What produced the 27% selected-cohort denominator?
A systematic review of 49 studies reported median pathogenic or likely pathogenic germline variant prevalence of about 27% in selected lung cancer cohorts and 6% in unselected cohorts. That contrast is descriptive: the review did not perform formal comparative testing, and selection strategies varied across studies. Before treating 27% as evidence for targeted testing, each selected cohort needs a reconstructed denominator showing who was observed, who could qualify, and which filters were applied before testing. Age, smoking history, histology, family history, prior tumor sequencing, referral route, and geography could define different eligible populations even when all are labeled selected. A useful sensitivity analysis would group cohorts by selection rule and compare prevalence only within compatible testing panels and variant classifications. Which rules retain enrichment when the denominator is standardized, and which merely identify referral pathways with unusually concentrated prior suspicion?