Minimizing over-diagnosis in lung cancer screening.

Overestimation of the frequency and impact of over-diagnosis bias in lung cancer screening has contributed to long delays in implementation of lung cancer screening programs. Literature review reveals little evidence of substantial numbers of over-diagnosed non-lethal lung cancer. There is now strong evidence that lung cancers that would not cause symptoms or kill during normal anticipated survival are uncommon and mostly limited to in situ adenocarcinomas, identifiable as CT non-solid nodules. Prevention of overtreatment is possible within well-constructed diagnostic algorithms.