemented based primarily on smoking-related eligibility criteria derived from Western populations. However, lung cancer epidemiology differs substantially in Asia, where a considerable proportion of cases occur among never-smokers, particularly women. We aimed to identify clinical factors associated with lung malignancy detection in a real-world Taiwanese LDCT cohort in routine clinical practice and to explore the applicability of current smoking-based screening paradigms. We conducted a retrospective cohort study including 7,319 adults who underwent LDCT between October 2018 and March 2023. Patients were categorized according to the LDCT referral pathway, and data on demographics, smoking history, comorbidities, and pathological outcomes were collected. Cause-specific Cox proportional hazards models were used for the outcome-specific analyses, with alternative mutually exclusive first pathological diagnoses censored at their diagnosis date and participants without a pathological diagnosis censored at their last available follow-up date. During follow-up, 156 participants had pathological confirmation of lung malignancy. In the cause-specific Cox model using binary smoking status (Model 2), lung malignancy was significantly associated with female sex (adjusted HR 1.95, 95% CI 1.33–2.86), age ≥ 45 years (adjusted HR 3.81, 95% CI 2.19–6.63), smoking history (adjusted HR 2.48, 95% CI 1.47–4.18), and chronic obstructive pulmonary disease (adjusted HR 11.82, 95% CI 6.46–21.63). Notably, a substantial proportion of lung cancer cases occurred in non-smokers, particularly women. In contrast, family history was not consistently associated with overall lung malignancy detection but was associated with specific pathological subtypes. In this real-world LDCT cohort, factors associated with lung malignancy detection included older age, female sex, smoking history, and COPD, with a notable burden among non-smoking women. These findings warrant further prospective studies to evaluate whether demographic and clinical factors could complement current smoking-based eligibility criteria in Asian populations.