LBA48 - Community-based mass screening with low-dose CT for lung cancer in Guangzhou

Background

This study aimed to investigate an approach for lung cancer (LC) screening in the general population and develop a risk prediction model to improve risk assessment.

Methods

This project is a community-based mass screening with LDCT for the early detection of LC. Participants were enrolled and screened from 2015 to 2021. Eligible subjects were between 40 and 74 years of age among the residents from four communities in Guangzhou, while exclusion criteria were persons diagnosed with LC within the past 5 years. A questionnaire including detailed demographic data and health conditions was used. Binary logistic regression analysis was used to screen potential risk factors. A multivariate model was built based on the participant characteristics combined with carcinoembryonic antigen (CEA) at a cutoff of 3.9. Model discrimination was evaluated by the area under the curve (AUC).

Results

11,708 participants were screened, comprising 5,452 males and 6,256 females with a median age of 59 (IQR, 51-65) years. 189 (1.6%) LCs were diagnosed, among which 162 (85.7%) cases were in stage 0-I. Only 37 (19.6%) and 105 (55.6%) of diagnosed cases met the criteria per NCCN and Chinese screening guidelines, respectively. We found seven independent risk/protective factors for LC through multivariate adjustment (Table). Using these variables combined with CEA, the model presented an AUC of 0.71 (95%CI, 0.67-0.75), which was significantly higher than that of guidelines in NCCN (0.52, 95%CI 0.50-0.55) and China (0.62 95%CI 0.58-0.67), respectively. Stratified analysis by smoking and stages showed that the AUCs were higher among smokers (0.77, 95%CI 0.71-0.83) and stage I to IV LC (0.74, 95%CI 0.70-0.78, excluding MIA) than in non-smokers (0.69, 95%CI 0.64-0.74) and preinvasive diseases (AIS and MIA, 0.64, 95%CI 0.57-0.72), respectively. Table: 000LBA48

Conclusions

Mass screening with LDCT can identify early-stage LC in Guangzhou, and a risk prediction model based on participant characteristics combined with CEA improves LC risk assessment.

Clinical trial identification

NCT04938804.

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Guangzhou Municipal Party Committee and Municipal Government and Guangzhou Municipal Health Commission.

Disclosure

All authors have declared no conflicts of interest.