Abstract Rationale There is a growing interest in the relationship between lung cancer (LC) rate (reflecting risk), lung cancer mortality (reflecting biology and stage at diagnosis) and lung cancer screening (LCS) outcomes (reflecting screening efficacy). Secondary analyses suggest specific comorbidities such as severe COPD (GOLD 3-4), chronic bronchitis and diabetes mellitus may be associated with attenuated benefits. We ask, “do other common cardiorespiratory comorbidities modify LCS screening outcomes?” Aim This study aimed to better understand the relationship between LC rate, LC mortality and LCS efficacy, where the latter is defined as LC deaths averted per 1000 screened (i.e., absolute reduction favouring those randomised to CT compared to CXR). Methods In a secondary analysis from the National Lung Screening Trial (N = 18,463), we sub-grouped subjects according to the 10 most common self-reported comorbidities (not mutually exclusive) and COPD-related phenotypes according to spirometry and clinical history (mutually exclusive). We compared LC rate, LC mortality and LCS efficacy across these different groups to identify which comorbidities influence LCS efficacy and how this might contribute to differences in LCS in COPD-related subgroups. Results We found LC rates were highest in those with the greatest airflow limitation (GOLD 3-4 and Diagnosed COPD) and lowest in those who were referent groups (no respiratory comorbidity). We found LC mortality was greatest in those where the LC rate (risk) was greatest and generally parallel to the LC rate, reflecting a roughly 50% LC mortality regardless of subgroup. LCS efficacy was highest for those with increased LC risk in those with mild-moderate airflow limitation (GOLD 1-2 and Undiagnosed COPD) but low (negative) for those with a higher prevalence of chronic bronchitis (CB) and diabetes mellitus (DM). Lower LCS efficacy was also seen in those with GOLD 3-4 (CB = 38%, DM = 12%), and GOLD 0 (CB = 65%, DM = 15%). Conclusion We conclude that LCS efficacy is greatest for those with moderately elevated LC risk, less severe COPD (CB or GOLD 3-4) and less non-LC mortality risk (DM). Based on these observations, we suggest that reduced LCS efficacy in this NLST sub-group is not related to comorbidity globally, but driven by very specific comorbidities (namely DM and CB). We propose that the mechanism underlying these observations may result from effects DM and CB on LC biology, competing mortality and possibly the ease with which CT-based screening identifies malignant nodules. The latter is the subject of further analysis. This abstract is funded by: None
Scott et al. (Fri,) studied this question.