P1.03-064 Chest X-Ray (CXR) Screening Improves Lung Cancer (LC) Survival in the Prostate Lung Colon and Ovary (PLCO) Randomized Population Trial (RPT)

Abstract

The effectiveness of CXR-screening for LC was estimated in the context of performing a cost-effectiveness analysis of LC-screening comparing CT, CXR, and no screening. CXR-screening has long been considered ineffective because no RPT has demonstrated a LC mortality reduction. However, CXR-screening has been shown to produce a significant survival advantage not attributable to overdiagnosis or other screening biases (JCO: 20, 1973, 2002). The lung portion of the PLCO trial, which compared CXR to no screening, reported no LC mortality reduction after 13-years follow-up (JAMA, 306, 1875, 2011). However, that analysis included all LCs diagnosed over 13-years, despite the fact that the active screening period lasted only 3-years. LC survival was not reported. Since screening is exceedingly unlikely to provide any advantage to individuals diagnosed many years after active screening is discontinued, and because sojourn time associated with CXR-screening is estimated to be up to 4 years, we evaluated outcomes of LCs diagnosed within 7-years of randomization in PLCO. PLCO randomized 77,445 subjects to an experimental group (EG) undergoing a prevalence CXR and 3 annual incidence CXRs and 77,456 others to an unscreened control group (CG). Using Kaplan-Meier methods in the intent-to-screen analysis of PLCO data, LC survival and mortality were calculated for all LCs diagnosed during the 13-year follow-up, as well as those diagnosed within 7 years of randomization. LC incidence and mortality were compared with Fisher’s exact test. Survival was compared with the log-rank test. All p-values are two-sided. After 13-years, 1,838 and 1,737 lung cancers were detected in EG and CG, respectively (RR=1.06; 95%CI 0.99-1.13; p=0.09). Median survival was 13.2-months vs. 11.5-months, and 5-year survival was 24% vs. 19% in EG and CG respectively (p=0.0008). There were 1,217 and 1,203 LC deaths, indicating no LC mortality reduction (RR=1.01; 95%CI: 0.93-1.09; p=0.77). Within 7-years of randomization, 1,072 and 1,022 lung cancers were detected in EG and CG, respectively (RR=1.05; 95%CI 0.96-1.14; p=0.27). Median survival was 15.4-months vs. 11.5months, and 5-year survival was 27% vs. 18% in EG and CG, respectively (p<0.0001). Among these cases, there were 764 and 811 LC deaths, indicating a trend toward reduced LC mortality that was not statistically significant (RR=0.94; 95%CI:0.85-1.04; p=0.24) In PLCO, randomization to CXR-screening produced a significant improvement in LC survival. This survival advantage cannot be attributed to any conventional screening bias including overdiagnosis. The benefit is diminished when lung cancers diagnosed well beyond the active screening interval are included in the analysis.