肺癌生存率的性别差异：一项使用PLCO试验的竞争风险模型

Purpose:Lung cancer causes more than 130,000 deaths each year in the United States. Although females with non-small cell lung cancer (NSCLC) appear to have better survival than males, many prior studies did not fully account for key confounders, including smoking history and competing causes of mortality. Moreover, large cohort analyses often relied on Cox models, which may overestimate cause-specific risks when competing mortality is substantial. Hence, in this study, we assessed whether sex-based survival differences in NSCLC persist after adjusting for these confounders and applied competing risk regression to adequately account for competing causes of death and more accurately quantify lung cancer-specific mortality. Methods:Using data from the Prostate, Lung, Colorectal, and Ovarian (PLCO) trial, we selected patients whose first-ever malignancy was primary NSCLC, excluding anyone with a prior other cancer diagnosis or secondary/metastatic lung tumors. Lung cancer-specific survival for females versus males was evaluated using Fine and Gray competing risk models, adjusting for sociodemographic factors, cancer location, histology, treatment, smoking status, and comorbidities. Because patients with NSCLC experience substantial mortality from non-cancer-related causes, we applied a competing risk model to accurately account for this. This approach avoids the more than twofold (2.25-fold) overestimation of cause-specific risk that can occur when Cox models are used in elderly populations. All analyses were performed using SAS 9.4. Results:Among 2,793 patients (59% male, 41% female), 2,006 (72%) experienced lung cancer-related death, 531 (19%) died from other causes, and 256 (9%) were alive at last follow-up. Female patients were more likely to be never smokers (14% vs 5%, p<0.0001), have adenocarcinoma (55% vs 42%, p<0.0001), and less likely to have cardiovascular disease (heart failure: 7% vs 18%, p<0.0001). Unadjusted Fine and Gray models showed (hazard ratio [HR]: 0.81, 95% confidence interval [CI]: 0.74-0.88) lower lung cancer-specific mortality among women. Adjusted analyses demonstrated that female sex (HR: 0.85, 95% CI: 0.74-0.98) was associated with decreased risk of lung cancer-specific death. A significant interaction between female sex and stage IV disease (b coefficient: 0.18, 95% CI: 0.99-1.45) showed that lung cancer-specific survival of women and men was similar among patients with advanced disease. Conclusions:Females with NSCLC demonstrate reduced lung cancer-specific mortality after adjustment for competing risks and confounders. These results could be suggestive of sex-based disparities in inherent biological differences in tumor biology, natural disease progression, or overall increased life expectancy in females, further supporting investigating molecular pathways including hormonal signaling and immune responses to guide optimal treatment strategies.