How big is too big for lung nodules on screening scans?BMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f1070 (Published 20 February 2013) Cite this as: BMJ 2013;346:f1070
Screening high risk adults for lung cancer in the US starts with low dose computed tomography of the chest, followed by further scans, a biopsy, or even resection for those with non-calcified nodules. A large minority of adults screened have these nodules, and the efficiency, effectiveness, and cost effectiveness of screening is intimately bound up with the size of nodule that triggers further investigation. A retrospective look at a large cohort of baseline screens suggests that increasing the trigger threshold from the current 5 mm to 7 mm, or even 8 mm, would cut the number of screened adults who need further investigations from 16% (3396/21 136) to 7.1% (95% CI 6.7% to 7.4%) or 5.1% (4.8% to 5.4%), respectively. The same move would delay diagnosis by up to nine months for 5.0% (1.1% to 9.0%) or 5.9% (1.7% to 10.1%) of adults with cancer—a total of six or seven people in this cohort of 21 136.
Would the delay matter? Possibly not, say the authors. Most screen detected cancers in this cohort were early stage adenocarcinomas, the least aggressive type.
We still need prospective testing of different thresholds in large populations, says a linked comment (p 289). Prediction models that incorporate patient characteristics and nodule features other than size would also help inform screening decisions. The appearance of the nodule (solid or semi-solid) and the presence (or absence) of spiculation are also important.
Cite this as: BMJ 2013;346:f1070