Chest X-ray useless for lung cancer screening? (PLCO Trial, JAMA)
The National Lung Screening Trial demonstrated annual screening with chest CT saves one life per 320 screened compared to yearly screening with chest radiography. However, there was no "usual care" group (no screening at all) to compare against.
Oken et al report additional results from the PLCO Trial (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial). PLCO enrolled 154,901 people aged 55-74 (the same ages as the NLST subjects) between 1993 and 2001. Half were randomized to get annual CXR screening for 3 years (and other stuff, like regular rectal exams); the others got "usual care," and 11% of them ended up receiving chest X-ray screening (slightly contaminating the results, but which was predicted ahead of time and allowed for). Follow-up was 13 years (continuing 10 years after screening stopped).
Results: Lung cancer incidence and mortality were identical between the two groups (1213 vs. 1230 lung cancer deaths, rate ratio 0.99).
Importantly, though, 45% of subjects were never-smokers, and only 10% were current smokers. To those who would exclaim, "These findings may not apply to smokers!" -- you'd be right, and the clever authors thought of that. They created a subset of 30,300 patients who would have been eligible for the NLST (same ages; median pack-year history of 52). Among this cohort, at six years of follow-up (about the same as NSLT):
CXR screening group: 518 lung cancer cases, 316 lung cancer deaths
Usual care group: 520 lung cancer cases, 334 lung cancer deaths
So ... chest X-ray screening of heavy smokers prevented 18 deaths, right? Maybe, maybe not. That's a ~5% absolute mortality reduction. Their rate ratio was 0.94 but crossed 1.0 (0.81 - 1.10). They calculated they had 77% power to detect a 20% mortality difference, but only 26% power to detect a mortality reduction of 10%.
Further suggestion of a potential tiny benefit to CXR screening: Among those with non-small cell lung cancer diagnosed during the trial, people in the screening group were slightly more likely to have their NSCLC be stage I (32% vs. 27%), and slightly less likely to be stage IV (35% vs. 38%).
Also, 998 of the 1696 total lung cancers were detected in people after they had completed 3 rounds of screening. As an editorialist makes clear, the longer the follow-up included after screening stops (in this "stop-screen" design), the harder it will be to detect a benefit of screening (as "interval" cancers mount after screening stops, equally in both groups due to the natural incidence of lung cancer, which pushes results toward null).
Reassuringly, of 12,778 people with a positive chest X-ray who underwent further diagnostics (repeat imaging, biopsies, etc.), only 54 (0.4%) had a complication. Roughly one-third of these were pneumothoraces.
Authors politely but firmly conclude that chest X-ray is ineffective as a screening test for lung cancer:
Annual screening with chest radiographs over a 4-year period did not significantly decrease lung cancer mortality compared with usual care neither in the PLCO as a whole nor in the subset of participants who would have been eligible to enroll in the NLST.
Oken MM et al (the PLCO Project Team). Screening by Chest Radiograph and Lung Cancer Mortality. JAMA 2011;306(17):1865-1873
Editorial by Harold Sox. JAMA 2011;306:1914-1915.
(Originally posted Oct 28 after E-pub ahead of print.)