Willie NELSON.

Top Line: Few things are more hotly debated on twitter than the utility of routine lung cancer screening.

The Study: The rub is that lung cancer is common and it’s deadly. And most of the other heavy hitters (e.g. breast, prostate, colon, and cervical cancers) all have widespread screening tools. But there’s yet to be a slam dunk lung cancer screening strategy that clearly saves lives. The Dutch phase 3 NELSON trial enrolled <13K men ages 50-74 with a history of at least 15-20 pack years who continued to smoke or quit less than 10 years ago. They were randomized to non-con CT screening with automated volumetric readings of lung nodules at baseline, 1 year, 3 years, and 5.5 years versus no screening. Indeed, just like the notorious NLST, lung cancer mortality was statistically reduced from 3.3 → 2.5 deaths per 1000 enrollees. That translates to a much more impressive relative reduction of 26%. Enter the critics who think lung cancer specific mortality is a garbage endpoint when so heavily outweighed by all other mortality (13% in both arms here), not to mention the potential vagueness of causes of death. Perhaps the most interesting distinction between the Dutch and US trials is the rate of “positive” findings on screening CT: 2% in NELSON (with a positive predictive value of 44%) versus 39% in NLST (with a positive predictive value of a measly <4%). The Dutch might rightly point out that such accuracy in interpreting screening scans obviates much of the concern with needless invasive work-ups. Finally, in case you’re wondering (because the initial trialists weren’t), women enrollees were added on as an afterthought with similar results in their gender-specific subanalysis.

TBL: Non-con volumetric CT screening for lung cancer decreases lung cancer deaths without a heavy undue burden of working-up false positives. | de Koning, N Engl J Med 2020